American Society of

Nuclear Cardiology

IMAGING GUIDELINES FOR NUCLEAR

CARDIOLOGY PROCEDURES
A Report of The American Society of
Nuclear Cardiology Quality Assurance Committee
Editor
E. Gordon DePuey, MD
Writing Group Chairs
James R. Corbett, MD Christopher L. Hansen, MD
John D. Friedman, MD Josef Machac, MD
Richard A. Goldstein, MD Kenneth J. Nichols, PhD
Milena J. Henzlova, MD Peter L. Tilkemeier, MD

Writing Group
Olakunle O. Akinboboye, MB, BS, MPH, MBA Ravi K. Garg, MD
Stephen L. Bacharach, PhD David K. Glover, MD, PhD
Timothy M. Bateman, MD Robert J. Gropler, MD
Jeroen J. Bax, MD Sean W. Hayes, MD
Robert Beanlands, MD Gary V. Heller, MD, PhD
Frank Bengel, MD Mark C. Hyun, CNMT
Steven R. Bergmann, MD, PhD Lynne L. Johnson, MD
Daniel S. Berman, MD Jonathan Links, PhD
Jeffrey S. Borer, MD Josef Machac, MD
Salvador Borges-Neto, MD John J. Mahmarian, MD
Elias H. Botvinick, MD April Mann, CNMT, NCT, RT(N)
Richard C. Brunken, MD Benjamin D. McCallister, Jr., MD
James Case, PhD Robert A. Pagnanelli, BSRT (R)(N), CNMT, NCT
Manuel D. Cerqueira, MD Randolph Patterson, MD
Keith B. Churchwell, MD Steven C. Port, MD
C. David Cooke, MSEE Heinrich R. Schelbert, MD
S. James Cullom, PhD Raymond Taillefer, MD
Seth T. Dahlberg, MD Mark Travin, MD
Dominique Delbeke, MD Serge Van Kriekinge, PhD
Marcelo F. DiCarli, MD Siu-Sun Yao, MD
Edward P. Ficaro, PhD Frans J. Th. Wackers, MD, PhD
James R. Galt, PhD R. Parker Ward, MD
Ernest V. Garcia, PhD

Approved by the American Society of Nuclear Cardiology Board of Directors

June 2, 2006.
Copyright © 2006 American Society of Nuclear Cardiology
All rights reserved.

e21
 Introduction to the Guidelines
E. Gordon DePuey, MD, Editor
ASNC Imaging Guidelines Committee Chair

These guidelines have been developed by the Quality Assurance Committee of the American Society of
Nuclear Cardiology (ASNC). They are an update and expansion of earlier guidelines that were also developed
by the ASNC Quality Assurance Committee.1-5
The task of the Committee has been to document state-of-the-art applications and protocols approved and
adopted by experts in the field and distribute these protocols to the nuclear cardiology community.
Subcommittees were formed to document guidelines for the particular applications addressed in this manual.
The final document was reviewed and approved by the American Society of Nuclear Cardiology Board of
Directors.
Within the document protocol items judged to be required are indicated as such. Standard means that the
parameter value listed represents methodology judged to be standard by the consensus of the committee; its
utilization is recommended, but other techniques may also be valid. Preferred means that the parameter value
listed is expected to provide the best results and its selection is strongly recommended. Techniques termed
optional indicate that the parameter value listed may be employed or another acceptable parameter may be
substituted. Techniques still considered primarily research applications and those not published in
peer-reviewed journals usually have not been included.
This manual is designed to provide imaging guidelines for those physicians and technologists who are
qualified in the practice of nuclear cardiology. Although care has been taken to ensure that information
supplied is accurate, representing the consensus of experts, it should not be considered as medical advice or
a professional service. The imaging guidelines described in this manual should not be utilized in clinical
studies at any institution until they have been reviewed and approved by qualified physicians from that
institution.
The experts who worked as the Chairs of the Writing Groups on the revision of these guidelines are listed
below by the protocol they contributed:

Instrumentation Quality Assurance and Performance Kenneth J. Nichols, PhD

First-pass Radionuclide Angiography (FPRNA) John D. Friedman, MD
Equilibrium Radionuclide Angiocardiography (ERNA) James R. Corbett, MD
Stress Protocols and Tracers Milena J. Henzlova, MD
Myocardial Perfusion Planar Imaging Peter L. Tilkemeier, MD
Myocardial Perfusion and Function SPECT Christopher L. Hansen, MD
Richard A. Goldstein, MD
PET Myocardial Perfusion and Glucose Metabolism Imaging Josef Machac, MD
Standardized Reporting of Myocardial Perfusion Images Peter L. Tilkemeier, MD

References
1. Updated imaging guidelines for nuclear cardiology procedures, Part 1. J Nucl Cardiol 2001;1:G5-G58.
2. Port SC, ed. Imaging guidelines for nuclear cardiology procedures. J Nucl Cardiol 1999;3:G47-G84.
3. Schelbert HR, Beanlands R, Bengel F, Knuuti J, DiCarli M, Machac J, Patterson R. Pet myocardial perfusion and glucose
metabolism imaging: part 2— guidelines for interpretation and reporting. J Nucl Card 2003;9:557-71.
4. Hendel RC, Wackers FJT, Berman DS, Ficaro E, DePuey EG, Klein L, Cerqueira M. Reporting of Radionuclide Myocardial
Perfusion Imaging Studies. J Nucl Cardiol 2003;10:705-8.
5. Tilkemeier PL, Cooke CD, Ficaro EP, Glover DK, Hansen CL, McCallister BD. Standarized reporting matrix for radionuclide
myocardial perfusion imaging. American Society of Nuclear Cardiology, July 2005, Available from: http://www.asnc.org/
yourpractice/standardizedreportingmatrix.pdf. Accessed: March 31, 2006.

e22
We acknowledge the following individuals who wrote the original, previously published
Imaging Guidelines for Nuclear Cardiology Procedures, Part 1 and Part 2, the PET Myocardial
Glucose Metabolism and Perfusion Imaging, Part 1 and Part 2, and the Reporting of
Radionuclide Myocardial Perfusion Imaging Studies consensus statement.
Stephen L. Bacharach, PhD, Timothy M. Bateman, MD, Jeroen J. Bax, MD, Robert Beanlands,
MD, Frank Bengel, MD, Daniel S. Berman, MD, Jeffrey S. Borer, MD, Salvador Borges-Neto,
MD, Ignasi Carrio, MD, James Case, PhD, Manuel Cerqueira, MD, S. James Cullom, PhD,
Dominique Delbeke, MD, E. Gordon DePuey, MD, Marcelo Di Carli, MD, Edward P. Ficaro,
PhD, John D. Friedman, MD, James R. Galt, PhD, Ernest V. Garcia, PhD, Guido Germano, PhD,
L. Stephen Graham, PhD, Mark Groch, PhD, Rory Hachamovitch, MD, Robert C. Hendel, MD,
Diwakar Jain, MD, Lynne L. Johnson, MD, Hosen S. Kiat, Larry Klein, MD, MD, Juhani
Knuuti, MD, Karen A. Kurdziel, MD, Scott M. Leonard, BS, CNMT, Josef Machac, MD,
William H. Martin, MD, John J. Mahmarian, MD, Kenneth J. Nichols, MD, Robert Pagnanelli,
BSRT(R)(N), CNMT, Randolph E. Patterson, MD, Dudley J. Pennell, MD, Pierluigi Pieri, MD,
Steven C. Port, MD, Janet Saffer, PhD, Heinrich R. Schelbert, MD, PhD, Douglas Schulman,
MD, Albert J. Sinusas, MD, Wendy R. Smeltzer, CNMT, Patricia Stewart-Henney, RT, H.
William Strauss, MD, Kenneth J. Tillman, CNMT, Kenneth Van Train, MS, Frans J. Th.
Wackers, MD, PhD, Denny D. Watson, PhD, Howard Weinstein, MD, Kim A. Williams, MD,
Michael Yester, PhD, Robert Zimmerman, MS, I. George Zubal, PhD

Publication of the Guidelines is supported by educational grants from:

Astellas Pharma US, Inc.
Bristol-Myers Squibb Medical Imaging
Bracco Diagnostics Inc.
Digirad
GE Healthcare
Mallinckrodt
MDS Nordion

e23
 IMAGING GUIDELINES FOR NUCLEAR CARDIOLOGY PROCEDURES

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . e22

Instrumentation Quality Assurance and Performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . e25

First-pass Radionuclide Angiography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . e42

Equilibrium Radionuclide Angiocardiography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . e56

Stress Protocols and Tracers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . e80

Myocardial Perfusion Planar Imaging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . e91

Myocardial Perfusion and Function Single Photon Emission Computed

Tomography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . e97

Positron Emission Tomography Myocardial Perfusion and Glucose Metabolism

Imaging. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . e121

Standarized Reporting of Myocardial Perfusion Images . . . . . . . . . . . . . . . . . . . . . . . . . e152

e24
 Instrumentation quality assurance and performance
Kenneth J. Nichols, PhD,a Stephen L. Bacharach, PhD,b
Steven R. Bergmann, MD, PhD,b S. James Cullom, PhD,b Edward P. Ficaro, PhD,b
James R. Galt, PhD,b Gary V. Heller, MD, PhD,b Jonathan Links, PhD,b
and Josef Machac, MDb

INTRODUCTION different energy positions in the spectrum of different

cameras. Thus, energy peak and window settings should
The proper choice of equipment to acquire clinical
be established for each individual camera based on the
data for specific purposes and a well-designed quality
energy spectrum display.1
assurance (QA) program are both essential requirements Parallel-hole collimation is standard. The low-
for optimizing diagnostic accuracy. The following guide- energy, high-resolution collimator is usually best for
lines are intended to provide reasonable suggestions as Tc-99m, although some “all-purpose” collimators give
to appropriate means of assessing equipment function excellent results. Imaging with Tl-201 is usually best
in conjunction with nuclear cardiology imaging. Be- with the low-energy, medium-resolution (all-purpose)
cause equipment manufacturers can vary considerably collimators because count statistics become limiting
with the optimal manner in which to perform specific when using high-resolution collimators. The differ-
tests, this document should be used as guidelines only ence in medium- and high-resolution collimators is
and is not intended to replace the recommendations by usually that the collimator depth (length of the colli-
manufacturers of specific models of imaging equip- mator hole) is greater in high-resolution collimators.
ment. They have similar near-field resolution. The “high-
resolution” collimator maintains good resolution at a
EQUIPMENT greater distance from the collimator face. The differ-
ence is more important in single photon emission
Planar imaging. The small–field-of-view (FOV)
computed tomography (SPECT) imaging, where the
scintillation camera is ideal for cardiac imaging. The
distance from patient to collimator is greater.2
10-inch FOV covers the heart and shows enough sur-
rounding area to evaluate lung uptake and to sample
extracardiac background activity. A 128 ⫻ 128 matrix Performance parameters for low energy
over this FOV results in pixel spacing of about 2 mm. A (<150 keV collimators)
camera with a 15-inch FOV should be zoomed using a
magnification factor of 1.2 to 1.5 so that the pixel size Resolution
is less than 3 mm and approximately equal to 2.5 or (FWHM at Relative
2.0 mm. Collimator type 10 cm) sensitivity
Energy windows should be symmetric about the
photopeak. A window of 20% is standard for technetium Ultra-high resolution 6 mm 0.3
99m. With the improved energy resolution of many High resolution 7.8 mm 0.6
modern cameras, a 15% window can be used with little All/general purpose 9.3 mm 1*
loss of primary gamma rays, with some improvement in High sensitivity 13.2 mm 2.1
contrast. The low energy and greater width of the
*A relative sensitivity of 1 corresponds approximately to a colli-
thallium 201 photopeak require a wider window. An mator efficiency of 2.7 ⫻ 10⫺4, whereas efficiency is defined as the
energy window setting of 30% is appropriate for the fraction of gamma rays and x-rays passing through the collimator
70-keV peak of Tl-201 and a 15% energy window for the per gamma ray and x-ray emitted by the source.
167-keV peak. The absolute energy calibration can be
unreliable at the low energy of Tl-201 falling at slightly Collimators. The selection of which collimator to
use is important to resultant image quality. A confound-
ing aspect of this selection is that collimators with the
same name (eg, “general purpose”) vary in performance
Chair. Member. from one manufacturer to another. The table above gives
J Nucl Cardiol 2006;13:e25-41.
1071-3581/$32.00 approximate values for collimator specifications. Refer
Copyright © 2006 by the American Society of Nuclear Cardiology. to specific imaging protocols for appropriate collimator
doi:10.1016/j.nuclcard.2006.08.005 selection.
e25
e26 Nichols et al Journal of Nuclear Cardiology
Instrumentation quality assurance and performance November/December 2006

SPECT imaging. SPECT detectors are scintillation acquisition modes such that detectors do acquire data as
cameras mounted on a gantry. Many variables dictate the they move. Studies using both computerized and phan-
performance of a SPECT imaging system, including the tom simulations have shown that image quality is essen-
number of detectors for a given device. Single-head tially the same whether data are collected continuously
cameras have been used widely for cardiac imaging. or in a step-and-shoot fashion.5
Adding more detectors is beneficial, since doubling the SPECT/transmission computed tomography imag-
number of detectors doubles acquired counts, if all other ing. Currently, there are two types of transmission
variables remain fixed. For cardiac SPECT studies in tomographic imaging systems for acquiring patient-spe-
which a 180° orbit is recommended, the preferred con- cific attenuation maps that can be used to correct SPECT
figuration is to have two detectors separated by 90° as images for photon attenuation. The first type, referred to
they rotate around the heart. For studies in which a 360° in these guidelines as transmission computed tomogra-
orbit is preferred, three detectors separated by 120° from phy (TCT), uses a sealed source (eg, gadolinium 153)
each other are preferred. with the standard collimated scintillation detectors used
The trend is to trade off some of the additional for SPECT imaging. The second type of transmission
counts that would be obtained with multiple detectors imaging system uses an x-ray tube in conjunction with a
equipped with general-purpose collimators for higher- computed tomography (CT) detector. The primary dif-
resolution imaging by using higher-resolution collima- ference between these classes of transmission imaging
tors. Fan beam collimators are available but are not now systems is the photon emission rate that dictates the
commonly employed in cardiac imaging. These collima- quality-control (QC) protocols that are required. Because
tors allow more of the crystal area to be used in imaging x-ray tubes can produce photon currents which can be
the heart, magnifying the image and increasing sen- counted with CT detectors at significantly higher rates
sitivity.2 In large patients or when using a converging than conventional sealed sources and scintillation detec-
geometry with a steep angle, there is a potential for tors, CT-acquired images can be acquired on the order of
cutting-off, or truncating, portions of the heart and/or seconds to a few minutes depending on the tube strength.
chest. This truncation may generate artifacts. As a result, x-ray CT projection data are not contami-
Another variable in SPECT systems is detector nated by scatter emission photons. The photon fluence
orbit. The traditional orbit used for SPECT acquisition from sealed-source TCT systems is much lower than that
has been circular, with a rotational range of 180° or 360° from x-ray CT, so the emission scatter or cross-talk into
and with step-and-shoot motion. Most systems today the transmission window cannot be ignored for TCT.
allow elliptical detector motion to follow the body Due to the low photon fluence and the non-negligible
contour, reducing the distance from the camera to the cross-talk scatter component from the injected radio-
body, thereby improving spatial resolution. For single- tracer, the TCT QC procedure is slightly more involved
head cameras, the 180° angular range from the 45° right and will be outlined separately from the QC for systems
anterior oblique to the 135° left posterior oblique orien- using x-ray tubes.
tation has become the preferred orbit, because the detec- SPECT x-ray CT imaging. Consistent with trends in
tor is closer to the heart in these views, resulting in positron emission tomography (PET)/CT systems, hybrid
higher spatial resolution and image contrast. There is SPECT systems have evolved, combining SPECT and
also less scatter and attenuation over this angular range. CT systems. While PET systems are generally complete
The 360° orbit usually results in better uniformity of rings or a partial-ring system, the SPECT components
normal myocardium.3,4 Some double- and triple-head are typically large-FOV variable-angle dual-detector sys-
systems acquire the entire 360° orbit because of the tems. These combined systems, in practice, demonstrate
fixed separation of the detectors. In these systems the a range of capability and integration. CT components
entire 360° orbit can be used. One concern in using range from non-diagnostic units suitable for use in
360° acquisitions is that reconstructed images can anatomical localization and attenuation correction to
have a different “normal distribution” compared to 16-slice systems capable of CT angiography. The
those from 180° orbits. This could cause a misdiag- SPECT detectors in SPECT/CT systems do not differ in
nosis if interpreted as a 180°-acquired study, since the any significant way from those of stand-alone SPECT
inferior wall usually has lower activity in normal systems. These systems may be viewed from a protocol
subjects acquired with a 180° orbit than when acquired perspective as stand-alone systems where an emission
with a 360° orbit.3,4 study is followed or preceded by a CT scan for
Most SPECT systems use the step-and-shoot mode attenuation correction. Depending upon the number of
of acquisition in which the detector does not acquire data CT slices acquired, the CT scanner may be used, as
as it moves from one angle to the next. Some systems with stand-alone CT scanners, for CT angiography and
offer either continuous or continuous step-and-shoot calcium scoring. The CT and SPECT components may
Journal of Nuclear Cardiology Nichols et al e27
Volume 13, Number 6;e25-41 Instrumentation quality assurance and performance

then be analyzed independently or in 3-dimensional other crystal types in use. Each can be used successfully
(3D) image registration, depending on the type of for cardiac imaging. BGO has the highest stopping
study. power, but it has the poorest energy resolution (useful for
PET imaging–2-dimensional versus 3D scanners. scatter reduction) and timing resolution (useful for min-
The majority of dedicated PET cameras consist of rings imizing random events). GSO and LSO both have better
of small detectors (typically a few millimeters on a side, timing resolution and, in theory, better energy resolution.
several 10s of millimeters deep). Coincidences between If 2D imaging is being contemplated, it is possible that
detectors in a single ring produce one tomographic slice GSO and LSO may not offer much advantage over BGO.
of data. Usually one or more adjacent rings may also For 3D imaging, GSO and LSO can better minimize
contribute to counts in that slice. In a so-called 2-dimen- random events, although some BGO machines can in
sional (2D) PET scanner, there is a lead or tungsten part compensate for this with better dead-time perfor-
septum (a 1-dimensional collimator) between adjacent mance. The better energy resolution of GSO or LSO and
rings. This septum partially shields coincidences from consequent reduction of scatter in 3D mode make these
occurring between detectors in one ring and detectors in detector types advantageous. At present, the theoretical
a non-adjacent ring. By minimizing coincidences be- energy resolution for these detectors does not seem to
tween a ring and its more distant neighboring rings, the have been realized in practice, leaving all three crystal
septum greatly reduces scattered events. Many manufac- types with similar energy-based scatter rejection and
turers have tried to increase the sensitivity of their making 2D imaging still the method of choice if scatter
scanners by removing the septa between adjacent rings. rejection is critical.
This permits coincidences between all possible pairs of PET imaging–attenuation correction. In most previ-
detectors, greatly increasing sensitivity but also greatly ous models of PET scanners, rotating rod sources of
increasing scatter. A scanner with no septa in place is germanium 68/gallium 68 were used to perform a trans-
referred to as a “3D” or “septa-out” scanner. The in- mission scan prior to or immediately after emission
creased sensitivity is greatest for the central slice and imaging. Current commercially available PET scanners
falls rapidly (and usually linearly) for slices more are PET/CT scanners, which rely on CT scans for
distant from the central slice. The edge slices have a attenuation correction. In 2D scanners, rotating rod
sensitivity about the same as in a 2D scanner but with sources work well for cardiac imaging, although it adds
greater scatter. Scatter as measured by the National 3 to 8 minutes to scan time. For scanners that can only
Electrical Manufacturers Association (NEMA) is typ- operate in 3D mode, a rotating rod source can be
ically on the order of 10% to 15% for 2D scanners and problematic, due to excessive count rates in the detec-
30% to 40% or more for 3D scanners. In chest slices tors. Sometimes, a non–positron-emitting source such as
encompassing the heart (as opposed to the relatively cesium 137 (shielded from the closest detectors) is used
small NEMA phantom), there is an even larger in- to reduce count rate problems, although this methodol-
crease in scattered counts for 3D imaging. For cardiac ogy has some difficulties, especially with scatter. The CT
applications, scatter tends to increase the counts in scanner in hybrid systems is usually used for the atten-
cold areas surrounded by higher-activity regions (eg, a uation correction transmission scan, though this can
defect surrounded by normal uptake). Some manufac- cause problems if the CT and PET scans are not precisely
turers have scanners that have retractable septa, per- aligned.
mitting the user to choose between 2D or 3D opera- PET/CT imaging. The latest trend in emission com-
tion. Many (but not all) PET/CT manufacturers have puted tomography (ECT), including both SPECT and
opted for scanners that operate only in 3D mode PET, is the addition of a CT system to the ECT scanner.
(which many companies feel is optimum for oncology These combined systems, in practice, demonstrate a
studies due to patient throughput concerns). range of integration. At one end of the spectrum, the
Situations in which 3D mode may be advanta- hardware and software of the CT system are completely
geous include those in which (1) whole-body patient integrated within the PET scanner. In this approach, a
throughput is important (eg, a busy oncology prac- common, unified gantry is used, and a single, unified
tice); (2) radiation exposure is critical, so reductions in software system with an integrated ECT/CT interface is
injected activity are desired; and (3) special (usually provided. At the other end of the spectrum, the hardware
research) radiopharmaceuticals are being used, which and software of the CT system stand alone. In this
can only be produced in sub-millicurie quantities. approach, a separate CT gantry is carefully placed in
PET imaging– crystal types. Three different crystal front of or behind the ECT gantry, and a separate
types are commonly employed—BGO (bismuth ger- workstation is used to control the CT system.
manate), GSO (gadolinium oxyorthosilicate), and LSO The commercially available PET/CT systems dem-
(lutetium oxyorthosilicate)—although there are also onstrate this range of integration. In all cases, the
e28 Nichols et al Journal of Nuclear Cardiology
Instrumentation quality assurance and performance November/December 2006

manufacturer starts with a state-of-the-art PET scanner, Laboratories applying for ICANL accreditation have
whose characteristics have been described in the section the options of accreditation for nuclear cardiology, gen-
above. The manufacturer then adds a CT system, con- eral nuclear medicine, or PET or a comprehensive
sisting of a 2-, 4-, 6-, 8-, or 16-slice scanner (with 32- and accreditation that includes two or more modalities. The
64-slice combined systems anticipated, given the exis- ICANL equipment guidelines also require proof via
tence of such stand-alone CT systems). In all cases, the written documentation of ongoing camera and non-
combined PET/CT scanner appears to have a unified imaging equipment QC that is reviewed during the site
gantry (because of the covers), although separate gantries visit. In addition to the camera-specific QC program that
for the PET and CT portions may actually be used. must include daily energy peaking and intrinsic or
Depending on the degree of system integration, separate extrinsic uniformity, weekly resolution and linearity,
workstations may be used to control the PET and CT monthly high count floods, and center of rotation (COR),
portions of the combined system, or a single, unified the laboratory must also submit camera- and computer-
workstation may be present. specific acquisition and processing protocols.7 Many
other written protocols and documentation of their ad-
herence are reviewed during the site visit and include but
ACCREDITATION ISSUES RELATED TO EQUIPMENT
are not limited to radiation safety and radioactive mate-
AND QA
rials handling, clinical and general protocol guidelines,
A well-designed, regularly followed QA plan for and administrative protocols. In addition to the protocols,
imaging devices ensures the best possible diagnostic all medical and technical personnel interpreting and
service to the patient population. Agencies that accredit performing imaging services within the laboratory are
facilities for medical use take QA programs seriously. required to be appropriately trained, experienced, and
For instance, the Joint Committee on the Accreditation of credentialed and/or licensed. A typical application sub-
Hospital Organization’s guidelines address the use of mission for accreditation consists of a compilation of
medical isotopes and associated radiation measuring and documentation, including the professional credentials
imaging equipment. of medical and technical staff, a list of imaging and
Over the past few years, there has been a growing non-imaging equipment, and written imaging and
trend among medical insurance companies to require that stress procedure protocols. Additionally, the laborato-
a facility be accredited by an external agency as a ries are required to submit selected patient studies in
prerequisite for reimbursement of medical imaging pro- digital format as well as copies of the final reports sent
cedures.6 to referring physicians. For nuclear cardiology, a
Agencies that accredit nuclear cardiology facilities, minimum of five myocardial perfusion studies is
specifically, include the American College of Radiology required, and all nuclear cardiology case studies must
(ACR) and the Intersocietal Commission for the Accred- be randomly selected using the guidelines outlined in
itation of Nuclear Medicine Laboratories (ICANL).7 The the application.
ACR equipment guidelines require proof of a QA pro-
gram specific to gamma cameras, SPECT cameras, and
QC PROCEDURES
PET cameras (as appropriate to a given facility), includ-
ing the submission of images of flood fields, bar phan- Planar imaging QC. Appropriate QC procedures are
toms, multipurpose Plexiglas SPECT phantoms, multi- necessary to ensure images of the highest possible
purpose Plexiglas PET phantoms, and equipment technical quality for the equipment used and thus allow
acceptance testing reports.6,8 An annual physicist’s re- the best possible diagnostic service to the patient popu-
port also must be submitted, which includes all of the lation.9
imaging examples just described, along with the results
of additional annual camera tests and the results of tests Planar QC procedures
of non-imaging equipment used to measure radiation
(survey meters, dose calibrators, and so on). An accept- For information,
able proof of proper credentialing of all physicians, Test Frequency see paragraph
technologists, and physicists must be submitted. Repre-
sentative clinical images also must be submitted. Panels Energy peaking Daily 1
of physicists review phantom images and physicists’ Uniformity test Daily 2
reports, and panels of physicians review clinical images, Resolution and Weekly 3
all of which must be judged to be of an acceptable quality linearity
in order for a facility to be eligible for accreditation. Sensitivity test Weekly or daily 4
Journal of Nuclear Cardiology Nichols et al e29
Volume 13, Number 6;e25-41 Instrumentation quality assurance and performance

Performance parameters for detectors MBq). For large rectangular cameras (such that the
point should be 7-8 feet away), 20 to 25 MBq is
For appropriate. In some cameras, obtaining intrinsic
information,
Parameter Standard Preferred see paragraph flood fields can be difficult. Some manufacturers
provide software to correct for non-uniformities
Integral uniformity ⬍5% ⬍3% 2
due to the necessity of having a point source closer
Differential uniformity ⬍5% ⬍3% 2 than 5 UFOV diameters. Because of these difficul-
Intrinsic resolution full ⬍6 mm ⬍4 mm 3 ties, it may be more practical to perform this test
width at half extrinsically using radioactive sheet flood sources.
maximum
To ensure a true response during acquisition, count
rates should be kept between 10 and 25 kcps. For
1. Energy peaking. Energy peaking is performed to some older systems, a lead ring should be used to
verify that the camera is counting photons having the shield the outermost tubes from the radiation to
correct energy. This test consists of either manually or prevent edge packing. Flood images that will be
automatically placing the correct pulse height analyz- inspected visually should be acquired as 256 ⫻ 256
er’s energy window over the photopeak energy to be matrices for 3 M counts (5 M for larger rectangular
used. Care must be taken that the technologist verifies detectors). Photographs of the flood field should be
the correct placement of the window and that a recorded and stored, both physically and digitally.
radioactive point source is used at a distance away of Flood images used for calculations of uniformity
greater than 5 useful FOV (UFOV) diameters from require two to three times more counts to reduce
the uncollimated camera; a sheet source is typically statistical noise. The recommended number of
used in front of a collimated camera.10 In either case, counts is at least 4,500 counts/cm2 (eg, 5.7 M
the full UFOV of the camera should be illuminated by counts for a 400-mm circular detector). NEMA
the source. Window verification should be done even recommends acquiring a minimum of 10,000
on automated systems where there are single buttons counts for the center (6.4 mm) pixel of the flood
or computer protocols to select for each energy; even
image.10
in these automated systems, the energy windows tend
The flood images should be examined each day for
to drift. These systems allow for window offsets to
each detector prior to use to verify that the detectors
correct for these drifts. This peaking test should
are properly peaked and that the floods are uniform.
indicate for each camera head whether the camera’s
In addition, several parameters are quantified from the
automatic peaking circuitry is working properly,
flood images, which should be computed and the
whether the peak appears at the correct energy, and
results recorded as part of the usual QA procedures.
whether the shape of the spectrum is correct. If cost,
In the event of power shortages and power outages,
time, and the equipment permit, photographs of the
the process of peaking and flooding the detectors
spectra with the superimposed energy window should
should be performed again to ensure proper function
be taken and stored.
before resuming patient imaging. Two uniformity
2. Uniformity test. Uniformity testing is performed to
parameters are computed—integral uniformity and
verify that the camera’s sensitivity response is
uniform across the detector’s face. Some manufac- differential uniformity. If the flood images are ac-
turers recommend that this test be performed in- quired in a larger matrix size, the pixel matrix should
trinsically (using a point source without collima- be reduced to 64 ⫻ 64 by summing or averaging
tors) while others recommend that this test be pixels prior to uniformity calculation. Integral unifor-
performed extrinsically (with the collimator in mity is a global parameter measuring uniformity over
place in conjunction with a sheet source, usually of an extended area of the detector, expressed as fol-
cobalt 57). This test consists of exposing the lows:
camera with a uniform source of radioactivity, a
process commonly referred to as “flooding” the
detector. If performed intrinsically, a radioactive Integral uniformity ⫽ 100% ⫻ (Max ⫺ Min)
point source is positioned at a distance at least five ⁄ (Max ⫹ Min)
times the crystal’s UFOV from the center of the where Max is the maximum count and Min is the
detector. This test is usually performed immedi- minimum count found in any pixel within the speci-
ately following peaking of the detector. The point fied area. Differential uniformity is a regional param-
source should consist of a small volume (approxi- eter that measures contrast over a small neighbor-
mately 0.5 mL) of fluid and low activity (7-11 hood. This measurement is performed using all 5 ⫻
e30 Nichols et al Journal of Nuclear Cardiology
Instrumentation quality assurance and performance November/December 2006

1–pixel areas in both the X and Y directions, ex- conducted, the phantom should be rotated 90° so that
pressed as follows: every fifth time the test is done, the pattern position
repeats. Bar phantom images should be recorded and
Differential uniformity ⫽ 100% ⫻ Largest deviation stored. These images should be assessed for how
(Max ⫺ Min) ⁄ (Max ⫹ Min) straight the lines imaged are and for intrinsic spatial
resolution. Change in resolution is assessed by docu-
It should be noted that manufacturers vary consider- menting the smallest bars that are discerned. Spatial
ably as to their recommendations regarding the ability resolution as measured by the full width at half
of a particular camera to use a flood field collected at
one energy (eg, 140 keV for Tc-99m) to correct the maximum may be approximated by multiplying 1.7
field of data acquired at a different energy (eg, 70 keV times the smallest bar size seen.9
for Tl-201).11 For some Anger cameras, it may be 4. Sensitivity test. This test is performed to document
essential to acquire flood fields separately for Tl-201, the sensitivity of the detector and, more importantly,
iodine 123, I-131, and so on. In that case, most users the change of sensitivity over time. The test consists
will perform these corrections intrinsically, not ex- of calculating detector sensitivity (expressed in terms
trinsically. Failure to apply an adequate overall flood- of counts per minute per megabecquerel) of a known
field correction is seen most strikingly on the uniform source, calibrated with a dose calibrator. The point
section of a multipurpose Plexiglas SPECT phantom source should always be located at exactly the same
(see below).9
distance in front of the camera for repeat measure-
3. Resolution and linearity test.This test is performed ments. A convenient means of measuring sensitivity
to document spatial resolution and its change over changes is by recording the time that it takes to
time as well as the detector’s ability to image straight acquire the preset counts for an intrinsic (or extrinsic,
lines. The test consists of imaging a flood source if more practicable) flood source.
intrinsically through a spatial resolution test phantom.
The flood source should be acquired as described in SPECT imaging QC. All of the QC procedures
the Uniformity test section. Most commercially avail- required of planar imaging instruments are also required
able bar phantoms are suitable for this test. These for SPECT imaging, since tomography depends on
include the parallel-line-equal-space bar phantoms acquiring accurate planar projections. In addition, proce-
and orthogonal hole or 4-quadrant phantoms. If the dures specific to SPECT imaging systems are discussed
4-quadrant phantom is used, each time the test is below.

SPECT QC procedures

For information
Test Requirement Frequency see paragraph

COR Mandatory Weekly 1

Uniformity correction Mandatory Weekly 2
Multipurpose Plexiglas phantoms Strongly recommended Quarterly-annually 3

1. COR. An alignment error between the electronic matrix some systems this means that two separate acquisitions
of the detector and the mechanical COR can potentially are required: one for each detector (with the other
result in a characteristic “doughnut” (if a 360° orbit and detector disabled). Many manufacturers require that a
a point source are used) or “tuning fork” artifact (if a specific protocol be followed for the determination and
180° orbit is used) in the transverse images.12 The recalibration of the COR. While some manufacturers
effects are most evident when the error is greater than limit the COR calibration to service engineers, all
two pixels in a 64 ⫻ 64 matrix. Errors less than this systems may be checked for correct COR calibration. If
reduce spatial resolution and image contrast through no specific COR acquisition protocol is recommended
blurring of the image and cause significant artifacts by the manufacturer, the COR may be determined
(particularly at the apex).13 The accuracy of COR through the acquisition of a point source of activity
alignment should be checked weekly for each camera (18-37 MBq) on the patient table 4 to 8 inches away
head, unless indicated otherwise by the manufacturer. In from the axis of rotation. SPECT data are acquired over
Journal of Nuclear Cardiology Nichols et al e31
Volume 13, Number 6;e25-41 Instrumentation quality assurance and performance

360° with equally spaced projections with a circular atic for daily use. Using the lower-energy correction
orbit. The same angular orientation, collimation, zoom, floods for higher-energy radionuclides can result in
matrix size, and energy window employed for the incorrect compensation and therefore is not recom-
patient study should be employed for the COR acquisi- mended.
tion. Five to ten seconds per frame for 64 views over 3. Multipurpose Plexiglas phantoms. It is strongly
360° is sufficient. COR correction values for each orbit recommended by NEMA that acquisition and recon-
are then computed, stored in the computer, and used to struction of a multipurpose Plexiglas phantom should
realign the projection data before reconstruction. It is be performed quarterly.15 For facilities intending to
essential that COR errors be checked for each collimator pursue laboratory ACR accreditation, performing
that is to be used clinically.14 It is recommended that these tests is mandatory. In particular, it should be
these measurements be performed weekly. New COR noted that if accreditation is sought from the ACR, it
calibrations should be performed after servicing of the is required to submit acceptable SPECT phantom
camera, after power surges or outages, and for the images for both Tc-99m and Tl-201 simulations.
computer after software upgrades. These SPECT phantoms are cylindrical or elliptical
2. High-count extrinsic flood-field uniformity correc- water baths into which radioactivity is injected and
tions. Manufacturers vary considerably as to their contain regions with solid spheres of different sizes,
recommended schedule and means of acquiring these regions with solid rods or bars of different sizes
high-count corrections. It has recently been noted that alternating with radioactive water, and regions con-
some cameras may not require the acquisition of taining only radioactive water.16 These phantoms are
extrinsic floods more often than annually to verify used to determine the 3D contrast, resolution, and
collimator integrity and that all uniformity corrections uniformity of the scanner, for which high activities
should be acquired intrinsically, so long as the camera (740-925 MBq) and “fine” sampling (128 ⫻ 128
is correctly tuned.11 For many systems, collimators matrices and 128 projections over 360°) generally are
are sufficiently well designed and manufactured that employed.9 Acquisitions are performed using typical
they do not degrade SPECT uniformity. Therefore, as Tc-99m energy settings, with detectors positioned as
with any of the procedures discussed in these guide- close to the phantom as is feasible throughout a 360°
lines, it is always important to follow the manufac- acquisition, so as to optimize spatial resolution.2 At
turers’ recommended QA protocols. least 30 million counts should be acquired, and data
In SPECT, it is implicitly assumed that the efficiency then are reconstructed for sections of the phantom to
of photon detection is constant across the surface of the enable assessment of contrast through the center of
collimated detector. Flood-field uniformity errors result spheres, 3D resolution through sections of rods, and
when the variation in efficiency is significant as com- uniformity through “blank” sections. Systemic prob-
pared to the performance parameters in the above table. lems that can be revealed in this fashion include
Anger cameras utilize stored flood-field correction maps suboptimal energy resolution through failure to dis-
to correct for variations in sensitivity across the FOV play adequate contrast, potential COR problems
before reconstruction. Deficiencies can lead to charac- through loss of resolution through solid rod sections,
teristic “ring” artifacts, most easily seen on the uniform and inappropriate or inadequate flood-field correc-
sections of multipurpose Plexiglas SPECT phantoms9 tions through the appearance of anomalous concentric
and myocardial perfusion artifacts. Daily checks of rings in uniform sections. These tests should be
flood-field extrinsic (with collimator) uniformity are performed quarterly, as well as following major
performed with a 3-million-count flood for a typical equipment repairs and installation of new software, to
FOV 128 ⫻ 128 or 256 ⫻ 256 matrix. To correct for verify the overall ability of the hardware and software
sensitivity variations due to the collimator, 30- to 100- to correctly perform tomographic reconstructions.
million-count images are acquired for each detector (128 When used in conjunction with standards established
⫻ 128 or 256 ⫻ 256 matrix) and stored for uniformity during acceptance testing, these quarterly tests can be
correction. It is essential to perform uniformity measure- helpful in signaling the point at which the manufac-
ments for each collimator and that the same collimator turer’s service representatives should be called to
that was used to acquire the flood and generate the further diagnose the causes of significant degradation
correction matrix be used to acquire the patient study. It of 3D system performance and to remedy these
is important that energy values similar to those being problems.11
used for clinical studies also be used for the flood source.
Co-57 solid sheet sources (122 keV) are commonly used QC Procedures for Sealed-Source SPECT/TCT Systems
and, more rarely, Tc-99m fillable sources for Tl-201 and SPECT parameters outlined in the SPECT imaging
Tc-99m imaging. A solid sheet source is less problem- QC section of these imaging guidelines should be used to
e32 Nichols et al Journal of Nuclear Cardiology
Instrumentation quality assurance and performance November/December 2006

ensure the quality of the emission tomographic data. In blank scan should be uniform, similar to uniformity
addition to those parameters, QC guidelines need to be floods. Stringent uniformity indices of ⫾ 10% are
followed to ensure that the transmission system using a not reasonable for the blank scan. Rather, the blank
sealed radioactive source is operating as designed. These scan should be inspected to ensure that there are no
tests are tabulated below. gross non-uniformity artifacts (ie, holes or bands of
pixels with no counts). For scanning-source sys-
QC procedures for sealed-source SPECT/TCT tems, the blank scans should not show discontinui-
systems ties or abrupt changes in pixel intensity in the axial
direction of the scanning source. The presence of
For information, these artifacts is consistent with improper scan-
Test Frequency see paragraph ning-detection alignment and should be checked by
a service engineer.
Energy peaking Daily 1 3. Source strength. For systems using a Gd-153 trans-
Transmission source Daily 2 mission source, photons collected in the transmission
mechanics window consist of primary transmission photons and
Source strength Monthly 3 scattered photons (cross-talk) from the emission ra-
diotracer. The ratio of these components, transmission
1. Energy peaking. This test is performed to verify that and cross-talk, is referred to as the transmission–to–
the camera is counting photons in the proper energy cross-talk ratio (TCR). This TCR value depends on
windows. Using a pulse height energy (z) analyzer, the transmission source strength, the injected radio-
which is available on all acquisition stations, the opera- pharmaceutical, the injected activity, and the body
tor should verify that the emission, transmission, and habitus. Transmission source decay, higher injected
scatter (if applicable) windows are properly set and that activities, and larger body sizes all tend to decrease
photons are being counted in each window. For some the TCR value. Lower TCR values result in recon-
systems, this may necessitate manually opening the structed attenuation maps with increased bias and
shutter to the transmission source. If this is not possible, noise. Since the TCR value will decrease as the
a quick “blank” scan (see next paragraph) can be source decays, its behavior should be trended over the
acquired to verify that transmission photons are being life of the source, which can guide the user as to when
properly counted. the sources should be replaced. This QC protocol
NOTE: Consult the vendor’s recommendation for en- should be performed at least monthly, with the base-
ergy peaking. Some vendors do not permit or require the line scan being performed when the TCT-ECT system
peaking of all energy windows simultaneously. is installed or the transmission sources have been
2. Transmission source mechanics. When patients are replaced. If the user suspects problems with the
not being imaged, the transmission source is shielded TCT-ECT system, a test should be performed imme-
and, on systems where the source translates across the diately prior to using the system for patient imaging.
FOV, left in the “parked” position. When a patient is Two protocols are provided, one using a cylinder
imaged, the shutter used to shield the source is phantom and the other using an anthropomorphic
opened, allowing transmission photons to be directed chest phantom. The chest phantom provides the more
toward and through the patient. For translating comprehensive check of the TCT-ECT system for
sources, the source will then translate axially along cardiac imaging compared to the cylinder for obvious
the axis of the body for each projection. To verify the reasons. For those sites that may not have access to a
operation of the source shutter and translating me- chest phantom, the cylinder protocol is provided
chanics, a reference “blank” transmission scan should which is capable of identifying potential problems
be acquired. This scan is required for all TCT proto- with a TCT-ECT system.
cols and is recommended to be acquired weekly and
possibly daily prior to the first use of the system for SPECT/TCT Protocol 1: Cylinder Phantom
that day. The frequency of this test will depend on the Required Equipment
half-life of the isotope of the transmission source and 18- to 20-cm-diameter fillable cylinder
the stability of the TCT system. Follow the manufac- 111 to 185 MBq (3-5 mCi) Tc-99m or 37 MBq
turer’s recommended acquisition protocol for acquir- (1 mCi) Tl-201
ing a transmission blank scan. When complete, visu- Acquisition Protocol
ally inspect planar images and check for artifacts (ie,
focal cold spots, bands of missing data, axial discon- 1. The cylinder is positioned with the long axis of the
tinuities). A common misconception is that the cylinder parallel to the table bed. Since processing
Journal of Nuclear Cardiology Nichols et al e33
Volume 13, Number 6;e25-41 Instrumentation quality assurance and performance

will involve summing slices, any tilt in the cylinder sources are weak, and it is recommended that either
should be minimized, as seen by the detector. the imaging times should be increased for the clinical
2. The vendor-recommended TCT-ECT acquisition pro- protocols or the transmission sources should be re-
tocol should be used. placed.
NOTE: The total acquisition time should not be less 2. Cross-talk correction is incorrect. While this correc-
than 12 minutes. tion is typically done automatically without user
interaction, it does have the potential for failure. To
Processing Protocol investigate if improper cross-talk correction is the
source of error, the measurement should be repeated
1. Emission cross-talk is removed from transmission
with no activity injected in the cylinder. This will
data (for most systems, this is done automatically and
necessitate refilling the tank with water (no activity)
does not need to be initiated by the user).
and repeating the protocol. If the new ROI value for
2. Attenuation maps are reconstructed from the trans-
the attenuation map without activity is within the
mission projection data. If a filter is applied to the
tabulated range, the service representative should be
reconstructed image data, the vendor-recommended
contacted to have the TCT-ECT system checked. The
filter should be used.
cross-talk estimate used to correct the transmission
3. Map slices from the center of the cylinder to provide
data is likely the problem, as the uncontaminated data
a single 5-cm-thick slice are summed.
are within the acceptable range and the cross-talk
4. A circular region of interest (ROI) is drawn centered
“compensated” values are not.
in the cylinder image that is approximately 90% of the
3. Bad blank scan: A new blank scan and a new cylinder
diameter of the cylinder.
phantom filled only with water should be acquired (no
5. The mean attenuation coefficient (␮) value in the ROI
activity injected into cylinder). If the new ROI value
is recorded. The following chart provides the ex-
is not within the tabulated range and your sources
pected range of ␮ values for the transmission isotope
have not expired, call the service representative. In
energy.
this case, there is likely a serious inconsistency
NOTE: Units for the attenuation coefficient (␮) may
problem between the transmission and blank scan
depend on the system and vendor.
data. If the new ROI value is within the acceptable
range, inject activity into the phantom and reacquire
the phantom.
Isotope Energy (keV) ␮ Expected Range
SPECT/TCT Protocol 2: Chest Phantom
Gd-153 100 0.160/cm–0.176/cm Required Equipment
Co-122 122 0.152/cm–0.168/cm An anthropomorphic chest phantom with a heart
Tc-99m 140 0.145/cm–0.161/cm insert (no defects in heart).
Barium 133 360 0.106/cm–0.117/cm
Injected Activity Concentrations
A simulated 1,110 MBq (30 mCi) sestamibi stress
6. Non-corrected (NC) and attenuation-corrected (AC)
study.
emission data are reconstructed. Slices are summed from
Heart: 250 kBq/mL (6.8 ␮Ci/mL)
the center of the cylinder to provide a 5-cm-thick slice.
Tissue: 25 kBq/mL (0.7 ␮Ci/mL)
The NC and AC images should be inspected together.
Liver: 150 kBq/mL (4.0 ␮Ci/mL)
The AC image should be more uniform than the uncor-
Lungs: 0 kBq/mL (0.0 ␮Ci/mL)
rected image. Due to the low activity levels injected into
the phantom, it is difficult to provide an acceptable Acquisition Protocol
quantitative range for an ROI.
1. The phantom is positioned on the imaging bed.
SPECT/TCT Protocol 1: Troubleshooting 2. The vendor-recommended TCT-ECT cardiac acquisi-
If the ROI value from step 5 falls outside of the tion protocol is used.
expected range, the system can potentially yield errone- NOTE: The total acquisition time should not be less
ous data. Possible sources of error are as follows: than 12 minutes.
1. Transmission sources are too weak. The acquisition Processing Protocol
should be repeated with longer scan duration (ie,
double the scan time). The processing steps should 1. Emission cross-talk is removed from transmission
be repeated and a new ROI value recorded for the data (for most systems, this is done automatically and
attenuation map. If the value improves, then the does not need to be initiated by the user).
e34 Nichols et al Journal of Nuclear Cardiology
Instrumentation quality assurance and performance November/December 2006

2. The attenuation maps are reconstructed from the source of error, the measurement should be re-
transmission projection data. peated with no activity. This will necessitate refill-
3. The transmission maps are visually inspected. Possi- ing the tank with water (no activity) and repeating
ble artifacts and their sources are identified. Obvious the protocol. Alternatively, the cylinder protocol
artifacts include image truncation (ring artifact on could be performed, as this protocol is ideally
periphery of imaging FOV) and cross-talk correction suited to determine cross-talk of blank scan prob-
errors (depressed pixel intensities in region of heart or lems. See troubleshooting comments 2 and 3 for the
liver). If the truncation artifact involves a significant cylinder protocol.
area of the imaging FOV, the phantom is repositioned
If the ROI ratio values from step 7 fall outside of the
and data are reacquired. If cross-talk errors are
accepted range, the system can potentially yield errone-
present, the troubleshooting section below should be
consulted. ous data. Possible sources of error are as follows:
4. Two small circular ROIs are drawn in the region of
the heart and the liver. The mean ␮ values are 1. Transmission sources are too weak. The acquisition
recorded for the heart and liver ROIs. The expected should be repeated with longer scan duration (ie,
ranges of ␮ values for the transmission isotope energy double the scan time). The processing steps should be
are presented above. repeated and results recorded with a new ROI value
NOTE: Units for the attenuation coefficient (␮) may for the attenuation map. If the value improves, the
depend on the system and vendor. sources are weak, and it is recommended that either
5. NC and AC emission data are reconstructed. NC and the clinical imaging time be increased or new trans-
AC polar maps are computed from the image data. mission sources be installed.
6. NC and AC short-axis and horizontal and vertical 2. Cross-talk correction is incorrect. While this correc-
long-axis images are reviewed in a comparative tion is typically done automatically without user
display. Visually, the AC images should be more interaction, it does have the potential for failure. To
uniform than the NC images. No region of the heart investigate if improper cross-talk correction is the
should be noticeably hotter than the rest. The apex of source of error, repeat the measurement with no
some phantom hearts may be cooler than the rest of activity. This will necessitate refilling the tank with
the phantom as the wall thickness of some phantom water (no activity) and repeating the protocol. If the
hearts does vary and can fall below the imaging new ROI value for the attenuation map without
resolution. In this case, the partial-volume effect can activity is within the tabulated range, the service
be attributed to the lower activity values. representative should be contacted to have the TCT-
7. The AC intensity values in the anterior, lateral, posterior, ECT system checked. If the new value is not within
and septal regions should be derived, using the ROI tool the tabulated range and your sources have not ex-
with either the polar maps or a mid–short-axis image, pired, the service representative should be called. If
and the results recorded. The anterior-posterior and the sources are near their expiration date, the sources
septal-lateral ratios should be 1.0 ⫾ 10%. should be replaced.

SPECT/TCT Protocol 2: Troubleshooting QC Procedures for X-Ray–Based SPECT/CT Systems

If the ROI value from step 5 falls outside of the All of the QC procedures required of SPECT
accepted range, the system can potentially yield errone- imaging are required for SPECT/CT imaging. Addi-
ous data. Possible sources of error are as follows: tionally, procedures specific to CT imaging are re-
1. Transmission sources are too weak. The acquisition quired and are listed below. In the current state of
should be repeated with a longer scan duration (ie, hybrid SPECT/CT and PET/CT development, the CT
double the scan time). The processing steps should be equipment is very similar or identical for these two
repeated and results recorded for a new ROI value for modalities. The tables in this section list the critical
the attenuation map. If the value improves, then the requirements.
transmission sources are weak, and it is recommended
that either the imaging times for clinical protocols CT QC procedures
should be increased or new transmission sources
should be installed. For information,
2. Cross-talk correction is incorrect. While this cor- Test Requirement see paragraph
rection is typically done automatically without user
interaction, it does have the potential for failure. To Calibration Mandatory 1
investigate if improper cross-talk correction is the Field uniformity Mandatory 2
Journal of Nuclear Cardiology Nichols et al e35
Volume 13, Number 6;e25-41 Instrumentation quality assurance and performance

1. Calibration. The reconstructed CT image must exhibit the two images must be in registration). Such registra-
accurate, absolute CT numbers (in Hounsfield units). tion is not easily achieved, because the SPECT and CT
This is critical for the use of CT images for SPECT portions of commercial combined SPECT/CT systems
attenuation correction, because the quantitative CT val- may not be coincident (ie, the SPECT and CT “slices”
ues are transformed (usually via a bilinear or trilinear are not in the same plane) and because SPECT and CT
function with one hinge at or near the CT value for gantries may be contiguous. In practice, this means that
water) to attenuation coefficients at 511 keV. Any errors the SPECT and CT acquisitions may not simultaneously
in CT numbers will be propagated as errors in estimated image the same slice. For contiguous gantries, the bed
radionuclide attenuation coefficients, which in turn will must travel different distances into the gantry to image
adversely affect the attenuation-corrected SPECT val- the same slice in the patient for SPECT versus CT,
ues. CT system calibration is performed with a special providing ample opportunity for misregistration, via x,
calibration phantom that includes inserts of known CT y, z misalignment of bed motion or (of perhaps even
numbers. This calibration is done by the manufacturer’s greater concern) because of differential “bed sag” for the
field service engineers. The CT calibration is then SPECT and CT portions. In addition, electronic drift can
checked daily with a water-filled cylinder (usually 24 influence the “position” of each image so that calibra-
cm in diameter; usually provided by the manufacturer). tions for mechanical registration can become inaccurate
In practice, if the error is greater than 5 Hounsfield units over time. Thus, it is imperative to check SPECT-to-CT
(ie, different than the anticipated value of 0 Hounsfield registration on an ongoing basis. This can be performed
units), the CT system is considered out of calibration. with a specific phantom or jig containing an array of
The technologist will usually then do an air calibration, point sources visible in both SPECT and CT. Errors in
to determine if this corrects the overall calibration (ie, co-location in the fused SPECT-CT images are assessed,
brings the CT number for water back to within 5 such as by means of count profiles generated across
Hounsfield units of 0). If it does not, the manufacturer’s transaxial slices. Such errors (after software registration
field service engineer must be called. On an annual corrections) should be not greater than the spatial reso-
basis, or after any major repair or calibration, calibration lution of the SPECT scan. It is important to image this
is checked by the manufacturer’s field service engineer registration jig in a number of positions along the bed. It
with a special phantom that includes plastic inserts that may also be helpful to place a weight on the end of the
simulate bone, muscle, fat, and water. bed (to simulate some sagging of the imaging bed) and
2. Field uniformity. The reconstructed CT image must repeat the assessment.
exhibit uniform response throughout the FOV. In 2. Attenuation correction accuracy. The use of the CT
practice, this means that a reconstructed image of a image for SPECT attenuation correction requires a
uniform water-filled cylinder must itself demonstrate transformation of the observed CT numbers (in
low variation in CT number throughout this image. In Hounsfield units) to attenuation coefficients at the
practice, small circular ROIs are placed at the four energies of the radionuclides. This transformation is
corners of the cylinder image, and the mean CT usually accomplished with a bilinear calibration
number is compared to that from a region in the center
curve, “hinged” at a CT value of 0 (ie, hinged at the
of the phantom; the difference in mean region CT
CT value for water). At a minimum, it is important to
number should not exceed 5 Hounsfield units. Non-
image a water-filled cylinder to assess SPECT field
uniformities greater than this may produce sufficient
uniformity and SPECT activity concentrations after
quantitative inaccuracies as to affect SPECT attenua-
CT-based SPECT attenuation correction. Errors in
tion correction based on the CT image.
CT-to-SPECT attenuation transformations are usually
manifested as a corrected SPECT image without a
Combined SPECT/CT QC procedures “flat” profile from edge to center (ie, the activity at the
edge is either too high or too low relative to that at the
For information, center of the phantom) and with resulting attenuation-
Test Requirement see paragraph corrected absolute SPECT values that are incorrect
(although these values depend on absolute SPECT
Registration Mandatory 1 scanner calibration as well as accurate CT-based
Attenuation Mandatory 2 SPECT attenuation correction). Moreover, if avail-
correction able, more sophisticated phantoms with variable at-
accuracy
tenuation (and variable activity distributions) can be
1. Registration. The reconstructed SPECT and CT im- used to more comprehensively assess any errors in
ages must accurately reflect the same 3D locations (ie, CT-based SPECT attenuation correction.
e36 Nichols et al Journal of Nuclear Cardiology
Instrumentation quality assurance and performance November/December 2006

PET imaging QC. The following table lists recom- tional procedures may be required by particular manu-
mended PET imaging QC schedules. Note that, unlike facturers (see paragraph 6 below). The procedures below
planar and SPECT imaging, there are no widely ac- should be suitable for ensuring overall proper basic
cepted, published QC procedures for PET. Some addi- operation of a PET scanner.

Suggested QC procedures: Dedicated PET imaging devices

For information,
Procedure Frequency see paragraph

Acceptance testing as per NEMA NU 2-2001 Once upon delivery and upon 1
major hardware upgrades
Sensitivity and overall system performance Daily (or at least weekly) 2
Transverse resolution Annually 3
Accuracy (corrections for count losses and randoms) Annually 4
Scatter fraction Annually 4
Accuracy of attenuation correction Annually 5
Image quality Annually 5
Measurements specified by the Manufacturer As per the manufacturer 6

1. Acceptance testing. It is recommended that the against which to track changes that may occur over
NEMA performance measurements (NU 2-2001) be time.
made before accepting the PET scanner.17 Many of 2. Sensitivity. Subtle changes in PET system sensitivity
these tests can be performed by the company supply- may occur slowly over time. More dramatic changes
ing the PET scanner. If so, it is recommended that in sensitivity may occur that are associated with
the purchaser’s representative work with the man- hardware or software malfunction. There are simple
ufacturer’s representatives during these tests, as the tests designed to monitor such changes in sensitivity,
manufacturer may not perform them as specified by as well as to check proper machine operation prior to
NEMA recommendations. These recommendations each day’s patient scans. Ideally, these tests should be
have been evolving in recent years, and the more performed daily, but not less frequently than weekly.
recent recommendations (NU 2-2001) have super- One recommended procedure to measure system sen-
seded older recommendations (NU 2-1994),18 pri- sitivity is the following:
marily to accommodate PET scanners that are a. A 20-cm-diameter ⫻ 20-cm-long (or longer)
operated in 3D mode.19 However, these newer water-filled phantom can be used with a uniform
recommendations also better reflect imaging of concentration of a water-soluble fluorine 18
objects of the size of a typical adult thorax region compound (eg, F-18 fluorodeoxyglucose) (eg,
by incorporating measurements of the International 1-2 mCi for a 2D scanner or a factor of 5 less for
Electrotechnical Commission (IEC) body phan- a 3D scanner). Alternatively, a uniform cylindri-
toms.20 Aside from this phantom, however, if a cal phantom of approximately the same size can
facility intends to operate the PET scanner only in be used containing Ge-68/Ga-68 (available as a
2D mode, the older measurements in the 1994 solid). The phantom is placed on the imaging
standard, which are somewhat easier to perform, table with the long axis of the cylinder parallel
may be satisfactory. There are some differences to the long axis of the scanning bed, centered in
between these NEMA standards (both 1994 and the transaxial field of the scanner. The phantom
2001) and the IEC standards. should be aligned as well as possible, with the
There are two reasons for performing these assistance of a laser-positioning device if avail-
measurements: (1) to ensure that the new PET able.
scanner meets specifications published by the man- b. Data are acquired for at least 15 minutes, which
ufacturer and (2) to provide a standard set of should be sufficient for a 2D or 3D mode scanner
measurements that allows the user to document the with the activities specified above, with attenua-
limitations of the scanner, providing a standard tion correction, using whichever mode of attenua-
Journal of Nuclear Cardiology Nichols et al e37
Volume 13, Number 6;e25-41 Instrumentation quality assurance and performance

tion correction is being used clinically in the performance in the presence of non-uniform attenu-
facility (eg, CT, rotating Ge-68 rod, or rotating ating substances.
Cs-137 rod attenuation corrections). 6. Variations among manufacturers. The above recom-
c. Phantom images should be reconstructed using a mendations regarding PET scanner QA are general
typical clinical protocol. guidelines. Each manufacturer has its own periodic QC
d. A circular ROI should be drawn on the central part recommendations for parameters such as “singles” sen-
of the transaxial slice, of a diameter on the order of sitivity, coincidence timing, energy calibration, and
15 cm. For each slice, the mean measured concen- overall system performance. These, by necessity, require
tration (in units of meqabecquerels per milliliter) very different measurement protocols that may vary
should be computed and divided by input concen- even between models for the same manufacturer. These
tration, which is known from the measured activity measurements must be performed as detailed by the
and phantom volume for an F-18 phantom or from manufacturer. However, the measurements specified
the specification sheet for a Ge-68 phantom. Val- above are not intended to replace these basic system-
ues across slices should be within approximately specific QC measurements.
5% to 10% of each other. The mean sensitivity PET/CT imaging QC. The PET portion of the
value should be computed over all slices (but combined system should be assessed as described above.
excluding the first and last 2-3 slices). These The CT portion should likewise be assessed via standard
values should be within 10% of unity and should CT QC procedures. With respect to its use in a combined
change only slowly over time. PET/CT system, the CT portion should undergo the
e. Each slice, as well as the sum of all slices but following tests:
excluding the first and last 2 to 3 slices, should be
inspected for artifacts. The count profile across the CT QC procedures
summed slice should be generated and examined,
which should be uniform to within the tolerance For information,
established during the initial acceptance testing of Test Requirement see paragraph
the PET scanner.
f. Note that if F-18 is used rather than Ge-68/Ga-68, Calibration Mandatory 1
the variation in measured mean values (meqabec- Field uniformity Mandatory 2
querels per milliliter/known meqabecquerels per
milliliter) will depend on both variations in the 1. Calibration. The reconstructed CT image must ex-
scanner and variations in the device used to mea- hibit accurate, absolute CT numbers (in Hounsfield
sure the injected activity. units). This is critical for the use of CT images for
PET attenuation correction, because the quantitative
3. Transverse resolution. This is measured using either CT values are transformed (usually via a bilinear or
the NEMA NU 2-2001 recommendations for a point trilinear function with one hinge at or near the CT
source or NEMA NU 2-1994 recommendations for a value for water) to attenuation coefficients at 511
rod source. keV. Any errors in CT numbers will be propagated as
4. Scatter fraction. This typically is measured accord- errors in estimated 511-keV attenuation coefficients,
ing to either NEMA NU 2-2001 (section 6) or NEMA which in turn will adversely affect the attenuation-
NU 2-1994 specifications, although it is recom- corrected PET values. CT system calibration is per-
mended that if the scanner is to be used in 3D mode, formed with a special calibration phantom that in-
NU 2-2001 is preferred. cludes inserts of known CT numbers. This calibration
5. Accuracy of attenuation correction and overall is done by the manufacturer’s field service engineers.
clinical image quality. Attenuation correction should The CT calibration is then checked daily with a
be assessed using the IEC phantom, as specified in the water-filled cylinder (usually 24 cm in diameter;
NU 2-2001 recommendations.17 If this phantom is not usually provided by the manufacturer). In practice, if
readily available, it is suggested that similar measure- the error is greater than 5 Hounsfield units (ie,
ments can be performed with a phantom approximat- different than the anticipated value of 0 Hounsfield
ing a typical human body shape and size (eg, a 20 ⫻ units), the CT system is considered out of calibration.
30 – cm elliptical phantom or anthropomorphic phan- The technologist will usually then do an air calibra-
tom) with at least one cold sphere or cylinder and one tion, to determine if this corrects the overall calibra-
hot sphere or cylinder, as well as at least some tion (ie, brings the CT number for water back to
material simulating lung tissue, to ensure proper within 5 Hounsfield units of 0). If it does not, the
e38 Nichols et al Journal of Nuclear Cardiology
Instrumentation quality assurance and performance November/December 2006

manufacturer’s field service engineer must be called. PET-to-CT registration on an ongoing basis. This is
On an annual basis, or after any major repair or usually performed with a specific phantom or jig
calibration, calibration is checked by the manufac- containing an array of point sources visible in both
turer’s field service engineer with a special phantom PET and CT. Errors in co-location in the fused
that includes plastic inserts that simulate bone, mus- PET-CT images are assessed, such as by means of
cle, fat, and water. count profiles generated across transaxial slices. Such
2. Field uniformity. The reconstructed CT image must errors (after software registration corrections) should
exhibit uniform response throughout the FOV. In be less than 1 mm. It is important to image this
practice, this means that a reconstructed image of a registration jig in a number of positions along the bed.
uniform water-filled cylinder must itself demonstrate It may also be helpful to place a weight on the end of
low variation in CT number throughout this image. In the bed (to produce some bed sag) and repeat the
practice, small circular ROIs are placed at the four assessment.
corners of the cylinder image, and the mean CT 2. Attenuation correction accuracy. The use of the CT
number is compared to that from a region in the center image for PET attenuation correction requires a
of the phantom; the difference in mean region CT transformation of the observed CT numbers (in
number should not exceed 5 Hounsfield units. Non- Hounsfield units) to attenuation coefficients at 511
uniformities greater than this may produce sufficient keV. This transformation is usually accomplished
quantitative inaccuracies so as to affect PET attenua- with a bilinear calibration curve, “hinged” at a CT
tion correction based on the CT image. value of 0 (ie, hinged at the CT value for water). At
a minimum, it is important to image a water-filled
In addition to the (independent) QC tests for the PET
cylinder to assess PET field uniformity and PET
and CT portions of the combined system, it is necessary
activity concentrations after CT-based PET attenu-
to perform additional tests that assess the combined use
ation correction. Errors in CT-to-PET attenuation
of PET and CT.
transformations are usually manifest as a corrected
PET image without a “flat” profile from edge to
Combined PET/CT QC procedures center (ie, the activity at the edge is either too high
or too low relative to that at the center of the
For information, phantom) and with resulting attenuation-corrected
Test Requirement see paragraph absolute PET values that are incorrect (although
these values depend on absolute PET scanner
Registration Mandatory 1 calibration as well as accurate CT-based PET
Attenuation Mandatory 2 attenuation correction). If possible, the CT-based
correction attenuation-corrected PET values should be com-
accuracy
pared with those from the rotating rod source–
based attenuation-corrected PET values in the same
1. Registration. The reconstructed PET and CT images phantom. Moreover, if available, more sophisti-
must accurately reflect the same 3D locations (ie, the cated phantoms with variable attenuation (and vari-
two images must be in registration). Such registration able activity distributions) can be used to more
is not as easily achieved, because the PET and CT comprehensively assess any errors in CT-based
portions of all commercial combined PET/CT sys- PET attenuation correction.
tems are not coincident (ie, the PET and CT “slices”
Clinical QA for Each Patient Procedure
are not in the same plane) and because the PET and
Whereas the recommended QA procedures de-
CT gantries are contiguous. In practice, this means
scribed in this document will reduce the frequency of
that the PET and CT acquisitions do not simulta-
SPECT and PET misdiagnosis due to technical problems,
neously image the same slice. In fact, because the bed
ongoing clinical QA for each patient procedure is also
must travel different distances into the gantry to
needed in order to obtain optimal imaging results. For
image the same slice in the patient for PET versus CT,
ongoing clinical PET QA recommendations, please see
there is ample opportunity for misregistration, via x,
the section of the guidelines entitled “Positron Emission
y, z misalignment of bed motion or (of perhaps even
Tomography Myocardial Perfusion and Glucose Metab-
greater concern) because of differential “bed sag” for
olism Imaging.” Recommendations for ongoing clinical
the PET and CT portions. In addition, electronic drift
SPECT QA are as follows:
can influence the “position” of each image, so that
calibrations for mechanical registration can become 1. Patient instructions. The imaging procedure should
inaccurate over time. Thus, it is imperative to check be explained to the patient before acquisition.
Journal of Nuclear Cardiology Nichols et al e39
Volume 13, Number 6;e25-41 Instrumentation quality assurance and performance

2. Patient positioning. When using 180° orbits for camera, if there is motion in the first frames but no
SPECT, the patient’s left arm must be positioned motion for the rest of the study, this will have little
away from his or her side. This is usually accom- effect. However, with multi-detector systems, the
plished by placing the arm above the head using an single episode of motion will be propagated through-
arm support device to maximize comfort. With some out other frames of the 180° arc. On the other hand,
arm support systems, it is more comfortable to raise a 1-pixel shift (vertical translation) over as few as 4
both arms into the support system. When using 360° frames in the middle of the projection set can
orbits, both arms need to be positioned away from produce a clinically significant artifactual defect.
the patient’s side. Also, it is important that patients Rotational angular motion is less frequently detected
be positioned similarly for additional studies includ- since it appears parallel to the motion of the heart in
ing consistent radius of rotation, such as for both the the projection images. Motion toward or away from
rest and stress or immediate and delayed portions of a detector cannot be detected. A horizontal line of
a study. reference drawn on the screen or in one’s mind can
3. Motion detection/correction. The patient should be be used as the reference point for evaluation of the
observed throughout the acquisition by a technolo- magnitude of motion. If motion is detected, an
gist to ensure that patient motion does not occur. assessment of its severity must be made. A rule of
Talking, head movement, and irregular breathing thumb is that translational motion (eg, the patient
should be avoided during the acquisition. Tomo- moves and remains there throughout the study) on
graphic data should be reviewed for patient motion the order of 1 pixel in a 64 ⫻ 64 matrix represents an
immediately following completion of an imaging approximate limit before significant artifacts can
session by the technologist who acquired the data. result. Motion of one-half pixel can often be de-
Acquisitions with motion should be corrected or tected but generally does not result in clinically
reacquired before reconstruction. Review of the significant artifacts and represents the upper limit of
originally acquired projection data as an endless cine motion detectable with automated methods. Trans-
loop is perhaps the single most useful mechanism lational or “drifting” motion along the patient axis in
available to physicians to detect patient motion, to dual- and multi-detector systems yields unique
assess the adequacy of motion correction, and to “jumps” in the position of the myocardium where
assess the integrity of the collected data, not only for the detector views are joined. The same criteria for
patient motion but also for a wide variety of imaging significance of motion for multiple-detector systems
artifacts.21 Critical to the accurate reconstruction of should be applied as those described above for
the tracer distribution is the fixed alignment of the single-detector systems. Once motion is detected,
detector coordinates and the organ being imaged. manual methods (if available) can be used to shift
The most common source of misalignment is the the frames back to the expected correct position.
motion of the patient or heart relative to the detector Manufacturer-specific automated methods can be
coordinates during acquisition. There is some phys- executed that track the myocardium across the FOV
iologic cardiac motion over the cardiac cycle, which and perform an approximate realignment.
is unavoidable and accepted as a known source of 4. Detection of inappropriate energy settings. Occa-
image degradation. Otherwise, heart motion can sionally, data may not be acquired with the correct
result from the patient moving or exaggerated dia- energy setting, such as when data are acquired on a
phragmatic excursion that can occur with heavy or Tc-99m setting for a patient who has been injected
erratic breathing. With patient or heart motion, count with Tl-201. This results in loss of counts, resolu-
values are erroneously placed back into the tomo- tion, and contrast. One way to recognize this prob-
graphic image, resulting in potential artifacts that lem after the fact is to compare information in the
can corrupt the accuracy of tracer representation. header part of the DICOM (Digital Imaging and
Therefore, it is critical that the rotating projection Communications in Medicine) images, which now
views be played in a cine format and evaluated for have become the standard image storage format,
motion. If significant motion is detected, a repeat regarding the energy setting of the camera versus the
study usually is required. Translational motion along protocol used for patient preparation.22
the patient axis (“up” and “down” motion) is the 5. Image count rate. Manufacturers have various
most frequently and most easily detected type of means of making this count rate information avail-
motion since it is perpendicular to the movement of able to the user. Some typical count rates for
the heart in the rotating cine views. Tomographic individual projection images are provided in the
image degradation depends on the frame in which table below, representing ranges of counts from
motion occurs. For instance, with a single-detector acceptable SPECT studies. These values represent
e40 Nichols et al Journal of Nuclear Cardiology
Instrumentation quality assurance and performance November/December 2006

average left anterior oblique counts in a planar 8. Arrhythmia detection. Equipment manufacturers
projection over the left ventricular (LV) myocar- vary considerably as to R-R information retained for
dium, using image acquisition protocols described subsequent display. If available, the QA screen that
above. displays the histogram of the patient’s R-wave
triggers received by the computer during data acqui-
Expected count ranges in left anterior sition for each patient should be reviewed. Some
oblique projection algorithms are available to analyze gated SPECT
data after the fact and report patterns of arrhythmias
Whole heart Maximum if detected from anomalies in curves of counts
(ⴛ1000) counts plotted versus projection angles from among the
different time bins used to acquire data.23
Tl-201 (111 MBq/LEHR/ 10–28 85–134 9. Image reconstruction QA. It is important to exam-
64 stops) ine the standard QA screens provided by most
Tc-99m (814 21–77 190–298 manufacturers to review the axes chosen for reori-
MBq/LEHR/64 stops) enting gated and ungated SPECT and PET data.
Tl-201 (111 MBq/LEAP/ 53–134 91–235 Those algorithms expect conventional input data to
32 stops)STRESS be oriented similarly for the stress and rest portions
Tl-201 (111 MBq/LEAP/ 33–100 66–185 of the scan. Some unusual clinical circumstances
32 stops)REST may confound reproducible axis selection for stress
LEHR, Low-energy high-resolution collimator; LEAP, low-energy, and rest tomographic reconstruction. In cases of
all-purpose or general-purpose collimator. severe transient ischemic dilation, the LV shape may
be genuinely distorted during stress compared to
6. Attenuation artifact detection/correction. Ideally, rest. More often, severe extensive perfusion defects
the use of attenuation-correcting hardware/software seen on stress that resolve at rest are the source of
will eliminate most attenuation artifacts encountered misalignments of one data set relative to the other.
clinically, including left breast and elevated dia- Some algorithms have been produced that attempt to
phragm or protuberant abdomen. However, users match data sets 3-dimensionally, but implementation
should not expect attenuation correction to totally of these will still require the user to examine the
correct patient attenuation, particularly in patients choices of reorientation axes for appropriateness.24
with unusually large body habitus. There can be 10. Polar and 3D myocardial perfusion maps. It is
more troublesome artifacts that are more difficult to important to review the QA screens supplied by the
correct, such as those due to metallic objects (eg, algorithms’ manufacturers to ensure that perfusion
metallic electrodes) interposed between the patient data have been sampled appropriately from LV apex
and the detectors. Some CT systems are sensitive to to base and to ensure that noncardiac activity (eg,
such objects; correcting for these may sometimes not from the spleen, bowel, and so on) has been ex-
be possible. However, review of the original rotating cluded. Two-dimensional and 3D perfusion maps
projection data should reveal a distinct shadow should convey the same impression regarding rela-
traversing the detector’s FOV at the location of the tive perfusion as is derived from the examination of
foreign metal object. short-axis tomographic sections.
7. COR error detection/correction. COR errors result 11. Functional computation QA. Review of manufac-
in a distinct “comet” pattern artifact on myocardial turer-supplied QA screens of segmented LV loca-
perfusion SPECT data when acquired from the right tions for all time bins for algorithms that compute
anterior oblique– 45° to left posterior oblique– 45° LV ejection fraction and volumes must be reviewed
projections. Unlike motion artifacts, these artifacts to verify appropriateness of these calculations.25 It is
are highly predictable.9 Re-acquisition of a point important to verify that computed endocardial bor-
source tomogram should provide a definitive answer ders closely match actual endocardial locations. This
as to whether the usual filtered backprojection re- can be assessed by visual examination of cinematic
construction process of data just collected has been displays of computed outlines superimposed on
compromised due to application of incorrect COR long-axis and short-axis tomographic sections.
data. If so, and if the manufacturer has provided the 12. Computer monitor QA. In addition to viewing QA
means to do so, it is often possible to correct such screens related to computations, it is important to
data after the fact by updating the patient’s data file include monitor QA on a daily or weekly basis. For
with the newly updated COR information and recon- example, if the monitor contrast is too high, perfu-
structing the data again. sion defects as well as artifacts will be exaggerated.
Journal of Nuclear Cardiology Nichols et al e41
Volume 13, Number 6;e25-41 Instrumentation quality assurance and performance

Typically, this is performed by the physicians visu- 13. Galt JR, Faber T. Principles of single photon emission computed
ally reviewing nuclear cardiology images by means tomography (SPECT) imaging. In: Christian PE, Bernier DR,
Langan JK, editors. Nuclear medicine and PET: technology and
of computer-generated high-resolution, high- techniques. St Louis: Mosby; 2003. p. 242-84.
contrast images that include a complete gray scale. 14. Cerqueira MD, Matsuoka D, Ritchie JL, Harp GD. The influence
In addition, standard film processor QA is a standard of collimators on SPECT center of rotation measurements: artifact
component of routine QA procedures performed by generation and acceptance testing. J Nucl Med 1988;29:1393-7.
all imaging laboratories. 15. Hines H, Kayayan R, Colsher J, Hashimoto D, Shubert R,
Fernando J, et al. National Electrical Manufacturers Association
recommendation for implementing SPECT instrumentation quality
References control. J Nucl Med 2000;41:383-9.
16. Greer KL, Jaszczak RJ, Coleman RE. An overview of a camera-
1. Bushberg JT, Seibert JA, Leidholt EM, Boone JM. The essential based SPECT system. Med Phys 1982;9:455-63.
physics of medical imaging. Baltimore: Williams & Wilkins; 1994. 17. National Electrical Manufacturers Association. NEMA standards
2. Cherry SR, Sorenson JA, Phelps ME. Physics in nuclear medicine. publication NU 2-2001: performance measurements of positron
Philadelphia: Saunders; 2003. emission tomographs. Washington, DC: National Electrical Man-
3. Shih W-J, Wierzbinski B. Cardiac SPECT: 360° circular acquisi- ufacturers Association; 2001.
tion may resolve defects of 180° data. J Nucl Med 2003;44:995-6. 18. National Electrical Manufacturers Association. NEMA standards
4. Liu Y-H, Lam PT, Sinusas AJ, Wackers FJT. Differential effect of publication NU 2-1994: performance measurements of positron
180° and 360° acquisition orbits on the accuracy of SPECT
emission tomographs. Washington, DC: National Electrical Man-
imaging: quantitative evaluation in phantoms. J Nucl Med 2002;
ufacturers Association; 1994.
43:1115-24.
19. Daube-Witherspoon ME, Karp JS, Casey ME, DiFilippo FP, Hines
5. Cao Z, Maunoury C, Chen CC, Holder LE. Comparison of
H, Muehllehner G, et al. PET performance measurements using the
continuous step-and-shoot versus step-and-shoot acquisition
NEMA NU 2-2001 standard. J Nucl Med 2002;43:1398-409.
SPECT. J Nucl Med 1996;37:2037-40.
20. International Electrotechnical Commission. Radionuclide imaging
6. American College of Radiology Web site. Available from: URL:
devices— characteristics and test conditions. Part 1: positron emis-
http://www.acr.org/s_acr/sec.asp?CID⫽1073&DID⫽1487. Accessed
sion tomographs. Geneva: International Electrotechnical Commis-
October 6, 2006.
sion; 1998.
7. Intersocietal Commission for the Accreditation of Nuclear Medi-
cine Laboratories Web site. Available from: URL: http://www. 21. DePuey EG. Artifact is SPECT myocardial perfusion imaging. In:
icanl.org. Accessed October 6, 2006. DePuey EG, Berman DS, Garcia EV, editors. Cardiac SPECT
8. American Association of Physicists in Medicine SPECT Task imaging. 2nd ed. Philadelphia: Lippincott Williams & Wilkins;
Group. Rotating scintillation camera SPECT acceptance testing 2001. p. 231-62.
and quality control. AAPM Report No. 22. New York: American 22. National Electrical Manufacturers Association. The DICOM stan-
Association of Physicists in Medicine; 1987. dard. Available from: URL: http://medical.nema.org/dicom/2004.
9. Nichols KJ, Galt JR. Quality control for SPECT imaging. In: html. Accessed October 6, 2006.
DePuey EG, Berman DS, Garcia EV, editors. Cardiac SPECT 23. Nichols K, Yao SS, Kamran M, Faber TL, Cooke CD, DePuey EG.
imaging. 2nd ed. Philadelphia: Lippincott Williams & Wilkins; Clinical impact of arrhythmias on gated SPECT cardiac myocar-
2001. p. 17-40. dial perfusion and function assessment. J Nucl Cardiol 2001;8:19-
10. National Electrical Manufacturers Association. NEMA standards 30.
publication NU 1-1994: performance measurements of scintillation 24. Slomka PJ, Nishina H, Berman DS, Akincioglu C, Abidov A,
cameras. Washington, DC: National Electrical Manufacturers As- Friedman JD, et al. Automated quantification of myocardial
sociation; 1994. perfusion SPECT using simplified normal limits. J Nucl Cardiol
11. Esser PD, Graham LS. A quality control program for nuclear 2005;12:66-77.
medicine cameras. In: Henkin RE, et al. Nuclear medicine. 2nd ed. 25. Germano G, Nichols KJ, Cullom SJ, Faber TL, Cooke CD. Gated
Philadelphia: Mosby; 2006. p. 246-56. perfusion SPECT: technical considerations. In: DePuey EG, Ber-
12. DePuey EG. How to detect and avoid myocardial perfusion man DS, Garcia EV, editors. Cardiac SPECT imaging. 2nd ed.
SPECT artifacts. J Nucl Med 1994;35:699-702. Philadelphia: Lippincott Williams & Wilkins; 2001. p. 103-15.
 First-pass radionuclide angiography
John D. Friedman, MD,a Daniel S. Berman, MD,b Salvador Borges-Neto, MD,b
Sean W. Hayes, MD,b Lynne L. Johnson, MD,b Kenneth J. Nichols, PhD,b
Robert A. Pagnanelli, CNMT, RT(R)(T), NCT,b and Steven C. Port, MDb

3. Because of bolus prolongation and fractionation, no

peripheral sites other than the antecubital (preferably
medial) and external jugular veins are suitable for
ACQUISITION PARAMETERS
FPRNA. The study should not be attempted if those
Purpose. To assess left ventricular (LV) and right sites are not available.
ventricular (RV) function at rest or during stress (evalu- 4. Some users prefer larger-bore cannulae in the 14- to
ation of wall motion, ejection fraction [EF], and other 16-gauge range, but they are optional and not highly
systolic and diastolic parameters). To assess and measure recommended because of the increased trauma.
left-to-right shunts. The cannula should be directly connected to a
suitable length of intravenous tubing, preferably 12
to 20 inches. The free end of the tubing should be
ACQUISITION PROTOCOLS attached to a three-way stopcock with a sufficiently
large bore to accommodate rapid injections. All
First-pass radionuclide angiography–LV function intravenous connections should be lock-type rather
than slip, to avoid contamination. A 10- to 20-mL
1. Technetium 99m diethylamine triamine pentaacetic saline bolus is used to flush the radionuclide bolus
acid (DTPA) is the radionuclide of choice for standard into the venous system. The saline bolus should be
first-pass radionuclide angiography (FPRNA) because injected at a continuous, uninterrupted rate so that
of its renal excretion, minimizing patient radiation the entire 10 to 20 mL is injected in 2 to 3 seconds.
exposure. Tc-99m pertechnetate may also be used. 5. The injection for LV studies must be rapid. The full
Other technetium-based compounds, such as the tech- width at half maximum (FWHM) of the bolus transit
netium perfusion agents sestamibi and tetrofosmin, are in the superior vena cava should be less than 1
suitable. The short-lived radionuclides gold 195m and second and, if possible, less than 0.5 second. That
iridium 191m have been used for FPRNA but are not will virtually guarantee a technically adequate
currently approved by the Food and Drug Administra- study, all other variables being equal.
tion. 6. The most common position used is the upright,
2. The standard dose given is 25 mCi for both rest and straight anterior view. Its advantages are the ease
exercise studies. A dose as low as 10 mCi may be used with which the chest may be stabilized against the
as an option in rest studies for multicrystal cameras. detector and the straightforward approach to posi-
Since the study is count dependent and the single tioning the patient so that the left ventricle will be in
crystal cameras have limited count rate capabilities, the field of view (FOV). A transmission source or a
doses of 20 to 25 mCi are typically recommended. 1-mCi tracer dose is recommended to ensure proper
When predicated by dosimetry considerations, lower positioning. The chief disadvantage of the anterior
doses may have to be used. Doses as low as 10 mCi view is the anatomic overlap that may occur with the
may be used with reasonable clinical success but run descending aorta and the basal portion of the infero-
the risk of inadequate count rates especially for wall septal wall and with the left atrium and the basal
motion analysis. A rule that should be applied to test if portion of the left ventricle. The shallow right
enough counts have been acquired for a diagnostic anterior oblique (RAO) view helps eliminate both of
clinical study is that the end-diastolic frame of the those sources of overlap but is more difficult to
representative cycle should have more than 2,500 standardize. A foam cushion cut at a 30° angle may
counts in the LV region of interest (ROI). be used for such positioning. The choice of upright

Chair.a Member.b
J Nucl Cardiol 2006;13:e42-55.
1071-3581/$32.00
Copyright © 2006 by the American Society of Nuclear Cardiology.
doi:10.1016/j.nuclcard.2006.08.006

e42
Journal of Nuclear Cardiology Friedman et al e43
Volume 13, Number 6;e42-55 First-pass radionuclide angiography

Table 1. Acquisition parameters

For information,
Rest Exercise see paragraph

Radiopharmaceutical dose Tc-99m Tc-99m Standard 1, 2, 13, 19

Multicrystal (LV and RV) 10–25 mCi 25 mCi Standard
Single crystal (LV and RV) 20–25 mCi/0.3–0.5 mL 20–25 mCi/0.3–0.5 mL Standard
Shunt 10–15 mCi/0.3–0.5 mL Standard
Injection site 3, 14
RV function Antecubital vein Antecubital vein Standard
External jugular External jugular Optional
LV function Antecubital vein Antecubital vein Standard
External jugular External jugular Optional
Shunt Antecubital vein Preferred
External jugular Optional
Intravenous cannula 18-gauge arm 18-gauge arm Standard 4
14-, 16-, or 20-gauge 14-, 16-, or 20-gauge Optional
neck neck
Injection rate 5, 15, 16, 20
RV function Slow (FWHM 2–3 s) Slow (FWHM 2–3 s) Preferred
LV function Rapid (FWHM ⬍1 s) Rapid (FWHM ⬍1 s) Standard
Shunt Rapid (FWHM ⬍1 s) Standard
Position 6, 17, 21
RV function Upright Upright Standard
Supine Supine Optional
20°–30° RAO 20°–30° RAO Preferred
Anterior Anterior Optional
LV function Upright Upright Preferred
Supine Supine Optional
Anterior Anterior Standard
RAO RAO Optional
Shunt Anterior Standard
ECG signal 7, 22
RV and LV function Multicrystal no Multicrystal no Standard
Single crystal yes Single crystal yes Standard
Shunt No Standard
Energy window frame time 120–160 keV 120–160 keV Standard 8, 9, 23
LV and RV function 25 ms 25 ms Standard
50 ms 50 ms Optional
Shunt 50 ms Standard
100 ms Optional
Total frames 10, 24
LV and RV function 2000 1500–2000 Standard
Shunt 2000 Standard
Matrix: Multicrystal 11, 25
RV and LV function 20 ⫻ 20 20 ⫻ 20 Standard
14 ⫻ 20 14 ⫻ 20 Optional
Shunt 20 ⫻ 20 or 14 ⫻ 20 Standard
Matrix: Single crystal 11, 25
RV function 64 ⫻ 64 64 ⫻ 64 Preferred
32 ⫻ 32 32 ⫻ 32 Optional
LV function 32 ⫻ 32 32 ⫻ 32 Preferred
64 ⫻ 64 64 ⫻ 64 Optional
Shunt 32 ⫻ 32 or 64 ⫻ 64 Standard
e44 Friedman et al Journal of Nuclear Cardiology
First-pass radionuclide angiography November/December 2006

Table 1. (Continued)
For information,
Rest Exercise see paragraph

Collimator: Single crystal 12, 18, 26

LV and RV function and shunt High sensitivity High sensitivity Standard
LV function Ultrahigh sensitivity Ultrahigh sensitivity Preferred
RV function Ultrahigh sensitivity Ultrahigh sensitivity Optional
Shunt Diverging Optional
Collimator: Multicrystal 12, 18, 26
(SIM 400) (RV, LV, and shunt) 18 mm 18 mm Standard
13 mm to 10 mCi 13 mm to 10 mCi Optional
27 mm to 25 mCi 27 mm to 25 mCi Optional
System 77 1 inch 1 inch Standard

versus supine positioning depends, to some degree, 8. The 140-keV photopeak of Tc-99m ⫾ 15% (140 ⫾ 21
on the clinical situation. The upright position is, in keV) is fairly standard and results in adequate count
general, preferred. Pulmonary background is re- rates. This corresponds very closely to the 120- to
duced in the upright position, which enhances study 160-keV window used. The window may be wid-
quality. Positioning of the patient during treadmill ened to ⫾ 30% for low-dose injections.
exercise is a critical issue since the FOV of the 9. Theoretically, the frame time should be varied to suit
detectors is small. It is suggested that a person the heart rate at the time of acquisition. The rela-
should be behind the patient during peak stress for tionship is fairly linear, with 50 milliseconds being
proper positioning of the chest in relation to the quite adequate at heart rates of less than 80
detector. This is of crucial importance so that the beats/min and 25 milliseconds for heart rates of
heart will be within the FOV during acquisition. A 125 to 175 beats/min. At very high heart rates, 10-
point source should also be placed in such a manner to-20 millisecond frame times should be considered
(left border of the sternum) that counts from the especially if diastolic function is of interest. In
point source will be acquired during injection of the practice, to avoid the potential errors that might
Tc-99m tracer. It is important to emphasize that for occur if the frame time was constantly being manip-
accurate studies of first pass during treadmill exer- ulated, a standard of 25 milliseconds per frame is
cise, the heart, lung, and point source should all be in recommended for all acquisitions.
the FOV throughout the entire acquisition. 10. Fifteen hundred to 2,000 frames should be sufficient,
7. In high count rate studies, as are typically acquired provided that these frames encompass the bolus
with the multicrystal camera, there are enough injection and capture the entire LV phase of bolus
counts at end diastole and end systole to reliably transit.
identify the end-diastolic and end-systolic frames 11. The matrix size of the multicrystal camera is fixed
without the aid of an electrocardiographic (ECG) due to the inherent design of the systems. However,
signal. However, for single-crystal gamma cameras, for single-crystal cameras, the matrix size will
count rates during the LV phase may occasionally be largely depend on the computer system being used
inadequate for reliable identification of the end- since most vendors do not offer many, if any,
diastolic frames. Acquisition of an ECG signal is choices for dynamic studies. Most systems that have
therefore highly recommended to facilitate data first-pass software support 64 ⫻ 64 acquisitions.
processing. Single-crystal cameras vary widely in Some systems also support 32 ⫻ 32 matrices. The
their count rate capabilities and thus in how appro- latter is preferable because count density per pixel is
priate they are for imaging a first-pass cardiac study. maximized. The 64 ⫻ 64 matrix works reasonably
Several state-of-the-art gamma cameras can count at well when count rates are high, but when counts are
least 150,000 counts per second at a 20% loss of suboptimal or when the LVEF is high, there is a
total counts. Use of cameras with lower count rate tendency for the end-systolic frame to have insuffi-
capabilities could lead to clinically significant inac- cient counts per pixel for assessment of regional wall
curacies in the determination of LVEF and particu- motion. The actual minimum number of counts per
larly in the assessment of wall motion. frame needed varies depending on the number of
Journal of Nuclear Cardiology Friedman et al e45
Volume 13, Number 6;e42-55 First-pass radionuclide angiography

frames per cardiac cycle, the actual LVEF, the a compromise, realizing that assessment of neither
amount of background radiation, and whether the ventricle is optimized.
study is performed to measure LVEF alone or in 17. A shallow (20° to 30°) RAO view is recommended
conjunction with an assessment of wall motion. to enhance right atrial–RV separation, which is the
12. The choice of collimators depends on the objectives chief advantage of FPRNA over gated equilibrium
of the study and the dose to be injected. For standard radionuclide angiography for the measurement of
rest and exercise studies using 20- to 25-mCi doses, RVEF.
the 18-mm-thick collimator provides a good com- 18. As for LV function studies, the choice of collimators
promise between sensitivity and spatial resolution depends on the objectives of the study and the dose
for multicrystal cameras. A thinner collimator sac- to be injected. For standard rest and exercise studies
rifices spatial resolution but may be necessary for using 10- to 25-mCi doses, the 18-mm collimator
lower dose injections. For single-crystal cameras, from multicrystal cameras provides a good compro-
most vendors offer high-sensitivity collimators and mise between sensitivity and spatial resolution. A
some offer ultrahigh-sensitivity collimators. It is thinner collimator sacrifices spatial resolution but
helpful to categorize the collimators quantitatively in may be necessary for lower dose injections. For
counts per millicurie per minute because one ven- single-crystal cameras, the high-sensitivity collima-
dor’s high-sensitivity collimator may be equivalent tor is the standard.
to another vendor’s ultrahigh-sensitivity collimator.
For the purposes of first-pass studies, a high-sensitivity FPRNA–shunt study
collimator should provide approximately 12,000 to
24,000 counts · s⫺1 · mCi⫺1. This number may vary 19. A 10- to 15-mCi dose of the Tc-99m radiopharma-
significantly depending on crystal thickness and ceutical is typically used.
dead time of the system. For large (L)FOV systems, 20. The injection should be rapid for shunt studies. The
it may be desirable to shield part of the peripheral premise of the shunt study is that the appearance in
FOV to reduce unwanted pulse pileup, which in- and clearance of the radionuclide bolus from the
creases the dead time of the system. pulmonary circulation are monoexponential in char-
acter. A delayed bolus may result in a pulmonary
curve that deviates enough from a monoexponential
FPRNA–RV function curve as to make the data uninterpretable. A 10- to
20-mL saline bolus is generally used to flush the
13. Since the injection bolus reaches the right ventricle radionuclide bolus into the venous system. The
without significant dispersion, lower doses may be saline bolus should be injected at a continuous,
adequate; 10- to 25-mCi doses are acceptable. Of uninterrupted rate so that the entire 10 to 20 mL is
note, tricuspid regurgitation will fragment and dis- injected in 2 to 3 seconds.
perse the radiotracer bolus and prolong RV transit 21. Acquisition in the anterior view is best for imaging
time, often rendering FPRNA evaluation of both the the lung fields, which are the areas of interest for the
left and right ventricles inadequate and invalid. shunt study. If both lung fields cannot be visualized
14. The use of the antecubital vein is appropriate for RV due to the detector size, the lung field of interest
studies. The external jugular vein may be used but, should be the right lung for suspected intracardiac
unlike the LV study, it may result in too rapid an shunts and the left lung for a suspected patent ductus
appearance and disappearance of the radionuclide arteriosus (PDA).
from the chamber. 22. Since only pulmonary data will be quantified, an
15. A 10- to 20-mL saline bolus is generally used to ECG signal is unnecessary.
flush the radionuclide bolus into the venous system. 23. Frame time is not crucial in a shunt study since the
The saline bolus should be injected at a continuous, data will ultimately be analyzed using curves whose
uninterrupted rate so that the entire 10 to 20 mL is data points do not require more than 100 millisec-
injected in 3 to 4 seconds. onds’ temporal resolution. Shorter frame times may
16. To optimize assessment of RV function, the FWHM be used since they may be added together during the
of the bolus in the superior vena cava should be 2 to analysis.
3 seconds, much slower than that for an LV study. 24. Total frames acquired should be 2,000.
The slower bolus increases the number of beats 25. For multicrystal systems, the matrix is not an option.
available for analysis. For assessment of biventricu- For single-crystal systems, a 64 ⫻ 64 matrix is
lar function, a bolus with an FWHM of 1 to appropriate. A 32 ⫻ 32 matrix may also be used in
2 seconds in the superior vena cava may be used as order to encompass both lungs in the FOV.
e46 Friedman et al Journal of Nuclear Cardiology
First-pass radionuclide angiography November/December 2006

26. Spatial resolution is much less important in the shunt below 50% of the maximum end-diastolic count should
study. Standard high-sensitivity collimators are ade- be excluded as long as this editing does not preclude a
quate. If available, a diverging collimator may be used. statistically adequate representative cycle. Premature
ventricular beats and post–premature ventricular contrac-
PROCESSING PROTOCOLS tion (PVC) beats should be excluded. If there are few
sinus beats, it may be difficult to generate a statistically
Measurement of ventricular function. Processing adequate representative cycle.
first-pass data has become increasingly automated and Beat editing is an optional routine in which individ-
considerably faster than in previous years. However, it is ual beats of varying duration may be time-corrected so
unlikely that processing of first-pass data can ever be that the final beat lengths are all identical. That approach
reliably, totally automated. There are too many varia- typically involves an interpolation of the data, but actual
tions in tracer transit due to technical and/or pathophys- end-systolic counts should always be preserved so that
iologic reasons for automated processing to be successful the EF is not altered by the time correction.
in all cases. The operator must be observant and careful at Background correction. Several approaches to
a few crucial steps in the processing to ensure consistently background correction have been proposed. They
accurate results. include the lung frame method, the count threshold
Preprocessing. Preprocessing of first-pass data is method, and the periventricular method. The lung
frequently performed, although it is not mandatory. method has been shown to give better results than the
Time-smoothing, uniformity correction, and dead-time other two and is thus the preferred method. The
correction are options that are typically applied when periventricular background region is used as a stan-
supported by the software. dard in many single-crystal camera systems.
Processing. First-pass data processing can be di- Lung frame method. In this approach, a frame of
vided into four major routines: creation of the initial data just before the appearance of activity in the LV ROI
time-activity curve (TAC), beat selection and creation of is chosen as representative of the distribution of the
the initial representative cycle, background correction, background (nonventricular activity). This is a crucial
and creation of the final representative cycle. A fifth step since variation in the background frame can sub-
optional routine is that of motion correction. stantially alter the calculated EF, volumes, and apparent
wall motion. The selected frame should be visualized.
That background “mask,” after appropriate normaliza-
The initial TAC
tion, is subtracted from the LV representative cycle. A
Grouping or reformatting. The raw or prepro- washout factor must be applied to the background since
cessed data should be grouped into 0.5- to 1.0-second the counts in the background are decreasing throughout
images to facilitate drawing an ROI around the ventricle. the LV phase. This approach has been shown to produce
If an ECG signal has been acquired, the raw or prepro- results that compare favorably with contrast angio-
cessed data may be cyclically added using the R wave graphic data and is somewhat better than either of the
to identify end-diastolic frames, thus creating a prelimi- other two approaches.
nary representative cycle. The end-diastolic frame of that Count threshold method. A frame of data just
preliminary representative cycle may then be used to before the appearance of the radionuclide in the left
draw an initial ventricular ROI. ventricle is identified, and the counts in that frame
Ventricular ROI. The operator or the computer become the new zero level for each subsequent frame of
should draw an initial ventricular ROI. This ROI need the LV phase.
not be highly accurate since it is only used for generating Periventricular method. This method is quite anal-
the initial TAC. A TAC of the raw or preprocessed data ogous to the periventricular background method in gated
using the initial ROI should be displayed using the equilibrium imaging. A horseshoe-shaped background
acquisition frame time. ROI is drawn around the ventricle, usually 2 to 3 pixels
Beat selection. Most first-pass software allows the wide and 1 to 2 pixels away from the LV border.
operator to identify the first and last beats to be included The final ROI. Once the background correction has
in the representative cycle. End-diastolic and end-sys- been applied to the initial representative cycle, the end-
tolic frames may be identified automatically by the diastolic and end-systolic frames should be displayed again
computer, but the operator must have the opportunity to and any necessary modifications of the initial ROI then
override the computer to select only appropriate beats. performed.
Because of variable mixing in the chamber, it is advis- Dual-ROI method. For first-pass studies acquired in
able to select beats both before and after the beat with the the anterior view, a separate ROI for the end-diastolic and
maximum counts. Beats whose end-diastolic counts are end-systolic frames is recommended. The operator must
Journal of Nuclear Cardiology Friedman et al e47
Volume 13, Number 6;e42-55 First-pass radionuclide angiography

manually draw the final ROI on both frames. In drawing the (End-diastolic counts ⫺ End-systolic counts)/End-diastolic
end-systolic ROI, left atrial activity must be excluded from counts. On occasion, it is impossible to correct accurately
the end-systolic counts. The LVEF calculated with the for background activity (very delayed bolus, markedly
dual-ROI approach tends to be higher than that calculated prolonged RV tracer transit, and so on). In that case, it is
with a single ROI because the valve plane is placed lower appropriate to report an estimated LVEF or LVEF range
during systole compared to diastole. based on the uncorrected data.
Single-ROI method. With the single-ROI method, Systolic emptying rates. Systolic emptying rates
only an end-diastolic ROI is used. This approach works such as the peak ejection rate may be calculated by
fairly well with studies acquired in the RAO projection, applying a Fourier filter (third- to fifth-order harmonic)
where there is better left atrial–LV separation. In the to the LV representative cycle curve and then taking the
anterior view, the single-ROI method may result in first derivative of that filtered curve. The peak ejection
spuriously low EFs because it can potentially include rate should be expressed in end-diastolic volumes per
extra ventricular counts in the end-systolic ROI. second.
Motion correction. Motion correction of first-pass Diastolic filling rates. Diastolic filling rates such as
data is occasionally necessary for studies acquired during the peak filling rate may be calculated by applying a
bicycle exercise and almost always necessary for studies Fourier filter (third- to fifth-order harmonic) to the LV
acquired during treadmill exercise. Motion correction representative cycle curve and then taking the first
may be performed using either or both of two methods, derivative of that filtered curve. The peak filling rate
the “single-isotope” method (internal correction method) should be expressed in end-diastolic volumes per second.
or “dual-isotope” method (external marker correction). The time–to–peak filling rate may be calculated and
The dual-isotope method is preferred for treadmill exer- expressed in milliseconds.
cise, while the single-isotope method is usually adequate Ventricular volumes. LV end-diastolic volume
for bicycle exercise. may be measured using either a geometric or count-
Single-isotope correction. The position of the ven- proportional technique. In the geometric approach, the
tricle is determined in each frame of the representative end-diastolic frame of the representative cycle is dis-
cycle by applying a center-of-mass algorithm within an played using a threshold for edge detection. The area of
operator-defined ROI. The latter should greatly exceed the the left ventricle is measured using the pixel area and the
actual size of the ventricle. By calculating the average x, y known size of a pixel. The longest length of the left
location of the center of mass, the location of the ventricle ventricle is identified by the operator and end-diastolic
xnyn in any frame may be reregistered to x, y. volume calculated using the modified Sandler and Dodge
Dual-isotope correction. An external point source equation. In the count-proportional approach the required
(americium 241, iodine 125) is applied to the chest, data are the total counts in the left ventricle, the counts in
usually midsternally or just to the right of the sternum. A the hottest pixel in the left ventricle, and the area of a
dual-energy acquisition is performed at peak exercise pixel (m) in centimeters. The end-diastolic volume is
such that two first-pass studies are acquired, one using then calculated as 1.38 m3/2 R3/2, where R is total LV
the external marker’s photopeak and the other using the counts/counts in the hottest pixel.
standard Tc-99m photopeak. After acquisition, the posi-
tion of the marker is determined in each frame using a
Left-to-right shunt study
center-of-mass algorithm. The average x, y location of
the marker is taken to represent the correct position of the The input function for the first-pass shunt study is a
marker had there been no motion. All data frames are high– count density TAC obtained from a pulmonary
then reregistered by the direction and magnitude of the ROI. In practice, either the left lung, the right lung, or
displacement of the marker in that frame. both may be used. In most cases, the right lung is
The final representative cycle. The finalized ROIs preferred because it is easier to create a pulmonary ROI
are then used to regenerate the TAC and the final that is free of contamination from the cyclic changes in
representative cycle is created from that curve using the counts in the left heart and great vessels. Regardless of
previously determined beat selection. It is this represen- the acquisition frame time, the pulmonary curve should
tative cycle that will be used to generate all the quanti- be displayed at a frame time of approximately 100 to 300
tative results describing ventricular function. milliseconds. That allows the operator to easily visualize
the entire curve for qualitative assessment of the pres-
ence or absence of a shunt. Not infrequently, the raw
FPRNA: Quantitation of results
curve requires time-smoothing to eliminate high-fre-
LVEF. The LVEF is calculated from the final quency contamination of the curve from cardiac chamber
background-corrected representative cycle as follows: or great vessel counts or from random noise.
e48 Friedman et al Journal of Nuclear Cardiology
First-pass radionuclide angiography November/December 2006

The first frame of the curve should unequivocally enough of the intermediate steps of processing to enable the
represent pulmonary activity rather than any superior physician to quickly review the processing either directly on
vena cava, right atrial, or RV activity because the shape the computer display or by reference to hard copy.
of the early part of the curve will determine the shape of
the subsequent mathematical fit. It may be helpful to
Display
mask out the superior vena cava and right heart from the
image before drawing the pulmonary ROI. Careful at-
27. The final representative cycle should be displayed in
tention to the statistical content of the pulmonary curve
a cinematic, endless loop format. Most authorities
and its freedom from contamination are crucial. The
use a color display in contrast to the recommended
operator may then apply either a gamma variate or an
display for equilibrium images. The lower pixel
exponential fit to the raw data. Qualitative assessment of
count density and the subtler change from cardiac
the closeness of the fitted curve to the raw curve is
cavities to background make a color display useful.
important. Varying the initial frame of the fit and the
The cine display is typically time-smoothed during
final frame of the fit may be necessary to produce the best
data processing and should not need additional
fitted curve possible. The fitted curve may then be
smoothing for display. Spatial smoothing may be
subtracted from the raw data to leave the shunt compo-
used after processing if the data are particularly
nent behind, which can, itself, be fitted with another
count-poor, but it should not be necessary for the
curve that represents the shunt component. The shunt
average study. It is preferable to normalize the
ratio Qp:Qs is then calculated as (A1 ⫹ A2)/A1, where
image to the peak activity in the ventricle because
A1 is the area under the primary fitted curve and A2 is
aortic or left atrial activity may be higher, thus
the area under the secondary (shunt) fitted curve.
making it more difficult to appreciate the count
changes in the ventricles. Cinematic displays of the
INTERPRETATION AND REPORTING bolus transit through the heart and great vessels are
helpful in analyzing aberrations of tracer transit that
General Comments may occur in patients with congenital anomalies.
28. Hard-copy displays are essential to study interpreta-
The interpretation of first-pass data should be per-
tion. TACs representing the bolus and the RV and
formed in a consistent, methodical manner, with partic-
LV phases of the bolus transit, as well as a final
ular attention to the quality of the data. Unlike equilib-
representative cycle time-activity (volume) curve,
rium radionuclide angiography, in which a quick
must be available for proper interpretation. Color
inspection of the cinematic display of the cardiac cycle is
hard-copy displays of parametric images may be
sufficient to reassure the interpreting physician of the
valuable aides in study interpretation. Such displays
adequacy of the data, the first-pass study requires con-
should not be used to the exclusion of the cinematic
siderably more attention to the details of data acquisition
display of the representative cycle.
and processing to provide consistently accurate interpre-
tations. The final representative cardiac cycle that is used
to generate both the quantitative results and the qualita- Quality Control
tive wall motion assessment can be affected by many
factors, including the adequacy of the injection bolus, the 29. The bolus. The adequacy of a bolus may be defined
count rate, the number and type of beats chosen for quantitatively by generating a TAC from an ROI that
inclusion, the manner in which background activity is includes the superior vena cava. The FWHM of such
determined, and occasionally, patient motion. Even in a curve should ideally be less than 1 second. As a
laboratories with extensive experience and well-defined routine quality check, it is helpful to inspect the
and well-executed acquisition and processing techniques, TAC of the bolus. Alternatively, one may inspect a
unavoidable patient-to-patient variability and different sequence of images from the early portion of the
cardiac and pulmonary physiology, as well as some study to qualitatively assess the bolus. Serial 1-sec-
degree of interobserver variability in processing, lead to ond images are useful for that purpose. The bolus
variability in the end product. The physician must may be assessed as good (FWHM ⬍1 second),
therefore exercise due diligence during interpretation of adequate (1 to 1.5 seconds), delayed (⬎1.5 seconds),
the results. or split (more than one discrete peak in the TAC).
Certain data must be routinely available so that the The split bolus is particularly problematic and may
interpreting physician may quickly assess the technical preclude accurate data processing. Identification of a
adequacy of the data and the accuracy of the processing. delayed or split bolus alerts the physician to the
Most commercial software routines automatically save possibility of oversubtraction of background and the
Journal of Nuclear Cardiology Friedman et al e49
Volume 13, Number 6;e42-55 First-pass radionuclide angiography

resultant spurious increase in LVEF, decrease in LV ably select beats whose end-diastolic counts are 70%
volume, and overestimation of regional wall motion. of the peak end-diastolic counts or greater.
30. Count statistics. The adequacy of the count rate 33. Background selection. The same curve used to
may be assessed by use of either the unprocessed confirm appropriate beat selection may be used to
data or the representative cycle. When examining confirm that an appropriate frame was chosen for
the unprocessed data, the count rate in the whole background correction. A frame as close to the
FOV during the RV phase of the study should beginning of the LV phase but not including LV
optimally be greater than 200,000 cps with a multi- activity is desired. Viewing the background frame
crystal system and greater than or equal to 150,000 image is helpful in determining that LV activity is
cps on a single-crystal system. When count rates not included and in visualizing any residual activity
drop below 100,000 cps, it is highly unlikely that in the right ventricle that could result in oversubtrac-
adequate studies will be obtained. Alternatively, and tion of background. On occasion, the lung frame
more accurately, the count rate of the representative method of background correction may not be accu-
cycle can be checked. This approach is more accu- rate because of poor RV-LV temporal separation. In
rate because it is the representative cycle that is used that case, the physician should demand that a repre-
to generate all quantitative results, and counts in the sentative cycle be created that has not been subjected
representative cycle may be inadequate even when to background correction. The uncorrected represen-
the count rate on the raw data is adequate if there are tative cycle is always generated during the process-
insufficient beats for analysis or background over- ing but may not be stored. Viewing this image in
subtraction. In general, LV end-diastolic counts in cine loop format after manually subtracting the
the representative cycle should not be less than background will allow an adequate assessment of
2,000 cps and should preferably exceed 4,000 cps. regional wall motion.
High-resolution wall motion images will require 34. Patient motion. Motion of the patient is rarely, if
greater than 5,000 cps. ever, a problem on a resting study. However, motion
31. Tracer transit. The transit of the radionuclide of the chest during acquisition of an exercise study is
should be inspected in every case. Alterations or seen frequently during treadmill exercise and occa-
anomalies of tracer transit may be detected visually sionally during bicycle exercise. Motion should be
by examining serial static images or by a cinematic suspected when typical distortions of the LV TAC
display of the bolus transiting the central circulation. are noted and should be confirmed by viewing a cine
For example, asymmetric pulmonary transit time or display of the bolus traveling through the chambers.
asymmetric maximal tracer concentration may indi- During treadmill exercise, chest wall motion may be
cate pulmonary vascular pathology or unilaterally corrected with the use of an external point source.
decreased pulmonary volume. A cine display may be The integrity of the point source and especially its
particularly helpful when the transit seems anoma- appearance in each frame of the study should be
lous. The most common disturbance of tracer transit confirmed.
is prolongation of the transit time through either or
both ventricles. Recognition of a prolonged transit
time is important because of the potential diagnostic Results
implications and because of the impact on back-
ground correction, which in turn affects EF, vol- 35. Cardiac rhythm and conduction. Interpretation of
umes, and wall motion. Physiologic causes of pro- the data may be influenced by the rhythm during the
longed tracer transit include valvular insufficiency, acquisition. For example, frequent PVCs, ventricular
severely depressed ventricular function, atrial fibril- bigeminy, or very irregular atrial fibrillation may
lation, and a left-to-right shunt. Combining image affect the EF or regional wall motion. In the setting
information such as an enlarged left atrial appendage of ventricular bigeminy, for example, no true sinus
(see below) and a prolonged tracer transit through beat EF can be determined. The diagnostic and
the left ventricle suggests mitral valve disease, prognostic significance of post-PVC beats is not
whereas prolonged LV tracer transit and an enlarged completely understood. With atrial fibrillation, the
ascending aorta suggest aortic valve disease. representative cycle may consist of beats with
32. Beat selection. A hard copy of the TAC should be widely varying R-R intervals and, hence, with dif-
generated during processing so that one may confirm ferent volumes and EFs. Pacemaker rhythm confers
that the appropriate beats have been selected for its own unique contraction sequence, which starts at
inclusion in the representative cycle. Unless the the apex and proceeds to the base. The latter can be
number of beats is very limited, one should prefer- recognized from the cinematic display of the repre-
e50 Friedman et al Journal of Nuclear Cardiology
First-pass radionuclide angiography November/December 2006

sentative cycle. A phase image may be helpful in

recognition of this pattern.
Both regional wall motion and LVEF are typically
altered by left bundle branch block. Because most
first-pass studies are acquired in the RAO or anterior
projections, paradoxical septal motion cannot be
detected; however, one may see what appear to be
inferoapical or anteroapical wall motion abnormali-
ties. A phase image may aid in recognizing this
phenomenon, although it is usually apparent on the
cinematic display of the representative cycle. Right
bundle branch block does not affect the LV contrac-
tion pattern.
36. Chamber sizes. Because the overwhelming major- Figure 1. FPRNA segmentation.
ity of first-pass studies are performed to evaluate the
left ventricle, the final representative cycle will show distinguished from the adjacent contractile myocar-
the left ventricle, the left atrium (in particular its dium. When available, previous studies should be
appendage), and the ascending aorta. Most of the left compared by use of side-by-side cine analysis.
atrium is overlapped with the ascending aorta, and Standardized nomenclature for the myocardial seg-
its size is difficult to assess. However, when the left ments visualized in the typical FPRNA study (ie,
atrium is very dilated, its appendage is quite prom- acquired in the anterior view) is shown in Figure 1.
inent in the anterior view, and the aortic root will If biplane FPRNA is performed, segmental analysis
appear to be dilated. Judging the size of the left may also be applied to the left anterior oblique
ventricle qualitatively is more difficult on a first-pass projection. It is recommended that the visualized
study than on an equilibrium study because one does segments be designated as the basal anterolateral,
not have all the surrounding chambers and great mid anterolateral, apical, mid inferoseptal, and basal
vessels in the same image as references. With inferoseptal segments. For a left anterior oblique
enough experience, moderate to severe degrees of acquisition, the visualized segments include the
LV chamber enlargement can be appreciated. Right septal, inferoseptal, inferoapical, inferolateral, and
atrial, RV, and pulmonary arterial sizes can be lateral segments.
evaluated on cinematic display or on serial static Even though the final representative cycle is cor-
0.5- to 1.0-second images from the raw data but are rected for background, it may be necessary to display
not particularly reliable. the cine with the lower level raised to 10% to 20%
Actual measurement of LV volume may be per- depending on the signal-to-noise ratio in the study. Any
formed with either geometric or count-based ap- one of a number of color schemes may be used to view
proaches and offers a more consistent and accurate the cine. Whether one is more representative of actual
assessment of chamber size. Normal values for the wall motion than another is speculative. The operator
left ventricle should be established for each laboratory should choose the scheme that works best clinically,
because they will vary depending on the type of but one should avoid color tables that condense all the
processing used, especially the type of background 3-dimensional information into a few colors that make
correction and the patient’s position during the acqui- the image appear 2-dimensional, as if all the informa-
sition (ie, supine, semisupine, or upright). tion were in the moving edges.
37. Regional wall motion. The cinematic display of the Many parametric images are available to the
representative cycle should be viewed and regional interpreting physician that may be used to reinforce
wall motion assessed qualitatively by use of the one’s subjective opinion. Occasionally, an abnor-
conventional terms of hypokinesis, akinesis, and mality will be evident on a parametric image that is
dyskinesis. For hypokinesis, the qualifiers of mild, not obvious on the cinematic display, especially
moderate, and severe are useful for communicating when the regional dysfunction is occurring in a plane
the severity of the abnormality. In addition, the that is perpendicular to the detector. The parametric
extent and location of the abnormality should also be images may be thought of as an independent, unbi-
reported, such as the basal (proximal) half or the ased observer similar to the way in which quantita-
apical (distal) quarter of the anterior wall. An aneu- tive displays of perfusion images are used. They are
rysm should be identified when an akinetic or also useful as quantitative measures of regional
dyskinetic segment can be clearly and discretely dysfunction. For example, one physician’s impres-
Journal of Nuclear Cardiology Friedman et al e51
Volume 13, Number 6;e42-55 First-pass radionuclide angiography

sion of moderate hypokinesis may be different than It is difficult to evaluate diastolic filling during
another physician’s, but a regional EF of 28% is exercise because the increase in heart rate usually
clear to anyone receiving the information. The most results in a loss of the transition between early peak
commonly used images are the regional EF image, the filling and atrial filling. At best, one can measure
stroke volume image, and the amplitude and phase peak diastolic filling, but without the requisite tem-
images. The latter three may be used in processing, as poral sampling necessary for high heart rates (ie, 10
well as in interpretation. One must keep in mind that to 20 ms/frame), any measured values may not be
the accuracy of the parametric image is highly depen- reliable. Some investigators have used filling frac-
dent on the statistics in the image and may be influ- tions (ie, the fraction of filling achieved during the
enced by translational movement of the heart; there- first third or first half of diastole). It is not clear
fore, the parametric image should not be used to the that such values offer any advantage over the
exclusion of the representative cycle cine because conventional values, and they are certainly a
the latter gives the operator the best visual feedback departure from the values typically measured in
on the statistical quality of the data. Very little gated equilibrium studies.
literature is available to document the accuracy of
parametric images for diagnosis.
Exercise/Intervention Studies
38. LVEF and RVEF. The LVEF is calculated from the
background-corrected end-diastolic and end-systolic 40. The representative cycles of both resting and exer-
counts in either ventricle. Published ranges for a cise or pharmacologic intervention studies should be
normal EF vary, but most laboratories accept a range viewed in a split-screen cinematic display. Each
of 50% to 80% for the left ventricle. The variability study should be normalized to itself.
of the LVEF has been reported to be ⫾ 4% at rest for 41. Regional wall motion of an exercise or intervention
the same individual studied on different days. It is study should be compared visually with regional wall
very important when interpreting the LVEF and motion of the resting study by use of standard qualita-
most especially when interpreting changes in LVEF tive or semiquantitative terms (see paragraph 11).
from one study to another to keep in mind that the During exercise or during administration of inotropic
LVEF is not a fixed number for any patient. It will or afterload-reducing agents, regional wall motion is
vary with the heart rate, the blood pressure, the level expected to increase. Regional wall motion may de-
of circulating catecholamines, position (upright vs crease during ischemia, during protocols that result in
supine), and medications. When there are an ade- abrupt increase in afterload such as isometric or sudden
quate number of beats to choose for the EF calcula- strenuous aerobic exercise, or during administration of
tion, it is preferable to select those beats whose drugs that acutely increase afterload. A semiquantita-
end-diastolic counts are 70% of the peak end- tive scoring system or quantitative regional EFs may be
diastolic counts or greater. The normal values for the useful for comparison of rest to exercise or interven-
RVEF vary with the type of processing used. With tional studies.
the use of separate end-diastolic and end-systolic 42. The size of the left ventricle may be qualitatively
ROIs, the lower limit of normal can be expected to evaluated on the cine displays. During exercise in
be 40%, ranging up to 65%. the upright position, LV volume usually increases.
39. Diastolic filling of the left ventricle may be assessed The magnitude of the increase is typically in the
by qualitative inspection of the TAC (volume curve) 10% to 20% range, although larger increases do
of the LV representative cycle. Obviously decreased occur in control subjects. When the volume in-
early rapid filling, a prolonged time to peak filling, creases by 50% or greater above baseline, coronary
and an increase in the atrial contribution to filling artery disease should be suspected even in the
may be recognized by visual inspection of the LV absence of a regional wall motion abnormality,
volume curve. The quantitative values for peak early especially if there is a concomitant, significant drop
filling and the time to peak filling should be ex- in LVEF. LV volume may fail to increase or may
pressed in end-diastolic volumes per second and in actually decrease even in the upright position in
milliseconds, respectively. The atrial contribution to patients with pericardial or valvular heart disease.
filling may be expressed as a ratio of the atrial to 43. During exercise in the upright position, one can
early peak filling or vice versa. The atrium typically anticipate that the EFs of both ventricles will in-
contributes 15% to 25% of total LV filling. The crease. At one point in time, failure to increase the
interpreting physician should not accept any dia- LVEF during exercise was invariably considered
stolic values without confirmation by visual inspec- pathologic. However, it is quite clear that some
tion of the LV volume curve. individuals may show a flat response to exercise and,
e52 Friedman et al Journal of Nuclear Cardiology
First-pass radionuclide angiography November/December 2006

occasionally, even a decrease in EF (especially particularly useful in avoiding pitfalls if the previous
elderly subjects) in the absence of coronary or study was technically suboptimal.
valvular heart disease. The higher the resting EF, the 46. The report should include the most important vari-
less of an increase one tends to see during exercise. ables from a stress or intervention that will help the
For diagnostic purposes, an absolute value of exer- referring physician assess the clinical significance of
cise EF may be more useful than the change from the findings. These variables are also important
rest to exercise. Most normal individuals will have a because they have independent diagnostic and prog-
peak exercise LVEF of 56% or greater. A decrease in nostic information.
LVEF to less than 56% should be considered abnormal 47. Overall study quality should be mentioned in the
in individuals aged younger than 70 years, but in the report. This serves to appropriately increase or
absence of regional dysfunction, the finding is not decrease the confidence of the physicians using the
specific for coronary artery disease. The change in EF report for clinical decision making. It is also useful
during exercise may also be influenced by the type of for subsequent screening of studies for inclusion in
exercise protocol used. A standard graded exercise research databases.
protocol should always be used. 48. The initial interpretation of the study should be made
RVEF typically increases during exercise but may without reference to clinical data to avoid bias. The
decrease in patients with pulmonary hypertension, physician should then correlate the findings and
including those in whom pulmonary hypertension interpretation with the clinical information to avoid
develops during exercise, such as those with mitral an obvious misinterpretation and to guarantee that
stenosis or severe exercise-induced LV dysfunction. the clinical question has been addressed. Including
In particular, patients with proximal right coronary the indication for the study in the report serves to
artery lesions may show decreases in RVEF during focus the interpreting physician’s attention to the
exercise. clinical question and is also useful for subsequent
44. LV tracer transit may be prolonged during exercise coding issues related to reimbursement. Studies
because of the appearance of mitral insufficiency should be classified as normal or abnormal. Categories
resulting from LV ischemia. This finding may be of probably normal, equivocal, and probably abnormal
recognized most readily on a TAC of the bolus may be added. Both the diagnostic and prognostic
transit through the left ventricle. Occasionally this contents of the data should be addressed. If perfusion
finding is accompanied by exercise-induced enlarge- scan data are available, a statement about the signifi-
ment of the left atrium. cance of the two data sets is appropriate.
49. Whenever previous studies are available, the cine
displays of the representative cycles should be
Conclusion
displayed side by side. When rest and exercise/
45. The radionuclide and doses used for the study, as intervention data are available, a quad-screen
well as the injection site, should be permanently display is optimal. Interpretation of serial changes
archived in the report. These are more important for in EFs should always take into account differ-
future reference in case a patient returns to the ences in the heart rates, blood pressures, and
laboratory for serial studies. Having the data is medications.

Table 2. First-pass radionuclide angiography: Guideline for interpretation

For information,
see paragraph

A. Display
1. Cinematic display of representative cycle Standard 27
a. Time smoothing Standard 27
b. Spatial smoothing Optional 27
c. Normalization Standard 27
2. Hard copy 2
a. Intermediate processing steps Standard 28
b. Functional images Optional 28
c. TACs Standard 28
Journal of Nuclear Cardiology Friedman et al e53
Volume 13, Number 6;e42-55 First-pass radionuclide angiography

Table 2. (Continued)
For information,
see paragraph

B. Quality control
1. Bolus Standard 29
2. Count statistics Standard 30
3. Tracer transit Standard 31
4. Beat selection Standard 32
5. Background selection Standard 33
6. Patient motion Standard 34
C. Results
1. Cardiac rhythm and conduction Standard 35
2. Chamber sizes
a. Qualitative Standard 36
b. Quantitative Preferred 36
3. Regional wall motion
a. Qualitative Standard 37
b. Quantitative Optional 37
4. LVEF Standard 38
5. RVEF Optional 38
6. LV diastolic filling
a. Qualitative Standard 39
b. Quantitative Optional 39
D. Exercise/intervention studies
1. Display Standard 40
2. Regional wall motion: Comparison to rest Standard 41
3. Chamber sizes: Comparison to rest Standard 42
4. EFs: Comparison to rest Standard 43
E. Conclusion
1. Comparison to previous studies Standard 44
2. Correlation with clinical findings Standard 45

Table 3. First-pass radionuclide angiography: Guideline for reporting

For information,
see paragraph

A. Demographics
1. Name Standard
2. Gender Standard
3. Age Standard
4. Date(s) of acquisition(s) Standard
5. Medical record identifier (inpatient) Standard
6. Height/weight taking into account BSA Standard
B. Acquisition parameters
1. Type(s) of acquisition(s) (rest/exercise/intervention) Standard
2. Radionuclide and doses Standard 45
3. Injection site Standard 45
4. Indication for study Standard 48
5. Study quality tandard 47
e54 Friedman et al Journal of Nuclear Cardiology
First-pass radionuclide angiography November/December 2006

Table 3. (Continued)
For information,
see paragraph

C. Results: Hemodynamic and exercise/intervention variables

1. Rest HR and BP, cardiac rhythm Standard 46
2. Exercise HR and BP, %MPHR, METS Standard 46
3. Exercise symptoms, reason for stopping Standard 46
4. Exercise ECG changes/arrhythmia Standard 46

D. Results: Resting RNA data

1. Chamber sizes
a. Qualitative Optional 36
b. Quantitative Standard 36
2. Regional wall motion
a. Qualitative Standard 37
b. Quantitative Optional 37
3. EFs Standard 38
4. Abnormalities of tracer transit Standard 31
E. Results: Exercise/intervention RNA data
1. LV size: Change from rest
a. Qualitative Standard 42
b. Quantitative Preferred 42
2. LV regional wall motion: Change from rest Standard 41
3. LVEF: Change from rest Standard 43
4. RVEF: Change from rest Optional 43
5. Abnormalities of tracer transit Standard 44
F. Conclusion
1. Normal/abnormal Standard 48
(definite, probable, equivocal) Optional 48
2. Assessment of severity of findings Standard 48
(diagnostic/prognostic) 48
3. Relationship to perfusion data if acquired Optional 48
4. Comparison to previous studies Standard 49

BSA, Body surface area; HR, heart rate; BP, blood pressure; %MPHR, maximum predicted heart rate; METs, metabolic equivalents;
RNA, radionuclide angiography.

Suggested Reading 5. Borges-Neto S, Coleman RE, Potts JM, Jones RH. Combined
exercise radionuclide angiocardiography and single photon
1. Aroney CN, Ruddy TD, Dighero H, et al. Differentiation of emission computed tomography perfusion studies for assess-
restrictive cardiomyopathy from pericardial constriction: assess- ment of coronary disease. Semin Nucl Med 1991;21:
ment of diastolic function by radionuclide angiography. J Am Coll 223-9.
Cardiol 1989;13:1007-14. 6. Borges-Neto S, Shaw L. The added value of simultaneous myo-
2. Berger HJ, Matthay RA, Loke J, et al. Assessment of cardiac cardial perfusion and left ventricular function. Curr Opin Cardiol
performance with quantitative radionuclide angiocardiography: RV 1999;14:460-3.
ejection fraction with reference to findings in chronic obstructive 7. Borges-Neto S. Perfusion and function assessment by nuclear
pulmonary disease. Am J Cardiol 1978;41:897-905. cardiology techniques. Curr Opin Cardiol 1997;12:581-6.
3. Berger H, Reduto L, Johnstone D, et al. Global and regional left 8. Campos CT, Chu HW, D’Agostino HJ Jr, Jones RH. Comparison
ventricular response to bicycle exercise in coronary artery disease: of rest and exercise radionuclide angiocardiography and exercise
assessment by quantitative radionuclide angiocardiography. Am J treadmill testing for diagnosis of anatomically extensive coronary
Med 1979;66:13-21. artery disease. Circulation 1983;67:1204-10.
4. Borges-Neto S, Coleman RE, Jones RH. Perfusion and function at 9. DePace NL, Iskandrian AS, Hakki A, et al. Value of left
rest and treadmill exercise using technetium-99m sestamibi: com- ventricular ejection fraction during exercise in predicting the
parison of one and two day protocols in normal volunteers. J Nucl extent of coronary artery disease. J Am Coll Cardiol 1983;1:
Med 1990;31:1128-32. 1002-10.
Journal of Nuclear Cardiology Friedman et al e55
Volume 13, Number 6;e42-55 First-pass radionuclide angiography

10. Foster C, Dymond DS, Anholm JD, et al. Effect of exercise ejection fraction using a single crystal nuclear camera. J Nucl Med
protocol on the left ventricular response to exercise. Am J Cardiol 1994;35:1292-300.
1983;51:859-64. 21. Nichols K, DePuey EG, Rozanski A. First-pass radionuclide
11. Friedman JD, Berman DS, Kiat H, et al. Rest and treadmill exercise angiocardiography using single crystal gamma cameras. J Nucl
first-pass radionuclide ventriculography: validation of left ventricular Cardiol 1997;4:61-73.
ejection fraction measurements. J Nucl Cardiol 1994;4:382-8. 22. Nickel O, Schad N, Andrews EJ, et al. Scintigraphic measurement
12. Gal R, Grenier RP, Carpenter J, et al. High count rate first-pass of left ventricular volumes from the count-density distribution.
radionuclide angiography using a digital gamma camera. J Nucl J Nucl Med 1982;23:404-10.
Med 1986;27:198-206. 23. Philippe L, Mena I, Darcourt J, French WJ. Evaluation of valvular
13. Gal R, Grenier RP, Schmidt DH, Port SC. Background correction regurgitation by factor analysis of first-pass angiography. J Nucl
in first-pass radionuclide angiography: comparison of several Med 1988;29:159-67.
approaches. J Nucl Med 1986;27:1480-6. 24. Potts JM, Borges-Neto S, Smith LR, Jones RH. Comparison of
14. Gal R, Grenier RP, Port SC, Dymond DS. Left ventricular volume bicycle and treadmill radionuclide angiocardiography. J Nucl Med
calculation using a count-based ratio method applied to first-pass 1991;32:1918-22.
radionuclide angiography. J Nucl Med 1992;33:2124-32. 25. Reduto LA, Wickemeyer WJ, Young JB, et al. Left ventricular
15. Johnson LL, Rodney RA, Vaccarino RA, et al. Left ventricular diastolic performance at rest and during exercise in patients with
perfusion and performance from a single radiopharmaceutical and coronary artery disease. Circulation 1981;63:1228-37.
one camera. J Nucl Med 1992;33:1411-6. 26. Updated imaging guidelines for nuclear cardiology procedures,
16. Jones RH, McEwen P, Newman GE, et al. Accuracy of diagnosis of part 1. J Nucl Cardiol 2001;8:G1-58.
coronary artery disease by radionuclide measurement of left ventric- 27. Upton MT, Rerych SK, Newman GE, et al. The reproducibility
ular function during rest and exercise. Circulation 1981;64:586-601. of radionuclide angiographic measurements of LV function in
17. Lee KL, Pryor DB, Pieper KS, Hasell FE Jr. Prognostic value of normal subjects at rest and during exercise. Circulation 1980;
radionuclide angiography in medically treated patients with coro- 62:126-32.
nary artery disease. Circulation 1990;82:1705-17. 28. Verani MS, Lacy JL, Guidry GW, et al. Quantification of left
18. Maltz DL, Treves S. Quantitative radionuclide angiocardiography: ventricular performance during transient coronary occlusion at various
determination of Qp:Qs in children. Circulation 1973;47:1048-56. anatomic sites in humans: a study using tantalum-178 and a multiwire
19. Morrison DA, Turgeon J, Quitt T. Right ventricular ejection fraction gamma camera. J Am Coll Cardiol 1992;19:297-306.
measurements: contrast ventriculography versus gated blood pool and 29. Williams KA, Bryant TA, Taillon LA. First-pass radionuclide angio-
gated first-pass method. Am J Cardiol 1984;54:651-3. graphic analysis with two regions of interest to improve left ventric-
20. Nichols K, DePuey EG, Gooneratne N, et al. First-pass ventricular ular ejection fraction accuracy. J Nucl Med 1998;39:1857-61.
 Equilibrium radionuclide angiocardiography
James R. Corbett, MD,a Olakunle O. Akinboboye, MB, BS, MPH, MBA,b
Stephen L. Bacharach, PhD,b Jeffrey S. Borer, MD,b Elias H. Botvinick, MD,b
E. Gordon DePuey, MD,b Edward P. Ficaro, PhD,b Christopher L. Hansen, MD,b
Milena J. Henzlova, MD,b and Serge Van Kriekinge, PhDb

I. PLANAR IMAGING 2).4 Heparin unfavorably affects labeling efficiency.

Thus, if at all possible, Tc-99m–pertechnetate should not
A. Purpose be injected in heparinized intravenous lines, or they
should be flushed thoroughly. Similarly, intravenous
Planar equilibrium radionuclide angiography (ERNA)
lines containing dextrose solution may alter labeling
is used to determine global and regional measures of
efficiency. In addition, antibodies against red blood cells
ventricular function (primarily left ventricular [LV]
may inhibit labeling. Antibodies may develop as a result
function) at rest and/or during exercise stress or pharma-
of drugs such as methyldopa and penicillin. Antibodies
cologic intervention. These measures of ventricular func-
may also be present in patients with chronic lymphocytic
tion may include evaluations of ventricular wall motion,
leukemia, non-Hodgkin’s lymphoma, and systemic lupus
ejection fraction (EF), and other parameters of systolic
erythematosus.
and diastolic function. The following sections provide a
Labeling efficiency is also diminished when “old”
technical description of the techniques to acquire and
Tc-99m–pertechnetate of low specific activity is used.
process the data necessary to assess parameters of
Tc-99m decays to Tc-99, which is no longer useful for
ventricular performance.
imaging but nevertheless competes with the radioactive
form for stannous ions. To prevent the presence of carrier
B. Radiopharmaceuticals Tc-99, the Tc-99m dose should be taken from relatively
fresh (⬍24 hours after elution) eluate from the
1. Inject the patient with technetium 99m–labeled red
generator.5
blood cells, with activity of approximately 20 to 25
mCi/70 kg body weight (11-13 MBq/kg) to provide
the radioisotope tag for resting studies. For exercise
studies, the activity can be increased to 25 to 35 C. Acquisition Parameters–Rest/Exercise Imaging
mCi/70 kg patient. The radiation dosimetry to the
patient from 20 mCi of Tc-99m–labeled red blood 1. Collimator. For resting studies, use of a parallel-
cells labeled in vitro is 0.3 to 0.52 rem effective dose hole, high-resolution collimator with spatial resolu-
equivalent. With in vivo labeling of red blood cells, tion of approximately 8 to 10 mm full width at
doses will run at the higher end of this range.1 half maximum (FWHM) (of a line spread function)
2. Labeling methods or better at 10 cm distance from the collimator, as
a. In vivo or modified in vivo/in vitro methods (eg, well as a sensitivity (for Tc) of approximately
using 2 to 3 mg stannous pyrophosphate 15 min- 4,000 to 5,000 counts · s⫺1 · mCi⫺1 (108 to 135
utes before injection of the radiopharmaceutical).2 counts · s⫺1 · MBq⫺1), is preferred. If a time-limited
b. Commercial in vitro kit.3 stress study (eg, bicycle exercise) is to be performed,
a higher-sensitivity (and therefore poorer–spatial
Quality control–labeling efficiency. Poor label- resolution) collimator may be considered, such as a
ing of red blood cells can be easily recognized on ERNA low-energy all-purpose (LEAP) collimator (typically
images. Free Tc-99m–pertechnetate accumulates in the 12 mm FWHM at 10 cm and sensitivity of approx-
mucosa of the stomach and in the thyroid gland. A imately 10,000 counts · s⫺1 · mCi⫺1, or approxi-
number of frequently used drugs and solutions are mately 280 counts · s⫺1 · MBq⫺1), or optionally a
known to interfere with red blood cell labeling (Table high-sensitivity collimator. In this case, it is espe-
cially important to keep the collimator– chest wall
distance minimized. It is recommended that the
Chair.a Memberb. same collimator be used for both the rest and stress
J Nucl Cardiol 2006;13:e56-79.
1071-3581/$32.00
study, with the parallel-hole LEAP collimator being
Copyright © 2006 by the American Society of Nuclear Cardiology. preferred. If caudal tilt is used (see paragraph 11), a
doi:10.1016/j.nuclcard.2006.08.007 slant-hole collimator may be considered instead of a
e56
Journal of Nuclear Cardiology Corbett et al e57
Volume 13, Number 6;e56-79 Equilibrium radionuclide angiocardiography

Table 1. Radiopharmaceuticals used for planar ERNA

For information
Rest Exercise see paragraph

Radiopharmaceutical Tc-99m–labeled RBCs Tc-99m–labeled RBCs Standard 1

Dose 20–25 mCi/70 kg 25–35 mCi/70 kg Preferred 1
Labeling method In vivo In vivo Not Recommended 2
Modified in vivo/in vitro Modified in vivo/in vitro Standard 2
In vitro In vitro Preferred 2

Table 2. Causes of poor red blood cell labeling

Cause Mechanism

Hydralazine Oxidation of stannous ion

Prazosin Decreases labeling rate
Propranolol Increases dissociation
Digoxin Decreases labeling rate
Doxorubicin ?
Iodinated contrast ?
Heparin Complexes with Tc-99m
Oxidation of stannous ion
Dextrose Complexes with Tc-99m
Methyldopa Induces RBC antibodies
Oxidation of stannous ion
Penicillin Induces RBC antibodies
Quinidine Induces RBC antibodies
Immune disorders Induces RBC antibodies
Prolonged generator ingrowth Increased carrier
Decreased hematocrit Relative increase in plasma which oxidizes stannous ion (?)
Excess stannous ion Tc-99m reduced outside RBC
Insufficient stannous ion Incomplete reduction with free Tc-99m–pertechnetate

RBC, Red blood cell

From Gerson MC. Cardiac nuclear medicine. 3rd ed. New York: McGraw-Hill; 1997.
Reproduced with permission of the McGraw-Hill Companies.

parallel-hole collimator to provide 10° to 15° of 3. FOV. The effective FOV is dependent on the cam-
caudal tilt while maintaining minimal collimator– era size and acquisition zoom utilized. Typically, an
chest wall distance. 18 ⫻ 18 – cm square FOV will be sufficient, but any
2. Pixel size. Any matrix size that results in a pixel size FOV size sufficient to encompass all four cardiac
less than approximately 4 mm/pixel (with approxi- chambers and at least 2 cm beyond the cardiac blood
mately 2 to 3 mm/pixel being preferred) can be used. pool (for positioning of a background region of
These acquisitions are usually performed using a interest [ROI]) is acceptable. Larger FOVs are ac-
zoomed 64 ⫻ 64 matrix of 16-bit (word) pixels. The ceptable, provided that they (1) do not inhibit min-
zoom required to meet the less than 4 mm/pixel imizing collimator-chest distance and (2) are not so
criteria will depend on the field of view (FOV) of the large as to cause increased gamma camera dead
camera used. Minimal to no zoom is recommended time. If large FOVs are used, care must be taken to
for small-FOV cameras (circular, 10 inch diameter, prevent acquisitions terminated on liver or spleen
or square, 8 inch) to zoom factors of 1.5 to 2.2 for overflow and to ensure that data are displayed with
large-FOV (rectangular, 21 inch) cameras. Smaller- cardiac structures at maximum intensity. Use of a
FOV cameras, if available, are preferred. In any lead apron as a shield for the liver/spleen may be
case, pixel size should not exceed 5 mm. See appropriate, positioned with the aid of a persistence
paragraph 5. scope.
e58 Corbett et al Journal of Nuclear Cardiology
Equilibrium radionuclide angiocardiography November/December 2006

Table 3. Camera/computer setup6-11

For information,
Rest Exercise see paragraph

Collimator Parallel-LEAP Parallel-LEAP Preferred 1

(rest/exercise) Parallel–high Sensitivity Parallel–high Sensitivity Optional 1
Collimator Parallel–high Resolution Preferred 1
(rest only) Parallel-LEAP Optional 1
Slant hole Slant hole Optional 1
Pixel size ⬍4 mm/pixel ⬍4 mm/pixel Standard 2, 3
2–3 mm/pixel 2–3 mm/pixel Preferred 2, 3
Energy window 140 keV ⫾ 10% 140 keV ⫾ 10% Standard 4
Bad beat Buffered beat Buffered beat Preferred 5
On-the-fly: On-the-fly: Standard 5
Reject beat and next beat Reject beat and next beat Preferred 5
Beat length window ⫾10%–15% ⫾10%–15% Standard 5
Check trigger Check trigger Preferred 5
Acquisition method Frame mode Frame mode Standard 7
List mode List mode Optional 7
Frame rate (EF) ⬎16 Frames/cycle 16–32 Frames/cycle Standard 8
24–32 Frames/cycle 24–32 Frames/cycle Preferred 8
Count density ⬎1800/pixel (3 mm) or Not applicable Standard 9
20,000/cm2 (high Resolution
collimator)
⬎3600/pixel (3 mm) or 2–3 min acquisition Standard 9
40,000/cm 2 (high
Sensitivity collimator)
Positioning LAO (best right/left separation) LAO (best right/left separation) Standard 10a
LAO (caudal tilt) LAO (caudal tilt) Optional 10a
Plus anterior, lateral Standard 10b and 10c
Quality Control View cine loop View cine loop Preferred 6
R-R histogram R-R histogram Preferred 6

4. Energy window. 140 keV, ⫾10% window. satory) beat, is preferred. The typical beat length
5. Bad beat/beat length window (arrhythmia re- window is ⫾10% to 15% but will vary depending
jection).8,12,13 If systolic function only (EF) is to on heart rate and rhythm. If the study is acquired
be assessed, accepting less than 10% to 15% with significant arrhythmias, poor statistics may
arrhythmic beats is acceptable. If the examination result, and accurate computation of EF may be
is performed to determine diastolic function, beat compromised. The beat length window may re-
length windowing (arrhythmia rejection) is neces- quire lengthening to improve statistics but will
sary. The preferred arrhythmia rejection mode is compromise measurement of diastolic function
buffered beat, where each beat is temporarily and may adversely affect cine-loop displays if end
stored in memory to assess whether its beat length frames are not corrected or deleted. See paragraph
is within the (typical) ⫾10% to 15% R-R beat 8.
length window. If the beat is outside the window, With regard to triggers, assessment of the ade-
it is rejected without contaminating the gated data quacy of the R-wave trigger prior to instigation of
acquired. Standard arrhythmia rejection methods the gated acquisition should be performed. It is
typically terminate data acquisition if a premature recommended that the electrocardiographic (ECG)
beat is seen outside of the (⫾10% to 15%) beat trigger point be checked to ensure that the ECG
length window (a portion of the bad beat will be gating circuitry is synchronized to the peak of the
acquired). Rejection of the short or long beat, ECG R wave. Checks can be either performed
along with rejection of the subsequent (compen- visually, with a dual-trace oscilloscope, or checked
Journal of Nuclear Cardiology Corbett et al e59
Volume 13, Number 6;e56-79 Equilibrium radionuclide angiocardiography

with a commercially available dynamic phantom. give variable results depending on collimator
Poor-quality or delayed triggers can adversely affect sensitivity. To calculate a typical acquisition time,
the ventricular volume curve. for a particular collimator and dose, sum all
6. Post-acquisition quality control of the ERNA frames in the study together, and using a small
study is also recommended. Visualization of the ROI at the center of the left ventricle, determine
beating cine loop after acquisition allows evaluation the time necessary to achieve the above-men-
of data drop-off due to inadequate triggers, signifi- tioned counts per square centimeter. During stress,
cant arrhythmias, rhythm disorders, poor tag, or poor acquisition time is often the limiting factor. When
count statistics. Review of the beat-length R-R this is the case, at least 2 minutes (preferably 3
interval histogram can be used to assess cardiac minutes) of acquisition time at peak stress is
rhythm abnormalities or determine if significant recommended using the LEAP or high-sensitivity
arrhythmias were present. Quality control can antic- collimator, as previously specified.
ipate errors associated with inadequate ERNA stud- 10. Camera positioning–imaging angles. To acquire
ies prior to the reporting of ventricular performance. the ERNA study, position the patient supine (for
7. Acquisition method. Frame mode gating is standard greatest patient comfort) or in the right lateral
(forward framing), although forward/backward, as decubitus position (to minimize interference from
well as forward/backward by thirds, methods are diaphragm and spleen). Three views should be
optional. If arrhythmias are present, the frames at the recorded for assessment of wall motion of the left
end of the cardiac cycle (containing data acquired ventricle.
over shorter total acquisition time) should be either a. LAO view. The LAO view is optimized to
corrected or deleted (preferred). List mode data visualize the septum (best septal view— usually
acquisition is optional and offers increased beat the 45° LAO, but the angle will depend on body
length windowing flexibility, particularly for analy- habitus and cardiac orientation) (Figure 1). In
sis of diastolic function.6 the LAO view, the orientation should be such
8. Frame rate.12 A frame rate of less than 50 millisec- that the long axis of the ventricle is approxi-
onds/frame is preferred for resting EF. A frame rate mately vertical, with the apex pointing down
of less than 30 milliseconds/frame is preferred if and left ventricle on the right side of the image.
ejection and filling rates are to be computed. For rest Caudal tilt of the LAO view, typically 10° to
studies, a minimum of 16 to 19 frames per cardiac 15°, is helpful to separate the atria from the
cycle is recommended, with 24 or 32 being preferred ventricle and may be particularly useful in
for EF calculation and for calculation of ejection and patients with vertical hearts. The degree of
filling parameters. For stress studies (R-R ⬍600 caudal tilt is limited by the detector yoke
milliseconds), 32 frames/cycle is preferred for EF suspension and the necessity to keep the cam-
calculation and calculation of peak ejection and era face as close as possible to the chest. As an
filling rates. alternative to achieve a 10° to 15° caudal tilt, a
9. Count density. Studies containing approximately slant-hole collimator may be used, if available.
20,000 counts/cm2 (1,800 counts/pixel for a 3-mm When using a caudal tilt, depending on the
pixel) over the center of the left ventricle at rest imaging conditions, the LV and left atrial
using high-resolution collimation, or about 40,000 separation still may not be apparent. This may
counts/cm2 at rest with the higher-sensitivity col- result in inclusion of more left atrium than
limators (LEAP or high sensitivity), as used in desired, if the atrial-ventricular border is diffi-
rest/exercise studies, are preferred (which corre- cult to discern, as the superior aspect of the
sponds to about 3-4 million counts in a 15 ⫻ ROI may encroach into the left atrium. If the
15– cm FOV with a high-resolution collimator). atrial-ventricular border is not evident, use the
This count density is measured by summing all standard LAO view. On a correctly angulated
images in the gated series together and determin- LAO projection, the photopenic area of the
ing the count density from a small ROI at the septum is more or less vertical. To the left is
center of the left ventricle in the left anterior the foreshortened right ventricle. The right
oblique (LAO) projection. Total counts are not a atrium is partially hidden behind and superior
reliable indicator, as they depend too strongly on to the right ventricle. At the viewer’s right, the
the FOV. Nevertheless, a practical rule has been to LV blood pool is well isolated from the sur-
acquire at least 200,000 counts per image frame rounding structures by the myocardium as a
for a 16-frame resting study using a high-resolu- photopenic halo. The appendage of the left
tion collimator. Similarly, acquisition time will atrium can occasionally be seen superior to the
e60 Corbett et al Journal of Nuclear Cardiology
Equilibrium radionuclide angiocardiography November/December 2006

Figure 1. Correct positioning to image left ventricle.8 A, LAO images that are optimal, too anterior,
and too lateral in obliquity. B, In an optimized LAO view the axis of the left ventricle should be
vertical. In images that are too anterior, there is a rightward tilting of the axis from base to apex. In
images that are too laterally positioned, there is a leftward tilting of the axis from base to apex.
(Reproduced with permission from DePuey.8)

left ventricle, usually separated from the LV quently completely obscured by the right ven-
blood pool by the photopenic area of myocar- tricular (RV) blood pool. Further structures that
dium. As a rule, LV blood pool is projected as can be evaluated on the anterior view are the
nearly a “short-axis” view. In other words, one pulmonary artery and the ascending aorta. The
looks “down the barrel” of the left ventricle. In myocardial segments typically visualized in the
some patients, the heart is vertically oriented. anterior view are the basal anterolateral, mid
This anatomic variant can be recognized be- anterolateral, apical, mid inferoseptal, and
cause of the elongated shape of both ventricles basal inferoseptal.
and visualization of both the right atrium and c. Lateral view. The lateral or left posterior
left atrium superior to the ventricles. The pul- oblique (LPO) ⫹45° more lateral (with patient
monary artery and ascending aorta can also be in right side cubitus) than the LAO angle
evaluated in the LAO view. The myocardial selected. These views are best acquired with
segments typically visualized in the LAO view the patient lying on his or her right side. An
are the septal, inferoseptal, inferoapical, infero- optimal left lateral view shows the long axis of
lateral, and lateral. The two other preferred the left ventricle. “Long axis” is defined as the
views are as follows: longest dimension from valve plane to apex.
b. Anterior view. The anterior view ideally Because there is individual variation among
should be ⫺45° more anterior than the LAO patients, either the straight left lateral or the
selected. On the anterior view, the border- LPO view may show the long axis of the left
forming contour on the right side of the heart ventricle best. In the left lateral projection, the
(left side of image) is the right atrium. To the left ventricle is superimposed on activity of the
viewer’s right, the border-forming contour of right ventricle. Anterior and superior to the left
the heart on this view is the left ventricle. ventricle, the RV outflow tract and the pulmo-
Because of the overlying activity of the right nary artery can be noted. The mitral valve plane
ventricle, only the anterolateral wall and apex is often well demarcated by a linear photopenic
can be evaluated. The inferior wall is fre- area caused by attenuation by fatty tissue in the
Journal of Nuclear Cardiology Corbett et al e61
Volume 13, Number 6;e56-79 Equilibrium radionuclide angiocardiography

Table 4. Quantitative parameters of ventricular function–planar ERNA6-8,10,11,16

For information, see
Parameter Method paragraph

LV volume curve Manual ROI at ED and ES Preferred 3

generation
Manual or automatic ROIs at each time point Optional 3
Manual or automatic ROIs at ED only Optional 3
Background Manual at ED Preferred 4
Automatic at ED Standard 4
Automatic or manual at ES Optional 4
LVEF From end-diastolic and end systolic ROIs Preferred 5
From Fourier-filtered curve Optional 5
From single end diastolic ROI (not recommended) Optional 5
LV Wall motion Visual assessment of cine loop Preferred 6a
Phase and amplitude analysis Optional 6b
Principal component or factor analysis Optional 6c
Regional EF Optional 6d
LV emptying Peak rate of emptying Preferred 7a
Average rate of emptying Optional 7b
Time to peak emptying rate Optional 7a
LV filling Peak rate of filling Preferred 8
Average rate of filling Optional 8
Time to peak filling rate Optional 8
LV volumes Counts based Optional 9
Geometric based Optional 9
RVEF Not widely accepted at equilibrium Optional 10
Heart/lung ratio Counts based LV/lung Optional 13

ED, End diastole; ES, end systole.

atrioventricular groove. Posterior to the left hottest pixel within the image over all time points. In
ventricle are the left atrium and the descending the presence of intense extracardiac activity, radio-
aorta. The spleen is usually visualized in the activity in the heart is at the darker (lower) end of the
right lower corner of the image. The myocar- gray scale and may be almost invisible. Rather than
dial segments typically visualized in the lateral using lead shielding, normalization of the cardiac
or LPO view are the basal anterior, mid ante- image to the hottest pixel within the heart usually
rior, apical, mid inferior, and basal inferior. deals adequately with this display problem. Alterna-
tively, the extracardiac activity may be subtracted or
D. Assessments of Ventricular Function–ERNA at “masked out.”
Rest 2. Smoothing.14,15 The smoothing process is designed
to remove statistical fluctuations from image data by
Image Display and Quantification modifying individual data points within the image.
1. Display.6,8,10,11 Multiple-view ERNA (LAO, ante- Multi-frame digitized ERNA data are often tempo-
rior, and left lateral or posterior oblique projections) rally and spatially smoothed. For temporal smooth-
are usually displayed simultaneously as endless-loop ing, pixels are modified by weighted averaging of
movies in quadrants of the computer screen. The data from preceding and following frames in time,
display should visualize the entire heart and its usually 3 to 5, with the center pixel replaced with
surroundings. ERNA images are best viewed by use this average value. For spatial smoothing, pixels are
of a linear gray scale. Color display is strongly modified by weighted averaging of counts from a
discouraged. Occasionally, intense extracardiac ac- group of neighboring pixels within the same image,
tivity may cause a problem with image display. usually 9, with the center pixel in the group replaced
Computer images are usually normalized to the with this weighted average value. This is referred to
e62 Corbett et al Journal of Nuclear Cardiology
Equilibrium radionuclide angiocardiography November/December 2006

as Gaussian nine-point weighted smoothing. The and that all time points may be averaged for good
exact number of temporal or spatial points used for statistics. An ROI adjacent to the end-systolic border
the smoothing will depend on the number of time can be used to estimate background, but care must
points acquired and the acquisition resolution. Im- be taken to use only those time points that do not
ages of adequate count density rarely require spatial include LV activity. A rule of thumb is that the
smoothing. background count rate/pixel is typically 30% to 70%
3. LV volume curve generation.17-20 Most parameters of the end-diastolic LV counts, and exceptions to
describing ventricular function are extracted from a this rule occur infrequently.
complete LV volume curve or time-activity curve 5. LVEF.6-8,13,16 Many methods are commercially
(TAC). This curve can be obtained either from a available for computation of LVEF. In general,
single ROI drawn at end diastole (and modified at LVEF is calculated based on the assumption that LV
end systole, if necessary, to include any dyskinetic volumes throughout the cardiac cycle are propor-
regions) or preferably using multiple ROIs drawn at tional to LV counts. LV counts at end diastole and at
each time point. ROIs should be edited on an end systole or throughout the cardiac cycle are
amplitude or difference image to exclude overlap- measured by constructing LV ROIs. The measured
ping atrial counts. Manually drawn ROIs are the LV counts within these LV ROIs are corrected for
most consistently accurate, though time-consuming. background scatter. Background is measured using
Many automatic techniques exist for drawing ROIs. ROIs constructed adjacent to the lateral or inferoapi-
It is important that the resulting ROIs be checked cal walls of the ventricle. LV ROI counts are
visually and altered manually if necessary. Irregu- corrected for background by subtracting background
larities in LV contours occasionally occur using counts from LV ROI counts. This is referred to as
automatic algorithms, especially for exercise studies, background correction (BkCorr). LVEF then is cal-
and for ROIs drawn near end systole. These irregu- culated using the usual equation: LVEF ⫽ ([End-
larities can have significant effects on parameters diastolic volume (EDV) ⫺ End-systolic volume
extracted from the LV curve. If EF only is to be (ESV)]/EDV) ⫻ 100 or, as applied to ERNA, LVEF
determined, the preferred method (and the simplest) ⫽ ([BkCorr end-diastolic counts ⫺ BkCorr end-
for LV volume curve generation is from manually systolic counts]/BkCorr end-diastolic counts) ⫻
drawn ROIs over end diastole and end systole, with 100.
the volume curve being processed by weighted Three basic approaches are commonly employed.
interpolation of curves from end-diastolic and end- (a) LVEF can be computed by selecting only the
systolic ROIs (weighted to end diastole near the end-diastolic and end-systolic frames and construct-
beginning and end of the curve and weighted to end ing the appropriate LV and background ROIs. This
systole at curve minimum). If ejection and filling is just as effective as using all frames from the entire
rates are to be computed, ROIs drawn on all frames cardiac cycle, and if regions are being constructed
are preferred. NOTE: Single ROI definition at end manually, this greatly decreases processing time. No
diastole may underestimate EF. Optionally, auto- matter what the approach, care must be taken to
matic edge detection may be used, if each frame is ensure that LV ROIs are appropriately fitted to
reviewed, and the ROI corrected, if necessary. include all LV counts, but only LV counts (ie, LV
4. Background. Background is critical for the mea- ROIs) should not extend to include the activity from
surement of many LV parameters. Usually an ROI 5 any adjacent structures such as the left atrium above,
to 10 mm away from the end-diastolic border, drawn right ventricle to the left, an adjacent intensely active
from approximately 2 o’clock to 5 o’clock, is used, spleen below and to the right, or a grossly ectatic
although the exact location used is less important descending aorta immediately to the right. Automat-
than consistent placement, ensuring that atrial ically generated LV ROIs may need to be edited or
counts, counts from the spleen or descending tho- redrawn by the operator if they do not tightly fit the
racic aorta, LV counts, or a gastric air bubble is apparent LV boundaries. This is especially the case
excluded. With automatic routines, visual verifica- at end systole, where automatically generated ROIs
tion of the background ROI is essential. A visual often do not fit well, extending irregularly beyond
examination of the TAC produced from the back- the ventricle frequently including the aortic root to
ground ROI (it should be flat) is useful to determine the left and beyond the apex above. If the end-
if LV activity or atrial activity is spilling into the systolic LV ROI is not correctly drawn, the calcu-
background ROI. If the background curve is flat, this lated LVEF will generally be underestimated. (b) If
indicates that the background ROI has not been all frames throughout the cardiac cycle are employed
positioned over any periodically beating structures to generated an LV TAC, one optional method that
Journal of Nuclear Cardiology Corbett et al e63
Volume 13, Number 6;e56-79 Equilibrium radionuclide angiocardiography

can be used fits two or four Fourier harmonics to the motion image, reducing the appearance of noise in
LV TAC, extracting the first and the minimum the cine. However, the effects of PCA analysis on
points, as the end-diastolic and end-systolic count the assessment of regional wall motion have not
values. Because this method uses the entire curve, it been completely evaluated. (b) Phase and amplitude
reduces statistical fluctuations, even for very short images have been reported to be of use for the
(stress) acquisitions. If the first point of the filtered detection and quantification of wall motion defects.
curve is used, one must ensure that the ECG gate is It is preferred that this method supplement, not
correctly set up, with gating occurring no later than replace, visual assessment of the cine loop. Analyses
the peak of the R wave, preferably during the of the phase image and phase histogram, as well as
upslope of the R wave to ensure that the first point visual assessment of a dynamic phase image, have
truly corresponds to end diastole. If the R-wave both been reported to be useful in reducing the
trigger is not precisely at end diastole (see section subjectivity associated with visual assessment of
I.C.5, Bad beat/beat length window [arrhythmia wall motion defects using cine loops. Also, in
rejection]), then the maximum value of the filtered patients with conduction abnormalities, phase anal-
curve can be used to identify the end-diastolic ysis has proven useful in identifying the pattern,
counts. Fourier fitting may fail to provide reliable location, and/or point of origin of arrhythmic foci.
results if there is “drop off” in TAC counts at the end (c) PCA or factor analysis creates a mathematically
of the curve due to heart rate variability from sinus derived set of functional images expressing signifi-
arrhythmia or other arrhythmias varying the R-R= cant motion components in the image. Displayed as
interval. This effectively reduces the time of acqui- amplitude images and associated time signatures,
sition of frames at the end of the diastolic phase of they may add to the assessment of regional wall
the cardiac cycle and the corresponding TAC. Fou- motion and can be viewed in a similar fashion to
rier filtering should be applied only to curves that phase and amplitude images or applied to process
have no drop off or that have been corrected for drop the cine-loop display (see above). This method is
off. (c) EF can also be computed from the LV counts clearly a supplement at this point and should be used
produced from a single end-diastolic ROI, applied to in conjunction with standard cine-loop assessment of
both the end-diastolic and the end-systolic images. wall motion. (d) Regional EF (ie, dividing the left
In this case, the EF will be consistently underesti- ventricle into 6-8 subregions and applying the con-
mated compared to the above methods. The single ventional formulation for EF) has been reported to
end-diastolic LV ROI approach is discouraged. aid in the assessment of regional or segmental wall
6. LV wall motion.14,15,21-27 (a) Visual assessment of motion. It is preferred that this method supplement,
all three standard views is preferred. It is critical that not replace, visual assessment of the cine loop.
the cine loop consist of approximately 12 to 16 Cardiac rhythm and conduction.8,13,28,29 Because
frames (fewer frames may lose temporal informa- the ERNA is formatted and displayed as an endless-
tion, greater may compromise statistics). If the loop cine of a single representative beat, the
acquisition is performed with more than 16 frames “rhythm” always looks regular. Abnormalities of
for improved temporal resolution, it is imperative rhythm can only be discerned by the relationship of
that the frames be appropriately added/recombined atrial to ventricular contraction. The most common
or filtered before visual assessment. Spatial filtering sustained disturbance of rhythm is atrial fibrillation.
and temporal filtering are often employed in cine- Atrial fibrillation or flutter can be assumed to be
loop presentations. Spatial filtering is typically a present when no atrial contraction is detected. Oc-
9-point spatial smooth and reduces the apparent casionally, one may diagnose flutter or atrial tachy-
quantum mottle or image noise. Temporal filtering cardia by a difference between the atrial and ven-
typically weights the current frame two parts and the tricular contraction rates—that is, 2 or 3 atrial
previous and post frames one part to recreate the contractions to each ventricular contraction. A pace-
cine-loop image with 16 frames or more. Both types maker rhythm is usually apparent because the LV
of filtering tend to make the cine loop appear more activation starts at the apex, and the wave front of
visually pleasing. If rest and stress are to be com- contractions proceeds to the base. Left bundle
pared, it is preferred to show both cine loops branch block can be diagnosed by the typical para-
simultaneously. Optionally, modern application of doxical pattern of septal motion.
principal components analysis (PCA) or factor anal- 7. LV emptying.12,30 (a) The maximal LV emptying
ysis can create a mathematically derived cine loop rate is determined by measuring the peak slope of
which separates various types of wall motion (atrial, the LV curve, expressed in units of EDV/s. The
ventricular), producing a more visually pleasing counts in the end-diastolic ROI are used to represent
e64 Corbett et al Journal of Nuclear Cardiology
Equilibrium radionuclide angiocardiography November/December 2006

the EDV, and the counts in subsequent frames are (tPFR) can also be measured from the LV TAC and
referenced to this value to compute EDV/s. The time is expressed in milliseconds. As with the PFR, the
to peak emptying (from the end-diastolic frame) may tPFR also varies from laboratory to laboratory but on
also be computed and expressed in milliseconds. average should be expected to be less than 180
Measurement error in the down slope of the LV milliseconds. The relative contribution of atrial fill-
volume curve is greatly amplified by statistical noise ing to LV filling may be quantified as the ratio of the
in the unprocessed (unfiltered) curve. For this rea- atrial peak to the peak of the rapid filling phase on
son, either the LV curve is generally first filtered or the first derivative curve. Ratios of less than 1:4 are
a small region of the curve is fitted to a polynomial, normal but may increase with aging.
or other similar techniques are employed to mini- 9. LV volumes.36-40 Acceptable results have been
mize noise without distorting the value of slope. reported in the literature using both counts-based
Measurements of peak emptying at exercise are and (to a lesser extent) geometrically based methods,
often considered too heart rate– dependent or statis- although counts-based methods are preferred. Geo-
tically inadequate to be of clinical use. When com- metric methods are based on the standard “area-
puting peak or maximal emptying rate, 32 frames length” methods and are hampered by the limited
per cardiac cycle is preferred to ensure accurate spatial resolution of ERNA. Both the counts-based
assessment of maximal rate. (b) The slope of a line and geometric methods may produce highly inaccu-
connecting the end-diastolic and end-systolic points rate results unless extraordinary attention is paid to
can be used as a measure of the average LV methodologic detail. These methods are not widely
emptying rate. Alternatively, methods that depend used. Assessments of LV volume are affected by
on the time it takes for the left ventricle to empty one photon attenuation and Compton scatter. Counts-
third (or any other arbitrary fraction) of the way based methods include the “aortic arch” method or
from end diastole to end systole have been reported. methods involving blood draws and calibration of
8. LV filling/diastolic function.8,12,31-35 The same the counts-to-volume ratio. The latter method is
techniques described in paragraph 7 above can be highly influenced by photon attenuation. Calculation
used to measure diastolic filling rates. All of the of absolute volumes is not recommended, except for
same considerations mentioned above for emptying laboratories that have the ability to independently
also hold for filling. Note that the gating require- validate their methodology.
ments for adequate representation of diastolic pa- 10. RVEF.41-44 Because of overlap with other cardiac
rameters are more stringent than for systolic ejection chambers, ERNA is not the procedure of choice for
parameters, due to data drop off at the end of the measurement of RVEF. Frequently, there is overlap
cardiac cycle. of right atrial activity during RV systole, which will
Qualitative. Visual analysis of the shape of the lead to an erroneously low calculation of RVEF.
LV TAC is frequently sufficient to detect gross Overlap of the right atrial activity with the right
abnormalities of diastolic filling. Prolongation of ventricle may be partially circumvented by acquiring
isovolumic relaxation, a delay in the onset of rapid a separate shallower LAO, about 20° LAO, chosen
filling, a decrease in the slope of the rapid filling to optimize the separation of the right ventricle from
phase, or an exaggerated contribution of atrial con- both the right atrium and left ventricle. Although
traction to LV filling may be readily apparent and optimal separation is generally impossible, im-
should be noted. Such findings are typical of hyper- proved separation usually is relatively easily accom-
trophic ventricles. Aging, pericardial disease, and plished. Rotating slant-hole collimators, if available,
restrictive myocardial disease are also associated may be quite useful in optimizing atrioventricular
with changes in the pattern of filling and a decrease separation. Either true first-pass or “gated first-pass”
in the rate of filling. radionuclide angiocardiography is the preferred ap-
Quantitative. Peak diastolic filling rate can be proach. Both of those techniques yield RVEF values
quantified from the first derivative of the diastolic that are higher than those measured on a standard
portion of the LV TAC. To obtain reliable values for ERNA study. The lower limit of normal with these
diastolic filling, the LV volume curve should have methods is 0.40.
sufficient temporal resolution. Values for normal 11. RV size. RV size is best evaluated in the anterior
studies vary from laboratory to laboratory, but a view. Assuming that ERNA images are routinely
generally accepted lower limit of normal for the acquired with the same gamma camera and same
peak diastolic filling rate (PFR) is 2.50 EDV/s. PFR zoom factor, abnormal enlargement of the right
tends to decrease with age in otherwise healthy older ventricle can be identified by visual inspection and
subjects. In addition to the PFR, the time to PFR mental comparison to normal studies. There is no
Journal of Nuclear Cardiology Corbett et al e65
Volume 13, Number 6;e56-79 Equilibrium radionuclide angiocardiography

reliable quantitative measurement method for the 20% during exercise and concomitant decreases in
RV volume with ERNA. ESV.
12. RV regional wall motion. RV regional wall motion 4. LVEF and RVEF changes from rest.12,43,47,48,53-57
is best assessed by use of the information from both Both LVEF and RVEF typically increase during
the anterior and LAO views. Any single view may exercise. Many authors have suggested that the nor-
be inadequate. Regional wall motion is usually mal response is an increase of at least 5%, or 5 EF
qualitatively graded as normal, mildly hypokinetic, units. That criterion is based on the reproducibility of
severely hypokinetic, akinetic, or dyskinetic. the ERNA measurements of LVEF. Nevertheless, this
13. Heart/lung ratio. Optionally, heart/lung ratio can be criterion does not hold true under all circumstances.
computed. The ratio of the counts in the cardiac For instance, with increasing age, the ability to
blood pool to the counts in the lung can be useful to augment LVEF during exercise decreases. The type
assess ventricular compensation. Pooling of blood in of exercise protocol, the subject’s gender, acquisition
the lungs has been reported to be indicative of LV during submaximal exercise, isometric exercise,
failure. markedly hypertensive responses to exercise, and
coexisting non-coronary heart disease may all alter
the response of the EF to exercise. Consequently, a
E. Assessment of Ventricular Function During
failure to increase LVEF of at least 5% is a sensitive
Exercise and Interventions
but very nonspecific criterion for diagnosis of coro-
Image Display and Quantification7,8,11,45,46 nary heart disease. In the absence of an exercise-
induced regional wall motion abnormality, changes in
1. Display. ERNA images acquired at baseline and LVEF alone are nonspecific. Coupled with a large
during exercise or pharmacologic interventions (⬎20%) increase in EDV during exercise, a drop in
should be displayed side by side on quadrants of the LVEF with normal regional wall motion during ex-
computer screen for evaluation of changes between ercise should be viewed as highly suspicious for
the two sets of images. coronary artery disease.
2. Regional wall motion changes from rest.45,47-49 A more important parameter is the absolute level of
One should expect an increase in regional excursion LVEF at peak exercise. Even if angiographic coro-
during exercise, during inotropic stimulation, and nary artery disease is documented, a peak exercise
during administration of afterload-reducing agents LVEF greater than 50% indicates a favorable prog-
such as nitroglycerin. The standard approach to de- nosis. It is important to ensure that acquisition of
tection of changes in regional wall motion between radionuclide data is performed during peak exercise.
two studies is to visually assess the change on the In most patients, with and without significant disease,
side-by-side display. A somewhat more rigorous ap- a significant increase in LVEF can be noted immedi-
proach is the semiquantitative method in which ven- ately after discontinuation of exercise. Abnormalities
tricular segments are assigned scores where 4 is in RVEF during exercise are most often seen in
normal, 3 is mildly hypokinetic, 2 is moderately patients with chronic pulmonary diseases and in
hypokinetic, 1 is severely hypokinetic, 0 is akinetic, particular in those with pulmonary hypertension. Pa-
and ⫺1 is dyskinetic. A significant change in regional tients with cardiomyopathy and bi-ventricular dys-
wall motion between two studies is defined as a function or proximal right coronary artery stenoses
change in score of 2 or greater. may also demonstrate abnormal RV responses.
3. Chamber size changes from rest.50-52 During inter- 5. Comparison to previous studies and correlation
ventions, changes in chamber size may occur. A mild with clinical data.45,55,58 When a patient has under-
increase in end-diastolic volume is normal during gone previous radionuclide studies, the results of this
physical exercise, especially with patients in the study should be compared with the previous ones.
upright position.50 During dobutamine stress, a de- Ideally, one should display the old and new studies
crease in LV chamber size may be observed. These side by side. Serial LVEF data are particularly impor-
changes may be too small to be appreciated by visual tant in patients undergoing chemotherapy for cancer
analysis but can be quantified either as a relative and also in patients with heart failure, myocarditis, or
change (from decay-corrected count changes) or ab- cardiomyopathy or after undergoing transplantation.
solute volume change. Visually, only moderate to For this reason, it is helpful to record and reproduce
severe dilation of the ventricles should be reported. the camera angles at which the three planar cardiac
Such marked volume changes are almost always views are acquired for each study. Interpretation of
abnormal. If volume is measured quantitatively, one ERNA data may be performed without knowledge of
should expect increases in LV EDV of only 10% to the clinical data; however, once an initial interpreta-
e66 Corbett et al Journal of Nuclear Cardiology
Equilibrium radionuclide angiocardiography November/December 2006

tion is made, the interpreting physician should always septal LAO, anterior, and lateral projections have
review the available clinical information to avoid been acquired. In patients with severe obstructive
obvious misinterpretation and to guarantee that the airways disease and patients with congenital heart
interpretation appropriately addresses the clinical disease, often previously undiagnosed, to mention
question that prompted the study. only a few causes of significant variability, size,
6. Study quality. Poor-quality studies cannot be inter- position, and rotation can differ greatly from the
preted with confidence and a high degree of repro- expected. The best septal projection can on rare
ducibility. Studies can be subjectively graded as (a) occasion be as shallow as straight anterior or as steep
excellent, (b) average, (c) suboptimal but interpret- as 10° to 30° LPO. If the technologist has not
able, and (d) uninterpretable. Placing such a designa- carefully identified and noted the angulation of the
tion in the report communicates a level of confidence best septal LAO projection and the standard “ante-
in the data that is helpful to recipients of the report. It rior” (⫺45°) and “lateral” (⫹45°) projections, the
may also be used to screen studies from inclusion in quantification of ventricular function will be signifi-
research data. If possible poor-quality (suboptimal cantly impaired and the interpretation of segmental
and uninterpretable) studies should be repeated but, if function of the left and right ventricles hampered by
for some reason, they cannot be, the subjective grade either chamber overlap or misidentification of seg-
of study quality will at least alert the referring ments or both. For example, if the interpreting phy-
clinician to the limited reliability of the reported sician does not note that the rotation of the heart in the
results. sagittal oblique plane is quite steep in a patient with
7. Type of exercise or intervention protocol.50,59 The severe chronic obstructive pulmonary disease and
type of exercise should be specified in the report: depressed diaphragm, that individual may interpret
physical exercise on treadmill or supine or upright motion at the base of the left ventricle in the LAO
bicycle. The exercise protocol should be stated. Ex- projection as anterior wall motion, as is more com-
ercise ERNA studies are performed with bicycle monly the case in patients without overinflated lungs,
ergometers, and the levels of stress during each image rather than motion of the mitral valve plane. In
acquisition are reported in watts or kilogram-meters patients with pulmonary hypertension the right ven-
of work and duration. Treadmill protocols such as the tricle is often greatly dilated and, with this, the septum
Bruce, modified Bruce, and Naughton protocol are is often rotated to a steep LAO, lateral, or LPO
not used with exercise ERNA. For pharmacologic projection. If the technologist does not recognize and
intervention, the generic name of the drug (eg, dobut- note this during image acquisition, chamber overlap
amine) and maximal dose (eg, 40 mg · kg⫺1 · min⫺1) may be so severe that the left ventricle cannot be
infused should be stated. Furthermore, whether drugs assessed at all.
were administered to either enhance or counteract the 2. Attenuation artifacts. As is the case for myocardial
effect of the pharmacologic stressor should be re- perfusion imaging, breast attenuation may also affect
ported. ERNA imaging. On the LAO view, the entire heart
8. Symptoms, heart rate and blood pressure re- may be in the “shadow” of the left breast. At times,
sponse, ECG changes, and endpoint of stress. this may give the illusion of a halo around the heart
Within the report of the radionuclide study, a succinct and suggest pericardial fluid. Lack of “swinging” of
description should be given of important clinical the heart and the typical configuration of the shadow
parameters: duration of exercise or stress protocol, may provide clues for the artifact.
baseline and peak stress heart rate, maximal workload 3. Activity outside the heart and great vessels. Any
(in metabolic equivalents when applicable), baseline vascular structure (tumor, etc) with a sufficient vol-
and peak stress blood pressure, symptoms during test, ume of red blood cells can be visualized by ERNA
(re)production of symptoms and chest pain, and ECG imaging. Therefore it is important to be attentive for
changes compared with baseline. any unusual radioactivity outside the heart and great
vessels and seek clinical correlation. Free Tc-99m–
pertechnetate accumulates in the thyroid gland and
F. Image Analysis/Interpretation7,8,10,11,39
gastric mucosa.
1. Overall cardiac assessment. The initial assessment 4. Chamber sizes.
of ERNA studies should include an overall general a. LV size. One can qualitatively assess the relative
assessment of the size, position, and rotation of the size of various cardiac chambers. This assumes
cardiac blood pool (heart) and proximal great arteries. that the same camera and magnification are used
In most patients there is not a great deal of variability routinely. Because in many patients the right
in regard to position and rotation if the standard best ventricle is normal in size and function, RV
Journal of Nuclear Cardiology Corbett et al e67
Volume 13, Number 6;e56-79 Equilibrium radionuclide angiocardiography

end-diastolic size may serve as a benchmark for rounded by a thick photopenic area—that is, the
qualitative assessment of the relative size of other hypertrophied myocardium. During systole, there can
cardiac structures. On a normal study, the right be almost complete LV cavity obliteration.
ventricle is usually somewhat larger than the left 6. Pericardial space. Pericardial fluid accumulation can
ventricle and the RV inferior wall and apex extend be identified on ERNA studies. When a large amount
below the left ventricle. A normally sized left of fluid is present, a photopenic area surrounds the
ventricle “fits” within the crescent of the right heart, extending up to the roots of the large vessels.
ventricle on the LAO view. The presence or On the cine display, a swinging motion of the heart
absence of marked LV hypertrophy (LVH) can be can be appreciated. When estimating the extent of the
estimated by qualitative assessment of the thick- photopenic area around the ventricular blood pool,
ness of the septum. The septum is well delineated one should account for both the thickness of the
by RV and LV blood pool, and thus the thickness myocardium and the presence of epicardial fat before
of the myocardium can be assessed. In severe deciding that pericardial fluid is present. Conse-
LVH, a thick photopenic halo typically surrounds quently, small amounts of fluid are impossible to
the LV blood pool. distinguish from normal variants. The shadow of a
b. RV size. RV size is best evaluated in the anterior large overlying breast may, particularly in the LAO
view. Assuming that ERNA images are routinely view, mimic pericardial fluid. Only swinging motion
acquired with the same gamma camera and same of the heart is a certain sign of a large amount of
zoom factor, abnormal enlargement of the right pericardial fluid. The preferred technique to assess
ventricle can be identified by visual inspection and pericardial fluid continues to be echocardiography.
mental comparison to normal studies. There is no
7. Size of pulmonary artery and aorta. The pulmonary
reliable quantitative measurement method for the
artery and the ascending and descending aorta can
RV volume with ERNA.
also be evaluated visually on good-quality ERNA
c. Atrial sizes. The right atrium forms the left lower
studies. Only qualitative assessments, such as dilation
border on the anterior view of the cardiac image.
of the pulmonary artery and dilation and tortuosity of
Size and contraction of the right atrium can be
the ascending aorta, aortic arch, or descending aorta,
evaluated in this view during ventricular systole.
can be made. The three conventional views allow for
The left atrium is best evaluated on the left lateral
visual assessment from different angles.
view during ventricular systole. Because of over-
8. Conclusion. It is important to summarize the results
lying and surrounding radioactivity, frequently no
clear outline of the left atrium is present. However, of the test as either “normal” or “abnormal.” Equiv-
the general size and contractility of the left atrium ocal statements should be avoided if at all possible. In
usually can be appreciated. The size of the left addition, the report should reflect the degree to which
atrium should be judged in comparison to the long the test is abnormal: “markedly,” “moderately,” or
axis of the left ventricle. The contraction of the left “mildly” abnormal. Finally, a comparison should be
atrium is appreciated as a change in count density made to previous results, if applicable. A serious
during ventricular diastole and is sometimes dis- attempt should be made to provide an answer to the
cernible on radionuclide studies. clinical questions and indication for study.
Atrial contraction does, however, occur at the
very end of the acquisition cycle. Atrial contrac- II. SINGLE PHOTON EMISSION COMPUTED
tion is shorter than the ventricular cycle. In the TOMOGRAPHY IMAGING
presence of ventricular ectopy or irregular rhythm,
the last frames have lower count density, resulting A. Purpose
in “flicker” of the endless-loop cine. During pro-
cessing, one or two frames at the end of the cycle Single photon emission computed tomography
may be “cut off’” for aesthetic reasons. As a result, (SPECT) ERNA is used to determine global and regional
atrial contraction may no longer be evaluable. measures of ventricular function (primarily LV function)
at rest and/or during pharmacologic intervention. These
5. LVH. The presence of LVH is best assessed in the measures of ventricular function may include evaluations
LAO view as more than normal thickening of the of ventricular wall motion, EF, and other parameters of
septum. This is a subjective evaluation that requires systolic and diastolic function. The following sections
familiarity with the normal appearance of the septum provide a technical description of the techniques to
on ERNA images acquired with a particular gamma acquire and process the data necessary to assess param-
camera. In severe LVH the LV blood pool is sur- eters of ventricular performance.
e68 Corbett et al Journal of Nuclear Cardiology
Equilibrium radionuclide angiocardiography November/December 2006

Table 5. Radiopharmaceuticals used for SPECT ERNA

For information,
see paragraph

Radiopharmaceutical Tc-99m–labeled RBCs Standard 1

Dose 25–35 mCi/70 kg Preferred 1
Labeling method In vivo Not recommended 2
Modified in vivo/in vitro Standard 2
In vitro Preferred 2

RBC, Red blood cells.

B. Radiopharmaceuticals FWHM or better at 10 cm and sensitivity [Tc] of

approximately 4,000-5,000 counts · s⫺1 · mCi⫺1
1. Inject the patient with Tc-99m–labeled red blood
[108-135 counts · s⫺1 · MBq⫺1]) are preferred
cells, with activity of approximately 25 to 30 mCi/70
when using multidetector SPECT systems. LEAP
kg body weight (14-17 MBq/kg) to provide the
collimators (typically 12 mm FWHM at 10 cm and
radioisotope tag. With in vivo labeling or red blood
sensitivity of approximately 10,000 counts · s⫺1 · mCi⫺1,
cells, doses will run at the higher end of this range.1
or approximately 280 counts · s⫺1 · MBq⫺1) may be
2. Labeling methods
a. In vivo or modified in vivo/in vitro methods (eg, preferred when using a single-head SPECT camera.
using 2 to 3 mg stannous pyrophosphate 15 min- 3. Pixel size. A 64 ⫻ 64 matrix size of 16-bit word
utes before injection of the radiopharmaceutical).2 pixels or a 128 ⫻ 128 matrix size of 8-bit byte pixels
b. Commercial in vitro kit.3 is preferred. Acquisition zooms that result in pixels
4.8 to 6.6 mm square in the acquired planar projec-
tion images and comparably sized cubic voxels in
C. Acquisition Parameters and the reconstructed SPECT image sets are standard of
Reconstructions–Gated SPECT ERNA practice. Acquisitions may be performed using ac-
1. Camera. Use of dual-detector (two-head) SPECT quisition zooms as high as 1.75 with large-FOV
cameras in the 90° configuration are the most com- cameras (1.75 for patients with small body habitus)
mon preferred camera setup (90° gantry rotation providing pixels as small as 4 mm square. However,
required), but three-head cameras in 120° configu- it is critical that the entire cardiac blood pool be in
rations are another preferred approach (120° rota- the FOV in all projections or severe truncation
tion). For 360° image acquisitions, three-head artifacts may result. If pixel sizes smaller than 4.8
SPECT systems (120° configuration) are the most mm are desired, a test rotation with observation of
efficient systems for image acquisition (120° rota- the blood pool using a persistence scope in all views
tion acquisitions) followed by dual-head SPECT is recommended. With large-FOV rectangular detec-
systems in a 180° configuration (180° rotation ac- tors, common on most modern and currently avail-
quisitions). Single-head SPECT cameras are the able SPECT systems, a test orbit is only required in
least efficient systems for both 180° and 360° patients with very severely dilated hearts. Most
acquisitions, requiring 180° and 360° rotations, re- quantification programs are validated at only one or
spectively, to acquire a full study. With multidetec- two pixel/voxel sizes, and it is generally recom-
tor SPECT systems, ERNA SPECT can be per- mended that an acquisition zoom (voxel size) be
formed in half the time, about 15 minutes, of a used on all patients in keeping with the image input
3-view planar ERNA series and are highly recom- requirements of the quantification software used.
mended. Dual-headed cameras in the 180° configu- 4. Energy window. 140 keV, ⫾ 15% window.
ration are optional but are not recommended unless 5. Bad beat/beat length window (arrhythmia rejec-
360° image acquisitions are desired (180° rotations tion). The preferred arrhythmia rejection mode is
are required). Single-head SPECT cameras with on-the-fly bad beat rejection. Typically, the standard
180° rotations are optional but are not recom- arrhythmia rejection methods interrupt data acquisi-
mended. Acquisition time with a single-head camera tion if a premature beat is detected outside of the
(180° rotation) is approximately 30 minutes. beat length acceptance window. Rejection of the
2. Collimator. High-resolution parallel-hole collima- short or long beat is typical, with rejection of
tors (resolution of approximately 8-10 mm subsequent beat preferred. The typical beat length
Journal of Nuclear Cardiology Corbett et al e69
Volume 13, Number 6;e56-79 Equilibrium radionuclide angiocardiography

Table 6. Camera and computer setup and reconstruction7,60-64

For information,
Data/method see paragraph

Camera and Computer Setup

Camera Dual-head–90° configuration Preferred 1
Three-head–120° configuration Optional 1
Single head Optional 1
Collimator Parallel–high Resolution (dual Preferred 2
head)
Parallel-LEAP (single head) Preferred 2
Pixel size 4.8-6.6 mm/pixel Standard 3
Energy window 140 keV ⫾15% Standard 4
Beat length window ⫾15%–35% Standard 5
Bad beat rejection Reject beat Standard 5
Reject beat and next beat Preferred 5
Acquisition method Frame mode Standard 6
Frame rate 16 frames/cycle over full cycle Standard 7
16 frames/cycle over partial cycle Optional 7
Number of views 60–64 Standard 8
30–32 Optional 8
30–32 (single head) Standard 8
Time per view 20–30 s Standard 8
40–60 s Optional 8
40–60 s (single head) Standard 8
Acquisition stop mode Time per view Standard 8
Cardiac cycles per view Preferred 8
Acquisition time per view Preferred 8
Rotation 180° Standard 9
360° Optional 9
Gated SPECT Reconstruction
Parameters
SPECT reconstruction: Filtered backprojection or iterative Standard 10
Reconstruction filter Butterworth 0.55 Nyquist cut-off
Order ⫽ 7 Standard 10
Butterworth 0.45 Nyquist cut-off Optional 10
Oblique reorientation Short-axis oblique Standard 11
Long-axis coronal Optional 11
Long-axis sagittal Optional 11

acceptance window for SPECT is the mean R-R= tions assume equal sampling (acquisition time) at
interval ⫾ 15% to 35% but will vary depending on each camera view. If several cardiac cycles are
heart rate and rhythm. This window is generally excluded due to cycle lengths falling outside the
somewhat larger than that used for planar ERNA, as acceptance window, severe reconstruction artifacts
significant arrhythmias may result in poor statistics, can occur including severe streaking of the recon-
which, depending on the acquisition stop mode structed images. For this reason, although not com-
employed, can adversely affect the quality of the monly employed, “acquisition stop for accepted
gated SPECT reconstruction. The beat length win- beats” with normalization of the actual acquisition
dow may require widening in some patients if there time at each projection to a common time (eg, 30
is significant variation in cycle length due to either seconds) and “acquisition stop for accepted time” at
sinus arrhythmia or premature beats. Regarding each projection (eg, 30 seconds) are the preferred
acquisition stop mode, SPECT image reconstruc- stop modes. Both stop for accepted beats and stop
e70 Corbett et al Journal of Nuclear Cardiology
Equilibrium radionuclide angiocardiography November/December 2006

for accepted time ensure virtually identical sampling 8-frame studies. LVEF values may be decreased
at each projection, whereas simple acquisition for significantly if 8 gated frames per cycle are acquired.
camera dwell time at each projection with bad beat The statistics of SPECT acquisitions typically pre-
rejection turned on can result in some projections clude the use of higher frame rates. An alternative to
having no accepted beats (in the worst case) if there higher frame rates per cycle is to acquire 16 frames,
is a run of frequent premature beats or a drift of heart over one half or two thirds of the cardiac cycle, if
rate outside the acceptance window. As with planar systolic ejection parameters and EF only are re-
studies, it is recommended that the ECG trigger quired. However, with this optional acquisition
point be checked to ensure that the ECG gating mode, diastolic function analyses will be compro-
circuitry is gating on the upslope or peak of the ECG mised if not precluded entirely.
R wave. The ECG gate setup should be checked with 8. Number of projections (views) and time per view.
either a dual-trace oscilloscope or strip chart re- When using a dual-head SPECT cameras, 60 or 64
corder output from the ECG gate on each patient. projections (30 or 32 projections per head) over a
Since most gating devices are designed to recognize 180° rotation (right anterior oblique to LPO) at
a rapid increase in QRS voltage, the optimal input to approximately 30 seconds per view are preferred
these devices is an ECG lead that is predominately a (total acquisition time of about 15 or 16 minutes).
large monophasic R wave accompanied by relatively Optionally, 30 or 32 views may be used (15 or 16
small P waves and T waves and an artifact-free views per head) at 60 seconds per view for enhanced
baseline. In most cases if the negative electrode statistics in very large or severely arrhythmic pa-
(typically the right arm lead) is placed just below the tients (total acquisition time of about 15 or 16
right clavicle in the mid-clavicular line, the positive minutes). With 3-headed systems, the total number
electrode (typically the left arm lead) is placed above of projections acquired over a 180° orbit and the
the costal margin 2 to 4 cm below the V4, V5, or V6 acquisition time per projection are the same as for
position so as to avoid overlapping the cardiac FOV, dual-head systems. This will require approximately
and the ground lead is placed in a similar position on 30% more total acquisition time for a 180° study.
the right lower chest, a good signal will be obtained. However, this 180° approach throws away half of
Care should be taken to carefully prepare the elec- the acquired data (ie, in the same time required for a
trode sites with alcohol or other skin preparation 180° study, a 360° study has actually been acquired).
materials to ensure a stable artifact-free signal. In Some laboratories take advantage of all the acquired
difficult cases, a quick review of a standard 12-lead projections and reconstruct with the full 360° pro-
electrocardiogram may be helpful in planning elec- jection data set. This permits reduced time per
trode placement, and usually only repositioning the projection and reduced total acquisition time with
positive electrode will the only required change. The the same or greater total acquired counts for 360°
ECG lead wires should be positioned so that they reconstructions. When using a single-head gamma
have no tension on them and so that neither the camera, 30 or 32 projections at 60 seconds per view
patient nor camera can either displace, snag, or are recommended (total acquisition time of about 30
bump them during acquisition setup and imaging, to 32 minutes).
resulting in interruption of gating and/or artifacts. 9. Rotation. With single- and dual-head SPECT cam-
Bad ECG gating devices including gating devices eras, 180° rotation is preferred; 360° rotation is
with excessive delays between QRS onset and out- optional with either a single- or dual-head camera
put of signal to the camera can adversely affect the but is not recommended unless a three-head SPECT
ventricular volume curve and severely lengthen camera is used.
image acquisition time or result in poor statistics. 10. SPECT reconstruction/filter. Filtered backprojection
ECG gating devices can be checked with commer- is the suggested reconstruction method. Iterative meth-
cially available dynamic phantoms. ods can also be used when available and are required if
6. Acquisition method. Frame mode (forward fram- attenuation correction is performed. Different SPECT
ing) is standard. However, when available, forward- reconstruction filters are preferred by different clinical
backward gating should be considered, especially in sites. The suggested filter for each of the 16 gated
patients with significant arrhythmia where the dia- frames is a Butterworth filter with 0.55 Nyquist fre-
stolic phase of the volume curve and cine displays of quency cutoff and order of 7. If the study is count-poor
wall motion may be severely distorted when only (due to significant arrhythmias, poor tag, or other
forward framing is used. technical reasons), a Butterworth filter with 0.45
7. Frame rate. Sixteen frames per cardiac cycle are Nyquist frequency cutoff and order of 7 may decrease
preferred because of the poor temporal resolution of statistical noise and improve the quality of the recon-
Journal of Nuclear Cardiology Corbett et al e71
Volume 13, Number 6;e56-79 Equilibrium radionuclide angiocardiography

struction. As algorithms become commercially avail- parameters extracted from the LV curve if not
able, iterative reconstruction methods will probably manually or semiautomatically corrected during
become the preferred approach. Temporal filling anal- user quality assurance. Definition of the mitral and
gous to that employed in planar imaging (I.D.2) should aortic valve planes is the most difficult step in this
be applied to raw data or to transverse reconstructed quantification–this is the result of the activities
images of count-poor studies. within the LV, left atrium, and aortic blood pools
11. Oblique reorientation. Preferably, each of the 16 all being generally of similar intensity and sepa-
gated frames’ transverse reconstructions are reoriented rated only by the relatively thin and sometimes
in short-axis oblique slices and, optionally, long-axis poorly defined mitral and aortic valve planes.
coronal slices, most commonly referred to as horizontal 2. Background. Background subtraction is not required
long-axis slices, and long-axis sagittal slices, most and generally not performed for SPECT imaging. The
commonly referred to as vertical long-axis slices. inherently 3D nature of reconstructed SPECT images
Typical 3-dimensional (3D) reconstructions of the and the 3D surface rendering of the LV blood pool
SPECT ERNA data are accomplished using the short- employed in most quantitative methods obviate the
axis oblique data only. However, long-axis views are need for background correction (see paragraph 1
often important when observing regional wall motion above).
by cine-loop display of the oblique reformatted slice 3. LVEF. As discussed in paragraph 3 above, LVEF can
data. See paragraph 5 in the next section. be computed by applying an end-diastolic ROI to
end-diastolic images and an end-systolic ROI to
end-systolic images reconstructed from slices
D. Assessment of Ventricular Function–Gated
summed to include the entire left ventricle (but
SPECT ERNA Imaging
excluding other cardiac chambers). These are essen-
Image Display and Quantification tially adaptations of gated SPECT ERNA data sets to
planar ERNA quantification programs. These are
1. LV volume curve generation.65-70 Most parame- outdated and to be discouraged. Automatic or semi-
ters describing ventricular function are extracted automatic programs which are inherently volumetric
from a complete LV volume curve in a manner and consider the left ventricle as a 3D object are the
similar to that employed for planar ERNA studies. current standard. LV volume curves can be generated
Techniques that obtain this curve from either a and LVEFs can be computed in a manner similar to
single 2-dimensional ROI drawn at end diastole what is done with planar ERNA. Although count-
(and modified at end systole as necessary) or using based methods can be used, with SPECT, volume-
multiple ROIs drawn at each time point over the based methods are preferred. Optionally, a geometric-
summed short-axis slices, which include the entire based method may be used to compute LVEF from
left ventricle but exclude the left atrium, have been EDVs and ESVs. LVEFs obtained from SPECT
used on occasion in the past but are outmoded and ERNA are likely to be higher than LVEF values
should be avoided. Most methods described for the determined from the planar ERNA method due to the
quantification of LV volumes from SPECT ERNA complete removal of all activity from the left atrium.
employ 3D regions encompassing the entire left Preliminary results indicate that SPECT ERNA
ventricle, but only the left ventricle, at each frame LVEFs are approximately 7 to 10 EF units higher than
throughout the cardiac cycle. Like methods used to those determined from planar studies.71 Fitting the
quantify LV volumes from gated SPECT perfusion LV curve with two or three harmonics and extracting
studies, these methods sum the calibrated voxels the maximum and minimum points as the EDV and
and partial voxels within the defined 3D LV volume ESV values are sometimes employed. Volumetric
throughout the cardiac cycle. Unlike planar ERNA methods are generally less affected by variations in
methods that employ TACs where background sampling at the end of the cardiac cycle due to cycle
subtraction is required, SPECT methods are inher- length variation. Therefore, the end-diastolic frame
ently volumetric and do not typically employ back- may occur at either the beginning or the end of the
ground subtraction. Several automatic or semiauto- cardiac frame cycle (frame 14, 15, or 16). If a
matic methods have been described that in most non-volumetric activity-based method is employed
patients are quick and accurate, although it is and the original TAC is produced from a single
important that the automatic results be checked end-diastolic ROI, the EFs will be consistently lower
visually and modified as necessary. Irregularities in than if the TAC is produced from multiple ROIs. In
the LV contour occasionally occur using automatic this case, SPECT ERNA LVEF values may be com-
algorithms and can have significant effects on the parable to or less than multi-ROI planar ERNA
e72 Corbett et al Journal of Nuclear Cardiology
Equilibrium radionuclide angiocardiography November/December 2006

Table 7. Quantitative parameters of ventricular function-gated SPECT ERNA imaging7,63,64,84

For information,
Parameter Method see paragraph

LV volume curve generation Automatic ROI’s at end diastole or at each Preferred 1

time point
Manual ROI at end diastole Standard 1
Background None Standard 2
LVEF Counts-based 2D/3D method Preferred 3
Geometrically based method Optional 3
Automated method Optional 3
From single end diastolic ROI Optional 3
RVEF Not well validated at this point Optional 4
Wall motion 3D visual assessment of movie loop Preferred 5
Slice-based visual assessment of movie loop Optional 5
Regional EF Optional 5
LV emptying Peak/average rate of emptying Optional 6
LV filling Peak/average rate of filling Optional 7
LV volumes Counts based Standard 8
Pixel based Optional 8

calculations. Care must be used when applying axis slices are standard on most commercial computer
SPECT ERNA LVEF values to the evaluation of systems as part of gated SPECT software packages.
chemotherapy patients where standards have been Wall motion analyses by SPECT ERNA may be the
established using planar methods. A thorough under- most useful application of this technique and can be
standing of the differences between LVEF normal performed on most commercial systems. Optionally,
values between SPECT and planar studies is required. regional EFs can be computed from the segmented
4. RVEF. Unlike planar ERNA studies, accurate com- left ventricle and have been shown to be helpful in
putation of RVEF may be possible with SPECT identifying wall motion defects in patients with cor-
ERNA due to the removal of chamber overlap and the onary artery disease. Alternatively, 2-dimensional
3D nature of SPECT. Automatic volume-based meth- planar projections can be easily generated from the
ods to date have not been as well validated as those 3D data sets to permit viewing of cine images from a
used for the left ventricle. Some studies validating variety of projection angles.
RVEF values from SPECT ERNA have yielded SPECT ERNA has been used for the assessment
positive results, although others have not.72-74 The ventricular activation sequences and the identification
same activity-based techniques described in para- of the sites of AV nodal bypass tracks, as well as LV
graphs 1 and 3 above can be applied for the measure- and RV arrhythmias.77-80 A newer application of
ment of RVEF, but as was the case for the left SPECT ERNA has been the study of activation
ventricle, they are also discouraged for RV analysis. sequences to calculate parameters of ventricular syn-
All of the same considerations mentioned above for chrony such as the site of last activation useful in
the left ventricle also hold for the right ventricle, but guiding resynchronization therapy for congestive heat
since the RV chamber is geometrically more complex failure.81 These uses have been recently reviewed, but
than the left ventricle, final results may vary. unfortunately, these applications currently remain
5. Wall motion.60,65,66,75,76 Regional wall motion in available only to research laboratories since there is
SPECT ERNA may be determined from cine displays no commercially available software that performs
of multiple long- and short-axis slices and from 3D these functions.64,81
displays of the cardiac chambers in cine-loop fashion 6. LV emptying. LV emptying, average and maximum,
(preferred). The 3D displays may be shaded-surface may be computed in similar fashion to planar ERNA
and/or wire-frame displays or, optionally, volume- methods, as can the systolic ejection period.
rendered displays. These 3D images are best dis- 7. LV filling. LV filling, average and maximum, may be
played in multiple cardinal views or rotated under computed in similar fashion to planar ERNA meth-
user control. Cine-loop displays of long- and short- ods, as can the diastolic filling period(s).
Journal of Nuclear Cardiology Corbett et al e73
Volume 13, Number 6;e56-79 Equilibrium radionuclide angiocardiography

8. LV volumes.65-70,82 LV volumes can be accurately standard system measurement on all modern

computed using SPECT ERNA techniques. LV SPECT systems), LV volumes at end diastole and
surface rendering methods are inherently volumet- end systole and throughout the cardiac cycle can be
ric and are less hampered by chamber overlap. easily computed. Although many individual studies
Moreover, the spatial distribution of the left ven- have compared LV volumes to those derived from
tricle is more clearly delineated. The most straight- cardiac contrast ventriculography going back as far
forward approach for LV volume determination as the early 1980s and, more recently, against
that involves no geometric assumptions is to sum cardiac contrast ventriculography and gated mag-
the 3D pixels (voxels) contained within the LV netic resonance imaging (see cautions discussed in
volume of interest, excluding all other structures, paragraph 3 above), as a caution, LV volumes have
and multiply the number of voxels by the calibrated not been validated extensively with any one of the
voxel volume. With the calibrated voxel volume (a commercially available quantitative programs.83

III. EQUILIBRIUM RADIONUCLIDE ANGIOCARDIOGRAPHY: GUIDELINE FOR INTERPRETATION

Table 8. Guideline for interpretation
For information, see paragraph

Planar SPECT

A. Display
1. Quad screen cinematic display Standard I.D.1 II.D.1 and .5
2. Time smoothing Standard I.D.2 II.C.10
3. Spatial smoothing Optional I.D.2 II.C.10
B. Quality control
1. Image quality
a. Statistics–qualitative Standard I.B.2 II.C.10
b. Statistics–quantitative Optional I.C.9 II.C.8 and .10
c. Labeling efficiency–qualitative Standard I.B.2 I.B.2
2. Appropriate imaging angles Standard I.C.10 II.C.8 and .9
3. Appropriate zoom Standard I.C.2 II.C.3
4. Attenuation Standard I.F.2
5. Processing accuracy
a. Ventricular ROIs Standard I.D.3 and .5 II.D.1 and .3
b. Background ROIs Standard I.D.4 II.D.2
c. Volume curve(s) Standard I.D.3 II.D.1
C. Image Analysis
1. Cardiac rhythm and conduction Standard I.D.6 II.C.5 and II.D.5
2. LV size
a. Qualitative Standard I.F.4.a II.D.8
b. Quantitative volume Preferred I.D.9 II.D.8
3. LV regional wall motion
a. Qualitative Standard I.D.6 II.D.5
b. Semiquantitative Optional I.D.6 II.D.5
c. Quantitative Optional I.D.6 II.D.5
4. LVEF Standard I.D.5 II.D.3
5. LV diastolic filling
a. Qualitative Standard I.D.8 II.D.7
b. Quantitative Preferred I.D.8 II.D.7
6. RV size
a. Qualitative Standard I.D.11 and I.F.4.b II.D.4
b. Quantitative Optional I.D.11 and I.F.4.b II.D.4
e74 Corbett et al Journal of Nuclear Cardiology
Equilibrium radionuclide angiocardiography November/December 2006

Table 8. (Continued)
For information, see paragraph

Planar SPECT

7. RV regional wall motion Standard I.D.12 II.D.5

8. RVEF Optional I.D.10 II.D.4
9. Atrial sizes Standard I.F.4.c —
10. Aortic and pulmonary artery sizes Standard I.F.7 —
11. LV hypertrophy Optional I.F.5 —
12. Pericardial space Standard I.F.6 —
13. Activity outside heart and great vessels Standard I.F.3 —
D. Exercise/intervention study
1. Display Standard I.E.1 —
2. Regional wall motion: Changes from rest Standard I.E.2 —
3. Chamber size: Changes from rest Standard I.E.3 —
4. LVEF, RVEF: Changes from rest Standard I.E.4 —
E. Conclusion
1. Correlation with clinical data Standard I.E.5 I.E.5
2. Comparison to previous studies Standard I.E.5 I.E.5

IV. EQUILIBRIUM RADIONUCLIDE ANGIOCARDIOGRAPHY: GUIDELINE FOR REPORTING

Table 9. Guideline for reporting
For information, see paragraph

Planar SPECT

A. Demographic data
1. Name Standard
2. Gender Standard
3. Age Standard
4. Ethnic background Optional
5. Date acquisition Standard
6. Medical record number for inpatient Standard
7. Height/weight body surface area Standard
B. Acquisition parameters
1. Type of study Standard
2. Radionuclide/dose Standard
3. Indication for study Standard I.A. and V. I.A. and V.
4. Study quality Optional I.B.2, I.C.6, and I.E.6 II.C.5 and .10
C. Results: Rest
1. LV size
a. Qualitative Standard I.F.4.a II.D.8
b. Quantitative Optional I.D.9 II.D.8
2. LV regional wall motion Standard I.D.6 II.D.5
3. LV hypertrophy Optional I.F.5 —
4. LVEF Standard I.D.5 II.D.3
5. LV diastolic function
a. Qualitative Standard I.D.8 II.D.7
b. Quantitative Preferred I.D.8 II.D.7
Journal of Nuclear Cardiology Corbett et al e75
Volume 13, Number 6;e56-79 Equilibrium radionuclide angiocardiography

Table 9. (Continued)
For information, see paragraph

Planar SPECT

6. RV size
a. Qualitative Standard I.D.11andI.F.4.b —
b. Quantitative Standard I.D.11 II.D.4
7. RV regional wall motion Standard I.D.12 II.D.5
8. RVEF Optional I.D.10 II.D.4
9. Atrial sizes Optional I.F.4.c —
10. Aortic and pulmonary artery size Optional I.F.7 —
D. Results: Exercise/intervention parameters
1. Type of exercise/intervention protocol Standard I.E.7 —
2. Symptom(s) Standard I.E.8 —
3. Peak heart rate and blood pressure Standard I.E.8 —
4. METS achieved or percent maximum heart rate Optional I.E.8 —
E. Results: Exercise/intervention ERNA data
1. LV size: Change from rest
a. Qualitative Standard I.E.3 II.D.5
b. Quantitative Preferred I.E.4 II.D.3
2. LV regional wall motion: Change from rest Standard I.E.2 II.D.5
3. LVEF exercise Standard I.E.4 NA
4. RV size: Change from rest Standard I.E.3 —
5. RV regional wall motion: Change from rest Standard I.E.2 —
6. RVEF exercise Optional I.E.4 —
F. Conclusion
1. Normal or abnormal Standard I.F.8 I.F.8
2. Diagnostic significance of rest/exercise response Standard I.F.8 NA
3. Prognostic significance of rest/exercise response Optional I.F.8 NA
4. Comparison to previous results Standard I.E.5 I.E.5

NA, Not applicable; METs, metabolic equivalents.

V. GUIDELINES FOR CLINICAL USE OF RADIONUCLIDE ANGIOCARDIOGRAPHY

Table 10. Indications for equilibrium radionuclide angiocardiography (rest and/or exercise)
Planar SPECT

A. Diagnosis of acute coronary end diastole syndromes

1. Suspected ACS in the ED with nondiagnostic electrocardiogram and biomarkers III III
2. Detection of AMI when conventional measures are nondiagnostic III III
3. Diagnosis of STE AMI III III
B. Risk assessment and assessment of prognosis and therapy after STE AMI
1. Rest LV function I I
2. Rest RV function after suspected RV infarction IIa IIa
3. Presence of stress-induced ischemia and myocardium at risk IIb III
4. Detection of infarct size and residual viable myocardium III III
5. Predicting improvement in regional and global LV function after revascularization IIb IIb
C. Diagnosis, prognosis, and assessment of therapy in patients with unstable
angina/NSTEMI
1. Measurement of LV function I I
e76 Corbett et al Journal of Nuclear Cardiology
Equilibrium radionuclide angiocardiography November/December 2006

Table 10. (Continued)

Planar SPECT

2. Identification of ischemia in distribution of culprit lesion or in remote areas IIb III

3. Identification of severity and extent of disease in stabilized patients on medical therapy IIb III
4. Identification of severity/extent of disease with ongoing ischemia and nondiagnostic IIb III
electrocardiogram
5. Diagnosis of myocardial ischemia when history and ECG changes are unreliable IIb III
D. Diagnosis of chronic CAD
1. Assessment of ventricular performance (rest/exercise) I/I I/III
2. Diagnosis of symptomatic and selected asymptomatic patients with myocardial IIb III
ischemia
3. Planning PTCA–identifying lesions causing ischemia, if not otherwise known III III
4. Risk stratification before noncardiac surgery in selected patients IIb III
5. Screening of asymptomatic patients with low likelihood of disease III III
E. Assessment of severity, prognosis, and risk stratification of chronic CAD
1. Assessment of LV performance I I
2. Identification of extent, severity, and localization of ischemia IIb IIb
3. Risk stratification of patients with intermediate risk Duke treadmill score III IIb
4. Assessment of functional significance of intermediate coronary stenosis IIb IIb
F. Assessment of interventions in chronic CAD
1. Assessment for restenosis after PCI (symptomatic) IIb III
2. Assessment of ischemia in symptomatic patients after CABG IIb III
3. Assessment 3-5 y after CABG or PCI in select patients, high-risk asymptomatic IIb III
4. Routine assessment of asymptomatic patients after PTCA or CABG III III
G. CHF
1. Determination of initial LV and RV performance I I
2. RV dysplasia IIa IIb
3. Initial or serial assessment of ventricular function with exercise IIb III
4. Routine serial assessment of LV and RV function at rest IIb IIb
5. Initial and follow-up evaluation of LV function in patients receiving cardiotoxic drugs I IIb
6. Assessment of myocardial viability in patients with CAD and LV dysfunction III III
7. Assessment of the co-presence of coronary heart disease in patients without angina IIb III
8. Detection of myocarditis III III
9. Diagnosis and serial monitoring of hypertensive hypertrophic heart disease IIb IIb
10. Diagnosis of CAD in patients with hypertrophic cardiomyopathy III III
11. Diagnosis and serial monitoring of hypertrophic cardiomyopathy III III
H. After cardiac transplantation
1. Assessment of ventricular performance I I
2. Detection and assessment of coronary vasculopathy IIb IIb
I. Valvular heart disease
1. Initial and serial assessment of LV and RV function at rest I I
2. Initial and serial assessment of LV and RV function with exercise IIb III
3. Assessment of concomitant CAD IIB III
J. Congenital heart disease
1. Initial and serial assessment of LV and RV function I I

Adapted from references 85 and 86.

ACS, Acute coronary syndrome; AMI, acute Myocardial infarction; STE, ST elevation; NSTEMI, non–ST-elevation Myocardial infarction; CAD,
coronary artery disease; PTCA, percutaneous transluminal coronary angioplasty; PCI, percutaneous coronary intervention; CABG, coronary
artery bypass grafting; CHF, congestive heart failure.
Journal of Nuclear Cardiology Corbett et al e77
Volume 13, Number 6;e56-79 Equilibrium radionuclide angiocardiography

The American College of Cardiology, American 19. Jackson PC, Allen-Narker R, Davies ER, Rees JR, Wilde P, Watt
Heart Association, and American Society of Nuclear I. The assessment of an edge detection algorithm in determining
left ventricular ejection fraction using radio-nuclide multiple gated
Cardiology have developed guidelines for the use of acquisition and contrast ventriculography. Eur J Nucl Med 1982;
radionuclide imaging of the heart. Table 10 has been 7:62-5.
adapted from these guidelines. Note that items may be 20. Groch MW, Erwin WD, Murphy PH, Ali A, Moore W, Ford P, et
categorized as class III if not enough data are presently al. Validation of a knowledge-based boundary detection algorithm:
available to substantiate routine clinical implementa- a multicenter study. Eur J Nucl Med 1996;23:662-8.
21. Zaret BL, Strauss HW, Hurley PJ, Natarajan TK, Pitt B. A
tion.85,86
noninvasive scintiphotographic method for detecting regional ven-
tricular dysfunction in man. N Engl J Med 1971;284:1165-70.
22. Pavel DG, Byron E, Bianco JA, Zimmer AM. A method for
References
increasing the accuracy of the radionuclide measurement of ejec-
tion fraction and left ventricular volume curve [abstract]. J Nucl
1. Recalculated dose data for 19 frequently used radiopharmaceuti- Med 1977;18:641.
cals from ICRP publication 53. Technetium-labelled erythrocytes 23. Maddox DE, Holman BL, Wynne J, Idoine J, Parker JA, Uren R,
(RBC) Tc-99m. Ann ICRP 1998;28:61.
et al. Ejection fraction image: a noninvasive index of regional left
2. Callahan RJ, Froelich JW, McKusick KA, Leppo J, Strauss HW. A
ventricular wall motion. Am J Cardiol 1978;41:1230-8.
modified method for the in vivo labeling of red blood cells with
24. Murphy PH. ECG gating: does it adequately monitor ventricular
Tc-99m: concise communication. J Nucl Med 1982;23:315-8.
contraction? J Nucl Med 1980;21:399-401.
3. UltraTag RBC. St Louis: Mallinckrodt Medical; 1992. [package
25. Cavailloles F, Bazin JP, Di Paola R. Factor analysis in gated
insert]
cardiac studies. J Nucl Med 1984;25:1067-79.
4. Gerson MC, editor. Cardiac nuclear medicine. 3rd ed. New York:
26. Wendt RE, Murphy PH, Treffert JD, Groch MW, Erwin WD,
McGraw-Hill, Health Professions Division; 1997.
Schneider PM, et al. Application and interpretation of principal
5. Kelly MJ, Cowie AR, Antonino A, Barton H, Kalff V. An
component analysis of gated cardiac images [abstract]. J Nucl Med
assessment of factors which influence the effectiveness of the
1993;34:175P.
modified in vivo technetium-99m-erythrocyte labeling technique
27. Wendt RE, Murphy PA, Schneider PM, Treffert JD, Groch MW,
in clinical use. J Nucl Med 1992;33:2222-5.
Ford PV, et al. Lossy compression of dynamic studies using
6. Bacharach SL, Green MV, Borer JS. Instrumentation and data
eigenimage methods [abstract]. J Nucl Med 1994;35:P178.
processing in cardiovascular nuclear medicine: evaluation of ven-
28. Rosenbush SW, Ruggie N, Turner DA, Von Behren PL, Denes P,
tricular function. Semin Nucl Med 1979;9:257-74.
Fordham EW, et al. Sequence and timing of ventricular wall
7. Groch MW. Cardiac function: gated cardiac blood pool and first
motion in patients with bundle branch block. Assessment by
pass imaging. St Louis: Mosby; 1996.
8. DePuey EG. Evaluation of cardiac function with radionuclides. In: radionuclide cineangiography. Circulation 1982;66:1113-9.
Gottschalk A, Hoffer PB, Potchen EJ, editors. Diagnostic nuclear 29. Watson DD, Liedholdt EM, Carabello ME, et al. Gated blood pool
medicine. Baltimore: Williams and Wilkins; 1998. imaging in patients with atrial fibrillation [abstract]. J Nucl Med
9. Garcia EV. Physics and instrumentation of radionuclide imaging. 1981;22:P153.
Philadelphia: Saunders; 1991. 30. Wagner RH, Halama JR, Henkin RE, Dillehay GL, Sobotka PA.
10. Garcia EV, Bateman TM, Berman DS, Maddahi J. Computer Errors in the determination of left ventricular functional parame-
techniques for optimal radionuclide assessment of the heart. ters. J Nucl Med 1989;30:1870-4.
Baltimore: Williams and Wilkins; 1988. 31. Bonow RO, Bacharach SL, Green MV, Kent KM, Rosing DR,
11. Wackers FJT. Equilibrium radionuclide angiocardiography. 3rd ed. Lipson LC, et al. Impaired left ventricular diastolic filling in
New York: McGraw Hill; 1997. patients with coronary artery disease: assessment with radionuclide
12. Bacharach SL, Green MV, Borer JS, Hyde JE, Farkas SP, Johnson angiography. Circulation 1981;64:315-23.
GS. Left ventricular peak ejection rate, filling rate and ejection 32. Mancini GB, Slutsky RA, Norris SL, Bhargava V, Ashburn WL,
fraction: frame rate requirements at rest and exercise. J Nucl Med Higgins CB. Radionuclide analysis of peak filling rate, filling
1979;20:189-93. fraction, and time to peak filling rate. Response to supine bicycle
13. Bacharach SL, Green MV, Bonow RO, Findley SL, Ostrow HG, exercise in normal subjects and patients with coronary disease.
Johnston GS. Measurement of ventricular function by ECG gating Am J Cardiol 1983;51:43-51.
during atrial fibrillation. J Nucl Med 1981;22:226-31. 33. Lee FA, Fetterman R, Zaret BL, Wackers FJT. Rapid radionuclide-
14. Miller TR, Goldman KJ, Epstein DM, Biello DR, Sampathkuma- derived systolic and diastolic cardiac function using cycle-depen-
ran KS, Kumar B, et al. Improved interpretation of gated cardiac dent background correction and Fourier analysis. Proc Comput
images by use of digital filters. Radiology 1984;152:795-800. Cardiol 1985:443-6.
15. Steckley RA, Kronenberg MW, Born ML, Rhea TC, Bateman JE, 34. Seals AA, Verani MS, Tadros S, Mahmarian JJ, Roberts R.
Rollo FD, et al. Radionuclide ventriculography: evaluation of Comparison of left ventricular diastolic function as determined by
automated and visual methods for regional wall motion analysis. nuclear cardiac probe, radionuclide angiography, and contrast
Radiology 1982;142:179-85. cineangiography. J Nucl Med 1986;27:1908-15.
16. Strauss HW, Zaret BL, Hurley PJ, Natarajan TK, Pitt B. A 35. Bauch TD, Rubal BJ, Lecce MD, Smith TE, Groch MW. S2
scintiphotographic method for measuring left ventricular ejection triggered gated blood pool imaging for assessment of diastole.
fraction in man without cardiac catheterization. Am J Cardiol Biomed Sci Instrum 1995;31:201-6.
1971;28:575-80. 36. Slutsky R, Karliner J, Ricci D, Kaiser R, Pfisterer M, Gordon D, et
17. Cahill PT, Ornstein E, Ho SL. Edge detection algorithms in nuclear al. Left ventricular volumes by gated equilibrium radionuclide
medicine. IEEE Trans Nucl Sci 1976;23:555-9. angiography: a new method. Circulation 1979;60:556-64.
18. Chang W, Henkin RE, Hale DJ, Hall D. Methods for detection of 37. Dehmer GJ, Firth BG, Lewis SE, Willerson JT, Hillis LD. Direct
left ventricular edges. Semin Nucl Med 1980;10:39-53. measurement of cardiac output by gated equilibrium blood pool
e78 Corbett et al Journal of Nuclear Cardiology
Equilibrium radionuclide angiocardiography November/December 2006

scintigraphy: validation of scintigraphic volume measurements by 52. Dehmer GJ, Firth BG, Nicod P, Lewis SE, Hillis LD. Alterations
a nongeometric technique. Am J Cardiol 1981;47:1061-7. in left ventricular volumes and ejection fraction during atrial
38. Bourguignon MH, Schindledecker JG, Carey GA, Douglass KH, pacing in patients with coronary artery disease: assessment with
Burow RD, Camargo EE, et al. Quantification of left ventricular radionuclide ventriculography. Am Heart J 1983;106:114-24.
volume in gated equilibrium radioventriculography. Eur J Nucl 53. Port S, Cobb FR, Coleman RE, Jones RH. Effect of age on the
Med 1981;6:349-53. response of the left ventricular ejection fraction to exercise. N Engl
39. Links JM, Becker LC, Shindledecker JG, Guzman P, Burow RD, J Med 1980;303:1133-7.
Nickoloff EL, et al. Measurement of absolute left ventricular 54. Port S, McEwan P, Cobb FR, Jones RH. Influence of resting left
volume from gated blood pool studies. Circulation 1982;65:82-91. ventricular function on the left ventricular response to exercise
40. Massardo T, Gal RA, Grenier RP, Schmidt DH, Port SC. Left in patients with coronary artery disease. Circulation 1981;63:
ventricular volume calculation using a count-based ratio method 856-63.
applied to multigated radionuclide angiography. J Nucl Med 55. Gibbons RJ, Lee KL, Cobb F, Jones RH. Ejection fraction response
1990;31:450-6. to exercise in patients with chest pain and normal coronary
41. Berger HJ, Matthay RA, Loke J, Marshall RC, Gottschalk A, Zaret arteriograms. Circulation 1981;64:952-7.
BL. Assessment of cardiac performance with quantitative radionu- 56. Port SC, Oshima M, Ray G, McNamee P, Schmidt DH.
clide angiocardiography: right ventricular ejection fraction with Assessment of single vessel coronary artery disease: results of
reference to findings in chronic obstructive pulmonary disease. exercise electrocardiography, thallium-201 myocardial perfu-
Am J Cardiol 1978;41:897-905. sion imaging and radionuclide angiography. J Am Coll Cardiol
42. Brent BN, Mahler D, Matthay RA, Berger HJ, Zaret BL. Nonin-
1985;6:75-83.
vasive diagnosis of pulmonary arterial hypertension in chronic
57. Wackers FJ, Berger HJ, Johnstone DE, Goldman L, Reduto LA,
obstructive pulmonary disease: right ventricular ejection fraction at
Langou RA, et al. Multiple gated cardiac blood pool imaging for
rest. Am J Cardiol 1984;53:1349-53.
left ventricular ejection fraction: validation of the technique and
43. Winzelberg GG, Boucher CA, Pohost GM, McKusick KA, Bing-
assessment of variability. Am J Cardiol 1979;43:1159-66.
ham JB, Okada RD, et al. Right ventricular function in aortic and
58. DePace NL, Iskandrian AS, Hakki AH, Kane SA, Segal BL. Value
mitral valve disease: relation of gated first-pass radionuclide
of left ventricular ejection fraction during exercise in predicting the
angiography to clinical and hemodynamic findings. Chest 1981;
extent of coronary artery disease. J Am Coll Cardiol 1983;1:1002-
79:520-8.
1010.
44. Maddahi J, Berman DS, Matsuoka DT, Waxman AD, Stankus KE,
59. Foster C, Dymond DS, Anholm JD, Pollock ML, Schmidt DH.
Forrester JS, et al. A new technique for assessing right ventricular
Effect of exercise protocol on the left ventricular response to
ejection fraction using rapid multiple-gated equilibrium cardiac
exercise. Am J Cardiol 1983;51:859-64.
blood pool scintigraphy. Description, validation and findings in
60. Moore ML, Murphy PH, Burdine JA. ECG-gated emission com-
chronic coronary artery disease. Circulation 1979;60:581-9.
puted tomography of the cardiac blood pool. Radiology 1980;134:
45. Pryor DB, Harrell FE, Lee KL, Rosati RA, Coleman RE, Cobb FR,
et al. Prognostic indicators from radionuclide angiography in 233-5.
medically treated patients with coronary artery disease. Am J 61. Corbett JR, Jansen DE, Willerson JT. Radionuclide ventriculog-
Cardiol 1984;53:18-22. raphy: I. Technical aspects. Am J Physiol Imaging 1987;2:33-
46. Lee KL, Pryor DB, Pieper KS, Harrell FE, Califf RM, Mark DB, 43.
et al. Prognostic value of radionuclide angiography in medically 62. Corbett JR, Jansen DE, Willerson JT. Radionuclide ventriculogra-
treated patients with coronary artery disease. A comparison with phy: II. Anatomic and physiologic aspects. Am J Physiol Imaging
clinical and catheterization variables. Circulation 1990;82:1705- 1987;2:85-104.
17. 63. Corbett JR. Gated blood-pool SPECT. In: DePuey EG, Garcia BD,
47. Borer JS, Bacharach SL, Green MV, Kent KM, Epstein SE, Ernest V, editors. Cardiac SPECT imaging. New York:Raven
Johnston GS. Real-time radionuclide cineangiography in the non- Press; 1995. p. 257-73.
invasive evaluation of global and regional left ventricular function 64. Ficaro EP, Corbett JR. Gated blood-pool SPECT. In: DePuey EG
at rest and during exercise in patients with coronary-artery disease. GE, Berman DS, editors. Cardiac SPECT imaging. 2nd ed.
N Engl J Med 1977;296:839-44. Philadelphia: Lippincott Williams & Wilkins; 2001. p. 321-37.
48. Borer JS, Kent KM, Bacharach SL, Green MV, Rosing DR, Seides 65. Gill JB, Moore RH, Tamaki N, Miller DD, Barlai-Kovach M,
SF, et al. Sensitivity, specificity and predictive accuracy of Yasuda T, et al. Multigated blood-pool tomography: new method
radionuclide cineangiography during exercise in patients with for the assessment of left ventricular function. J Nucl Med
coronary artery disease. Comparison with exercise electrocardiog- 1986;27:1916-24.
raphy. Circulation 1979;60:572-80. 66. Corbett JR, Jansen DE, Lewis SE, Gabliani GI, Nicod P, Filipchuk
49. Turner DA, Shima MA, Ruggie N, Von Behren PL, Jarosky MJ, NG, et al. Tomographic gated blood pool radionuclide ventricu-
Ali A, et al. Coronary artery disease: detection by phase analysis of lography: analysis of wall motion and left ventricular volumes in
rest/exercise radionuclide angiocardiograms. Radiology 1983;148: patients with coronary artery disease. J Am Coll Cardiol 1985;6:
539-45. 349-58.
50. Poliner LR, Dehmer GJ, Lewis SE, Parkey RW, Blomqvist CG, 67. Faber TL, Stokely EM, Templeton GH, Akers MS, Parkey RW,
Willerson JT. Left ventricular performance in normal subjects: a Corbett JR. Quantification of three-dimensional left ventricular
comparison of the responses to exercise in the upright and supine segmental wall motion and volumes from gated tomographic
positions. Circulation 1980;62:528-34. radionuclide ventriculograms. J Nucl Med 1989;30:638-49.
51. Dehmer GJ, Lewis SE, Hillis LD, Corbett J, Parkey RW, Willerson 68. Groch MW, Leidholdt EM, Marshall RA, et al. Gated blood pool
JT. Exercise-induced alterations in left ventricular volumes and the SPECT imaging: sources of artifacts [abstract]. Clin Nucl Med
pressure-volume relationship: a sensitive indicator of left ventric- 1991;16:717.
ular dysfunction in patients with coronary artery disease. Circula- 69. Chin BB, Bloomgarden DC, Xia W, Kim HJ, Fayad ZA, Ferrari
tion 1981;63:1008-18. VA, et al. Right and left ventricular volume and ejection fraction
Journal of Nuclear Cardiology Corbett et al e79
Volume 13, Number 6;e56-79 Equilibrium radionuclide angiocardiography

by tomographic gated blood-pool scintigraphy. J Nucl Med tomographic gated blood pool images. J Nucl Cardiol 1998;5:245-55.
1997;38:942-8. 79. Le Guludec D, Gauthier H, Porcher R, Frank R, Daou D,
70. Groch MW, Schippers DJ, Marshall RC, Barnett C. A quantitative Benelhadj S, et al. Prognostic value of radionuclide angiography in
program for gated blood pool SPECT imaging [abstract]. Clin Nucl patients with right ventricular arrhythmias. Circulation 2001;
Med 1991;16:713. 103:1972-6.
71. Bartlett ML, Srinivasan G, Barker WC, Kitsiou AN, Dilsizian V, 80. Le Guludec D, Slama MS, Frank R, Faraggi M, Grimon G,
Bacharach SL. Left ventricular ejection fraction: comparison of Bourguignon MH, et al. Evaluation of radionuclide angiography in
results from planar and SPECT gated blood-pool studies. J Nucl diagnosis of arrhythmogenic right ventricular cardiomyopathy.
Med 1996;37:1795-9. J Am Coll Cardiol 1995;26:1476-83.
72. Nichols K, Humayun N, De Bondt P, Vandenberghe S, Akinbob- 81. Botvinick EH. Scintigraphic blood pool and phase image analysis:
oye OO, Bergmann SR. Model dependence of gated blood pool the optimal tool for the evaluation of resynchronization therapy.
SPECT ventricular function measurements [see comment]. J Nucl J Nucl Cardiol 2003;10:424-8.
Cardiol 2004;11:282-92. 82. Caputo GR, Graham MM, Brust KD, Kennedy JW, Nelp WB.
73. Nichols K, Saouaf R, Ababneh AA, Barst RJ, Rosenbaum MS, Measurement of left ventricular volume using single-photon emis-
Groch MW, et al. Validation of SPECT equilibrium radionu- sion computed tomography. Am J Cardiol 1985;56:781-6.
clide angiographic right ventricular parameters by cardiac 83. Cahill JM, Chen MY, Ficaro EP, Corbett JR, Quaife RA. Valida-
magnetic resonance imaging [see comment]. J Nucl Cardiol tion of 4D-MSPECT analysis method for Tc-99m gated blood pool
2002;9:153-60. tomography: comparison of LV ejection fractions and volumes to
74. Slart RH, Poot L, Piers DA, van Veldhuisen DJ, Jager PL. magnetic resonance imaging [abstract]. J Nucl Cardiol 2003;10:
Evaluation of right ventricular function by NuSMUGA software: S20.
gated blood-pool SPECT vs. first-pass radionuclide angiography. 84. Quaife RA, Corbett JR. Radionuclide ventriculography. In:
Int J Cardiovasc Imaging 2003;19:401-7. McGhie AI, editor. Handbook of non-invasive cardiac testing.
75. Underwood SR, Walton S, Ell PJ, Jarritt PH, Emanuel RW, New York: Oxford University Press; 2001. p. 55-98.
Swanton RH. Gated blood-pool emission tomography: a new 85. Lee TH, Udelson JE. Nuclear cardiology. In: Zipes DP, Libby P,
technique for the investigation of cardiac structure and function. Bonow RO, Braunwald E, editors. Braunwald’s heart disease: a
Eur J Nucl Med 1985;10:332-7. textbook of cardiovascular medicine. 7th ed. Philadelphia: Saun-
76. McGhie AL, Faber TL, Willerson JT, Corbett JR. Evaluation of left ders; 2005.
ventricular aneurysm after acute myocardial infarction using tomographic 86. Klocke FJ, Baird MG, Bateman TM, Berman DS, Carabello BA,
radionuclide ventriculography. Am J Cardiol 1995;75:720-4. Cerquerira MD, et al. ACC/AHA/ASNC guidelines for the
77. Daou D, Lebtahi R, Faraggi M, Petegnief Y, Le Guludec D. Clinical Use of Cardiac Radionuclide Imaging. A report of the
Cardiac gated equilibrium radionuclide angiography and multihar- American College of Cardiology/American Heart Association
monic Fourier phase analysis: optimal acquisition parameters in Task Force of Practice Guidelines (ACC/AHA/ASNC Commit-
arrhythmogenic right ventricular cardiomyopathy. J Nucl Cardiol tee to Revise the 1995 Guidelines for the Clinical Use of
1999;6:429-37. Radionuclide Imaging) 2003. American College of Cardiology
78. Botvinick EH, O’Connell JW, Kadkade PP, Glickman SL, Dae web site. Available from: URL: http://www.acc.org/quality
MW, Cohen TJ, et al. Potential added value of three-dimensional andscience/clinical/topic/topic.htm#cardiacimaging. Accessed Oct-
reconstruction and display of single photon emission computed ober 23, 2006.
 Stress protocols and tracers
Milena J. Henzlova, MD,a Manuel D. Cerqueira, MD,b John J. Mahmarian, MD,b
and Siu-Sun Yao, MDb

EXERCISE STRESS TEST otherwise stable and pain-free, can undergo exercise
stress testing.
Exercise is the preferred stress modality in patients
2. Decompensated or inadequately controlled conges-
who are able to exercise to an adequate workload (at least
tive heart failure.
85% of age-adjusted maximal predicted heart rate and
3. Uncontrolled hypertension (blood pressure ⬎200/
five metabolic equivalents).
110 mm Hg).
4. Uncontrolled cardiac arrhythmias (causing symp-
A. Indications toms or hemodynamic compromise).
5. Severe symptomatic aortic stenosis.
1. Detection of obstructive coronary artery disease
6. Acute pulmonary embolism.
(CAD) in the following:
7. Acute myocarditis or pericarditis.
a. Patients with an intermediate pretest probability of
8. Acute aortic dissection.
CAD based on age, gender, and symptoms. 9. Severe pulmonary hypertension.
b. Patients with high-risk factors for CAD (eg, diabetes 10. Acute myocardial infarction (⬍4 days).
mellitus, peripheral or cerebral vascular disease).
2. Risk stratification of post–myocardial infarction pa-
tients before discharge (submaximal test at 4-6 days) C. Relative Contraindications
and early (symptom-limited at 14-21 days) or late
(symptom-limited at 3-6 weeks) after discharge. 1. Known left main coronary artery stenosis.
3. Risk stratification of patients with chronic stable CAD 2. Moderate aortic stenosis.
into a low-risk category that can be managed medi- 3. Hypertrophic obstructive cardiomyopathy or other
cally or into a high-risk category that should be forms of outflow tract obstruction.
considered for coronary revascularization. 4. Significant tachyarrhythmias or bradyarrhythmias.
4. Risk stratification of low-risk acute coronary syndrome 5. High-degree atrioventricular (AV) block.
patients (without active ischemia and/or heart failure 6. Electrolyte abnormalities.
6-12 hours after presentation) and of intermediate-risk 7. Mental or physical impairment leading to inability to
acute coronary syndrome patients 1 to 3 days after exercise adequately.
presentation (without active ischemia and/or heart 8. If combined with imaging, patients with complete left
failure symptoms). bundle branch block (LBBB), permanent pacemakers,
5. Risk stratification before noncardiac surgery in pa- and ventricular pre-excitation (Wolff-Parkinson-White
tients with known CAD or those with high-risk syndrome) should preferentially undergo pharmacologic
factors for CAD. vasodilator stress test (not dobutamine stress test).
6. To evaluate the efficacy of therapeutic interventions
(anti-ischemic drug therapy or coronary revascular- D. Limitations
ization) and in tracking subsequent risk based on
serial changes in myocardial perfusion in patients Exercise stress testing has a limited value in patients
with known CAD. who cannot achieve an adequate heart rate and blood
pressure response due to a noncardiac physical limitation
such as pulmonary, peripheral vascular, or musculoskel-
B. Absolute Contraindications etal abnormalities or due to lack of motivation. These
1. High-risk unstable angina. However, patients with patients should undergo pharmacologic stress with myo-
suspected unstable angina at presentation, who are cardial perfusion imaging.

E. Exercise Modalities
Chair,a Member.b
J Nucl Cardiol 2006;13:e80-90.
1071-3581/$32.00 1. Treadmill exercise is the most widely used stress
Copyright © 2006 by the American Society of Nuclear Cardiology. modality. Several treadmill exercise protocols are
doi:10.1016/j.nuclcard.2006.08.011 available which differ in the speed and grade of
e80
Journal of Nuclear Cardiology Henzlova et al e81
Volume 13, Number 6;e80-90 Stress protocols and tracers

treadmill inclination. The Bruce and modified Bruce 6. Excessive ST-segment depression (⬎2 mm).
protocols are the most widely used exercise protocols. 7. ST elevation (⬎1 mm) in leads without diagnostic Q
2. Upright bicycle exercise is commonly used in Europe. waves (except for leads V1 or aVR).
This is preferable if dynamic first-pass imaging is 8. Sustained supraventricular or ventricular tachycar-
planned during exercise. Supine or semi-supine exercise dia.
is relatively suboptimal and should only be used while 9. Development of bundle branch block (LBBB) or
performing exercise radionuclide angiocardiography. intraventricular conduction delay that cannot be
distinguished from ventricular tachycardia.
F. Test Procedure 10. Drop in systolic blood pressure of greater than 10
mm Hg from baseline, despite an increase in work-
1. Patient preparation: nothing by mouth (NPO) for 4 to load, when accompanied by other evidence of isch-
6 hours. emia.
2. A large-bore (18- to 20-gauge) intravenous (IV) 11. Hypertensive response (systolic blood pressure ⬎250
cannula should be inserted for radiopharmaceutical mm Hg and/or diastolic pressure ⬎115 mm Hg).
injection during exercise. 12. Technical difficulties in monitoring the electrocar-
3. The electrocardiogram should be monitored continu- diogram or systolic blood pressure.
ously during the exercise test and for at least 5
minutes into the recovery phase or until the resting
heart rate is less than 100 beats/min and/or dynamic PHARMACOLOGIC VASODILATOR STRESS
exercise-induced ST-segment changes have resolved. NOTE: Some of the pharmacologic stress protocols
A 12-lead electrocardiogram should be obtained at described in this section fall outside of manufacturer
every stage of exercise, at peak exercise, and at the package insert guidelines but have been documented in
termination or recovery phase. the literature and are now used commonly in the clinical
4. The heart rate and blood pressure should be recorded
practice of nuclear cardiology.
at least every 3 minutes during exercise, at peak
exercise, and for at least 5 minutes into the recovery
phase. Adenosine
5. All exercise tests should be symptom-limited.
1. Mechanism of action. Adenosine is a nonselective
Achievement of 85% of maximum, age-adjusted,
A2 receptor agonist which induces direct coronary
predicted heart rate is not an indication for termina-
arteriolar vasodilation through specific activation of
tion of the test.
the A2A receptor. This results in a 3.5- to 4-fold
6. The radiopharmaceutical should be injected as close
increase in myocardial blood flow. Myocardial re-
to peak exercise as possible. Patients should continue
to exercise for an additional 1.5 to 2 minutes after the gions supplied by stenotic coronary arteries have an
radiotracer injection. attenuated hyperemic response. Depending upon the
7. In patients who cannot exercise adequately and are severity of coronary stenosis and coronary flow re-
being referred for a diagnostic stress test, the radio- serve limitation, a relative flow heterogeneity is in-
tracer should not be injected at peak exercise and the duced. Adenosine generally does not cause myocar-
patients should be evaluated for a pharmacologic dial ischemia since myocardial blood flow increases
stress test. to a variable degree in all coronary artery vascular
8. Blood pressure medication(s) with antianginal prop- beds with minimal or no increase in rate-pressure
erties (␤-blocker, calcium channel blocker, and ni- product (ie, myocardial oxygen demand). However,
trates) should be discontinued for at least 48 hours in a small percentage of patients with severe CAD,
prior to a diagnostic stress test. In patients with true ischemia may also be induced because of a
established CAD, medication(s) discontinuation should coronary steal phenomenon. Since the myocardial
be left to the discretion of the referring physician. tracer uptake is proportional to the regional myocar-
dial blood flow, an unequal distribution of radiotracer
occurs in the myocardium. Nonselective activation of
G. Indications for Early Termination of Exercise A1, A2b, and A3 receptors may cause undesirable
1. Moderate to severe angina pectoris. side effects of adenosine infusion: AV block (A1
2. Marked dyspnea or fatigue. receptor), peripheral vasodilation (A2b receptor), and
3. Ataxia, dizziness, or near-syncope. bronchospasm (A2b and A3 receptors).
4. Signs of poor perfusion (cyanosis and pallor). 2. Adenosine dose. Adenosine should be given as a
5. Patient’s request to terminate the test. continuous infusion at a rate of 140 ␮g · kg⫺1 · min⫺1
e82 Henzlova et al Journal of Nuclear Cardiology
Stress protocols and tracers November/December 2006

over a 6-minute period. A shorter-duration adenosine 5. Procedure

infusion, lasting 4 minutes, has been found to be ● Patient preparation: NPO for 4-6 hours; no caf-
equally effective for the detection of CAD compared feine-containing beverages or medications for at
to the standard 6-minute infusion. least 12 hours prior to testing.
3. Indications. The indications for adenosine stress ● An infusion pump is required for adenosine to be
perfusion imaging are the same as for exercise myo- administered at a constant infusion rate.
cardial perfusion imaging and in the presence of the ● An IV line with a dual-port Y-connector is required
following conditions: for the injection of the radiotracer during adenosine
● Inability to perform adequate exercise due to non- infusion.
cardiac physical limitations (pulmonary, peripheral ● ECG monitoring should be carried out as with
vascular, musculoskeletal, or mental conditions) or exercise stress testing. A 12-lead electrocardiogram
due to lack of motivation. Of note, as with exercise will be recorded every minute during the infusion.
testing, anti-ischemic cardiac medications (includ- ● Blood pressure should be monitored every minute
ing ␤-blockers, nitrates, and calcium antagonists) during infusion and 3 to 5 minutes into recovery.
should be discontinued for at least 48 hours prior to Adenosine infusion should be given at a rate of 140
performing a diagnostic imaging test. ␮g · kg⫺1 · min⫺1 for 3 minutes followed by the
● Baseline electrocardiographic (ECG) abnormali- injection of the radiotracer. The infusion should be
ties: LBBB, ventricular pre-excitation (Wolff-Par- continued for another 3 minutes. For patients
kinson-White syndrome), and permanent ventricu- deemed to be at a higher risk for complications
lar pacing. Falsely positive imaging results are (borderline hypotension, controlled asthma), aden-
much less frequent with adenosine (approximately osine infusion may be started at a lower dose (70
10%) as compared to stress imaging with exercise ␮g · kg⫺1 · min⫺1). If this dose is tolerated well for
1 minute, the infusion rate should be increased to
(approximately 50%).
140 ␮g · kg⫺1 · min⫺1 and should be continued for
● Risk stratification of clinically stable patients into
4 minutes. The radiotracer should be injected 1
low- and high-risk groups very early after acute
minute after starting the 140 –␮g · kg⫺1 · min⫺1
myocardial infarction (ⱖ1 day) or following pre-
dose.
sentation to the emergency department with a
presumptive acute coronary syndrome. 6. Hemodynamic effects. Adenosine results in a modest
4. Contraindications increase in heart rate and a modest decrease in both
● Asthmatic patients with ongoing wheezing should
systolic and diastolic blood pressures.
not undergo adenosine stress testing. However, 7. Early termination of adenosine infusion. The aden-
patients with adequately controlled asthma can osine infusion should be stopped early under any of
undergo an adenosine stress test and can have the following circumstances:
pre-treatment with 1 to 2 puffs of albuterol or a ● Severe hypotension (systolic blood pressure ⬍80
comparable inhaler. mm Hg).
● Greater than first-degree AV block without a pace- ● Development of symptomatic, persistent second-
maker or sick sinus syndrome. degree or complete heart block.
● Systolic blood pressure less than 90 mm Hg. ● Wheezing.
● Use of dipyridamole, dipyridamole-containing ● Severe chest pain associated with ST depression of
medications, or aminophylline in the last 24 hours 2 mm or greater.
or ingestion of caffeinated foods (eg, chocolate) or ● Signs of poor perfusion (pallor, cyanosis, cold
beverages (eg, coffee, tea, sodas) within the last 12 skin).
hours. Pentoxifylline need not be stopped prior to ● Technical problems with the monitoring equip-
adenosine infusion. Pentoxifylline is a methylxan- ment.
thine derivative that increases the theophylline ● Patient’s request to stop.
level in some patients taking theophylline-contain- 8. Side effects of adenosine
ing drugs. ● Minor side effects are common and occur in ap-
● Relative contraindication: Profound sinus bradycar- proximately 80% of patients. The common side
dia (heart rate ⬍40/min). effects are flushing (35%-40%), chest pain (25%-
● Known hypersensitivity to adenosine or dipyridam- 30%), dyspnea (20%), dizziness (7%), nausea
ole. (5%), and symptomatic hypotension (5%). Chest
● Unstable acute myocardial infarction or acute cor- pain is nonspecific and is not necessarily indicative
onary syndrome. of the presence of CAD.
Journal of Nuclear Cardiology Henzlova et al e83
Volume 13, Number 6;e80-90 Stress protocols and tracers

● AV block occurs in approximately 7.6% of cases. ● The drug is infused intravenously over 4 minutes.
However, the incidence of second-degree AV block Although an infusion pump is desirable, dipyrid-
is only 4%, and that of complete heart block is less amole can also be administered by hand injection.
that 1%. Most cases (⬎95%) of AV block do not The radiotracer is injected 3 to 5 minutes after the
require termination of the infusion. completion of dipyridamole infusion. The half-life
● ST-segment depression of 1 mm or greater occurs of dipyridamole is approximately 30 to 45 minutes.
in 5% to 7% of cases. However, unlike chest pain,
this is usually indicative of significant CAD. 6. Hemodynamic effects. Dipyridamole results in sim-
● Fatal or nonfatal myocardial infarction is extremely ilar hemodynamic changes as seen with adenosine
rare, and the reported incidence is less than 1 in with a modest increase in heart rate and a modest
1000 cases. decrease in both systolic and diastolic blood pres-
● Due to an exceedingly short half-life of adenosine sures.
(⬍10 seconds), most side effects resolve in a few 7. Side effects. Over 50% of patients develop side
seconds after discontinuation of the adenosine in- effects (flushing, chest pain, headache, dizziness, or
fusion, and aminophylline infusion is only very hypotension). The frequency of these side effects is
rarely required. less than that seen with adenosine, but they last for a
longer period of time (15-25 minutes) and may vary
9. Combination of low-level exercise with adenosine significantly in individual patients. Aminophylline
infusion. The combination of low-level upright tread- (125-250 mg intravenously) is often required to re-
mill exercise (1.7 mph, 0% grade) during the adeno- verse these side effects. The incidence of AV block
sine infusion has been found to be safe. This results in with dipyridamole is less than that observed with
a significant reduction in the side effects of adenosine adenosine (2%). Aminophylline should also be used
(flushing, dizziness, nausea, and headache) and atten- in the presence of ischemic ECG changes after
uates the adenosine-induced drop in blood pressure. dipyridamole.
Image quality is improved by decreasing high hepatic 8. Combination of low-level exercise with dipyridam-
and gut radiotracer uptake, which is common with ole infusion. Patients who are ambulatory may un-
pharmacologic stress perfusion imaging. Therefore dergo low-level treadmill exercise (1.7 mph, 0%
low-level exercise may be performed in combination grade) for 4 to 6 minutes soon after the completion of
with pharmacologic stress. However, since it is desir- dipyridamole infusion. Radiotracer is injected during
able not to increase the heart rate of patients with this low-level exercise, and the exercise continues for
LBBB undergoing pharmacologic stress, low-level 2 additional minutes to allow for tracer uptake in the
exercise supplementation should not be used in pa- myocardium. This significantly reduces the side ef-
tients with LBBB. fects and improves image quality. Low-level exercise
supplementation is not recommended for patients
with LBBB.
Dipyridamole

1. Mechanism of action. Dipyridamole is an indirect Dobutamine

coronary artery vasodilator that increases the tissue
levels of adenosine by preventing the intracellular 1. Mechanism of action. Dobutamine infusion results
reuptake and deamination of adenosine. Dipyridamole- in direct ␤1 and ␤2 stimulation with a dose-related
induced hyperemia lasts for more than 15 minutes. increase in heart rate, blood pressure, and myocardial
2. Dipyridamole dose. Dipyridamole is administered at contractility. Dobutamine increases regional myocar-
0.56 mg/kg intravenously over a 4-minute period. dial blood flow based on physiologic principles of
3. Indications. The indications for dipyridamole stress coronary flow reserve. A similar dose-related increase
perfusion imaging are the same as for adenosine in subepicardial and subendocardial blood flow oc-
myocardial perfusion imaging. curs within vascular beds supplied by normal coro-
4. Contraindications. The contraindications for dipy- nary arteries. However, blood flow increases mini-
ridamole stress testing are the same as with adeno- mally within vascular beds supplied by significantly
sine. In patients taking oral dipyridamole, IV dipyrid- stenosed arteries, with most of the increase occurring
amole may be administered safely and efficaciously. within the subepicardium rather than the subendocar-
5. Procedure dium. However, at a dose of 20 ␮g · kg⫺1 · min⫺1,
● Patient preparation: NPO for 4 to 6 hours and no dobutamine-induced coronary flow heterogeneity is
caffeine-containing beverages or medication at significantly less than that induced by adenosine or
least 12 hours prior to testing. dipyridamole.
e84 Henzlova et al Journal of Nuclear Cardiology
Stress protocols and tracers November/December 2006

Figure 1. Stress/redistribution/reinjection/18- to 24-hour Tl-201 imaging protocol.

2. Dobutamine dose. Dobutamine is infused incremen- ●Severe aortic stenosis.

tally starting at a dose of 5 to 10 ␮g · kg⫺1 · min⫺1, ●Atrial tachyarrhythmias with uncontrolled ventric-
which is increased at 3-minute intervals to 20, 30, and ular response.
40 ␮g · kg⫺1 · min⫺1. The half-life of dobutamine is ● Prior history of ventricular tachycardia.
approximately 2 minutes. It is customary to use ● Uncontrolled hypertension (blood pressure ⬎200/
atropine in patients in whom the heart rate does not 110 mm Hg).
increase beyond 120 beats/min with the maximum ● Patients with aortic dissection or large aortic aneu-
dose of dobutamine while performing stress echocar- rysm.
diography. As with exercise stress, achieving greater ● Patients who are on ␤-blockers where the heart rate
than 85% of the predicted heart rate is desirable. and inotropic responses to dobutamine will be
(Dobutamine should not be used for patients with attenuated.
LBBB.) 5. Procedure
3. Indications ● Patient preparation: NPO for 4 to 6 hours.
● Dobutamine is a secondary pharmacologic stressor ● An infusion pump is necessary for dobutamine
that is used only in patients who cannot undergo administration.
exercise stress and who also have contraindications ● An IV line with a dual-port Y-connector is required
to pharmacologic vasodilator stressors (mainly for injecting radioisotope during dobutamine infu-
bronchospastic airway disease). sion.
● Dobutamine perfusion imaging has not been stud- ● ECG monitoring and blood pressure monitoring
ied as extensively as adenosine or dipyridamole should be performed as with other pharmacologic
perfusion imaging in the evaluation and prognosti- stressors.
cation of patients with CAD. ● Dobutamine infusion should start at a dose of 10
4. Contraindications ␮g · kg⫺1 · min⫺1. The dobutamine dose should
Dobutamine stress testing should not be used under then be increased at 3-minute intervals up to a
the following circumstances: maximum of 40 ␮g · kg⫺1 · min⫺1. The radiotracer
● Recent (⬍1 week) myocardial infarction. should be injected at 1 minute into the highest
● Unstable angina. dobutamine dose, and dobutamine infusion should
● Hemodynamically significant left ventricular out- be continued for 2 minutes after the radiotracer
flow tract obstruction. injection.
Journal of Nuclear Cardiology Henzlova et al e85
Volume 13, Number 6;e80-90 Stress protocols and tracers

Figure 2. Tc-99m imaging protocols: Two-day exercise stress/rest.

6. Early termination of dobutamine. The indications Tl-201

for early termination of dobutamine are similar to
Tl-201 is an analog of potassium (monovalent cat-
those for exercise stress. Termination for ventricular
ion), with a physical half-life of 74 hours, gamma decay
tachycardia or ST-segment depression is more likely
with 80-keV emission energy, high first-pass extraction
with dobutamine than with other stressors.
(85%), active membrane transport into the myocyte,
7. Side effects. Side effects occur in about 75% of
patients. The common side effects are palpitation rapid clearance from the intravascular space, and mono-
(29%), chest pain (31%), headache (14%), flushing exponential washout (redistribution) which starts 10 to
(14%), dyspnea (14%), and significant supraventric- 15 minutes after injection. Washout depends on initial
ular or ventricular arrhythmias (8%-10%). Ischemic tracer concentration in the myocyte and on myocardial
ST-segment depression occurs in approximately one blood flow. Clearance occurs via the kidneys.
third of patients undergoing dobutamine infusion.
Severe side effects may require IV administration of a Imaging Protocols
short-acting ␤-blocker (esmolol, 0.5 mg/kg over
1 minute). A single dose of 2.5 to 4.0 mCi of Tl-201 is
injected prior to peak exercise stress or at peak
pharmacologic vasodilatation, and SPECT imaging
RADIOTRACERS starts 10 to 15 minutes later. Redistribution (rest)
Currently utilized myocardial perfusion tracers for imaging is done 2.5 to 4.0 hours later. In cases where
single photon emission computed tomography (SPECT) standard stress-redistribution imaging shows a fixed or
imaging include thallium 201 and two technetium 99m minimally reversible perfusion abnormality, myocardial
agents (Tc-99m sestamibi and Tc-99m tetrofosmin). viability can be assessed with a rest image at 18 to 24
Characteristics for each tracer, typical protocol, and hours or following reinjection of an additional 1 to 2 mCi
dosimetry are presented. dose of Tl-201. An alternative method for viability
NOTE: Some of the radiopharmaceutical doses assessment is injection of 3 to 4 mCi of Tl-201 at rest
described in this section fall outside of the manufacturer followed by 3- to 4-hour redistribution imaging. Protocol
package insert guidelines but are now commonly used in options and timing for assessment of perfusion and
the clinical practice of nuclear cardiology. viability are shown in Figure 1.
e86 Henzlova et al Journal of Nuclear Cardiology
Stress protocols and tracers November/December 2006

Figure 3. Rest Tl-201/stress Tc-99m separate-acquisition dual-isotope protocol.

Figure 4. Tc-99m imaging protocols: One-day exercise stress/rest.

Tc-99m–labeled Tracers emission energy, first-pass extraction less than Tl-201,

Tc-99m sestamibi and Tc-99m tetrofosmin have uptake and mitochondrial retention dependent on blood
very similar characteristics: lipid-soluble, cationic, phys- flow and transmembrane energy potentials. Their myo-
ical half-life of 6 hours, gamma decay with 140-keV cardial washout (redistribution) is clinically negligible.
Journal of Nuclear Cardiology Henzlova et al e87
Volume 13, Number 6;e80-90 Stress protocols and tracers

Figure 5. Tc-99m imaging protocols: One-day rest/exercise stress.

Figure 6. Tc-99m imaging protocols: One-day rest/adenosine pharmacologic stress.

e88 Henzlova et al Journal of Nuclear Cardiology
Stress protocols and tracers November/December 2006

Figure 7. Tc-99m imaging protocols: One-day rest/dipyridamole pharmacologic stress.

These agents are extracted via the hepatobiliary system and pharmacologic stress. This requires administration of
and excreted into the gastrointestinal tract. Lacking a low dose (one fourth of the total dose, or 8 to 12 mCi)
redistribution, Tc-99m–labeled tracers require two sepa- for the first study and a larger dose (three fourths of the
rate injections at stress and rest. The two agents have total dose, or 24-36 mCi) for the second study. One-day
sufficiently similar characteristics that the recommended rest/stress Tc-99m protocols are now performed almost
protocols use similar camera setup and acquisition times universally with no delay between the rest and subse-
and vary only in the optimal time for image acquisition quent stress images. The initially proposed 1990 protocol
following rest, exercise, and pharmacologic stress. Op- specified a 2-hour delay between rest and stress to allow
timal validation of imaging times has not been exten- the rest dose to decay in order to maximize the stress/rest
sively studied, and factors such as camera availability count density ratio and minimize rest-to-stress “shine-
and the presence of liver and gastrointestinal activity through” or “cross-talk.” However, simply increasing the
usually dictate when imaging is performed. In the figures stress dose provides the same stress/rest count density
a range of imaging times is suggested. ratio achieved by letting the rest dose decay (20% in 2
hours). Thus a 3:1 stress/rest dose ratio with a 2-hour
delay and a 3.5-4:1 ratio with no delay provide the same
Imaging Protocols
result. Note that the 2-hour delay is the total time
Two-day protocol. Ideally, stress and rest imag- between rest injection and starting the post-stress imag-
ing with Tc-99m agents should be performed on two ing. Thus the waiting time to stress is not fixed but depends
separate days, as shown in Figure 2, to avoid having on the sum of intervals from rest injection to the post-stress
residual activity from the first study contaminate the imaging. In contrast, for the 1-day stress/rest protocol,
second study. In overweight patients (ie, ⬎250 lb or wherein the stress scan is performed first and relative tracer
body mass index ⬎30) or in female patients where uptake in the myocardium is increased consequent to the
significant breast attenuation is anticipated, a low dose of stress-induced coronary hyperemia, a delay between the
Tc-99m radiotracer may result in suboptimal images and stress injection and the subsequent resting injection is
a 2-day imaging protocol is preferable. essential and should be as close to 4 hours as possible
One-day protocols. For logistical reasons, stress between the stress injection and starting the post-rest
and rest studies are usually performed using a 1-day imaging. Stated doses are commonly used in the United
protocol as shown in Figures 3, 4, 5, and 6 for exercise States, but total dose limits vary in other countries. Issues
Journal of Nuclear Cardiology Henzlova et al e89
Volume 13, Number 6;e80-90 Stress protocols and tracers

regarding the imaging sequence (stress vs rest first) and the 6. Braat SH, Leclercq B, Itti R, Lahiri A, Sridhara B, Rigo P.
minimum time interval between the two radiotracer injec- Myocardial imaging with technetium-99m-tetrofosmin: compari-
son of one-day and two-day protocols. J Nucl Med 1994;35:
tions are not fully settled. 1581-5.
In patients without a prior history of CAD with an 7. Candell-Riera J, Santana-Boado C, Castell-Conesa J, et al. Simul-
intermediate pre-test likelihood based on risk factors, a taneous dipyridamole/maximal subjective exercise with Tc-99m-
low-dose stress/high-dose rest Tc-99m protocol may be MIBI SPECT: improved diagnostic yield in coronary artery dis-
preferable since a significant percentage of these patients ease. J Am Coll Cardiol 1997;29:531-6.
8. Cerqueira MD, Verani MS, Schwaiger M, Heo J, Iskandrian AS.
will have normal stress studies, thereby avoiding addi- Safety profile of adenosine stress perfusion imaging: results from
tional radiation exposure from a rest study. Due to higher Adenoscan Multicenter Trial Registry. J Am Coll Cardiol 1994;
tracer uptake during low-dose stress, the waiting time to 23:384-90.
high-dose rest imaging needs to be longer (3-4 hours) or 9. Dakik HA, Wendt JA, Kimball K, Pratt CM, Mahmarian JJ.
the rest dose needs to be higher to achieve a rest/stress Prognostic value of adenosine Tl-201 myocardial perfusion imag-
ing after acute myocardial infarction: results of a prospective
ratio of at least 1:4.
clinical trial. J Nucl Cardiol 2005;12:276-83.
Tracer-specific imaging times. For Tc-99m ses- 10. Eagle KA, Berger PB, Calkins H, et al. ACC/AHA guideline
tamibi, minimum delays of 15 to 20 minutes for exercise, update on perioperative cardiovascular evaluation for noncardiac
45 to 60 minutes for rest, and 60 minutes for pharmaco- surgery. American College of Cardiology/American Heart Asso-
logic stress are optimal. For Tc-99m tetrofosmin, mini- ciation Task Force on Practice Guidelines (Committee to Update
the 1996 Guidelines on Peroperative Cardiovascular Evaluation for
mum delays of 10 to 15 minutes for exercise, 30 to 45
Noncardiac Surgery). Circulation 2002;105:1257-67.
minutes for rest, and 45 minutes for pharmacologic stress 11. Elliott MD, Holly TA, Leonard SM, Hendel RC. Impact of an
are optimal. Since there is minimal redistribution with abbreviated adenosine protocol incorporating adjunctive treadmill
these agents, longer delays, up to 2 hours, between the exercise on adverse effects and image quality in patients undergoing
radiotracer injection and imaging can be used for logis- stress myocardial perfusion imaging. J Nucl Cardiol 2000;7:584-9.
tical reasons. 12. Gibbons RJ, Balady GJ, Bricker JT, et al. ACC/AHA 2002
guideline update for exercise testing: summary article: a report of
Dual-isotope imaging. Use of Tl-201 for initial the American College of Cardiology/American Heart Association
rest imaging and a Tc-99m–labeled tracer for stress Task Force on Practice Guidelines (Committee to Update the 1997
perfusion imaging, as shown in Figure 7, allows a shorter Exercise Testing Guidelines). Circulation 2002;106:1883-92.
duration of the entire imaging protocol, but there is a 13. Hays JT, Mahmarian JJ, Cochran AJ, Verani MS. Dobutamine
significantly higher radiation dose to the patient. Use of thallium-201 tomography for evaluating patients with suspected
coronary artery disease unable to undergo exercise or vasodilator
rest/3- to 4-hour redistribution Tl-201 imaging prior to
pharmacologic stress testing. J Am Coll Cardiol 1993;21:1583-90.
the stress Tc-99m study provides valuable information 14. Higley B, Smith FW, Smith T, et al. Technetium-99m-1,2
on myocardial viability and should be considered in bis[bis(2-ethoxyethyl)phosphino]ethane: human biodistribution,
patients with prior infarction or heart failure. dosimetry and safety of a new myocardial perfusion imaging agent.
J Nucl Med 1993;34:30-8.
15. Holly TA, Satran A, Bromet DS, Mieres JH, Frey MJ, Elliott MD,
et al. The impact of adjunctive adenosine infusion during exercise
Suggested Reading myocardial perfusion imaging: results of the Both Exercise and
Adenosine Stress Test (BEAST) trial. J Nucl Cardiol 2003;
1. Abbott BG, Afshar M, Berger AK, Wackers FJT. Prognostic 10:291-6.
significance of ischemic electrocardiographic changes during aden- 16. Iskandrian AS, Verani MS, Heo J. Pharmacologic stress testing:
osine infusion in patients with normal myocardial perfusion mechanism of action, hemodynamic responses, and results in
imaging. J Nucl Cardiol 2003;10:9-16. detection of coronary artery disease. J Nucl Cardiol 1994;1:
2. Amanullah AM, Berman DS, Erel J, et al. Incremental prognostic 94-111.
value of adenosine myocardial perfusion single-photon emission 17. Jain D. Technetium labeled myocardial perfusion imaging agents.
computed tomography in women with suspected coronary artery Semin Nucl Med 1999;29:221-36.
disease. Am J Cardiol 1998;82:725-30. 18. Jain D, Wackers FJTh, Mattera J, McMahon M, Sinusas AJ, Zaret
3. Amanullah AM, Berman DS, Hachamovitch R, et al. Identification BL. Biokinetics of technetium-99m-tetrofosmin: myocardial per-
of severe or extensive coronary artery disease in women by fusion imaging agent: implications for a one day imaging protocol.
adenosine technetium-99m sestamibi SPECT. Am J Cardiol 1997; J Nucl Med 1993;34:1254-9.
80:132-7. 19. Jamil G, Ahlberg AW, Elliott MD, Hendel RC, Holly T, McGill
4. Aqel RA, Zoghbi GJ, Trimm JR, Baldwin SA, Iskandrian AE. CC, et al. Impact of limited treadmill exercise on adenosine
Effect of caffeine administered intravenously on intracoronary- Tc-99m sestamibi single-photon emission computed tomographic
administered adenosine-induced coronary hemodynamics in pa- myocardial perfusion imaging in coronary artery disease. Am
tients with coronary artery disease. Am J Cardiol 2004;93:343-6. J Cardiol 1999;84:400-3.
5. Berman DS, Kiat H, Friedman JD, et al. Separate acquisition 20. Kelley JD, Forster AM, Higley B, et al. Technetium-99m-
rest thallium-201/stress technetium-99m sestamibi dual-isotope tetrofosmin a new radiopharmaceutical for myocardial perfu-
myocardial perfusion single-photon emission computed tomog- sion imaging. J Nucl Med 1993;34:222-7.
raphy: a clinical validation study. J Am Coll Cardiol 1993;22: 21. Klodas E, Miller TD, Christian TF, Hodge DO, Gibbons RJ.
1455-64. Prognostic significance of ischemic electrocardiographic changes
e90 Henzlova et al Journal of Nuclear Cardiology
Stress protocols and tracers November/December 2006

during vasodilator stress testing in patients with normal SPECT 31. Stein RA, Chaitman BR, Balady GJ, et al. Safety and utility of
images. J Nucl Cardiol 2003;10:4-8. exercise testing in emergency room chest pain centers: an advisory
22. Lette J, Tatum JL, Fraser S, Miller DD, et al. Safety of dipyrid- from the Committee on Exercise, Rehabilitation, and Prevention,
amole testing in 73,806 patients: the Multicenter Dipyridamole Council on Clinical Cardiology, American Heart Association.
Safety Study. J Nucl Cardiol 1995;2:3-17. Circulation 2000;102:1463-7.
23. Mahmarian JJ, Shaw LJ, Olszewski GH, Pounds BK, Frias ME, Pratt 32. Taillefer R. Radionuclide myocardial perfusion imaging protocols.
CM; INSPIRE Investigators. Adenosine sestamibi SPECT post- In: Heller GV, Hendel RC, editors. Nuclear cardiology: practical
infarction evaluation (INSPIRE) trial: a randomized, prospective applications. New York: McGraw-Hill Medical Publishing Divi-
multicenter trial evaluating the role of adenosine Tc-99m sestamibi sion; 2004.
SPECT for assessing risk and therapeutic outcomes in survivors of 33. Taillefer R, Laflamme L, Dupras G, Picard M, Phaneuf DC,
acute myocardial infarction. J Nucl Cardiol 2004;11:458-69. Leveille J. Myocardial perfusion imaging with 99mTc-methoxy-
24. Mahmood S, Gunning M, Bomanji JB, et al. Combined rest isobutyl-isonitrile (MIBI): comparison of short and long time
thallium-201/stress technetium-99m-tetrofosmin SPECT: feasibil- intervals between rest and stress injections. Preliminary results.
ity and diagnostic accuracy of a 90-minute protocol. J Nucl Med Eur J Nucl Med 1988;13:515-22.
1995;36:932-5. 34. Thomas GS, Prill NV, Majmundar H, Fabrizi RR, Thomas JJ,
25. Mieres JH, Shaw LJ, Arai A, Budoff MJ, Flamm SD, Hundley Hayashida C, et al. Treadmill exercise during adenosine infusion is
WG, et al. Role of noninvasive testing in the clinical evaluation of safe, results in fewer adverse reactions, and improves myocardial
women with suspected coronary artery disease. Circulation 2005; perfusion image quality. J Nucl Cardiol 2000;7:439-46.
111:682-96. 35. Trueth MG, Reyes GA, He ZX, Cwaig E, Mahmarian JJ, Verani
26. O’Keefe JH, Bateman TM, Handlin LR, Barnhart CS. Four- versus MS. Tolerance and diagnostic accuracy of an abbreviated
6-minute infusion protocol for adenosine thallium-201 single adenosine infusion for myocardial scintigraphy: a randomized,
photon emission computed tomography imaging. Am Heart J prospective study. J Nucl Cardiol 2001;8:548-54.
1995;129:482-7. 36. Vaduganathan P, He ZX, Raghavan C, Mahmarian JJ, Verani MS.
27. Pennell DJ, Mavrogeni S, Forbat SM, Karwatowski SP, Under- Detection of left anterior descending coronary artery stenosis in
wood SR. Adenosine combined with dynamic exercise for myo- patients with left bundle branch block: exercise, adenosine or
cardial perfusion imaging. J Am Coll Cardiol 1995;25:1300-9. dobutamine imaging. J Am Coll Cardiol 1996;28:543-53.
28. Pennell DJ, Underwood SR, Ell PJ. Safety of dobutamine stress for 37. Vitola JV, Brambatti JC, Caligaris F, Lesse CR, Nogueira PR,
thallium myocardial perfusion tomography in patients with asthma. Joaquim AI, et al. Exercise supplementation to dipyridamole
Am J Cardiol 1993;71:1346-50. prevents hypotension, improves electrocardiogram sensitivity, and
29. Pennell DJ, Underwood SR, Swanton RH, Walker JM, Ell PJ. increases heart-to-liver activity ratio on Tc-99m sestamibi imag-
Dobutamine thallium myocardial perfusion tomography. J Am ing. J Nucl Cardiol 2001;8:652-9.
Coll Cardiol 1991;18:1471-9. 38. Wackers FJTh, Berman DS, Maddahi J, et al. Technetium-99m
30. Samady H, Wackers FJTh, Joska TM, Zaret BL, Jain D. Pharma- hexakis 2-methoxyisobutyl isonitrile: human biodistribution, do-
cologic stress perfusion imaging with adenosine: role of simulta- simetry, safety and preliminary comparison to thallium-201 for
neous low-level treadmill exercise. J Nucl Cardiol 2002;9:188-96. myocardial perfusion imaging. J Nucl Med 1989;30:301-11.
 Myocardial perfusion planar imaging
Peter L. Tilkemeier, MD,a and Frans J. Th. Wackers, MD, PhDb

Although single photon emission computed tomog- severity of inducible myocardial ischemia, submaximal
raphy (SPECT) is preferable for myocardial perfusion exercise can provide clinically relevant information.2
scintigraphy, in a minimal number of circumstances, pla- Positioning. The most important part of positioning
nar imaging may be useful or may be the only modality is the ability to reproduce the same position on initial and
available. delayed (or rest) images. Even slight differences in
Purpose. To evaluate regional myocardial perfusion angulation of the camera, positioning of breasts or other
and function. Planar imaging is an acceptable method for soft tissue, or the pressure of the camera on the chest wall
myocardial perfusion imaging. The anatomy of the heart can produce artifacts and inaccuracies in comparing rest
is sufficiently simple that the imaging specialist can and stress images. It is vital to bring the camera head as
comprehend the location and extent of defects from close to or touching the patient’s chest in order to get the
multiple projections without need of computer recon- highest possible count rates, as opposed to SPECT imag-
struction. Although SPECT imaging is presently consid- ing, where some distance is necessary to avoid collisions
ered state-of-the-art for myocardial perfusion imaging with the patient during rotational acquisition.
and preferable, planar imaging still has a role in the daily The standard imaging positions are supine ante-
routine of a laboratory. Imaging at the bedside of acutely rior, supine 45° left anterior oblique (LAO), and a
ill patients, or instrumented patients, can only be per- right-side decubitus 90° left lateral (LL). The 90° LL
formed using planar imaging technique and portable decubitus view provides optimal visualization of the
gamma cameras. Planar views can be quickly repeated if inferior wall and reduces subdiaphragmatic and breast
the patient moves during acquisition. Planar imaging attenuation artifacts. Admittedly, the right-side decu-
may be the only way to acquire images in very obese bitus position is less stable than the supine position,
patients, who are too heavy for the imaging table of a making it somewhat more difficult to obtain identical
SPECT camera. It may also be the only way to acquire repositioning. An alternate LL view is the shallow 70°
images in severely claustrophobic patients. LAO position. The latter position is suboptimal at
Electrocardiography (ECG)– gated planar images times due to frequent occurrence of artifacts: subdia-
can be obtained using standard software for equilibrium phragmatic attenuation of the inferior wall and breast
radionuclide angiography studies. Finally, planar imag- attenuation of the anterior wall.3
ing is the basis for good SPECT imaging. The ability to The LAO view should be chosen in such a way that
obtain high-quality planar images is an essential skill, the right ventricle and left ventricle are well separated by
even for those who routinely use SPECT imaging.1 a vertically visualized septum (ie, “best septal” view).
One should be aware that in individual patients the heart
PROCEDURE may not always be in the same position. Hearts may be
rotated clockwise or counterclockwise so that a “straight”
Exercise. Adequate exercise is most important if the 45° LAO will not always display the desired image. It is
aim of the study is to detect coronary artery disease preferred to search for the “best septal” view instead of
(CAD). In patients with mild and moderate CAD, myo- a straight 45° LAO. The angulation of the detector head
cardial blood flow may become abnormal only at high for acquisition of anterior and LL views should then be
heart rates or at high double products. At lower heart correspondingly modified. The advantage of this option
rates, myocardial blood flow may be normal and perfu- is that it provides greater standardization of the imaged
sion images will be correspondingly normal. In patients left ventricle, which will simplify quantitative analysis.
with known CAD who are being evaluated for extent and The disadvantage of this option is the increased complex-
ity, and it carries the potential for non-reproducible posi-
Chair.a Member.b tioning.
J Nucl Cardiol 2006;13:e91-6.
1071-3581/$32.00
Female patients should be imaged consistently with
Copyright © 2006 by the American Society of Nuclear Cardiology. the bra off and without camera pressure, which might
doi:10.1016/j.nuclcard.2006.08.012 produce variable tissue displacement.

e91
e92 Tilkemeier and Wackers Journal of Nuclear Cardiology
Myocardial perfusion planar imaging November/December 2006

Positioning frames of the ECG-gated myocardial perfusion images

can be summed to produce a single static planar image for
Patient Detector conventional visual and quantitative analysis.4
View position position Alternative The total imaging time for static planar Tl-201
imaging should be 8 to 10 minutes per view, whereas the
Shallow LAO Supine 45° “Best septal” total imaging time for static planar Tc-99m–labeled
Anterior Supine 0° “Best septal” agents can be reduced to 5 minutes per view. When ECG
minus 45° gating is used, imaging time per view using Tc-99m
Steep LAO Supine 70° “Best septal” agents should be extended to 8 to 10 minutes. For
plus 25°
ECG-gated image acquisition with Tl-201, image acqui-
Left lateral Right decubitus 0° Not
sition time per view should be at least 10 minutes per
applicable
view. Planar images acquired with a 10-inch-diameter
Positions of cameras and patients must be exactly gamma camera should have at least 600,000 counts
reproduced for rest and stress imaging. (1 million counts preferred) in the field of view. Alterna-
tively, optimal count density can be defined as at least
100 counts in the pixel with maximal counts in the left
IMAGE ACQUISITION
ventricle.5,6
Gamma camera positions are as shown in the pre- In female patients an optional l-minute image of a
vious table. Using either thallium 201 (T1-201) or line source marker that delineates the contour of the breast
technetium 99m (Tc-99m)–labeled agents, one may op- can be acquired to aid in identifying breast attenuation
tionally use ECG-synchronized gating and acquire in artifacts.
16-frame, multiple-gated acquisition (equilibrium radio- Cardiac shields or other masking devices should
nuclide angiography) mode. No beat rejection should be not be used. The extracardiac background cannot be
used. In patients with atrial fibrillation, ECG gating determined correctly unless both the cardiac activity
should not be performed. ECG-gated acquisition allows and extracardiac activity have been recorded in the
for cine review of wall thickening and motion. The multiple raw data.

Acquisition
Tc-99m Tl-201

Collimator High resolution Low energy, medium resolution

Field of view Full 10-inch field of small camera or 1.2-1.5 Same
zoom with large–field-of-view camera
Matrix 128 ⫻ 128 128 ⫻ 128
Window 140 keV 20% centered 78 keV, 30% centered
Gating Optional 16 frames/cardiac cycle Same
Imaging time (per view) 5 min (10 min ECG-gated) 8-10 min (10 min ECG-gated)
Imaging counts At least 1 million At least 1 million

QUANTITATIVE PROCESSING OF PLANAR IMAGES comparing defects in stress and rest images and to detect
Quantitative processing includes using the computer subtle defect reversibility. Registration of the stress and
to produce standardized raw images for visual evalua- rest images also can be performed to ensure that the same
tion. The gray scale should be fully utilized to display the myocardial region is being sampled. A normal database
heart normalized to maximal left ventricular (LV) count also may be incorporated in the quantitative program so
density, and not scaled to visceral activity. Background that segmental tracer activity can be compared with the
subtraction is performed. The background-subtracted average obtained from the normal database.7-11
images are useful for visual evaluation and are used for There is no single “generic” description for what all
measurements of myocardial activity. These measure- good computer programs should offer. There are com-
ments provide quantitative determination of a suspected mon features among several successful programs. In the
defect so that consistent standards can be set for defect following section we review these features and comment
detection. The measurements are especially useful in on acceptable variations.
Journal of Nuclear Cardiology Tilkemeier and Wackers e93
Volume 13, Number 6;e91-6 Myocardial perfusion planar imaging

Regions of interest. The first step in quantification isons. There are several ways of doing this. Maximizing
is to locate the heart by placing reference regions of the cross-correlation coefficient between the two images
interest around the heart. Regions can be rectangular or is a robust method that can be performed by the com-
elliptical. Elliptical regions fit the heart better. Rectan- puter without operator intervention.
gular regions are best used by having the operator set the Profile generation. After subtraction of the refer-
boundaries by touching the apparent “edge” of the heart ence (background) plane to compensate for tissue cross-
and then moving the region outward approximately talk and registration of the images to allow precise
4 pixels. This operation is highly reproducible. The comparisons, quantification becomes the relatively sim-
reference boundaries are used to separate the region ple matter of finding a convenient way of indicating
containing myocardial activity from background activity. image count density. Again, there are several ways of
Background subtraction. Background subtraction doing this. One basic way is to display count profiles
is the removal of the background or, more precisely, the across the heart. Four profiles will sample the myocardial
“tissue cross-talk” from the raw image. For each image, count distribution adequately within the limitations of
a background image is generated from the smoothed image resolution (each profile represents an average of
image using the above-mentioned reference background about a 1-cm-wide slice across the heart) and produce an
region. The background is then subtracted from the intelligible display. A more commonly used alternative
unsmoothed raw image, leaving behind the myocardial for graphic display of myocardial activity is the circum-
activity. ferential count distribution profile. The circumferential
Background subtraction is essential to planar imag- profile method provides a more compact and dense
ing both for valid quantitation and to restore defect single-curve display of counts sampled around the myo-
contrast adequate for visual assessment. Background cardial “rim” and allows the simple plot of a second
correction is in fact more critical for planar imaging with profile indicating normal limits. Either method, trans-
Tc-99m–labeled agents than for imaging with Tl-201.
verse count profiles or circumferential count profiles,
The relative tissue distribution of Tc-99m–labeled agents
will provide an adequate and ultimately equivalent quan-
at rest and after exercise may be markedly different
titative representation of myocardial activity. Either
compared to that of Tl-201.12
method facilitates standardization and reproducibility of
A modified version of the conventional interpolative
image interpretation.14
background algorithm has worked well for both Tl-201
A more fundamental choice is what parameter to use
and Tc-99m sestamibi planar images. The modification
to quantify myocardial activity. Many programs, includ-
allows the background-defining regions to cross regions
ing most methods used to generate bull’s-eye maps for
of intense extracardiac activity without causing signifi-
SPECT imaging, perform a search across the myocardium
cant background error in the background-subtracted car-
diac image.13 for the maximum pixel count in a transmural myocardial
Rescaling the image gray scale. When Tl-201 is sample. The other choice is to take an average of counts
used as the imaging agent, the heart is usually the organ in the transmural sample. The advantage of the latter
with the most intense activity. When using myocardial method is that it reduces statistical noise because it is an
tracers labeled with Tc-99m, activity in the abdominal average, and it is intuitively more representative in regions
viscera often exceeds that of the heart. This normally of subendocardial scar or ischemia. The disadvantage is
causes the computer to scale image intensities to the that the transmural average is quite sensitive to the
extracardiac activity, which will cause erratic and sub- accurate definition of endocardium and epicardium.
optimal visualization of the heart. Any computer pro- Variability in locating the epicardial and endocardial
gram for quantitative image processing should have a limits probably nullifies the gain in precision from the
convenient mechanism to suppress activity outside the averaging of more pixels. The use of maximum counts
heart if it becomes greater than the activity in the heart. provides a quantitative parameter that is less sensitive to
When comparing initial and delayed images or images edge location. This parameter has been used extensively
obtained after reinjection, each image should be individ- and has been reasonably robust in practice. Either
ually scaled to the area of most intense cardiac activity. method is usable. Normal standards and normal limits
If the images were scaled to different maxima, the will be quite different for those using the transmural
appearance of defect reversibility would be distorted. average. They are not comparable with values based on
Image registration. Comparison of rest and exer- transmural maximum.15
cise images to detect redistribution or reversibility can be Normal database. Data from “normal” subjects
accurate only if the same myocardial segments are being may be incorporated into the computer program and
compared. Image registration so that stress and rest indicated in the output as normal limits. Because of
images are precisely aligned facilitates accurate compar- variations in equipment and positioning, the normal
e94 Tilkemeier and Wackers Journal of Nuclear Cardiology
Myocardial perfusion planar imaging November/December 2006

database should not be used until it has been validated between image sets. Images are typically normalized to
in-house using standardized imaging protocols. themselves. Planar images should be interpreted without
Along with the average values obtained from the any processing. Background-subtracted images may be
normal database, the standard deviations (SDs) also need generated for quantitative analysis but should not be
to be obtained. Different myocardial segments will have interpreted without viewing the unsubtracted images as
different degrees of normal variability, which should be well.
accounted for in deciding if a segment is outside normal Evaluation for technical sources of artifacts. The
limits. Individual segments may be flagged using limits images should then be carefully inspected for potential
of 2.0 to 2.5 SDs. The computer may also flag reversible image artifacts, the most common of which is attenua-
segments, but this is a more complex operation. The tion.17 Suspected soft-tissue attenuation should be thor-
database must have SDs comparing stress and rest oughly evaluated and its effect on the interpretation
segments. If a segment has a significant stress defect, carefully considered. The use of breast marker images
reversibility may be indicated if that defect changes may be helpful in distinguishing true perfusion defects
toward normal by more than 1 SD. Additional “expert from breast attenuation.18 Attenuation of the inferior
logic” may also be incorporated to scan for secondary wall by the diaphragm or an enlarged right ventricle
segments with nonsignificantly reduced initial uptake should also be considered. Other sources of attenuation
and significant reversibility.1,16 (eg, pleural effusions or infiltrates, foreign objects, other
Limitations. Well-trained readers consistently out- soft tissues) should be noted.
perform readings even from relatively sophisticated Adjacent subdiaphragmatic activity, as is frequently
computer programs. The programs are valuable in stan- seen in the liver and bowel, may create overlap artifacts
dardizing the images and image processing and in that spuriously increase the activity in the inferior wall,
maintaining consistent interpretive standards. However, creating the appearance of a relative paucity of counts in
the judgment of a well-trained reader should override the other myocardial segments. Intense noncardiac activity
computer logic. Computer programs that dogmatically may cause scaling artifacts in the myocardial images.
indicate normal and abnormal scans or scan segments Techniques for masking such noncardiac activity are
can be intimidating and misleading. Readers must be available.
prepared to disagree and overrule the computer. Other- Count-poor studies are subject to misinterpretation.
wise, the readers will be no better than the computer. Apparent perfusion abnormalities may resolve when a
statistically adequate study is available. The factors
leading to suboptimal count statistics include body hab-
INTERPRETATION AND REPORTING
itus, radionuclide dose, collimation, window, acquisition
Images should be assessed initially for technical ade- time, and the level of myocardial blood flow.
quacy, including target-to-background activity, splanch- Patient motion is rarely a problem because of the
nic tracer uptake, count adequacy in the cardiac region of short duration of imaging.
interest, adequacy of count normalization and masking, Initial perfusion image analysis and interpreta-
appropriate orientation of the planar projections, and tion. The initial interpretation of the perfusion scan should
registration of the stress and delay (Tl-201) or stress and be performed without any clinical information other than
rest (Tc-99m agents) planar projections and appropriate the patient’s gender, height and weight, and presence of left
location and alignment of the ventricular region of bundle-branch block. This approach minimizes the bias in
interest utilized for quantitative analysis, if performed. study interpretation. All relevant clinical information should
Display. Planar perfusion images may be displayed be reviewed after a preliminary impression is formed.
by use of the computer screen, x-ray film, or paper copy. Before segmental analysis of myocardial perfusion, the
The use of the computer screen is strongly recommended reader should note whether there is LV cavity dilation at
and is the preferred medium. Counts should preferably rest or during exercise or pharmacologic challenge. Dilation
be represented by a linear gray scale. If a color table is on both the stress and resting studies suggests LV dysfunc-
used, the scale should be simultaneously displayed on tion but may occur in volume-overload states with normal
screen. Otherwise, the adequacy of the display medium ventricular function. Transient ischemic dilation is a marker
(film or paper) should be established by inspection of a for multivessel CAD. It is typically described qualitatively
standard test pattern, which provides testing of both but may be quantified.19
resolution and gray scale. The presence of increased lung uptake should be
The initial set of images is typically displayed noted by comparison of the pulmonary to myocardial
together with the subsequent set of images aligned counts. This is especially important when imaging with
adjacent to or underneath it. The interpreting physician Tl-201. Although qualitative assessment is standard,
should confirm that the imaging angles have not changed calculation of lung-to-heart ratios is preferred. Lung
Journal of Nuclear Cardiology Tilkemeier and Wackers e95
Volume 13, Number 6;e91-6 Myocardial perfusion planar imaging

uptake may be increased on stress images if the exercise

was submaximal. Lung uptake during dipyridamole and
adenosine imaging tends to be higher than that seen
during adequate exercise. Both the likelihood of CAD
and the risk of an adverse outcome are increased when
lung uptake is increased.
Right ventricular (RV) intensity is normally about
50% of the peak LV intensity. Increased RV uptake is
indicative of RV pressure overload, most typically be-
cause of pulmonary hypertension. RV intensity may also
appear increased when LV uptake is globally reduced.
The images should be examined for the presence of
activity in organs other than the heart and pulmonary Figure 1. Planar MPI segmentation.
vasculature. Thyroid uptake, breast uptake, pulmonary pa-
renchymal (tumor) uptake, or uptake in other thoracic or
upper abdominal structures should be noted. Hepatic or
splenic enlargement and the apparent absence of a gallblad- objectively designate segments as normal or abnormal. A
der on technetium studies may be of clinical significance. 5-point model has been recommended for semiquantita-
A statement regarding the overall quality of the tive scoring of segments.
study is helpful for two reasons. First, it alerts physicians
using the report to any shortcomings that might reduce Category Score
the accuracy of the data and their interpretation. Second,
it is useful as an inclusion/exclusion criterion for subse- Normal perfusion 0
quent acceptance of the study for research or other Mild reduction in counts—not definitely 1
statistical analyses.20 abnormal
Segmental perfusion assessment. The detection Moderate reduction in counts—definitely 2
and localization of perfusion defects is typically per- abnormal
formed by use of visual analysis of the unprocessed Severe reduction in counts 3
images. Perfusion defects are assigned to a particular Absent uptake 4
myocardial segment or segments. Standardized nomen-
clature is recommended (Figure 1). In the anterior planar The summed scores may also be calculated by
image, the anterolateral and inferior walls are divided adding all segment scores together to provide an inte-
into basal (1 and 5) and mid (2 and 4) segments, and the grated index of severity and extent. This may be done on
apex (3) separates the inferior and anterolateral walls. In the initial and resting or delayed images, and a summed
the LAO view, the septum is divided into anteroseptal (6) difference score could then be used to represent the
and inferoseptal (7) segments, whereas the lateral wall is amount of defect reversibility.21
divided into inferolateral (9) and anterolateral (10) seg- Perfusion defect severity and extent: Quantitative.
ments, and the inferior segment (8) separates the septum Quantitative analysis can be useful to supplement visual
from the lateral wall. In the lateral view, the anterior and interpretation of the images. In planar perfusion imaging,
inferior walls are divided into basal (11 and 15) and mid quantitative analysis requires that a background subtrac-
(12 and 14) segments, whereas the apex (13) separates tion be applied. The physician should review the back-
the anterior and inferior walls. Perfusion defect severity ground-subtracted images for technical adequacy. Quan-
is typically expressed qualitatively with terms such as titative data may be displayed as graphs of counts versus
mild, moderate, or severe. Defect extent may be charac- angular sampling location or as profiles that compare
terized qualitatively as small, medium, or large. counts in contralateral walls of each image. Supporting
Rather than the qualitative evaluation of perfusion data exist for both approaches. When counts are graphed
defects, it is preferred that the physician apply a semi- versus radial sampling location, the resultant curve is
quantitative method on the basis of a validated segmental typically compared with a gender-matched, reference
scoring system. This approach standardizes the visual database. The 2.0- to 2.5-SD curve is usually graphed
interpretation of scans, reduces the likelihood of over- with the patient’s curve, and segments where the pa-
looking significant defects, and provides an important tient’s curve drops below the reference curve are consid-
semiquantitative index that is applicable to diagnostic ered abnormal. The area between the patient’s curve and
and prognostic assessments. Furthermore, semiquantita- the reference curve is a quantitative index of the com-
tive scoring can be used to more reproducibly and bined extent and severity of the perfusion abnormality.7-9
e96 Tilkemeier and Wackers Journal of Nuclear Cardiology
Myocardial perfusion planar imaging November/December 2006

Separate gender-based reference databases are rec- for detection of coronary artery disease. J Nucl Med 1981;22:
ommended for thallium and technetium-based perfusion 585-93.
8. Garcia E, Maddahi J, Berman D, Waxman A. Space/time quanti-
agents because the myocardial distribution and, in par- tation of thallium-201 myocardial scintigraphy. J Nucl Med 1981;
ticular, the extracardiac activity are significantly different. 22:309-17.
Separate databases are also preferred, when available, for 9. Kaul S, Chesler DA, Boucher CA, Okada RD. Quantitative aspects
pharmacologic and exercise studies. Quantitative analy- of myocardial perfusion imaging. Semin Nucl Med 1987;17:
sis of the data should not be used as a surrogate for visual 131-44.
10. Van Train KF, Berman DS, Garcia EV, Berger HJ, Sands MJ,
analysis but rather as an expert opinion that may be used Friedman JD, et al. Quantitative analysis of stress thallium-201
to modify the physician’s impression. myocardial scintigrams: a multicenter trial. J Nucl Med 1986;27:
Reversibility of perfusion defects may be categorized 17-25.
qualitatively as minimal, partial, or complete. Reversibil- 11. Wackers FJ, Fetterman RC, Mattera JA, Clements JP. Quantitative
ity can be defined quantitatively as defects in which planar thallium-201 stress scintigraphy: a critical evaluation of the
method. Semin Nucl Med 1985;15:46-66.
pixels improve to fewer than 2.5 SDs from the normal 12. Sinusas AJ, Beller GA, Smith WH, Vinson EL, Brookeman V,
reference distribution at that location. The number of Watson DD. Quantitative planar imaging with technetium-99m
pixels that must show improvement for reversibility to be methoxyisobutyl isonitrile: comparison of uptake patterns with
deemed present is arbitrary. thallium-201. J Nucl Med 1989;30:1456-63.
So-called reverse redistribution may be seen in 13. Koster K, Wackers FJ, Mattera JA, Fetterman RC. Quantitative
analysis of planar technetium-99m-sestamibi myocardial perfusion
stress-delayed thallium imaging sequences. Reverse re- images using modified background subtraction. J Nucl Med 1990;
distribution refers to segments with decreased or normal 31:1400-8.
intensity on the initial set of images that show even less 14. Kaul S, Chesler DA, Okada RD, Boucher CA. Computer versus
relative intensity on the delayed images. The interpreta- visual analysis of exercise thallium-201 images: a critical
tion of the finding remains controversial. appraisal in 325 patients with chest pain. Am Heart J 1987;114:
1129-37.
15. Sigal SL, Soufer R, Fetterman RC, Mattera JA, Wackers FJ.
Reproducibility of quantitative planar thallium-201 scintigraphy:
References quantitative criteria for reversibility of myocardial perfusion de-
fects. J Nucl Med 1991;32:759-65.
1. Maddahi J, Rodrigues E, Berman DS, Kiat H. State-of-the-art 16. Reisman S, Maddahi J, Van Train K, Garcia E, Berman D.
myocardial perfusion imaging. Cardiol Clin 1994;12:199-222. Quantitation of extent, depth, and severity of planar thallium
2. Mahmarian JJ, Verani MS. Exercise thallium-201 perfusion scin- defects in patients undergoing exercise thallium-201 scintigraphy.
tigraphy in the assessment of coronary artery disease. Am J Cardiol J Nucl Med 1986;27:1273-81.
1991;67:2D-11D. 17. Wackers FJ. Artifacts in planar and SPECT myocardial perfusion
3. Port SC, Oshima M, Ray G, McNamee P, Schmidt DH. Assess- imaging. Am J Card Imaging 1992;6:42-57.
ment of single vessel coronary artery disease: results of exercise 18. Schechter D, Bocher M, Karger H, Gotsman MS, Chisin R. Breast
electrocardiography, thallium-201 myocardial perfusion imaging artifact in planar cardiac imaging: a comparison of 201T1 to
and radionuclide angiography. J Am Coll Cardiol 1985;6:75-83. sestamibi-99mTc. Cardiology 1994;84:339-44.
4. Verzijlbergen JF, Suttorp MJ, Ascoop CA, Zwinderman AH, 19. Weiss AT, Berman DS, Lew AS, Nielsen J, Potkin B, Swan HJ,
Niemeyer MG, van der Wall EE, et al. Combined assessment of et al. Transient ischemic dilation of the left ventricle on stress
technetium-99m SESTAMIBI planar myocardial perfusion images thallium-201 scintigraphy: a marker of severe and extensive
at rest and during exercise with rest/exercise left ventricular wall coronary artery disease. J Am Coll Cardiol 1987;9:752-9.
motion studies evaluated from gated myocardial perfusion studies. 20. Wackers FJ, Bodenheimer M, Fleiss JL, Brown M. Factors
Am Heart J 1992;123:59-68. affecting uniformity in interpretation of planar thallium-201 imag-
5. Watson DD, Smith WH. Sestamibi and the issue of tissue ing in a multicenter trial. The Multicenter Study on Silent Myo-
crosstalk. J Nucl Med 1990;31:1409-11. cardial Ischemia (MSSMI) Thallium-201 Investigators. J Am Coll
6. Smith WH, Watson DD. Technical aspects of myocardial planar Cardiol 1993;21:1064-74.
imaging with technetium-99m sestamibi. Am J Cardiol 1990;66: 21. Porenta G, Dorffner G, Kundrat S, Petta P, Duit-Schedlmayer J,
16E-22E. Sochor H. Automated interpretation of planar thallium-201-
7. Berger BC, Watson DD, Taylor GJ, Craddock GB, Martin RP, dipyridamole stress-redistribution scintigrams using artificial neu-
Teates CD, et al. Quantitative thallium-201 exercise scintigraphy ral networks. J Nucl Med 1994;35:2041-7.
 Myocardial perfusion and function single photon
emission computed tomography
Christopher L. Hansen, MD,a Richard A. Goldstein, MD,a Daniel S. Berman, MD,b
Keith B. Churchwell, MD,b C. David Cooke, MSEE,b James R. Corbett, MD,b
S. James Cullom, PhD,b Seth T. Dahlberg, MD,b James R. Galt, PhD,b
Ravi K. Garg, MD,b Gary V. Heller, MD, PhD,b Mark C. Hyun, CNMT, NCT, RT(N,R),b
Lynne L. Johnson, MD,b April Mann, CNMT, NCT, RT(N),b
Benjamin D. McCallister, Jr, MD,b Raymond Taillefer, MD,b R. Parker Ward, MD,b
John J. Mahmarian, MDc

Purpose. To evaluate regional myocardial perfusion 2. Position. Factors influencing patient position in-
and function. clude camera/gantry design, minimization of arti-
facts, and patient comfort. The supine position is
ACQUISITION PROTOCOLS routinely used for SPECT imaging in all protocols in
most currently available systems. Prone imaging has
Protocols for the various nuclear cardiology single been reported to produce less patient motion and less
photon emission computed tomography (SPECT) acqui- inferior wall attenuation than supine imaging. The
sition studies are presented in the following pages. For combination of supine and prone images may be
each of the protocols, the acquisition parameters are helpful in identifying breast attenuation and attenu-
listed along with their corresponding value for exercise ation due to excessive lateral chest-wall fat, due to
and rest. Implementation of these protocol acquisition the shift in position of the attenuating structures that
parameters has been shown to provide acceptable images occur in the prone position. In some laboratories the
of good quality for routine clinical interpretation and advantages of prone imaging in clarifying artifactual
quantitation. However, protocol parameters other than defects have led to a routine use of the combination
those listed may be preferred at some institutions, and of supine followed by prone acquisitions. However,
ongoing research into corrections for attenuation, scatter, since the prone position may cause an artifactual
and camera response depth dependence may result in anteroseptal defect secondary to increased sternal
new optimal parameters in the future. Therefore these attenuation in this position, prone imaging is used as
protocols should be viewed as the current suggested list an adjunct to, not a replacement for, supine imaging.
of acquisition parameters. A description for each of the When being used in this fashion, gating is generally
acquisition parameters is listed below. not performed (unless the supine gated images were
1. Dose. The doses for each of the protocols represent deemed uninterpretable for patient motion), and the
standard doses commonly used clinically. The stan- acquisition time for the secondary (prone) image set
dard doses described are given for an average 70-kg is reduced by 20% to 40%. Using a dual-detector
patient. Doses may be adjusted upward for heavier camera with a 25- to 30-mCi Tc-99m dose, supine
patients by 0.04 mCi/kg for thallium 201 and by 0.31 acquisitions are performed for 25 seconds per stop
mCi/kg for technetium 99m. Another option is and prone acquisitions for 15 seconds per stop, with
increased imaging times or use of multidetector 30 to 32 stops per detector being obtained per
systems. Similar Tl-201 doses are used for ungated acquisition. With Tl-201, imaging should be begun
and gated studies. Gated Tl-201 imaging times can approximately 10 minutes after stress testing, and if
be adjusted based on the counts acquired for a soft tissue attenuation or patient motion compro-
preliminary 4-minute planar study in order to ensure mises a study, the benefit of repeating the acquisition
acquiring at least 500,000 background-subtracted is questionable. In contrast, Tc-99m sestamibi or
myocardial counts. Tc-99m tetrofosmin permits stress testing and tracer
injection to take place at a location remote from the
imaging laboratory and image acquisition can sim-
Co-Chair.a Member.b Board Reviewer.c ply be repeated when patient motion, soft tissue
J Nucl Cardiol 2006;13:e97-120.
1071-3581/$32.00
attenuation, or another artifact is considered to be
Copyright © 2006 by the American Society of Nuclear Cardiology. responsible for the production of a perfusion defect.
doi:10.1016/j.nuclcard.2006.08.008 Some camera/gantry designs require the patient to be
e97
e98 Hansen et al Journal of Nuclear Cardiology
Myocardial perfusion and function SPECT November/December 2006

positioned in a more upright position. Changes in oblique) compared to a 360° orbit. The recommen-
patient positioning from those described above will dation of which orbit will depend on the camera
likely cause changes in the distribution of adjacent configuration; it does not seem to be worthwhile to
soft tissue attenuation and need to be considered in increase imaging time to complete a 360° orbit since
image interpretation. New normal databases will much better count statistics will be obtained if that
most likely need to be generated for these configu- time is used to increase acquisition on a 180° orbit.
rations. A 360° orbit is appropriate for 3-headed cameras
3. Delay time. These times are listed as ranges; spe- with a 360° configuration where a 360° orbit is
cific values are optional. The objectives are to allow acquired in the same time as a 180° orbit. The utility
the patient to recover fully from exercise, thus of the posterior 180° of a 360° orbit is much greater
allowing the heart rate to return to baseline (reducing for higher-energy radioisotopes (such as technetium)
gating artifact), avoiding “upward creep” from compared to low-energy radioisotopes (such as thal-
changes in respiratory patterns while dyspnea re- lium).
solves and minimizing interference from hepatic 7. Orbit type. The main orbit options in SPECT
uptake. Provided that imaging times fall within the myocardial perfusion imaging are circular versus
specified ranges, clinically useful SPECT images noncircular (elliptical or body-contoured) orbits.
should result. Noncircular orbits follow the contour of the patient,
4. Energy windows. Energy window position is deter- bringing the camera closer to the patient, thereby
mined by the radioisotope employed: 140 keV for optimizing spatial resolution. Circular orbits main-
technetium-based perfusion agents and 70 keV for tain a fixed radius of rotation and on average result
thallium. It is reasonable to simultaneously acquire in the detector being further from the patient. In
the higher energy peaks of thallium (135 and 167 keV) general, there is reduced spatial resolution with
on cameras that are capable of doing this. The circular orbits since the detector-to-source distance
window sizes are determined largely by convention is greater with this technique. Imaging artifacts have
and reflect the tradeoff between image counts and been observed when noncircular orbits are used, due
resolution. The values shown are the most com- to increased variation of source-to-detector distance,
monly used and have been found to be acceptable for resulting in variation of spatial resolution. Circular
most cameras. On systems offering improved energy acquisitions continue to be the most frequently used
resolution, the window size may be reduced, result- option, but some manufacturers do provide noncir-
ing in decreased scatter and improved image reso- cular orbit capability.
lution, so long as imaging times are extended to 8. Pixel size. The SPECT protocols listed here specify
acquire the same clinically useful number of counts. a 6.4 ⫾ 0.4 –mm pixel size for a 64 ⫻ 64 image
The same energy windows used in performing pa- matrix. This size offers satisfactory image resolution
tient studies should be used for routine daily quality for interpretation and quantitation of both Tl-201
control (QC). and Tc-99m tomograms.
5. Collimator. Parallel-hole collimators are most com- 9. Acquisition type. The most widespread mode of
monly employed for cardiac SPECT acquisitions. tomographic acquisition is the “step-and-shoot” meth-
They fall into two categories: low-energy all-pur- od. In this approach, the camera acquires a projection
pose (LEAP), used mostly for Tl-201 studies, and and then stops recording data when moving to the next
low-energy high-resolution (LEHR), used for Tc- angle; this results in a small amount of dead time since
99m studies. Compared with LEAP collimators, the camera is not acquiring data while it is moving. An
LEHR collimators have longer bores, thinner septa, alternative is “continuous” mode, where the camera
and smaller holes, which provide better resolution at moves continuously and acquires each projection over
the expense of reduced sensitivity. Therefore, to use an angular increment. This eliminates dead time and
LEHR collimators, imaging agents providing high thus increases image counts at the expense of a small
count rates are required (ie, Tc-99m agents). Gener- amount of blurring due to the motion of the camera
ally, LEAP collimators are used for 3-mCi Tl-201 while acquiring. It seems likely that the increase in
studies, including gated SPECT acquisitions. For count statistics more than offsets the small amount of
dual-isotope studies, LEHR collimators are sug- blurring due to camera motion.
gested. 10. Number of projections. The optimal number of
6. Orbit. Due to the anterior position of the heart in the projections for emission studies depends on match-
left hemithorax, much higher count rates are ob- ing the number of projections to the resolution of the
tained per given period of imaging time for a 180° system. A thallium SPECT acquisition with a LEAP
orbit (45° right anterior oblique to 45° left posterior collimator is a relatively low-resolution study, for
Journal of Nuclear Cardiology Hansen et al e99
Volume 13, Number 6;e97-120 Myocardial perfusion and function SPECT

which 32 projections over 180° are sufficient. A performance” (subsection Clinical QA for Each
higher-resolution study using Tc-99m agents should Patient Procedure) but less than a maximum total
be collected with a high-resolution collimator; this imaging time of 30 minutes.
requires at least 60 to 64 projections over 180° to 16. The following acquisition parameters are recommended
prevent loss of resolution. Larger numbers of pro- for the imaging protocols described in the “Stress proto-
jections are not warranted at this time but could cols and tracers” section of these guidelines.
become beneficial if technical innovations result in
improved overall system resolution.
PROCESSING PROTOCOLS
11. Matrix. The standard matrix size for emission
SPECT is 64 ⫻ 64 pixels. 1. Filtering. Image filtering is a very complex topic that
12. Time/projection. The emission acquisition times encompasses techniques for image enhancement, recon-
listed have been found to produce images of accept- struction, and feature extraction. The main area of
able and comparable quality for rest and stress concern for an interpreter of SPECT studies is image
studies. enhancement by reducing noise prior to image recon-
13. Total time. For single-detector systems, the total struction. All forms of imaging are plagued by statistical
time for an emission acquisition ultimately is based variation in the acquired image counts commonly re-
on how long a patient can tolerate the procedure ferred to as noise. The quality of an image can be
without moving, balanced by the need to acquire described by the signal-to-noise ratio, which describes
sufficient counts. The maximum practical time is on the relative strength of the signal component (what is
the order of 30 minutes. For biplane systems, this actually being imaged) compared to noise. The signal-
time can be halved, and many laboratories obtain to-noise ratio is much higher at lower spatial frequencies
gated perfusion SPECT studies in only 12 to 15 (broad features that are constant over many pixels) and
minutes using biplane cameras. Consideration may decreases at higher spatial frequencies (features that
be made for increasing imaging time in patients change over few pixels such as edges). In general, the
likely to have lower count statistics (eg, obese greater the count statistics, the better the signal-to-noise
patients) if it is felt that they can tolerate this. ratio. A low-pass filter is generally used to reduce noise
14. Gated SPECT. Incorporation of wall motion infor- because it allows low spatial frequencies to pass through
mation from gated SPECT has been shown to and attenuates the high frequencies where image noise
increase specificity and confidence by helping to predominates. Low-pass filters such as the Hanning and
differentiate breast and diaphragmatic attenuation Butterworth can be characterized by a cutoff frequency
artifacts from true perfusion defects. Likewise, as- where they begin to affect the image. The cutoff fre-
sessment of regional wall motion and/or thickening quency can be adjusted, depending on the signal-to-
can be a valuable tool for detecting viability within noise ratio, to preserve as much of the signal and
a stress-induced perfusion defect. Left ventricular suppress as much noise as possible. If the cutoff is too
(LV) ejection fractions (EFs) and volumes, as well high, there is significant noise in the image; if the cutoff
as regional wall motion and thickening, now are is too low, significant information in the signal is
computed routinely from gated SPECT data using suppressed. Nuclear Cardiology images, because of their
commercially available software. The majority of relatively low count statistics, tend to have greater
stress myocardial perfusion radionuclide studies cur- amounts of image noise and filtered back projection, and
rently are acquired as gated SPECT data. However, because of its dependence on ramp filtering, tends to
there is mounting evidence that the information amplify this noise. The optimal filter for a given image
content of the post-stress acquisition may be differ- depends on the signal-to-noise ratio for that image;
ent from that of the resting data most likely due to underfiltering an image leaves significant noise in the
post-ischemic stunning of myocardium. Providing image, and overfiltering unnecessarily blurs image de-
that there is adequate count density, particularly with tail; both overfiltering and underfiltering can reduce
regard to the lower-dose acquisitions, both stress and image accuracy. Software reconstruction packages are
rest SPECT perfusion studies may be acquired as set with default filter selection and cutoff values that are
gated data sets. Optimizing protocols for which both optimized for the average patient. Adjustment of the
stress and rest gated data are acquired remains an filter cutoff can be done in patients with poor count
area of investigation. statistics (eg, obese patients) to optimally filter their
15. Multidetector systems. It is recommended for mul- images. However, this is discouraged unless the physi-
tidetector systems that total imaging time be ad- cian is thoroughly familiar with filter adjustment and the
justed to obtain greater than the minimum counts potential effects. Changing the filter cutoff may have
listed in “Instrumentation quality assurance and unexpected effects on the output of commercially avail-
e100 Hansen et al Journal of Nuclear Cardiology
Myocardial perfusion and function SPECT November/December 2006

Table 1. Patient protocol: Same-day rest-stress Tc-99m acquisition

For
information,
Rest study Stress study see paragraph

Dose 8–12 mCi 24–36 mCi Standard 1

Position Supine Supine Standard 2
Prone Prone Optional 2
Upright/semiupright Upright/semiupright Optional
Delay time (intervals)
Injection ¡ imaging 30–60 min 15 min to 1 h Standard 3
Rest ¡ stress 30 min to 4 h Standard 3
Acquisition protocol
Energy window 15%–20% symmetric Same Standard 4
Collimator LEHR Same Preferred 5
Orbit 180° (45° RAO to 45° LPO) Same Preferred 6
Orbit type Circular Same Standard 7
Non-circular Same Standard 7
Pixel size 6.4 ⫾ 0.4 mm Same Standard 8
Acquisition type Step and shoot Same Standard 9
Continuous Same Optional 9
No. of projections 60–64 Same Standard 10
Matrix 64 ⫻ 64 Same Standard 11
Time/projection 25 s 20 s Standard 12
ECG gated Optional Standard Preferred 14
Frames/cycle 8 8 Standard 14
16 16 Optional 14
R-to-R window 100% 100% Preferred 14

RAO, Right anterior oblique; LPO, left posterior oblique.

able analysis programs, especially those that employ of these algorithms is that the process of generating
edge detection such as defect quantitation and LV projections from the guess can be made as sophisti-
volumes and EF. cated as desired and can incorporate such variables as
2. Reconstruction. The traditional method of image re- attenuation, scatter, and depth-dependent blur. The
construction has been filtered backprojection, a tech- main disadvantage is the computational intensity of
nique based on a mathematical proof, which assumes the algorithm; it takes many times longer to complete
no attenuation, no scatter, and an infinite number of than filtered backprojection. However, due to contin-
projections. It is relatively straightforward and com- ual increases in computer processor speed, these
paratively fast. The vast majority of clinical experi- algorithms can now be completed in an acceptable
ence is based upon it, and it has withstood the test of time for routine clinical use. Nonetheless, iterative
time despite its inability to model attenuation and techniques have not yet been proven to be unequivo-
scatter. There is a different class of reconstruction cally superior to filtered backprojection.
algorithms that are based on iterative techniques. 3. Reorientation. A critical phase of myocardial pro-
These algorithms start with a rudimentary guess of the cessing is reorientation of tomographic data into the
distribution, generate projections from the guess, and natural approximate symmetry axes of an individual
compare these projections to the acquired projections. patient’s heart. This is performed either by an ob-
The guess is refined based on the differences between server or automatically and results in sectioning the
the generated and actual projections, and the process data into vertical long-axis, horizontal long-axis, and
is repeated (hence the term “iterative”) usually for a short-axis planes. Long-axis orientation lines should
fixed number of iterations but can also be repeated be parallel to long-axis walls of the myocardium and
until the error between the generated and actual should be consistent between rest and stress studies.
projections is acceptably small. The main advantage Inappropriate plane selections can result in mis-
Journal of Nuclear Cardiology Hansen et al e101
Volume 13, Number 6;e97-120 Myocardial perfusion and function SPECT

Table 2. Patient protocol: Same-day stress-rest Tc-99m acquisition

For
information,
Exercise study Rest study see paragraph

Dose 8–12 mCi 24–36 mCi Standard 1

Position Supine Supine Standard 2
Prone Prone Optional 2
Upright/semiupright Upright/semiupright Optional
Delay time (intervals)
Injection ¡ imaging 15 min to 1 h 30–60 min Standard 3
Stress ¡ rest 3–4 h Standard 3
Acquisition protocol
Energy window 15%–20% symmetric Same Standard 4
Collimator LEHR Same Preferred 5
Orbit 180° (45° RAO to 45° LPO) Same Preferred 6
Orbit type Circular Same Standard 7
Non-circular Same Standard 7
Pixel size 6.4 ⫾ 0.4 mm Same Standard 8
Acquisition type Step and shoot Same Standard 9
Continuous Same Optional 9
No. of projections 60–64 Same Standard 10
Matrix 64 ⫻ 64 Same Standard 11
Time/projection 25 s 20 s Standard 12
ECG gated Optional Standard Preferred 14
Frames/cycle 8 8 Standard 14
16 16 Optional 14
R-to-R window 100% 100% Preferred 14

RAO, Right anterior oblique; LPO, left posterior oblique.

aligned myocardial walls between rest and stress data detection of gating errors due to some types of
sets, potentially resulting in incorrect interpretation. It transient arrhythmias, a full display of all count-
is crucial that all axis choices be available as QC versus-projection curve data is required. Cine reviews
screens and that these are reviewed by the technolo- occasionally show abnormalities in the abdomen or
gist and the physician who reads each study to verify thorax such as renal cysts or abnormal uptake that
that axes were selected properly. may be suspicious for neoplasm. The clinical utility
4. Display– cine review. The most important post-ac- of these findings has not been clearly established.
quisition QC procedure is to view the raw tomo- 5. Display–study review. It is strongly recommended
graphic data in cine mode. This presentation offers a that physicians use the active computer screen for
sensitive method for detecting patient and/or heart reviewing images and use film and paper hard copies
motion, “upward creep,” breast shadow due to atten- only for record-keeping purposes. Images produced
uation, diaphragmatic attenuation, and superimposed by formatters onto transparency film or photographic
abdominal visceral activity, all of which can create paper can have variable contrast, also termed gamma,
artifacts in the reconstructed images. Review of the and result in inconsistent image interpretation. Com-
raw tomograms in cine mode is performed twice: puter screen outputs are relatively more stable and
once by the technologist immediately after the acqui- always have readily available monochromatic con-
sition and again by the physician during image trast bars or color code bars to the side of the images,
interpretation. For gated studies, usually it is only the enabling consistent viewing conditions. In addition,
sum of all gated tomograms that is reviewed in this computer screens offer rapid sequential and/or cine-
manner; this will alert the observer to most types of matic displays of image data. For all of these reasons,
gating errors due to arrhythmias manifested by an screen interpretation is strongly recommended over
intermittent flashing of the images. However, for the relying on interpretations from hard copies.
e102 Hansen et al Journal of Nuclear Cardiology
Myocardial perfusion and function SPECT November/December 2006

Table 3. Patient protocol: Two-day exercise Tc-99m acquisition

For
information,
Exercise study Rest study see paragraph

Dose 30 mCi 30 mCi Standard 1

Position Supine Supine Standard 2
Prone Prone Optional 2
Delay time (intervals)
Injection ¡ imaging 15 min to 1 h 30–60 min Standard 3
Acquisition protocol
Energy window 15%–20% symmetric Same Standard 4
Collimator LEHR Same Preferred 5
Orbit 180° (45° RAO to 45° LPO) Same Preferred 6
Orbit type Circular Same Standard 7
Non-circular Same Standard 7
Pixel size 6.4 ⫾ 0.4 mm Same Standard 8
Acquisition type Step and shoot Same Standard 9
Continuous Same Optional 9
No. of projections 60–64 Same Standard 10
Matrix 64 ⫻ 64 Same Standard 11
Time/projection 20 s 20 s Standard 12
ECG gated Optional Standard Preferred 14
Frames/cycle 8 8 Standard 14
16 16 Optional 14
R-to-R window 100% 100% Preferred 14

RAO, Right anterior oblique; LPO, left posterior oblique.

PERFUSION QUANTITATION radius is then rotated slightly (analogous to going to the

next spoke of the wagon wheel), and the activity is again
The display medium and translation table employed
measured. This is repeated until the entire circumference
can have a significant impact on image interpretation
has been traversed. Determination of activity along each
going as far as to make a normal perfusion scan appear
radius can be as simple as identifying the maximal pixel
abnormal or vice versa. Quantitative analysis is a direct
or as complicated as fitting the profile to a Gaussian
way of measuring relative uptake of a perfusion tracer
curve.
that is independent of the display medium and translation
2. Normalization and scaling. Myocardial perfusion
table and can thus greatly reduce variations in interpre-
imaging can only measure relative uptake. In order to
tation due to subjective analysis and inconsistent image
compare different studies or compare with normal
display. Quantitative analysis also allows for the com-
databases, the images must be normalized to a certain
parison of a study with a gender-specific normal data-
value. Each slice can be normalized to its own
base. For this reason, quantitative analysis is recom-
maximum, but it is much more common to normalize
mended as part of image interpretation. Quantitative
to the maximal pixel in the ventricle. All values are
analysis consists of the following steps:
multiplied by 100/value of the maximum pixel, which
1. Image quantitation. This measures the relative activity, sets the maximum pixel to 100 and all others to a
most often in a short-axis slice, by generating a circum- fraction of that. Finally, the ventricle is “scaled” to a
ferential profile also known as a radial plot. This is done constant number of curves; smaller ventricles with
by first identifying the center of the lumen and the inner fewer short-axis slices are interpolated up, and larger
and outer boundaries of the ventricle. The activity in ventricles with more short-axis slices are decimated
each part of the slice is determined by measuring the down to a constant number of curves. Thus relative
activity in each pixel lying on a radius between the inner activity can be compared for any location in two
and outer boundaries, analogous to moving along a ventricles regardless of their absolute tracer uptake or
spoke of a wagon wheel from the hub to the rim. The chamber size.
Journal of Nuclear Cardiology Hansen et al e103
Volume 13, Number 6;e97-120 Myocardial perfusion and function SPECT

Table 4. Patient protocol: Separate dual-isotope acquisition

For
information,
Rest study Exercise study see paragraph

Dose 2.5–3.5 mCi Tl-201 30 mCi Tc-99m Standard 1

Position Supine Supine Standard 2
Prone Prone Optional
Upright/semiupright Upright/semiupright Optional
Delay time (intervals)
Injection ¡ imaging 10–15 min 15 min to 1 h Standard 3
Rest ¡ stress No delay Standard 3
Acquisition protocol
Energy window 25%–30% symmetric, 70 keV 15%–20% symmetric, 140 keV Standard 4
20% symmetric, 167 keV
Collimator LEHR Same Preferred 5
Orbit 180° (45° RAO to 45° LPO) Same Preferred 6
Orbit type Circular Same Standard 7
Non-circular Same Standard 7
Pixel size 6.4 ⫾ 0.4 mm Same Standard 8
Acquisition type Step and shoot Same Standard 9
Continuous Same Optional 9
No. of projections 32–64 60–64 Standard 10
Matrix 64 ⫻ 64 Same Standard 11
Time/projection 40 s (32 fr) 25 s (64 fr) 20 s Standard 12
ECG gated Optional Standard Preferred 14
Frames/cycle 8 8 Standard 14
16 16 Optional
R-to-R window 100% 100% Preferred 14

RAO, Right anterior oblique; LPO, left posterior oblique.

3. Polar plot generation (optional). Each radial plot corresponds to a constant fraction of the volume of
generated in step 1 generates a series of numbers the left ventricle in a process known as “volume
that can be displayed as a graph. The entire LV weighting.”
volume would be represented by a series of graphs, 4. Database construction and analysis. It is fre-
which is difficult to assimilate. Alternatively, these quently difficult to differentiate true perfusion de-
plots can be used to make a “bulls-eye” or polar fects from soft tissue attenuation; women tend to
plot. Instead of generating a graph, the radial have anterior defects from breast attenuation, and
profiles are used to make a series of rings of men tend to have inferior wall defects from dia-
activity, where the intensity of the ring corresponds phragmatic attenuation. A range for “normal” soft
to the relative activity of the corresponding polar tissue attenuation can be defined by creating a
plot and the diameter of the rings grows moving normal database; gender-specific normal databases
toward the base of the heart. These can then be are usually employed due to the different attenua-
fused into a single image where each successive tion patterns for men and women. These can be
ring is formed around the one preceding it, analo- generated by performing radial plot analysis on
gous to the rings of a tree. This creates a 2-dimen- “normals”—that is, patients proven not to have
sional image that reflects the relative activity in the coronary disease or, more often, patients with an
3-dimensional left ventricle with the apex at the acceptably low probability of having coronary dis-
center of the polar plot and the base as the outer ease (usually ⬍5% or ⬍1%). The mean and stan-
ring. The width of the rings can be adjusted to dard deviation of activity for each point of the
reflect the relative size of the corresponding short- ventricle are calculated for the male and female
axis images so that a given area of the polar plot normals; the result is the normal database for each
e104 Hansen et al Journal of Nuclear Cardiology
Myocardial perfusion and function SPECT November/December 2006

Table 5. Patient protocol: Stress/redistribution Tl-201 acquisition

For
information,
Stress study Redistribution rest see paragraph

Dose 2.5–3.5 mCi Tl-201 Not applicable Standard 1

Position Supine Supine Standard 2
Prone Prone Optional
Upright/semiupright Upright/semiupright Optional
Delay time
Injection ¡ imaging 10–15 min* Not applicable Standard 3
Stress ¡ rest 3–4 h Standard 3
Acquisition
Energy window 30% symmetric, 70 keV Same Standard 4
20% symmetric, 167 keV
Collimator LEAP Same Preferred 5
Orbit 180° (45° RAO to 45° LPO) Same Preferred 6
Orbit type Circular Same Standard 7
Non-circular Same Standard 7
Pixel size 6.4 ⫾ 0.4 mm Same Standard 8
Acquisition type Step and shoot Same Standard 9
Continuous Same Optional 9
No. of projections 32 Same Standard 10
Matrix 64 ⫻ 64 Same Standard 11
Time/projection 40 s 40 s Standard 12

RAO, Right anterior oblique; LPO, left posterior oblique.

*An anterior planar image may be acquired during this interval to evaluate Tl-201 lung uptake.

gender. The normalized and scaled activity for each strates which areas show improved perfusion at rest; it
point of the ventricle of a patient is compared to the is called a reversibility map.
mean of the corresponding gender-based normal
database; if the activity is more than a predeter-
mined number of standard deviations below the GATED SPECT
mean of the normals (2.5 standard deviations is
most frequently used), it is considered abnormal. It Acquisition. The introduction of technetium-based
should be realized that this definition of normal is perfusion tracers has resulted in images with sufficient
statistical, not absolute. There will still be overlap count density to allow for cardiac gating adding parameters
between normal and abnormal uptake. of wall motion and EF to myocardial perfusion imaging.
5. Parameters. Different parametric images can be Gating requires a stable and consistent heart rhythm as well
generated to show the results of quantitative analysis, as sufficient temporal resolution to correctly characterize
such as “blackout,” “severity,” and “reversibility” the cardiac cycle. A stable heart rate and rhythm can be
maps. A blackout map generates a polar plot for the achieved by rejecting heartbeats that fall out of range at the
patient marking those points that fall a predetermined expense of a slight increase in image time; this is routine in
number of standard deviations below the gender- planar imaging but becomes impractical in SPECT due to
based normal limits and thus demonstrates the size or memory and time constraints. These constraints also limit
extent of the perfusion defect. Another type of para- the temporal resolution. Most laboratories gate the heart for
metric image can be generated which shows the 8 frames per cycle, although some laboratories have re-
number of standard deviations that activity falls ported good results with 16 frames per cycle. For either 8-
below the mean of normals and thus demonstrates or 16-frame gating, the recommendations are to avoid beat
severity. A third can be generated which quantitates rejection. Processing is done using commercially available
the activity on the stress and rest images and demon- software.
Journal of Nuclear Cardiology Hansen et al e105
Volume 13, Number 6;e97-120 Myocardial perfusion and function SPECT

Table 6. Patient protocol: Stress/reinjection/redistribution

For
(Redistribution) information,
Stress study Reinjection Rest study see paragraph

Dose 3 mCi 1.5 mCi Not applicable Standard 1

Position Supine Supine Standard 2
Prone Prone Optional 2
Upright/semiupright Upright/semiupright Optional
Delay time (intervals)
Injection ¡ imaging 10–15 min Not applicable Standard 3
Stress ¡ redistribution 3–4 h Standard 3
Reinjection ¡ imaging 20–30 min Standard 3
(MI)
24-h imaging Optional 3
Acquisition protocol
Energy window 30% symmetric, 70 keV Same Standard 4
20% symmetric, 167 keV
Collimator LEAP Same Preferred 5
Orbit 180° (45° RAO to 45° Same Preferred 6
LPO)
Orbit type Circular Same Standard 7
Non-circular Same Standard 7
Pixel size 6.4 ⫾ 0.4 mm Same Standard 8
Acquisition type Step and shoot Same Standard 9
Continuous Same Optional 9
No. of projections 32 Same Standard 10
Matrix 64 ⫻ 64 Same Standard 11
Time/projection 40 s 40 s Standard 12

RAO, Right anterior oblique; LPO, left posterior oblique.

INTERPRETATION AND REPORTING Display

General Comments 1. Recommended medium for display. It is strongly

suggested that the reading physician use the com-
The interpretation of myocardial perfusion puter monitor screen rather than hard copy to inter-
SPECT images should be performed in a systematic pret the study since a monitor screen is capable of
fashion to include (1) repeat evaluation of the raw displaying more variations in gray scale or color
tomographic images to determine the presence of (making it easier to discern smaller variations in
technical sources of abnormalities and extracardiac uptake) and is more consistent than film. Moreover,
activity; (2) acquisition of additional views when it is not possible to properly view moving images
appropriate; (3) interpretation of images with respect (eg, raw tomographic data or gated images) on hard
to the extent and severity of perfusion abnormalities copy. A linear gray-scale translation table is gener-
that are believed to be present, as well as chamber ally preferred to color tables since it demonstrates
sizes and, when appropriate, pulmonary uptake; (4) consistent grades of uptake compared to pseudocon-
incorporation of the results of quantitative analysis; touring seen with color scales, but this is also
(5) consideration of functional data obtained from the dependent on how familiar the reader is with a given
study; and (6) consideration of clinical factors that translation table. The reader should be aware that the
may have influenced the presence of any findings. All appearance of an image can change significantly
of those factors contribute to the production of a final when changing from one translation table to another.
clinical report. A linear scale is preferred to nonlinear (eg, sigmoi-
e106 Hansen et al Journal of Nuclear Cardiology
Myocardial perfusion and function SPECT November/December 2006

dal) scales since it most faithfully characterizes Evaluation of Images for Technical Sources
uptake over the range of activity. A logarithmic of Error
scale may be used for evaluating soft tissue uptake
or the right ventricle but should never be used for 4. Patient motion. The interpreting physician should
interpreting LV uptake. also review the raw tomographic images for possible
2. Conventional slice display of SPECT images. sources of artifact. Images should again be inspected
Three sets of images should be displayed: (1) a for the presence of patient motion. A cine display of
view generated by slicing perpendicular to the the planar projection data is highly recommended
long axis of the left ventricle (short axis), (2) a because motion in both the craniocaudal and hori-
view of long-axis tomograms generated by slicing zontal axes is readily detectable. Additionally, a
in the vertical plane (vertical long axis), and (3) a static “sinogram” and sometimes “linogram” may be
view of long-axis tomograms generated by slicing used to detect motion. Software routines are avail-
in the horizontal plane (horizontal long axis). The able for quantitation and correction of motion. With
short-axis tomograms should be displayed with experience, the reader will quickly become accus-
the apical slices to the far left with progression of tomed to the normal appearance of raw tomograms
slices toward the base in a left-to-right fashion. and be able to identify motion artifact. In patients
The vertical long axis should be displayed with who have had a technetium-based perfusion agent,
septal slices on the left and progression through consideration should be made for repeating the
the midventricular slices to the lateral slices in a image acquisition where feasible. Alternatives, such
left-to-right fashion. Similarly, the horizontal as planar imaging or prone positioning, may be
long-axis tomographic display should proceed left considered as well. The effect of patient motion on
to right from the inferior to the superior (anterior) the final reconstructions is complex. Generally, up-
surface. It is also recommended that, for purposes and-down motion (especially when the heart returns
of interpretation and comparison of sequential to the same baseline) is less deleterious to image
images (eg, stress and rest, rest and redistribu- quality than sideways motion. Also, up-and-down
tion), these images be displayed aligned and motion is much easier to correct either manually or
adjacent to each other serially. There are two with semiautomated software. Rotation currently
widely used approaches to image normalization. cannot be corrected either manually or with avail-
Each series (vertical, horizontal, short axis) may able software. Corrected raw tomographic images
be normalized to the brightest pixel in the entire should be evaluated in the same way for adequacy of
series. That is considered to provide the most the correction.
intuitively easy way to evaluate the extent and 5. Attenuation and attenuation correction. The cine
severity of perfusion defects. The drawbacks of display of the planar projection images is also
this approach are its sensitivity to focal hot spots, recommended for the identification of sources of
the frequently poor visualization of normal struc- attenuation, the most common being diaphragmatic
tures at the base and apex of the left ventricle, and in men and the breast in women. Breast attenuation
the lack of an ideal display of each individual can sometimes be improved by repeating the acqui-
slice. sition with the breast repositioned. Breast position-
The other approach is “frame normalization” in ing should be similar on the stress and rest images to
which each image is normalized to the brightest optimize accurate interpretation. This can be as-
pixel within the frame. That method provides opti- sessed by reviewing the raw tomograms. Diaphrag-
mal image quality of each slice. The drawback of matic attenuation may be minimized by imaging the
this approach is that the brightness of each slice is patient prone. Hardware and software for attenuation
unrelated to the peak myocardial activity in the and scatter correction are commercially available
entire series such that gradations in activity between and may obviate or at least mitigate these common
slices of a series may be lost. That drawback is attenuation artifacts. The evaluation of attenuation-
mitigated by the display of three orthogonal planes. corrected (AC) images is performed with the same
3. Three-dimensional display. Most commercial soft- approach as that used for non-AC images. As with
ware programs allow creation of 3-dimensional dis- the interpretation of non-AC studies, it is essential
plays. These may help less experienced readers that the interpreting physician be familiar with the
identify coronary distributions associated with per- segment-by-segment normal variation of uptake of
fusion defects. These should be used only as an radioactivity at stress and rest associated with the
adjunct to, not a replacement for, the conventional specific attenuation correction system that is being
image formatting described above. used. AC images are displayed in the same manner
Journal of Nuclear Cardiology Hansen et al e107
Volume 13, Number 6;e97-120 Myocardial perfusion and function SPECT

as uncorrected images. Because the currently avail- subendocardial ischemia. It is typically described
able correction algorithms are imperfect, it is rec- qualitatively but may be quantified.
ommended that the AC data be read simultaneously 9. Lung uptake. The presence of increased lung up-
with the uncorrected data. take after thallium perfusion imaging has been
6. Reconstruction artifacts. Superimposed bowel described as an indicator of poor prognosis and
loops or liver activity may create artifactually in- should therefore be evaluated in all patients when
tense uptake in the overlapped myocardium that using this perfusion agent. No clear consensus has
could mask a real perfusion defect or be misinter- emerged as to the significance of lung uptake with
preted as reduced uptake in adjacent or contralateral technetium-based perfusion agents.
segments. Non-superimposed but adjacent extracar- 10. Right ventricular uptake. Right ventricular (RV)
diac activity may also affect the reconstructed myo- uptake may be qualitatively assessed on the raw
cardial images. Intense activity in bowel loops or projection data and on the reconstructed data. There
adjacent liver may cause a negative reconstruction are no established quantitative criteria for RV up-
artifact, resulting in an apparent reduction in activity take, but in general, the intensity of the right ventri-
in the adjacent myocardial segments. There is cur- cle is approximately 50% of peak LV intensity. RV
rently no reliable correction for such artifacts, al- uptake increases in the presence of RV hypertrophy,
though they may be less prominent with iterative as most typically because of pulmonary hypertension.
opposed to filtered backprojection techniques. They The intensity of the right ventricle may also appear
can often be eliminated by repeating the acquisition relatively increased when LV uptake is globally
after the activity level in the adjacent structure has reduced. Regional abnormalities of RV uptake may
decreased. be a sign of right coronary artery stenosis. The size
of the right ventricle should be noted.
7. Myocardial statistics. Many factors are involved in
11. Noncardiac findings. Both thallium- and techne-
the final count density of perfusion images includ-
tium-based agents can be concentrated in tumors,
ing body habitus, exercise level, radiopharmaceuti-
and uptake outside the myocardium may reflect
cal dose, acquisition time, energy window, and
unexpected pathology. However, the accuracy and,
collimation. The interpreting physician should make
in particular, the specificity of myocardial perfusion
note of the count density in the planar projection
imaging for diagnosing non-cardiac conditions have
images because the quality of the reconstructed data
not been established.
is a direct reflection of the raw data. Perfusion
12. Perfusion defect location. Myocardial perfusion
defects can be artifactually created simply because defects should be identified by use of visual analysis
of poor statistics. As a general rule, peak pixel of the reconstructed slices. The perfusion defects
activity in the LV myocardium in an anterior planar should be characterized by their location as they
projection should exceed 100 counts in a Tl-201 relate to specific myocardial walls—that is, apical,
study and 200 counts in a Tc-99m study. anterior, inferior, and lateral. The term posterior
should probably be avoided because it has been
Initial Image Analysis and Interpretation variably assigned to either the lateral wall (circum-
flex distribution) or to the basal inferior wall (right
The initial interpretation of the perfusion study coronary distribution), thereby leading to confusion.
should be performed without any clinical information Standardization of segment nomenclature is highly
other than the patient’s gender, height and weight, and recommended. (See the segmentation models de-
peak exercise heart rate. Such an approach minimizes the scribed below.)
bias in study interpretation. All relevant clinical data 13. Perfusion defect severity and extent: Qualitative.
should be reviewed after a preliminary impression is Defect severity is typically expressed qualitatively
formed. as mild, moderate, or severe. Mild defects may be
identified by a decrease in counts compared to
8. Ventricular dilation. Before segmental analysis of adjacent activity without the appearance of wall
myocardial perfusion, the reader should note whether thinning, moderate defects demonstrate wall thin-
there is LV enlargement at rest or during stress. ning, and severe defects are those that approach
Dilation on both the stress and resting studies background activity. Defect extent may be qualita-
usually indicates LV dysfunction, although it may be tively described as small, medium, or large. In
seen in volume overload states with normal ventricular semiquantitative terms, small represents 5% to 10%,
function. Transient ischemic dilation is a marker for medium represents 15% to 20%, and large repre-
multivessel disease and is most likely due to diffuse sents 20% or greater of the left ventricle. Alterna-
e108 Hansen et al Journal of Nuclear Cardiology
Myocardial perfusion and function SPECT November/December 2006

Figure 1. SPECT myocardial perfusion imaging: 17-segment model.

tively, defect extent may also be estimated as a ventricle and one vertical long-axis slice to better
fraction such as the “basal one half” or “apical one represent the LV apex. In both the 17- and 20-
third” of a particular wall or as extending from base segment models, the basal and mid short-axis slices
to apex. Defects whose severity and extent do not are divided into 6 segments. In the 17-segment
change between image sets (eg, stress and rest) are model, the apical short-axis slice is divided into 4
typically categorized as “fixed” or nonreversible. segments, whereas in the 20-segment model, the
When changes do occur, a qualitative description of apical short-axis slice is divided into 6 segments. In
the degree of reversibility is required. the 17-segment model, a single apical segment is
14. Perfusion defect severity and extent: Semiquan- taken from the vertical long-axis slice, whereas in
titative. In addition to the qualitative evaluation of the 20-segment model, the apex is represented by
perfusion defects, it is preferred that the physician two segments. Each segment has a specific name. In
may also apply a semiquantitative method on the order to facilitate consistency of nomenclature with
basis of a validated segmental scoring system. This other imaging modalities, the 17-segment model has
approach standardizes the visual interpretation of become the preferred nomenclature.
scans, reduces the likelihood of overlooking signif- Seventeen-segment nomenclature (Figure 1). Seg-
icant defects, and provides an important semiquan- ments 1, 7, and 13 represent the basal (1), mid (7),
titative index that is applicable to diagnostic and and apical (13) anterior segments. Segments 4, 10,
prognostic assessments. It is generally considered and 15 represent the basal (4), mid (10), and apical
preferable to use a system with at least 16 segments. (15) inferior segments. The septum contains 5 seg-
The quality assurance (QA) committee of the Amer- ments, the basal anteroseptal (2), the basal infero-
ican Society of Nuclear Cardiology has considered septal (3), the mid anteroseptal (8), the mid infero-
several models for segmentation of the perfusion septal (9), and the apical septal (14). Similarly, the
images and has previously recommended either a lateral wall is divided into the basal anterolateral (6),
17- or 20-segment model for semiquantitative visual the basal inferolateral (5), the mid anterolateral (12),
analysis. The models use three short-axis slices the mid inferolateral (11), and the apical lateral (16).
(apical, mid, and basal) to represent most of the The long-axis apical segment is called the apex.
Journal of Nuclear Cardiology Hansen et al e109
Volume 13, Number 6;e97-120 Myocardial perfusion and function SPECT

Figure 2. SPECT myocardial perfusion imaging: 20-segment model.

Twenty-segment nomenclature (Figure 2). Seg-

ments 1, 7, and 13 represent the basal (1), mid (7),
and apical (13) anterior segments. Segments 4, 10,
and 16 represent the basal (4), mid (10), and apical
(16) inferior segments. The septum contains 6 seg-
ments, the basal (2), mid (8), and apical (14)
anteroseptal and the basal (3), mid (9), and apical
(15) inferoseptal. Similarly, the lateral wall contains
6 segments, the basal (6), mid (12), and apical (18)
anterolateral and the basal (5), mid (11), and apical
(17) inferolateral segments. The apex from the
vertical long-axis slice is divided into anteroapical
(19) and inferoapical (20) segments.
The myocardial segments may be roughly as-
signed to coronary arterial territories as indicated in
Figure 3 as long as the reader realizes that there can
Figure 3. SPECT myocardial perfusion imaging: coronary
be considerable variation among patients especially artery territories. LAD, Left anterior descending artery; RCA,
in the inferior and inferolateral segments of the left right coronary artery; LCX, left circumflex artery.
ventricle due to the variable extent of the circumflex
and right coronary artery territories.

both of those variables contain independent prog-

Semiquantitative Scoring System:
nostic power. Furthermore, semiquantitative scor-
The Five-Point Model
ing can be used to more reproducibly and objec-
The use of a scoring system provides a reproduc- tively designate segments as normal or abnormal.
ible semiquantitative assessment of defect sever- Points are assigned to each segment in direct
ity and extent. A consistent approach to defect proportion to the perceived count density of the
severity and extent is clinically important because segment.
e110 Hansen et al Journal of Nuclear Cardiology
Myocardial perfusion and function SPECT November/December 2006

encountered in patients without disease. Because of

Category Score artifacts during imaging and also the nature of
coronary blood flow, there will always be an overlap
Normal perfusion 0 between normals and patients with mild perfusion
Mild reduction in counts—not definitely 1 defects; this overlap can be reduced but not com-
abnormal
pletely eliminated by careful attention to image
Moderate reduction in counts—definitely 2
acquisition and reconstruction. Therefore quantita-
abnormal
tive analysis should only be used as an adjunct to
Severe reduction in counts 3
and not a substitute for visual analysis.
Absent uptake 4
Defect extent may be quantitatively expressed as a
In addition to individual scores, it has been recom- percentage of the entire left ventricle or as a percent-
mended that summed scores be calculated. The age of individual vascular territories, the latter being
summed stress score equals the sum of the stress less reliable because of the normal variations in
scores of all the segments, and the summed rest coronary anatomy. Defect severity may be quantita-
score equals the sum of the resting scores or redis- tively expressed as the number of standard devia-
tribution scores of all the segments. The summed tions by which the segment varies from the normal
difference score equals the difference between the range for that particular segment or segments. De-
summed stress and the summed resting (redistribu- fect reversibility may also be expressed as a percent-
tion) scores and is a measure of reversibility. In age of the entire left ventricle or of a vascular
particular, the summed stress score has been shown territory.
to have significant prognostic power. Before scor- 16. Reversibility. Reversibility of perfusion defects
ing, it is necessary for the interpreting physician to may be categorized qualitatively as partial or com-
be familiar with the normal regional variation in plete, the latter being present when the activity in the
count distribution of myocardial perfusion SPECT. defect returns to a level comparable to surrounding
normal myocardium. The semiquantitative scoring
15. Perfusion defect severity and extent: Quantitative. system may be used to define reversibility as a
Quantitative analysis is useful to supplement visual ⱖ2-grade improvement or improvement to a score
interpretation. Most techniques of quantitative anal- of 1. Reversibility on a quantitative polar plot or on
ysis are based on radial plots of short-axis slices. 3-dimensional displays will depend on the particular
Different techniques analyze the apex separately. software routine in use and the normal reference
These plots are then normalized to allow creation of databases used in the program. Areas of reversibility
or comparison to normal databases. Defects can be are typically described by pixels that improve to less
defined as where activity falls a given amount below than 2.5 SDs from the normal reference redistribu-
the mean of a normal database to evaluate size and tion or resting database. How many pixels have to
severity of defects. Quantitation of the stress images improve for reversibility to be deemed present is
is compared to the rest images to assess the degree of arbitrary.
ischemia versus infarction. It is customary to gener- So-called reverse redistribution may be seen in
ate separate normal databases based on gender as stress delayed thallium imaging sequences and has
well as the perfusion agent used. This quantitative been described in rest delayed technetium sestamibi
analysis is usually displayed as a “bulls-eye” or sequences. Reverse redistribution refers to segments
polar plot. The quantitative programs are effective in with decreased or normal intensity on the initial set
providing an objective interpretation that is inher- of images that show even less relative intensity on
ently more reproducible than visual analysis, elimi- the delayed images. The interpretation of the finding
nates the variability of the appearance of a defect remains controversial, but in certain clinical situa-
when viewed in different media (with different tions it seems to represent segments with a mixture
gammas) and different translation tables, and is of viable and nonviable myocardium that are fre-
particularly helpful in describing changes between quently supplied by patent infarct-related arteries.
two studies in the same patient. Quantitative analysis
also serves as a guide for the less experienced
GATED MYOCARDIAL PERFUSION SPECT
observer who may be uncertain about normal vari-
ations in uptake. Quantitative programs are by no Because of the comparatively low additional cost
means sophisticated enough to unequivocally differ- and substantial benefit of the information obtained,
entiate perfusion defects from artifact but help in gated studies of ventricular function should be a
understanding the range of uptake that can be routine part of myocardial perfusion SPECT. A
Journal of Nuclear Cardiology Hansen et al e111
Volume 13, Number 6;e97-120 Myocardial perfusion and function SPECT

systematic approach to display and interpretation of minimize the effect that gating errors have on the
the ventricular function derived from gated SPECT summed image.
is important. 19. Gated SPECT: Regional wall motion and thick-
ening. Regional wall motion should be analyzed by
17. Gated SPECT display. Multiple ventricular slices use of standard nomenclature: normal, hypokinesis,
should be evaluated. At a minimum, a quad-screen akinesis, and dyskinesis. Hypokinesis may be further
display of apical and mid-basal short-axis, a mid- qualified as mild, moderate, or severe. A semiquan-
ventricular horizontal long-axis, and a mid-ventric- titative scoring system is recommended where 0 is
ular vertical long-axis slice should be viewed. Other normal, 1 is mild hypokinesis, 2 is moderate hypo-
slices may be viewed for completeness or to resolve kinesis, 3 is severe hypokinesis, 4 is akinesis, and 5
a discrepancy between the clinical impression and is dyskinesis.
what is seen on the four standard cine views. Ideally, This is comparable to the 5-point scoring system
the software should allow the user to scroll through used in both contrast and radionuclide ventriculog-
any of the slices in any axis in cine mode. Each view raphy. As in any assessment of regional ventricular
should be normalized to the series of end-diastolic to function, one must be cognizant of expected normal
end-systolic slices to maintain the count density and abnormal variations such as the reduced wall
changes that reflect wall thickening. When available, excursion at the base compared with the apex, the
software routines that automatically define epicar- greater excursion of the basal lateral wall compared
dial and endocardial borders and that subsequently with the basal septum, and paradoxical septal motion
calculate ventricular volumes and EF should be resulting from left bundle branch block.
applied. Normal myocardial wall thickness is below the
Regional wall motion should be interpreted with a resolution of currently available SPECT systems.
gray-scale display. When computer edge analysis
However, regional wall thickening can be estimated
software is available, the physician may choose to
by use of the count increase from end diastole to end
analyze wall motion by use of the assigned endocar-
systole. Visually, it is not as easy to assign degrees
dial and epicardial contours, but reference should
of abnormality of thickening as it is to wall motion.
also be made to the wall motion without computer-
However, the evaluation of thickening with gated
derived edges. Regional wall thickening may be
perfusion SPECT lends itself to quantitation because
analyzed in gray scale or in a suitable color scheme,
it is characterized by count changes. Quantitative
although color displays may make it easier to see
normal databases are now available for assessment
changes in count intensity.
18. Gated SPECT QC. All the QA procedures for of regional wall thickening.
routine SPECT are applicable to gated SPECT with Wall motion and wall thickening are generally
the addition of the evaluation of the adequacy of the concordant. The principle exception to this occurs in
electrocardiographic (ECG) gate. The most common patients who have undergone cardiac surgery where
manifestation of poor gating is the appearance of a septal wall motion is frequently abnormal (paradox-
flashing pattern on the rotating planar projection ical) but there is normal wall thickening. Rather than
images that results from count fluctuations from separately scoring wall motion and wall thickening,
frame to frame. Ideally, a heart rate histogram it is commonly accepted to incorporate the two
should also be viewed to verify beat length unifor- findings into a single score while noting the presence
mity. Inspecting a time-volume curve, although of discordance in wall motion and wall thickening
useful with gated blood pool studies, is less useful when it occurs. In addition to noting LV wall
with gated SPECT since volumes are determined motion, wall thickening, and EF, the size of the left
geometrically and not by counts. As yet, there is no and right ventricles should be observed, and the
clear consensus on the beat length window for gated function of the right ventricle should be noted.
SPECT acquisitions. As in blood pool imaging, the 20. LVEF and LV volume. LVEF and LV and RV
narrower the window, the more physiologic the data, chamber sizes should routinely be evaluated quali-
but this runs the risk of compromising the quality of tatively. EF may be categorized as normal, mildly,
the SPECT perfusion images. Another important moderately, or severely reduced. Volume may be
aspect of QC is a visual or quantitative determina- categorized as normal, mildly, moderately, or se-
tion that the number of counts acquired in each verely increased. Alternatively, LVEF and end-
frame of the gated study was adequate for assess- diastolic and end-systolic volumes may be calcu-
ment of function. Software that collects all counts lated with geometric models applied to the
into a separate bin for the summed image can reconstructed data set. Several software routines that
e112 Hansen et al Journal of Nuclear Cardiology
Myocardial perfusion and function SPECT November/December 2006

correlate well with contrast and other radionuclide 24. Myocardial viability: Quantitative assessment. An
measurements are commercially available. alternate and perhaps more rigorous approach to the
21. Integration of perfusion and function results. The assessment of the viability of any segment is the
results of the perfusion and gated SPECT data sets quantitative determination of ischemic-to-normal ra-
should be integrated into a final interpretation. The tios. Regions of interest may be placed over the
wall motion is particularly helpful in distinguishing segment in question and over the most normal
real nonreversible perfusion defects from attenuation segment of the myocardium in that particular series
or motion artifacts. Fixed perfusion defects that do of images. The analysis should be applied to the
not show a corresponding abnormality of motion or resting images for technetium images or to the
thickening are more likely to be due to artifacts, resting, redistribution, or reinjection images for thal-
especially if the clinical data do not support prior lium. When this approach is used, one must take into
infarction. account the normal count variations such as the
relatively reduced counts in the normal inferior wall.
Segments with ratios less than 0.30 are considered
Myocardial Viability nonviable. Areas with ratios greater than 0.50 are
considered viable, whereas areas with ratios of 0.30
22. Viability: Qualitative assessment. The assessment to 0.50 are equivocal for viability. As indicated
of myocardial viability is a complex issue made even above for the semiquantitative approach, such re-
more difficult by the lack of consensus in the gions require additional data such as wall motion of
literature of the precise meaning of the term viabil- the region, exercise perfusion in the region, the
ity—whether it refers merely to the absence of scar change in perfusion or wall motion after nitroglyc-
or requires improvement in wall motion after revas- erin, the response of regional function to low-dose
cularization. It is, however, clear that the quantita- dobutamine, or myocardial metabolic imaging with
tive uptake of radionuclides such as Tl-201 and the fluorine 18 fluorodeoxyglucose.
available technetium agents does correlate with It is also important to recognize that viability of a
myocardial viability as defined by post-revascular- given segment does not necessarily equate to im-
ization improvement in both tracer uptake and re- provement in clinical outcome after revasculariza-
gional function. Myocardial segments with normal tion unless enough segments that are viable are
or mildly reduced tracer uptake at rest or on delayed available. The critical number of segments necessary
imaging almost invariably prove to be viable. The to justify revascularization strategies has not been
majority of myocardial segments in which there is adequately studied.
unequivocal improvement of uptake on either redis-
tribution or resting images also prove to be viable.
The more difficult challenge for the single photon Modification of Interpretation by Relevant
assessment of viability is in segments with severely Clinical Information
reduced tracer uptake. 25. Due to imperfections in the technology as well as the
23. Myocardial viability: Semiquantitative assessment. gradual impairment of coronary blood flow as ste-
The semiquantitative scoring system described noses become hemodynamically significant, there
above may be used to assess viability as follows. will always be overlap between normal and mildly
Segments with rest, reinjection, or redistribution abnormal perfusion scans. In these patients it is often
scores of 0 (normal perfusion) and 1 (slight reduc- particularly helpful to incorporate other clinical
tion in counts) are considered viable. Segments with information (eg, symptoms, risk factors, ST-segment
rest, redistribution, or reinjection scores of 2 (mod- changes, exercise tolerance) as well as prognostic
erately decreased perfusion) are consistent with a information in order to help the referring physician
combination of viable and nonviable myocardium, make the most appropriate management decisions
and segments with scores of 3 and 4 are generally for the patient. Homogeneous perfusion images of
nonviable. Segments with final scores of 4 are consid- patients who have other markers of severe ischemia,
ered nonviable. such as marked ST-segment changes, should be
Segments with a final score of 2 or 3 may be carefully evaluated for adjunctive markers of isch-
further characterized by use of a nitroglycerin pro- emia such as transient ischemic dilation or increased
tocol or a different modality such as positron emis- lung uptake (with thallium) in order to identify those
sion tomography imaging or low-dose dobutamine patients with balanced ischemia. The majority of
echocardiography. Low-dose dobutamine SPECT artifacts encountered will produce mild defects;
may also be useful. therefore moderate or severe defects, in the absence
Journal of Nuclear Cardiology Hansen et al e113
Volume 13, Number 6;e97-120 Myocardial perfusion and function SPECT

of dramatic artifacts, should be considered as reflect- rate, blood pressure and maximal blood pressure and
ing pathology. Finally, it needs to be understood that workload achieved (estimated metabolic equiva-
not all pathology detected by perfusion imaging lents), and the magnitude (in millimeters) of any
reflects epicardial coronary artery disease (CAD). ST-segment deviation.
If only one report is used for both the exercise or
pharmacologic study and the perfusion results, then
Reporting of SPECT Myocardial Perfusion
more detail about the stress test should be included
Scan Results
such as time of onset, duration and exact ECG leads
26. Subject information. The age, gender, height, with ST-segment changes, the type of chest pain
weight, and body surface area should be included in (typical, atypical, non-anginal) and its severity
the report because they may directly affect the image (mild, moderate, severe), and the presence of ar-
results and interpretation. For medical records pur- rhythmia.
poses, any identification number should be included. 31. Overall study quality. Including a statement about
Pertinent medications that may influence the results the quality of the study is helpful as it alerts the
may be included. physicians using the report to any shortcomings that
27. Type of study. The imaging protocol should be might reduce the accuracy of the data and their
specified, including the radiopharmaceutical and interpretation.
dose, imaging technique (gated vs ungated, supine, 32. Diagnosis and prognosis of CAD. The probability
or prone), and imaging sequence (stress/4-hour re- of CAD may be determined with available algo-
distribution, 1-day or 2-day rest/stress or stress/rest, rithms that use the pre-scan likelihood of CAD as
and so on). The date(s) of study acquisitions should determined by age, gender, character of chest pain,
also be included. the number of coronary risk factors, and the results
28. Indication for study. Placing the indication for the of the stress electrocardiogram. The perfusion data
study in the report helps focus the interpreting are then added to the model to produce a probability
physician on the clinical question raised by the of CAD. A qualitative probability may be reported
ordering physician and may be subsequently impor- on the basis of the definite presence or absence of a
tant for reimbursement issues. perfusion defect, the severity and extent of any
29. Resting ECG findings. Inclusion of ECG findings perfusion defects, and the presence of other markers
that may have a direct bearing on the study interpre- of CAD such as transient LV dilation, post-stress
tation should be included such as the presence of left stunning, or increased lung uptake. When CAD is
bundle branch block or LV hypertrophy. known to be present, the likelihood of stress-induced
30. Summary of stress data. The type of stress (bicycle ischemia is reported instead of the likelihood of
or treadmill) and the protocol should be identified significant CAD.
(Bruce, modified Bruce, Naughton, manual). For Although not widely available, some large labo-
pharmacologic stress, the agent, route of administra- ratories have enough internal follow-up data to be
tion, and dose should be indicated. The reason for able to statistically predict outcomes (death and
stopping the test should be noted. All symptoms nonfatal myocardial infarction) on the basis of per-
experienced by the patient during stress (eg, chest fusion image scores. If such data are available,
pain, dyspnea, claudication, dizziness) should be incorporation of the likelihood of an adverse event in
mentioned. the report is desirable. Otherwise, a qualitative
If a separate stress test report is generated, then statement about risk may be appropriate because the
the stress variables that could impact on the perfu- likelihood of an adverse event increases with the
sion study quality or findings should be included in presence of any of the following: perfusion defects
the perfusion scan report. At a minimum, the report in multiple vascular territories, transient LV dilation,
should include the total exercise duration, maximal increased lung uptake, and decreased LV systolic
heart rate and percent of predicted maximal heart function.
e114 Hansen et al Journal of Nuclear Cardiology
Myocardial perfusion and function SPECT November/December 2006

Myocardial perfusion SPECT: Guidelines for interpretation

For information,
see paragraph

A. Display
1. Medium
a. Computer screen Preferred 1
b. Film hard copy Discouraged 1
2. Format
a. Conventional slice display Preferred 2
i. Frame normalization Optional 2
ii. Series normalization Preferred 2
b. Three-dimensional display Optional 3
B. Technical sources of error
1. Motion Standard 4
2. Attenuation Standard 5
a. Attenuation correction Optional 5
3. Reconstruction artifacts Standard 6
4. Myocardial statistics Standard 7
C. Initial image interpretation
1. Ventricular dilation
a. Qualitative Standard 8
b. Quantitative Optional 8
2. Lung uptake
a. Qualitative Standard 9
b. Quantitative Preferred 9
3. Non-cardiac Standard 11
4. Perfusion defect assessment
a. Location Standard 12
b. Extent/severity
i. Qualitative Standard 13
ii. Semiquantitative Optional 14
iii. Quantitative Optional 15
5. Reversibility Standard 16
D. Gated SPECT
1. Display Standard 17
2. QC Standard 18
3. Regional wall motion Standard 19
4. Regional wall thickening Standard 19
5. LV EF
a. Qualitative Standard 20
b. Quantitative Preferred 20
6. LV volume
a. Qualitative Standard 20
b. Quantitative Optional 20
E. Integration of perfusion and function results Standard 21
F. Myocardial viability
1. Qualitative Standard 22
2. Semiquantitative Optional 23
3. Quantitative Preferred 24
G. Modification of interpretation Preferred 25
Journal of Nuclear Cardiology Hansen et al e115
Volume 13, Number 6;e97-120 Myocardial perfusion and function SPECT

Myocardial perfusion SPECT: Guideline for reporting

For information,
see paragraph

A. Demographic data
1. Name Standard 26
2. Gender Standard 26
3. Age Standard 26
4. Date(s) of acquisition(s) Standard 26
5. Medical record identification Standard 26
6. Height/weight (body surface area) Standard 26
7. Relevant medications Optional 26
8. Indication for study Standard 28
B. Acquisition parameters
1. Type(s) of studies Standard 27
2. Radionuclide(s) and doses Standard 27
C. Results: Exercise/intervention data
1. Resting ECG findings Standard 29
2. Exercise/intervention parameters
a. Heart rate, blood pressure, % maximal predicted Standard 30
heart rate, metabolic equivalents
b. Symptoms Standard 30
c. Reason for terminating Standard 30
d. ECG changes with exercise Standard 30
D. Results: Perfusion scan data
1. Potential sources of error
a. Motion Standard 4
b. Attenuation Standard 5
c. Adjacent/overlapping uptake Standard 6
2. Chamber sizes Standard 8
3. Lung uptake (thallium)
a. Qualitative Standard 9
b. Quantitative Preferred 9
4. Initial defect location Standard 12
5. Initial defect severity and extent
a. Qualitative Standard 13
b. Semiquantitative Preferred 14
c. Quantitative Optional 15
6. Reversibility
a. Qualitative Standard 16
b. Semiquantitative Preferred 16
c. Quantitative Optional 16
7. RV uptake Standard 10
8. Abnormal noncardiac uptake Standard 11
E. Results: Gated SPECT
1. Regional wall motion Standard 19
2. Regional wall thickening Standard 19
3. EF
a. Qualitative Standard 20
b. Quantitative Optional 20
4. LV volume Optional 20
F. Overall study quality Optional 31
e116 Hansen et al Journal of Nuclear Cardiology
Myocardial perfusion and function SPECT November/December 2006

Myocardial perfusion SPECT: Guideline for reporting (Continued)

For information,
see paragraph

G. Conclusion
1. Normal/abnormal
a. Three categories Standard 16
b. Five categories Preferred 16
2. Probability of CAD Optional 32
3. Estimated risk of adverse events Optional 32

33. Clinical interpretation of AC SPECT studies. The and women with similar heart sizes are compared,
interpretation of AC SPECT myocardial perfusion the difference disappears. The unsuspecting ob-
images follows a similar approach to that used for server may mistake this expected normal activity
uncorrected myocardial perfusion images, but there reduction for a distal or mid and distal left anterior
are differences and these should be taken into descending coronary perfusion defect. In general,
account in order to obtain good results. The normal the success of AC SPECT appears related to the
distributions of perfusion tracer uptake are signifi- diligence of the clinical laboratory in following
cantly different with AC compared to uncorrected recommended procedures for image acquisition, re-
studies, and because of this, it is important that the construction, QA, display, and quantification. Al-
interpreting physician have available and learn da- though it may seem reasonable that AC SPECT
tabases of normal tracer distribution(s). Although, should simply provide better images, like those
from system to system, these normal distributions without correction, but with the bothersome artifacts
are generally relatively similar, there can be differ- that often result in false-positive tests removed, the
ences that are dependent on the geometry of the normal distributions are significantly different and
imaging system, acquisition protocol, and process- must be accounted for to achieve optimal clinical
ing algorithms. There can also be differences in benefit. AC SPECT can correct many if not all
normal distribution(s) related to patient gender and attenuation artifacts, but normal is different. The
ventricular volume. The interpreting physician must normal mean activity distribution is different, and
be aware of these differences if they exist for their the variances in the normal distribution are different.
imaging system(s) and take them into account on a If these differences are not understood by interpret-
patient-by-patient basis when assessing clinical ing physicians, AC SPECT will be unreliable. Like-
studies. wise, quantification and display programs without
AC SPECT studies generally have more uniform appropriate normal AC databases should not be used
regional activity in the anterior, septal, inferior, and for quantification as spurious results will occur.
lateral walls, but mild reductions in apical and distal QA requirements are more demanding with AC
anterior activity are typical of the normal AC image than non-AC images and should be carefully as-
distribution. This apical and distal anterior activity sessed for each patient study. Artifacts due to
reduction is similar to that seen with positron emis- movement, either respiration or patient movement,
sion tomography myocardial perfusion imaging. and extracardiac radiotracer uptake can be amplified
This finding becomes more prominent when resolu- by the iterative algorithms that are employed in AC
tion recovery and scatter correction are included in reconstructions and processing. The quality and
the AC processing workflow and is often more registration of the transmission mu maps with the
prominent in patients with larger hearts (ie, larger- emission image data are additional key factors that
volume hearts high in the normal range and larger must be ensured, and if they cannot be ensured, the
will generally have greater accentuation of the associated AC images should be either disregarded
normal apical and distal anterior activity reduction or read with extreme caution. QA tools to aid these
compared to smaller-volume normal hearts). In low- assessments of registration and mu map quality are
likelihood normal patients, this reduction in distal still not uniformly available, but this should improve
activity is generally more prominent in men than in the near future.
women, as men generally have larger hearts. If men For the clinical interpretation of AC SPECT
Journal of Nuclear Cardiology Hansen et al e117
Volume 13, Number 6;e97-120 Myocardial perfusion and function SPECT

myocardial perfusion images, it is recommended tion with reference to findings in chronic obstructive pulmonary
that AC and non-AC images be displayed side by disease. Am J Cardiol 1978;41:897-905.
4. Berman DS, Kiat H, Friedman JD, Wang FP, van Train K, Matzer
side with displays of the normal activity distribu- L, et al. Separate acquisition rest thallium-201/stress technetium-
tion(s) and their variance distribution available as 99m sestamibi dual-isotope myocardial perfusion single-photon
required for comparison. This requires the avail- emission computed tomography: a clinical validation study. J Am
ability of normal databases specific for the imag- Coll Cardiol 1993;22:1455-64.
ing device, imaging protocol, and processing ap- 5. Berman DS, Kiat H, Van Train K, Garcia E, Friedman J, Maddahi
J. Technetium 99m sestamibi in the assessment of chronic
proach employed clinically. Extracardiac activity
coronary artery disease. Semin Nucl Med 1991;21:190-212.
especially when combined with respiratory and/or 6. Borges-Neto S, Coleman RE, Potts JM, Jones RH. Combined
patient movement can introduce artifacts and/or exercise radionuclide angiocardiography and single photon emis-
normalization errors that may require renormal- sion computed tomography perfusion studies for assessment of
ization or abandonment of the AC images alto- coronary artery disease. Semin Nucl Med 1991;21:223-9.
7. Brent BN, Mahler D, Matthay RA, Berger HJ, Zaret BL.
gether. Artifactual reductions in activity most
Noninvasive diagnosis of pulmonary arterial hypertension in
often affecting the apparent anterior and/or lateral chronic obstructive pulmonary disease: right ventricular ejection
wall perfusion tracer uptake can occur when there fraction at rest. Am J Cardiol 1984;53:1349-53.
is misregistration of SPECT and mu map images 8. Brooks RA, Chiro GD. Principles of computer assisted tomogra-
such that the myocardial activity from the SPECT phy (CAT) in radiographic and radioisotopic imaging. Phys Med
images is matched with the relatively low attenu- Biol 1976;21:689-732.
9. Burow RD, Pond M, Schafer AW, Becker L. “Circumferential
ation coefficients for adjacent lung tissue in the profiles:” a new method for computer analysis of thallium-201
mu maps. Some attenuation correction– capable myocardial perfusion images. J Nucl Med 1979;20:771-7.
SPECT systems acquire the SPECT and the trans- 10. Caldwell JH, Williams DL, Harp GD, Stratton JR, Ritchie JL.
mission image data sequentially rather than simul- Quantitation of size of relative myocardial perfusion defect by
taneously. If there is a change in position of single-photon emission computed tomography. Circulation 1984;
70:1048-56.
arm(s) or breasts between emission and transmis-
11. Callahan RJ, Froelich JW, McKusick KA, Leppo J, Strauss HW.
sion imaging, even though there may be perfect A modified method for the in vivo labeling of red blood cells with
registration of the heart in the emission and Tc-99m: concise communication. J Nucl Med 1982;23:315-8.
transmission images, there can be artifactual de- 12. Calnon DA, Kastner RJ, Smith WH, Segalla D, Beller GA,
fects introduced into the SPECT perfusion images Watson DD. Validation of a new counts-based gated single
by the misregistration of tissues outside the tho- photon emission computed tomography method for quantifying
left ventricular systolic function: comparison with equilibrium
rax. radionuclide angiography. J Nucl Cardiol 1997;4:464-71.
The currently available SPECT/computed to- 13. Choi JY, Lee KH, Kim SJ, Kim SE, Kim BT, Lee SH, et al.
mography systems require sequential emission Gating provides improved accuracy for differentiating artifacts
and transmission imaging with a change in bed from true lesions in equivocal fixed defects on technetium 99m
position between acquisitions. Although the qual- tetrofosmin perfusion SPECT. J Nucl Cardiol 1998;5:395-401.
14. Chouraqui P, Rodrigues EA, Berman DS, Maddahi J. Signifi-
ity of the mu maps with these systems will
cance of dipyridamole-induced transient dilation of the left
consistently far exceed the quality of sealed- ventricle during thallium-201 scintigraphy in suspected coronary
source transmission system mu maps, the greatly artery disease. Am J Cardiol 1990;66:689-94.
improved resolution of the mu maps they provide 15. Chua T, Kiat H, Germano G, Maurer G, van Train K, Friedman
make it even more important that registration of J, et al. Gated technetium-99m sestamibi for simultaneous assess-
ment of stress myocardial perfusion, postexercise regional ven-
emission and transmission reconstructions be ex-
tricular function and myocardial viability. Correlation with echo-
act to a tolerance of less than 1 pixel. cardiography and rest thallium-201 scintigraphy. J Am Coll
Cardiol 1994;23:1107-14.
16. Cooper JA, Neumann PH, McCandless BK. Effect of patient
Suggested Reading motion on tomographic myocardial perfusion imaging. J Nucl
Med 1992;33:1566-71.
1. Bateman TM, Berman DS, Heller GV, Brown KA, Cerqueira 17. Cullom SJ, Case JA, Bateman TM. Electrocardiographically
MD, Verani MS, et al. American Society of Nuclear Cardiology gated myocardial perfusion SPECT: technical principles and
position statement on electrocardiographic gating of myocardial quality control considerations. J Nucl Cardiol 1998;5:418-25.
perfusion SPECT scintigrams. J Nucl Cardiol 1999;6:470-1. 18. DePace NL, Iskandrian AS, Hakki AH, Kane SA, Segal BL.
2. Berger BC, Watson DD, Taylor GJ, Craddock GB, Martin RP, Value of left ventricular ejection fraction during exercise in
Teates CD, et al. Quantitative thallium-201 exercise scintigraphy predicting the extent of coronary artery disease. J Am Coll
for detection of coronary artery disease. J Nucl Med 1981;22: Cardiol 1983;1:1002-10.
585-93. 19. DePasquale EE, Nody AC, DePuey EG, Garcia EV, Pilcher G,
3. Berger HJ, Matthay RA, Loke J, Marshall RC, Gottschalk A, Bredlau C, et al. Quantitative rotational thallium-201 tomography
Zaret BL. Assessment of cardiac performance with quantitative for identifying and localizing coronary artery disease. Circulation
radionuclide angiocardiography: right ventricular ejection frac- 1988;77:316-27.
e118 Hansen et al Journal of Nuclear Cardiology
Myocardial perfusion and function SPECT November/December 2006

20. DePuey EG, Nichols K, Dobrinsky C. Left ventricular ejection 37. Germano G, Erel J, Kiat H, Kavanagh PB, Berman DS. Quanti-
fraction assessed from gated technetium-99m-sestamibi SPECT. tative LVEF and qualitative regional function from gated thalli-
J Nucl Med 1993;34:1871-6. um-201 perfusion SPECT. J Nucl Med 1997;38:749-54.
21. DePuey EG, Parmett S, Ghesani M, Rozanski A, Nichols K, 38. Germano G, Erel J, Lewin H, Kavanagh PB, Berman DS.
Salensky H. Comparison of Tc-99m sestamibi and Tl-201 gated Automatic quantitation of regional myocardial wall motion and
perfusion SPECT. J Nucl Cardiol 1999;6:278-85. thickening from gated technetium-99m sestamibi myocardial
22. Dilsizian V, Perrone-Filardi P, Arrighi JA, Bacharach SL, perfusion single-photon emission computed tomography. J Am
Quyyumi AA, Freedman NM, et al. Concordance and discor- Coll Cardiol 1997;30:1360-7.
dance between stress-redistribution-reinjection and rest-redistri- 39. Germano G, Kavanagh PB, Su HT, Mazzanti M, Kiat H,
bution thallium imaging for assessing viable myocardium. Com- Hachamovitch R, et al. Automatic reorientation of three-dimen-
parison with metabolic activity by positron emission tomography. sional, transaxial myocardial perfusion SPECT images. J Nucl
Circulation 1993;88:941-52. Med 1995;36:1107-14.
23. Eisner R, Churchwell A, Noever T, Nowak D, Cloninger K, Dunn 40. Germano G, Kiat H, Kavanagh PB, Moriel M, Mazzanti M, Su
D, et al. Quantitative analysis of the tomographic thallium-201 HT, et al. Automatic quantification of ejection fraction from gated
myocardial bullseye display: critical role of correcting for patient myocardial perfusion SPECT. J Nucl Med 1995;36:2138-47.
motion. J Nucl Med 1988;29:91-7. 41. Gibbons RJ, Lee KL, Cobb F, Jones RH. Ejection fraction
24. Eisner RL, Nowak DJ, Pettigrew R, Fajman W. Fundamentals of response to exercise in patients with chest pain and normal
180 degree acquisition and reconstruction in SPECT imaging. coronary arteriograms. Circulation 1981;64:952-7.
J Nucl Med 1986;27:1717-28. 42. Gibbons RJ, Verani MS, Behrenbeck T, Pellikka PA,
25. Esquerre JP, Coca FJ, Martinez SJ, Guiraud RF. Prone decubitus: O’Connor MK, Mahmarian JJ, et al. Feasibility of tomographic
a solution to inferior wall attenuation in thallium-201 myocardial 99mTc-hexakis-2-methoxy-2-methylpropyl-isonitrile imaging
tomography. J Nucl Med 1989;30:398-401. for the assessment of myocardial area at risk and the effect of
26. Faber TL, Akers MS, Peshock RM, Corbett JR. Three- treatment in acute myocardial infarction. Circulation 1989;80:
dimensional motion and perfusion quantification in gated 1277-86.
single-photon emission computed tomograms. J Nucl Med 43. Gill JB, Ruddy TD, Newell JB, Finkelstein DM, Strauss HW,
1991;32:2311-7. Boucher CA. Prognostic importance of thallium uptake by the
27. Faber TL, Cooke CD, Folks RD, Vansant JP, Nichols KJ, DePuey lungs during exercise in coronary artery disease. N Engl J Med
EG, et al. Left ventricular function and perfusion from gated 1987;317:1486-9.
SPECT perfusion images: an integrated method. J Nucl Med 44. Goerres GW, Burger C, Kamel E, Seifert B, Kaim AH, Buck A,
1999;40:650-9. et al. Respiration-induced attenuation artifact at PET/CT: techni-
28. Ficaro EP, Fessler JA, Shreve PD, Kritzman JN, Rose PA, cal considerations. Radiology 2003;226:906-10.
Corbett JR. Simultaneous transmission/emission myocardial per- 45. Greer KL, Jaszczak RJ, Coleman RE. An overview of a camera-
fusion tomography. Diagnostic accuracy of attenuation-corrected based SPECT system. Med Phys 1982;9:455-63.
99mTc-sestamibi single-photon emission computed tomography. 46. Grossman GB, Garcia EV, Bateman TM, Heller GV, Johnson
Circulation 1996;93:463-73. LL, Folks RD, et al. Quantitative Tc-99m sestamibi attenua-
29. Foster C, Dymond DS, Anholm JD, Pollock ML, Schmidt DH. tion-corrected SPECT: development and multicenter trial val-
Effect of exercise protocol on the left ventricular response to idation of myocardial perfusion stress gender-independent
exercise. Am J Cardiol 1983;51:859-64. normal database in an obese population. J Nucl Cardiol
30. Fricke H, Fricke E, Weise R, Kammeier A, Lindner O, Burchert 2004;11:263-72.
W. A method to remove artifacts in attenuation-corrected myo- 47. Hachamovitch R, Berman DS, Shaw LJ, Kiat H, Cohen I, Cabico
cardial perfusion SPECT Introduced by misalignment between JA, et al. Incremental prognostic value of myocardial perfusion
emission scan and CT-derived attenuation maps. J Nucl Med single photon emission computed tomography for the prediction
2004;45:1619-25. of cardiac death: differential stratification for risk of cardiac death
31. Friedman J, Berman DS, Van Train K, Garcia EV, Bietendorf and myocardial infarction. Circulation 1998;97:535-43.
J, Prigent F, et al. Patient motion in thallium-201 myocardial 48. Hansen C. Digital image processing for clinicians, part I: basics
SPECT imaging. An easily identified frequent source of of image formation. J Nucl Cardiol 2002;9:343-9.
artifactual defect. Clin Nucl Med 1988;13:321-4. 49. Hansen C. Digital image processing for clinicians, part II:
32. Friedman J, Van Train K, Maddahi J, Rozanski A, Prigent F, filtering. J Nucl Cardiol 2002;9:429-37.
Bietendorf J, et al. “Upward creep” of the heart: a frequent source 50. Hansen C. Digital image processing for clinicians, part III:
of false-positive reversible defects during thallium-201 stress- SPECT reconstruction. J Nucl Cardiol 2002;9:542-9.
redistribution SPECT. J Nucl Med 1989;30:1718-22. 51. Hansen CL, Cen P, Sanchez B, Robinson R. Comparison of
33. Friedman JD, Berman DS, Kiat H, Bietendorf J, Hyun M, Van pulmonary uptake with transient cavity dilation after dipyridam-
Train KF, et al. Rest and treadmill exercise first-pass radionuclide ole Tl-201 perfusion imaging. J Nucl Cardiol 2002;9:47-51.
ventriculography: validation of left ventricular ejection fraction 52. Hansen CL, Sangrigoli R, Nkadi E, Kramer M. Comparison of
measurements. J Nucl Cardiol 1994;1:382-8. pulmonary uptake with transient cavity dilation after exercise
34. Gal R, Grenier RP, Carpenter J, Schmidt DH, Port SC. High thallium-201 perfusion imaging. J Am Coll Cardiol 1999;33:
count rate first-pass radionuclide angiography using a digital 1323-7.
gamma camera. J Nucl Med 1986;27:198-206. 53. Iskander S, Iskandrian AE. Risk assessment using single-photon
35. Garcia EV, Cooke CD, Van Train KF, Folks R, Peifer J, DePuey emission computed tomographic technetium-99m sestamibi im-
EG, et al. Technical aspects of myocardial SPECT imaging with aging. J Am Coll Cardiol 1998;32:57-62.
technetium-99m sestamibi. Am J Cardiol 1990;66:23E-31E. 54. Iskandrian AE, Germano G, VanDecker W, Ogilby JD, Wolf N,
36. Garcia EV, Van Train K, Maddahi J, Prigent F, Friedman J, Mintz R, et al. Validation of left ventricular volume measure-
Areeda J, et al. Quantification of rotational thallium-201 myocar- ments by gated SPECT 99mTc-labeled sestamibi imaging. J Nucl
dial tomography. J Nucl Med 1985;26:17-26. Cardiol 1998;5:574-8.
Journal of Nuclear Cardiology Hansen et al e119
Volume 13, Number 6;e97-120 Myocardial perfusion and function SPECT

55. Iskandrian AS, Chae SC, Heo J, Stanberry CD, Wasserleben V, myocardial distribution and washout of thallium-201.
Cave V. Independent and incremental prognostic value of exer- Circulation 1981;64:924-35.
cise single-photon emission computed tomographic (SPECT) 71. Mahmarian JJ, Boyce TM, Goldberg RK, Cocanougher MK,
thallium imaging in coronary artery disease. J Am Coll Cardiol Roberts R, Verani MS. Quantitative exercise thallium-201 single
1993;22:665-670. photon emission computed tomography for the enhanced diagno-
56. Johnson LL, Rodney RA, Vaccarino RA, Egbe P, Wasserman L, sis of ischemic heart disease. J Am Coll Cardiol 1990;15:318-29.
Esser PD, et al. Left ventricular perfusion and performance from 72. Maniawski PJ, Morgan HT, Wackers FJ. Orbit-related variation
a single radiopharmaceutical and one camera. J Nucl Med in spatial resolution as a source of artifactual defects in thallium-
1992;33:1411-6. 201 SPECT. J Nucl Med 1991;32:871-5.
57. Johnson LL, Verdesca SA, Aude WY, Xavier RC, Nott LT, 73. Massardo T, Gal RA, Grenier RP, Schmidt DH, Port SC. Left
Campanella MW, et al. Postischemic stunning can affect left ventricular volume calculation using a count-based ratio method
ventricular ejection fraction and regional wall motion on post- applied to multigated radionuclide angiography. J Nucl Med
stress gated sestamibi tomograms. J Am Coll Cardiol 1997;30: 1990;31:450-6.
1641-8. 74. Mochizuki T, Murase K, Tanaka H, Kondoh T, Hamamoto K,
58. Kaul S, Chesler DA, Boucher CA, Okada RD. Quantitative Tauxe WN. Assessment of left ventricular volume using ECG-
aspects of myocardial perfusion imaging. Semin Nucl Med gated SPECT with technetium-99m-MIBI and technetium-99m-
1987;17:131-44. tetrofosmin. J Nucl Med 1997;38:53-7.
59. Kaul S, Chesler DA, Okada RD, Boucher CA. Computer versus 75. Nallamothu N, Bagheri B, Acio ER, Heo J, Iskandrian AE.
visual analysis of exercise thallium-201 images: a critical ap- Prognostic value of stress myocardial perfusion single photon
praisal in 325 patients with chest pain. Am Heart J 1987;114: emission computed tomography imaging in patients with left
1129-37. ventricular bundle branch block. J Nucl Cardiol 1997;4:487-93.
60. Kelly MJ, Cowie AR, Antonino A, Barton H, Kalff V. An 76. Nichols K, DePuey EG, Krasnow N, Lefkowitz D, Rozanski A.
assessment of factors which influence the effectiveness of the Reliability of enhanced gated SPECT in assessing wall motion of
modified in vivo technetium-99m-erythrocyte labeling technique severely hypoperfused myocardium: echocardiographic valida-
in clinical use. J Nucl Med 1992;33:2222-5. tion. J Nucl Cardiol 1998;5:387-94.
61. Kiat H, Berman DS, Maddahi J, De Yang L, Van Train K, 77. Nichols K, DePuey EG, Rozanski A. Automation of gated
Rozanski A, et al. Late reversibility of tomographic myocardial tomographic left ventricular ejection fraction. J Nucl Cardiol
thallium-201 defects: an accurate marker of myocardial viability. 1996;3(Pt 1):475-82.
J Am Coll Cardiol 1988;12:1456-63. 78. Nichols K, Dorbala S, DePuey EG, Yao SS, Sharma A, Rozanski
62. Kiat H, Van Train KF, Friedman JD, Germano G, Silagan G, A. Influence of arrhythmias on gated SPECT myocardial perfu-
Wang FP, et al. Quantitative stress-redistribution thallium-201 sion and function quantification. J Nucl Med 1999;40:924-34.
SPECT using prone imaging: methodologic development and 79. Nichols K, Tamis J, DePuey EG, Mieres J, Malhotra S, Rozanski
validation. J Nucl Med 1992;33:1509-15. A. Relationship of gated SPECT ventricular function parameters
63. King MA, Glick SJ, Penney BC, Schwinger RB, Doherty PW. to angiographic measurements. J Nucl Cardiol 1998;5:295-303.
Interactive visual optimization of SPECT prereconstruction fil- 80. Nuyts J, Mortelmans L, Suetens P, Oosterlinck A, de Rou M.
tering. J Nucl Med 1987;28:1192-8. Model-based quantification of myocardial perfusion images from
64. King MA, Schwinger RB, Doherty PW, Penney BC. Two- SPECT. J Nucl Med 1989;30:1992-2001.
dimensional filtering of SPECT images using the Metz and 81. O’Keefe JH, Bateman TM, Ligon RW, Case J, Cullom J,
Wiener filters. J Nucl Med 1984;25:1234-40. Barnhart C, et al. Outcome of medical versus invasive treatment
65. Klein JL, Garcia EV, DePuey EG, Campbell J, Taylor AT, strategies for non-high-risk ischemic heart disease. J Nucl Cardiol
Pettigrew RI, et al. Reversibility bull’s-eye: a new polar bull’s- 1998;5:28-33.
eye map to quantify reversibility of stress-induced SPECT thal- 82. Philippe L, Mena I, Darcourt J, French WJ. Evaluation of valvular
lium-201 myocardial perfusion defects. J Nucl Med 1990;31: regurgitation by factor analysis of first-pass angiography. J Nucl
1240-6. Med 1988;29:159-67.
66. Koster K, Wackers FJ, Mattera JA, Fetterman RC. Quantitative 83. Port S, Cobb FR, Coleman RE, Jones RH. Effect of age on the
analysis of planar technetium-99m-sestamibi myocardial perfu- response of the left ventricular ejection fraction to exercise.
sion images using modified background subtraction. J Nucl Med N Engl J Med 1980;303:1133-7.
1990;31:1400-8. 84. Port S, McEwan P, Cobb FR, Jones RH. Influence of resting left
67. Lee KL, Pryor DB, Pieper KS, Harrell FE, Califf RM, Mark DB, ventricular function on the left ventricular response to exercise in
et al. Prognostic value of radionuclide angiography in medically patients with coronary artery disease. Circulation 1981;63:856-
treated patients with coronary artery disease. A comparison with 63.
clinical and catheterization variables. Circulation 1990;82:1705- 85. Port SC, Oshima M, Ray G, McNamee P, Schmidt DH. Assess-
17. ment of single vessel coronary artery disease: results of exercise
68. Levy R, Rozanski A, Berman DS, Garcia E, Van Train K, electrocardiography, thallium-201 myocardial perfusion imaging
Maddahi J, et al. Analysis of the degree of pulmonary thallium and radionuclide angiography. J Am Coll Cardiol 1985;6:75-83.
washout after exercise in patients with coronary artery disease. 86. Port SC, Wackers FJ. Clinical application of radionuclide angiog-
J Am Coll Cardiol 1983;2:719-28. raphy. J Nucl Cardiol 1995;2:551-8.
69. Liu YH, Lam PT, Sinusas AJ, Wackers FJ. Differential effect of 87. Proceedings of the 4th Invitational Wintergreen Conference.
180 degrees and 360 degrees acquisition orbits on the accuracy of Wintergreen, Virginia, USA. July 12-14, 1998. Abstracts. J Nucl
SPECT imaging: quantitative evaluation in phantoms [see com- Cardiol 1999;6(Pt 1):93-155.
ments]. J Nucl Med 2002;43:1115-24. 88. Pryor DB, Harrell FE, Lee KL, Rosati RA, Coleman RE, Cobb
70. Maddahi J, Garcia EV, Berman DS, Waxman A, Swan HJ, FR, et al. Prognostic indicators from radionuclide angiography in
Forrester J. Improved noninvasive assessment of coronary medically treated patients with coronary artery disease. Am J
artery disease by quantitative analysis of regional stress Cardiol 1984;53:18-22.
e120 Hansen et al Journal of Nuclear Cardiology
Myocardial perfusion and function SPECT November/December 2006

89. Raiker K, Sinusas AJ, Wackers FJ, Zaret BL. One-year prognosis 100. Van Train KF, Areeda J, Garcia EV, Cooke CD, Maddahi J, Kiat
of patients with normal planar or single-photon emission com- H, et al. Quantitative same-day rest-stress technetium-99m-
puted tomographic technetium 99m-labeled sestamibi exercise sestamibi SPECT: definition and validation of stress normal
imaging. J Nucl Cardiol 1994;1(Pt 1):449-56. limits and criteria for abnormality. J Nucl Med 1993;34:1494-
90. Reisman S, Maddahi J, Van Train K, Garcia E, Berman D. 502.
Quantitation of extent, depth, and severity of planar thallium 101. Van Train KF, Berman DS, Garcia EV, Berger HJ, Sands MJ,
defects in patients undergoing exercise thallium-201 scintigraphy. Friedman JD, et al. Quantitative analysis of stress thallium-201
J Nucl Med 1986;27:1273-81. myocardial scintigrams: a multicenter trial. J Nucl Med 1986;27:
91. Rogers WL, Clinthorne NH, Harkness BA, Koral KF, Keyes JW. 17-25.
Field-flood requirements for emission computed tomography 102. Van Train KF, Garcia EV, Maddahi J, Areeda J, Cooke CD, Kiat
with an Anger camera. J Nucl Med 1982;23:162-8. H, et al. Multicenter trial validation for quantitative analysis of
92. Sabia PJ, Powers ER, Ragosta M, Smith WH, Watson DD, Kaul S. same-day rest-stress technetium-99m-sestamibi myocardial to-
Role of quantitative planar thallium-201 imaging for determining mograms. J Nucl Med 1994;35:609-18.
viability in patients with acute myocardial infarction and a totally 103. Van Train KF, Maddahi J, Berman DS, Kiat H, Areeda J, Prigent
occluded infarct-related artery. J Nucl Med 1993;34:728-36. F, et al. Quantitative analysis of tomographic stress thallium-201
93. Santoro GM, Sciagra R, Buonamici P, Consoli N, Mazzoni V, myocardial scintigrams: a multicenter trial. J Nucl Med 1990;31:
Zerauschek F, et al. Head-to-head comparison of exercise stress 1168-79.
testing, pharmacologic stress echocardiography, and perfusion
104. Wackers FJ. Artifacts in planar and SPECT myocardial perfusion
tomography as first-line examination for chest pain in patients
imaging. Am J Card Imaging 1992;6:42-57, discussion 58.
without history of coronary artery disease. J Nucl Cardiol
105. Wackers FJ, Berger HJ, Johnstone DE, Goldman L, Reduto
1998;5:19-27.
LA, Langou RA, et al. Multiple gated cardiac blood pool
94. Segall GM, Davis MJ. Prone versus supine thallium myocardial
imaging for left ventricular ejection fraction: validation of the
SPECT: a method to decrease artifactual inferior wall defects.
technique and assessment of variability. Am J Cardiol 1979;
J Nucl Med 1989;30:548-55.
43:1159-66.
95. Shanoudy H, Raggi P, Beller GA, Soliman A, Ammermann EG,
106. Watson DD, Campbell NP, Read EK, Gibson RS, Teates CD,
Kastner RJ, et al. Comparison of technetium-99m tetrofosmin and
Beller GA. Spatial and temporal quantitation of plane thallium
thallium-201 single-photon emission computed tomographic im-
aging for detection of myocardial perfusion defects in patients myocardial images. J Nucl Med 1981;22:577-84.
with coronary artery disease. J Am Coll Cardiol 1998;31:331-7. 107. Weiss AT, Berman DS, Lew AS, Nielsen J, Potkin B, Swan
96. Sigal SL, Soufer R, Fetterman RC, Mattera JA, Wackers FJ. HJ, et al. Transient ischemic dilation of the left ventricle on
Reproducibility of quantitative planar thallium-201 scintigraphy: stress thallium-201 scintigraphy: a marker of severe and
quantitative criteria for reversibility of myocardial perfusion extensive coronary artery disease. J Am Coll Cardiol 1987;9:
defects. J Nucl Med 1991;32:759-65. 752-9.
97. Sinusas AJ, Beller GA, Smith WH, Vinson EL, Brookeman V, 108. Williams KA, Taillon LA, Draho JM, Foisy MF. First-pass
Watson DD. Quantitative planar imaging with technetium-99m radionuclide angiographic studies of left ventricular function
methoxyisobutyl isonitrile: comparison of uptake patterns with with technetium-99m-teboroxime, technetium-99m-sestamibi and
thallium-201. J Nucl Med 1989;30:1456-63. technetium-99m-DTPA. J Nucl Med 1993;34:394-9.
98. Smanio PE, Watson DD, Segalla DL, Vinson EL, Smith WH, 109. Winzelberg GG, Boucher CA, Pohost GM, McKusick KA,
Beller GA. Value of gating of technetium-99m sestamibi single- Bingham JB, Okada RD, et al. Right ventricular function in aortic
photon emission computed tomographic imaging. J Am Coll and mitral valve disease. Chest 1981;79:520-8.
Cardiol 1997;30:1687-92. 110. Stewart RE, Schwaiger M, Molina E, Popma J, Gacioch GM,
99. Smith WH, Watson DD. Technical aspects of myocardial planar Kalus M, et al. Comparison of rubidium-82 positron emission
imaging with technetium-99m sestamibi. Am J Cardiol 1990;66: tomography and thallium-201 SPECT imaging for detection of
16E-22E. coronary artery disease. Am J Cardiol 1991;67:1303-10.
 Positron emission tomography myocardial perfusion
and glucose metabolism imaging
Josef Machac, MD,a Stephen L. Bacharach, PhD,b Timothy M. Bateman, MD,b
Jeroen J. Bax, MD,b Robert Beanlands, MD,b Frank Bengel, MD,b
Steven R. Bergmann, MD, PhD,b Richard C. Brunken, MD,b James Case, PhD,b
Dominique Delbeke, MD,b Marcelo F. DiCarli, MD,b Ernest V. Garcia, PhD,b
Richard A. Goldstein, MD,b Robert J. Gropler, MD,b Mark Travin, MD,b
Randolph Patterson, MD,b Heinrich R. Schelbert, MDb

INTRODUCTION to be standard by the consensus of the committee; its

utilization is recommended, but other techniques may also
Cardiac positron emission tomography (PET) imaging
be valid. “Acceptable” means that the parameter value
is a well-validated, reimbursable means to assess myocar-
listed is a valid alternative to “Standard.” Techniques
dial perfusion, left ventricular (LV) function, and viability.
termed “Optional” indicate that the parameter value listed
Presently, there is a proliferation of PET and PET/com-
may be used or another acceptable parameter may be
puted tomography (CT) instrumentation as well as an
substituted.
increase in educational programs that specifically address
Part 2, “Interpretation and Reporting,” provides a
PET and PET/CT imaging. Technologists performing PET
systematic approach to QC, display, interpretation, and
scans as well as physicians interpreting them should have a
reporting of cardiac PET scans. Both subjective and objec-
sound knowledge of recommended standards for the per-
tive semiquantitative interpretive methods to evaluate myo-
formance, interpretation, and quality control (QC) of car-
cardial perfusion and viability are described. The Commit-
diac PET in order to provide accurate and clinically relevant
tee recognizes that all of these options may not be available
information to referring physicians, facilitating optimal
on computer workstations presently provided commer-
patient management.
cially. Therefore such recommendations may be considered
These guidelines are an update of an earlier version of
as general guidelines to direct the nuclear physician’s scan
these guidelines that have been developed by the Quality
interpretation in a detailed and organized fashion.
Assurance Committee of the American Society of Nuclear
This publication is designed to provide imaging guide-
Cardiology.1,2 The task of the Committee has been to
document state-of-the-art PET applications and protocols lines for those physicians and technologists who are quali-
approved by experts in the field and distribute these proto- fied in the practice of nuclear cardiology. Although care has
cols to the nuclear cardiology community. It is recognized been taken to ensure that information supplied is accurate,
that PET and PET/CT imaging is evolving rapidly and that representing the consensus of experts, it should not be
these recommendations may need further revision in the considered as medical advice or a professional service,
near future. since specific guidelines are partly instrument-dependent,
Part 1, “Patient Preparation and Data Acquisition,” and the technology of PET and PET/CT imaging is evolv-
addresses the instrumentation and protocols recommended ing rapidly. The imaging guidelines described in this
to yield technically adequate and clinically meaningful publication should not be used in clinical studies at any
cardiac PET scans. This section includes detailed explana- institution until they have been reviewed and approved by
tions of patient preparation options, recommended QC qualified physicians and technologists from that institution.
parameters, and scan acquisition and processing techniques.
Within this document, items judged to be required are PART 1: PATIENT PREPARATION AND DATA
indicated as such. “Preferred” means that the parameter ACQUISITION
value listed is expected to provide the best results and its
selection is strongly recommended. “Standard” means that A. General Comments
the parameter value listed represents methodology judged
There is now an extensive infrastructure in PET
imaging, with more than 1000 installed PET and PET/CT
Chair.a Member.b cameras in North America, spurred by their successful
J Nucl Cardiol 2006;13:e121-51.
1071-3581/$32.00
use in clinical oncology. A large number of PET scan-
Copyright © 2006 by the American Society of Nuclear Cardiology. ners are continuing to be installed. Myocardial PET
doi:10.1016/j.nuclcard.2006.08.009 perfusion imaging with rubidium 82, reimbursed by
e121
e122 Machac et al Journal of Nuclear Cardiology
PET and glucose metabolism imaging November/December 2006

Centers for Medicare & Medicaid Services (CMS) since multicrystal ring PET and PET/CT detector systems. We
1995, can be performed with a commercially available have not included coincidence gamma camera systems or
Rb-82 generator, obviating the need for a cyclotron. All (noncoincidence) collimated single photon emission
metropolitan areas in North America now have at least computed tomography (SPECT) systems. Gamma cam-
one commercial fluorine 18 fluorodeoxyglucose (FDG) era coincidence systems (“hybrid” PET) were not con-
supplier. FDG PET imaging is now reimbursed for sidered, because it appeared that the impact of factors
myocardial viability imaging. More recently, CMS reim- such as linearity of counts with activity, scatter correc-
bursement became available for myocardial perfusion tion, attenuation correction, and so on had not yet been
imaging with nitrogen 13 ammonia, for those centers that addressed sufficiently to allow standardized guidelines to
do have a cyclotron, that could until recently offer PET be proposed. Noncoincidence collimated SPECT with
myocardial perfusion imaging only for patients who FDG presented a different problem. Collimated FDG
could afford to pay for the procedure or who participated imaging has the advantage of permitting simultaneous,
in a funded research study. This has made cardiac PET dual-isotope perfusion and metabolism measurements, a
imaging possible at many clinical institutions, with the valuable feature for viability measurements and a feature
use of dedicated PET and PET/CT systems. The wide- that is not possible with coincidence PET. Despite this
spread installation of PET/CT cameras represents an- advantage, relatively few institutions utilize noncoinci-
other significant change. The use of CT for attenuation dence FDG imaging. Therefore it is not clear that a
correction substantially shortens the acquisition time for “standard” method of acquisition could be determined at
a clinical study. The present guidelines try to address the this time. Both gamma camera coincidence systems and
use of PET/CT cameras as best as possible with present 511-keV SPECT systems may well play a role in future
knowledge and experience. It should be borne in mind, cardiac FDG viability measurements. When a larger
however, that only a handful of centers use PET/CT clinical experience has been gained, the development of
cameras for cardiac imaging, with few published studies guidelines for these modalities may prove worthwhile.
providing guidance. The field is undergoing a rapid Attenuation effects are much more significant with
evolution, and the present guidelines for PET/CT imag- PET than SPECT because the 511-keV photons are
ing may soon be superceded or modified with newer attenuated along the entire path between the two detec-
advances. tors connected by the coincidence circuit detecting an
N-13 ammonia and Rb-82 are well-accepted myo- event. Therefore attenuation correction must be per-
cardial perfusion agents3-9 for rest and stress PET myo- formed for accurate interpretation of PET images of an
cardial imaging or in combination with FDG for deter- asymmetrical organ such as the heart. The algorithms to
mination of viability. FDG uptake in the myocardium has correct for attenuation are more accurate with PET than
been well validated as an indicator of myocardial viabil- SPECT because the length of the path of attenuation is
ity.10-12 The methods of acquiring perfusion images and constant and known for PET (distance between any two
FDG images are given in Part 1. The manner in which the detectors) and variable for SPECT. Accurate correction
obtained images are interpreted and used clinically are for attenuation represents a significant advantage of PET
described in Part 2. Section B, subsection iii, of Part 1 compared to SPECT, especially in patients with large
describes N-13 ammonia perfusion acquisitions (Table 1), body habitus. In addition, accurate attenuation correction
and section B, subsection iv, describes Rb-82 perfusion and calibration of the PET systems allow for absolute
imaging (Tables 2-4). Section C of Part 1 (Tables 5-7) measurements of radioactivity in selected regions of
describes FDG imaging. Note that all information below interest and absolute quantification of myocardial blood
is applicable only to adult patients. flow both at rest and during stress using compartmental
Oxygen 15–labeled water is often considered the modeling and kinetic analysis.
ideal tracer for measurement of myocardial blood flow. Various methods have been developed with mea-
Its use is not covered in this document for two reasons. sured attenuation using radioactive transmission sources,
First, it is currently not a Food and Drug Administration– such as germanium 68. Typically, radioactive sources
approved drug. Second, it does not usually produce rotate around the patient while transmission images are
clinically interpretable perfusion images. Instead, a acquired.
(well-validated) mathematical model must be used to Several manufacturers are now offering integrated
produce numeric values of flow in each region of the PET/CT imaging systems combining state-of-the-art
myocardium. Likewise, tracers of myocardial metabo- dedicated PET tomographs side by side with multidetec-
lism such as carbon 11 acetate, C-11 fatty acids, or tor CT units, with a common imaging bed. The CT units
tracers suitable for myocardial receptor imaging are not in these integrated systems were first offered with 2 and
covered, due to their present investigational status. 4 rows of detectors, are now available with 8 and 16 rows
We discuss only the use of so-called dedicated, of detectors, and will soon advance to 32 to 64 rows of
Journal of Nuclear Cardiology Machac et al e123
Volume 13, Number 6;e121-51 PET and glucose metabolism imaging

detectors, opening the horizon for cardiac applications quality image.” It is critical for the user to have a good
combining myocardial perfusion and viability with cal- understanding of the characteristics of his or her PET
cium scoring and CT coronary angiography.13 However, scanner (eg, plots of noise-equivalent counts vs activity
the latter applications are outside the scope of this concentration, dead time and randoms measurements,
review. scatter fraction) in order to know how the machine will
With these integrated systems, a CT scan and a PET behave under the circumstances of cardiac imaging. See,
scan can be acquired sequentially with the patient lying for example, “Performance Measurements of Positron
on the imaging table and simply being translated be- Emission Tomographs” (NU 2-2001) from the National
tween the two systems. Accurate calibration of the Electrical Manufacturers Association.15
position of the imaging table and the use of common Three-dimensional acquisition (ie, septa-out) is, in
parameters in data acquisition and image reconstruction principle, many times more sensitive than 2D (septa-in).
permit the fusion of images of anatomy, perfusion, and However, this often is only true at low doses. Randoms,
metabolism from the same region of the body that are dead time, and scatter can greatly reduce the effective
registered in space and only slightly offset in time. sensitivity of 3D acquisitions and at high doses. Three-
Because of the high photon flux of x-rays, the CT scan is dimensional acquisitions can (depending on the scanner
actually a high-resolution transmission map. With proper characteristics) actually produce poorer-quality images
care, these data can be used to perform a high-quality than 2D imaging for the same imaging time. Therefore in
attenuation correction during image reconstruction of the the past, when using 3D imaging with a BGO crystal
emission data. camera, 3D imaging has often only been used when the
One advantage of CT attenuation maps over the dose must be minimized (eg, normal volunteers, in
radioactive sources is the short duration for acquisition children, or when multiple studies are planned). The 3D
of the transmission images, in the range of 10 to 60 acquisition of cardiac images is an option that should be
seconds for one bed position over the heart. An adequate considered only by those institutions that are able to
CT transmission map can actually be obtained with very carefully monitor and assess randoms, dead time, and
low current (10 mA),14 but the low resolution of these scattered events. Note that 3D imaging is more practical
images makes them suboptimal for anatomical localiza- with the advent of LSO- and GSO-based PET scanners
tion. and even with BGO scanners with new-generation opti-
PET scanner instrumentation and design are contin- mized photo multiplier/crystal coupling schemes and
ually evolving. New crystal materials such as lutetium high-speed electronics. Still, the use of 3D cardiac
oxyorthosilicate (LSO) and gadolinium oxyorthosilicate imaging with these new-generation machines remains to
(GSO) are now available as well as others. These crystals be fully characterized, and preliminary reports have been
have higher light output and shorter dead time than the mixed. Use of 3D imaging is highly dependent on the
conventional bismuth germanate (BGO) crystals but ability to minimize and accurately correct for dead time,
have reduced stopping power for 511-keV photons. For randoms, and scatter. Typically, there is much greater
some manufacturers, LSO- and GSO-based systems have scatter with 3D (ie, septa-out) operation than with 2D
superceded their conventional BGO systems. In certain (septa-in) operation, for all crystal types. The newer
applications (eg, 3-dimensional [3D], septa-out imaging) crystals (LSO or GSO) and newer-generation electronics
systems with these new crystals may offer improved (with BGO) may, in principle, permit reduction of
performance. In addition, changes in electronics, crystal/ randoms and dead time and therefore may permit shorter
photo multiplier tube arrangements, and scatter and imaging time to be achieved. The degree to which any of
randoms corrections recently have been utilized to im- these improvements can be achieved in practice for
prove performance of BGO-based instruments in 3D cardiac imaging remains unknown at this writing. Scatter
mode. Designing a PET scanner always involves making remains much higher in 3D mode than in 2D mode even
choices and tradeoffs between various instrument char- for new-generation 3D machines. The user must care-
acteristics (eg, sensitivity and scatter). As a result, there fully evaluate plots of noise-equivalent counts and other
are variations in scanner characteristics between various system parameters to determine the optimum dose of
models of machine from the same manufacturer, as well FDG in 3D mode. Note that some new scanners only
as between manufacturers. permit septa-out operation.
Therefore Tables 1, 2, 3, 4, and 7 should only be
taken as guides to appropriate image acquisition, not as
B. PET Perfusion Imaging: N-13 Ammonia and
representing hard-and-fast rules. This is especially true
Rb-82
when assessing 3D versus 2-dimensional (2D) acquisi-
tions, as well as when determining the appropriate N-13 ammonia has been used in scientific investi-
number of total true events necessary to create a “good- gations in cardiac PET imaging over the past two
e124 Machac et al Journal of Nuclear Cardiology
PET and glucose metabolism imaging November/December 2006

decades. Recently, its use has been approved for clinical scanners are shown in Tables 2, 3, and 4. The short
imaging. Its 10-minute half-life requires an onsite cyclo- half-life of Rb-82 poses a challenge for achieving opti-
tron and radiochemistry synthesis capability. The N-13 mal image quality. For that reason, optimal acquisition
nitrogen decays by positron emission. The daughter parameters differ among the several main types of PET
product is C-13 carbon, which is stable. Myocardial scanners. Because of the short half-life of Rb-82 and the
uptake of N-13 ammonia depends on flow, extraction, need for the patient to lie still in the camera during the
and retention. First-pass myocardial extraction of N-13 study, stress imaging of this agent is limited to pharma-
ammonia is very high (95%).16 Following this initial cologic stress, although studies have obtained service-
extraction across the capillary membrane and via passive able Rb-82 images with supine bicycle exercise or even
diffusion into the myocyte, N-13 ammonia is either treadmill exercise.28
incorporated into the amino acid pool as N-13 glutamine i. Preparation. Patient preparation is similar to
or back-diffuses into the blood. The myocardial tissue preparation for stress and rest myocardial SPECT imag-
retention is an adenosine triphosphate– dependent meta- ing (Tables 1-4). This includes an overnight fast of 6
bolic process.17 Thus uptake and retention can both be hours or more, avoidance of caffeinated beverages for at
altered by changes in the metabolic state of the myocar- least 12 hours, and avoidance of theophylline-containing
dium. A suggested acquisition protocol is given in Table 1, medications for 48 hours.29
which can be performed by a wide variety of PET gamma ii. Stress testing. This document does not address
cameras. methods for performing stress studies (eg, protocols for
The radiation dosimetry from N-13 ammonia in an administration of pharmacologic stress agents). These
adult is 0.148 rad total effective dose from 20 mCi.18 The protocols are, for the most part, generic for all perfusion
critical organ is the urinary bladder, which receives 0.60 agents.29 The specific differences in acquisition proto-
rad from 20 mCi.19 The dosimetry is relatively low, due cols for N-13 ammonia or Rb-82 imaging are related to
to the short half-life of N-13 and the low energy of the the duration of uptake and clearance by these radiophar-
emitted positrons. maceuticals and their physical half-lives.
Rb-82 is a monovalent cationic analog of potassium iii. N-13 ammonia acquisition protocol: Introduc-
and is produced in a commercially available generator by tion to Table 1. Table 1 summarizes the recommended
decay from strontium 82 attached to an elution column. guidelines for performing N-13 ammonia perfusion
Sr-82 has a half-life of 25.5 days and decays to Rb-82 by scans with dedicated, multicrystal PET or PET/CT cam-
electron capture. Rb-82 is eluted from the generator with eras for rest and stress myocardial PET perfusion imag-
saline. Rb-82 decays with a physical half-life of 75 ing for diagnosis or evaluation of coronary artery disease
seconds by emission of a positron. The daughter product (CAD) or as part of an assessment of myocardial
is krypton 82, which is stable. The Sr-82– containing viability. N-13 ammonia is a valuable agent for measuring
generator is replaced every 4 weeks. either absolute or relative myocardial blood flow.3,4,6-9 For
Rb-82 is eluted with 10 to 50 mL normal saline by measurements of absolute flow, dynamic acquisition
a computer-controlled elution pump, connected by intra- from time of injection is required, followed by fitting to
venous (IV) tubing to the patient. While the generator is one of several possible physiologic models. Absolute
fully replenished every 10 minutes, experiments have flow measurements will not be discussed here, because
shown that 90% of maximal available activity can be they are time-consuming and require a high level of
obtained within 5 minutes after the last elution.20 Thus, expertise, which has led to their being performed primar-
serial imaging can be performed every 5 to 6 minutes. ily in a research setting. Relative perfusion measure-
While the short half-life of Rb-82 taxes the performance ments, described below, are often used clinically in the
limits of PET scanners, it facilitates the rapid completion evaluation of myocardial perfusion at rest and stress or in
of a series of resting and stress myocardial perfusion the determination of viability. Methods for normalizing
studies. Rb-82 is a very efficient imaging agent for and interpreting these images are discussed in Part 2
routine clinical usage. below.
Rb-82 is extracted from plasma with high efficiency
by myocardial cells via the Na⫹/K⫹ adenosine triphos-
Notes for Table 1
phatase pump. Myocardial extraction of Rb-82 is similar
to thallium 20121,22 and less than N-13 ammonia. Ex- 1. Uptake is relatively rapid (typically often nearly
traction decreases with increasing blood flow.23,24 Rb-82 complete in 90 seconds), and radioactive decay (10-
extraction can be decreased by severe acidosis, hypoxia, minute half-life) is fast. Typically, uptake images are
and ischemia.25-27 Thus, uptake of Rb-82 is a function of acquired no sooner than about 90 seconds after the
blood flow, metabolism, and myocardial cell integrity. end of infusion. Therefore a very slow infusion will
Acquisition protocols for different types of PET require static imaging to be delayed, potentially
Journal of Nuclear Cardiology Machac et al e125
Volume 13, Number 6;e121-51 PET and glucose metabolism imaging

Table 1. N-13 ammonia cardiac perfusion studies

For details, see
Feature Technique note in text

Patient preparation Overnight fast (⬎6 h) Preferred

No caffeine or caffeinated beverages for 12 h Acceptable
No theophylline-containing medications for 48 h Preferred
Dose 10-20 mCi (typical) (370-740 MBq) Standard 1
Bolus or ⬍30-s infusion Preferred 1
Imaging acquisition Static Standard
Start time: 1.5-3 min after end of infusion Standard 2
Duration: 5-15 min Standard 3
Pixel size (reconstructed) 2-3 mm Preferred 4
4 mm Optional
Attenuation correction
PET Measured attenuation correction: immediately before Standard 5
scan
Measured attenuation correction: immediately after scan Optional 5
PET/CT CT-based measured attenuation correction: immediately Standard 5
after scan
Reconstruction method FBP or iterative expectation maximization (eg, OSEM) Standard 4
Gating ECG gating of myocardium Preferred 6
Patient positioning Arms out Preferred 7
Arms in (but no movement) Optional 7
Calcium scoring Gated CT acquisition separate from CTAC Optional 8

resulting in count loss because of the 10-minute 5. Measured attenuation by a rotating 511-keV source is
half-life. Large patients may benefit from higher preferred for dedicated 2D scanners. Either prescan or
(25-30 mCi) doses. postscan is satisfactory, providing it has been verified
2. The static image should not include the initial rapidly that the user’s attenuation correction software can
changing uptake portion of the study. Therefore a adequately correct for residual emission activity.
minimum of 90 seconds should typically elapse be- Attenuation correction simultaneous with emission
tween the end of infusion and the beginning of the scan is not recommended unless data become avail-
static scan. In fact, the arterial blood concentration of able to indicate that the high count rate and rapidly
ammonia is often still quite significant even at 90 changing distribution of the isotope will not adversely
seconds after a rapid bolus injection. Nonetheless, affect the transmission scan. See notes on attenuation
many published data are based on only a 90-second correction in Table 7. For patients undergoing PET/
delay before the start of imaging. CT, two separate CT-based transmission scans should
3. After an initial period of rapidly changing activity be performed for correction of the rest scan (either
levels during the uptake period, the decay-corrected before or after the emission scan) and stress (after the
ammonia concentration subsequently usually emission scan is preferred to prevent misregistration
changes only very slowly. However, the 10-minute artifacts on the corrected ammonia images).
N-13 half-life makes acquisition duration longer 6. If myocardial wall motion information is desired, it is
than 20 minutes of limited value unless total counts indeed possible to achieve sufficiently high-quality
are very low. gated ammonia scans to accurately evaluate wall
4. It is desirable to keep reconstruction parameters motion and ejection fraction (EF).30
similar to those used for the FDG portion of a 7. Ideally, the patient should be positioned supine, with
viability study (see notes for Table 7) in order that the arms out of the camera field of view. This can be
perfusion and metabolism are affected by reconstruc- tolerated by nearly all patients, provided some care is
tion parameters in the same way. This permits more given to a method to support the arms. Alternatively,
accurate comparison between the two image sets. an overhead bar has often been used as a hand-hold
e126 Machac et al Journal of Nuclear Cardiology
PET and glucose metabolism imaging November/December 2006

Table 2. Rb-82 rest/stress myocardial perfusion imaging guideline for BGO PET imaging systems
For details, see
Feature Technique note in text

2D dose: BGO systems 40-60 mCi (1480-2220 MBq) Standard 1

3D dose: BGO systems 10-20 mCi (370-740 MBq) Optional 1
Patient positioning
PET Use scout scan: 10-20 mCi Rb-82 (370-740 MBq) Standard 2
Use transmission scan Optional 2
PET/CT CT scout Standard 2
Injection rate Bolus of ⱕ30 s Standard
Imaging time 3-6 min Standard 3
Imaging delay after injection LVEF ⬎50% ⫽ 70-90 s Acceptable 3
LVEF ⬍50% ⫽ 90-130 s
130-150 s if LVEF is not known Optional
Imaging mode Phased/dynamic (no delay after injection) Preferred 3
Gated acquisition (delay after injection) Optional
Rest attenuation correction Measured attenuation correction, before or after Standard 4
Stress testing Pharmacologic agents Standard 5
Stress attenuation correction
PET Measured attenuation correction before or after Optional 6
stress to ensure image registration
PET/CT Measured attenuation correction after stress to Standard 6
ensure image registration
Reconstruction method FBP or iterative expectation maximization Standard 7
(eg, OSEM)
Reconstruction filter Sufficient to achieve desired Standard 7
resolution/smoothing, matched stress to rest
Pixel size (reconstructed) 2-3 mm Preferred 7
Coronary calcium scoring Gated CT acquisition separate from CTAC Optional 8

for arm support. In those very few cases in which may be useful to guide patient management. How-
arms-out positioning is not possible (eg, patients with ever, because the added prognostic value of this
very severe arthritis), the arms can be in the field of approach has not been proven, calcium scoring re-
view. In this case the transmission scan time may mains optional.
have to be increased, and it is of critical importance
iv. Rb-82 perfusion acquisition protocol. Intro-
that the arms not move between transmission and
duction to Tables 2, 3, and 4. Tables 2, 3, and 4
emission, or artifacts will result. Note that when
summarize the acquisition parameters necessary to ac-
performing ammonia/FDG perfusion/metabolism stud-
quire an Rb-82 perfusion study with a dedicated PET or
ies, it is best to keep patient positioning similar for both
PET/CT camera.4,9 Because of the divergent operating
studies, and this is often very difficult or impossible to
characteristics of BGO, LSO, and GSO PET imaging
accomplish with a change (arms in/out) in arm position.
systems, recommendations are considered separately for
In patients undergoing PET/CT imaging, arms in (down)
each type of system.
the field of view result in beam-hardening artifacts on
the CT-based transmission scan, which usually lead to
streak artifact of the corrected emission scans. Notes for Tables 2 and 3
8. With PET/CT systems, it is possible to acquire a
separate (from the CT-based attenuation correction 1. Scout scanning: Scout scanning is recommended
[CTAC]) gated CT with adequate breath-holding for before each injection to ensure that the patient is
measuring coronary calcium scores. Calcium scores correctly positioned and is not unnecessarily exposed
provide a measure of atherosclerotic burden that, in to radiation. This can be done with a fast transmission
combination with the stress perfusion PET images, image or with a low-dose Rb-82 injection (10-20 mCi).
Journal of Nuclear Cardiology Machac et al e127
Volume 13, Number 6;e121-51 PET and glucose metabolism imaging

Table 3. Rb-82 rest/stress myocardial perfusion imaging guideline for LSO PET imaging systems
For details, see
Feature Technique note in text

3D dose: LSO systems 30-40 mCi (1110-1480 MBq) Standard 1

2D dose: LSO systems 40-60 mCi (1480-2220 MBq) Optional 1
Patient positioning
PET Use scout scan: 10-20 mCi Rb-82 (370-740 MBq) Preferred 2
Use transmission scan Optional 2
PET/CT CT scout Standard 2
Injection rate Bolus of ⱕ30 s Standard 2
Imaging time 3-6 min Standard 3
Imaging delay after injection LVEF ⬎50% ⫽ 70-90 s Acceptable 3
LVEF ⬍50% ⫽ 90-130 s
130-150 s if LVEF is unknown Optional
Imaging mode Phased/dynamic (no delay after injection) Preferred 3
Gated acquisition (delay after injection) Optional
Rest attenuation correction Measured attenuation correction, before or after Standard 4
Stress testing Pharmacologic agents Standard 5
Stress attenuation correction
PET Measured attenuation correction before or after Optional 6
stress to ensure image registration
PET/CT Measured attenuation correction after stress to Standard 6
ensure image registration
Reconstruction method FBP or iterative expectation maximization Standard 7
(eg, OSEM)
Reconstruction filter Sufficient to achieve desired Standard 7
resolution/smoothing, matched stress to rest
Pixel size (reconstructed) 2-3 mm Preferred 7
Coronary calcium scoring Gated CT acquisition separate from CTAC Optional 8

Note that the low-dose Rb-82 scout scan is also used addition, the very large increase in scattered events
to estimate circulation times and cardiac blood pool may also obviate some of the sensitivity advantages
clearance times, which assist in selection of the of 3D imaging. Newer imaging crystals (eg, LSO,
optimum injection to imaging delay time between GSO, and others) allow imaging at higher count rates,
Rb-82 injection and initiation of acquisition of myo- as does the new generation of electronics coupled
cardial Rb-82 images (see 3 below). With PET/CT with BGO systems. Count rate issues are especially
systems, the CT scout scan is routinely used for critical to Rb-82 imaging.
patient positioning. NOTE: See previous discussion of 2D versus 3D
2. General dose considerations: In determining appro- imaging. For Rb-82 imaging, 3D imaging, even with
priate patient dosages,5 the following issues should be new-generation scanners, must be used with care.
considered: (a) Patient exposure is typically low 3. Rest imaging time: Rest imaging should be per-
relative to SPECT because of the short half-life of the formed before stress imaging to reduce the impact
isotope. (b) Staff exposure could be high unless care of residual stress effects (eg, stunning, steal).
is taken to avoid close proximity to the generator and About 80% of the useful counts are acquired in the
the patient during injection, because of the limited first 3 minutes, 95% of the useful counts are
effectiveness of shielding and the higher dosages used obtained in the first 5 minutes, and 97% are
in Rb-82 PET. (c) Three-dimensional imaging re- obtained in the first 6 minutes. The patient should
quires less dosage than 2D imaging because of the be infused with Rb-82 for a maximum of 30
improved sensitivity of the system; however, the dead seconds. After the dose is delivered, patients with
time and increased randoms of the camera may not normal ventricular function (LVEF ⬎50%) are
allow one to utilize the improved sensitivity. (d) In typically imaged starting 70 to 90 seconds after the
e128 Machac et al Journal of Nuclear Cardiology
PET and glucose metabolism imaging November/December 2006

injection. For those with reduced ventricular func- images. Many laboratories perform the transmis-
tion (LVEF 30%-50%), imaging usually is begun sion scan between the rest and stress images,
90 to 110 seconds after termination of the infusion, although some perform the transmission scan at
and those with poor function (LVEF ⬍30%) are the very beginning. Loghin et al33 have deter-
typically imaged at 110 to 130 seconds. These mined that if the transmission scan is performed
times can be estimated from observations of the in the beginning of the study, greater misregis-
scout Rb-82 scan, if used. Ideally, patients should tration occurs with the later stress images, pos-
be imaged by a dynamic acquisition to allow for sibly due to gradual upward creep of the dia-
retrospective removal of phases that have Rb-82 in phragm, due to pressure from visceral fat.
the blood pool, in which case imaging can begin b. CT based: For PET/CT systems, x-ray computed
within a few seconds after the start of the rubidium tomography can be used for acquiring a transmis-
infusion. Excessive blood pool counts can scatter sion map for attenuation correction. This technique
into myocardial defects, making them artifactually may be advantageous because the transmission
look milder and thus potentially reversible. Exces- map can be acquired rapidly (15-60 seconds), can
sive blood pool counts can also make the LV cavity be repeated for rest and stress and metabolic
appear smaller, especially at rest, leading to a false imaging, and has independent diagnostic informa-
perception of LV cavity dilatation during stress. tion, such as coronary calcium visualization and
The dynamic data set is also useful to estimate non-cardiac anatomic information. To acquire a
coronary flow reserve. Electrocardiographic (ECG) CT-based transmission scan, it is necessary to first
gating can also be used with Rb-82. Images can be acquire a planar scout CT acquisition. This scan is
acquired by 2D or 3D imaging modes (but see 2D used to measure the axial limits of the CT acqui-
vs 3D notes to Table 7). If simultaneous ECG sition. Following this acquisition, the CT transmis-
gating is used, the standard delay delineated above
sion scan is acquired. The best approach for CT
is applied. List mode acquisitions, which allow for
transmission imaging is still evolving, and there-
simultaneous phase/dynamic and ECG-gated acqui-
fore this guideline can only suggest some consid-
sitions, are optional as tools for sorting and recon-
erations. Some of the considerations for CT scan-
struction of list mode data but are not yet ready for
ning are as follows:
routine clinical use.
i. If CT is either for attenuation correction or
4. Rest transmission imaging: Rb-82 myocardial perfu-
anatomical evaluation, this will have an effect
sion should only be performed with attenuation cor-
on the kV and mAs used in the acquisition. A
rection.31 Attenuation correction can be accomplished
with a rotating line source in a dedicated PET system transmission scan usually requires only a low
or with CT in a PET/CT system. CT current, as opposed to calcium scoring or
a. Dedicated PET: Two techniques are typically CT angiography, which require higher CT
used for creating the patient-specific transmis- currents. An effort must be made to minimize
sion maps: direct measurement of patient atten- radiation exposure.
uation with a rotating line source of either Ge-68 ii. Breathing protocols are not clearly settled.
or cesium 137 or segmentation of patient-spe- Data seem to be in favor of free breathing and
cific attenuation maps. The former are very a slow CT scan. Current practice discourages
sensitive to the choice of reconstruction algo- breath-holding, particularly in end inspiration
rithm and, depending on reconstruction algo- because of its potential for causing uncorrect-
rithm used, could require 60 to 600 seconds’ able misregistration. A very rapid transmission
acquisition time to produce a reasonable atten- CT scan performed at the same speed as for
uation map.32 2D imaging generally requires whole-body PET/CT images frequently pro-
longer attenuation map imaging time compared duces artifacts at the lung-liver interface and
to 3D imaging. Segmentation algorithms are can sample parts of the heart and diaphragm in
relatively insensitive to noise but are very de- different positions, causing misregistration and
pendent on the quality of the program used for an artifact where pieces of the diaphragm
performing the transmission scan segmentation appear to be suspended in the lung. Although
and are influenced by lung attenuation inhomo- specifics vary among laboratories, the duration
geneities (eg, partial-volume effects from liver). of the CT transmission scan is typically from
Transmission data are typically performed se- 10 to 60 seconds.
quentially, so it is essential that the patient iii. Metal artifacts can present a challenge for
remain still between transmission and emission the reconstruction algorithm and must be
Journal of Nuclear Cardiology Machac et al e129
Volume 13, Number 6;e121-51 PET and glucose metabolism imaging

Table 4. Rb-82 rest/stress myocardial perfusion imaging guideline for GSO PET imaging systems
For details, see
Feature Technique note in text

3D dose: GSO systems 20 mCi (740 MBq) Standard 1

Patient positioning
PET Use scout scan: 10-20 mCi Rb-82 (370-740 MBq) Standard 2
Use transmission scan Optional 2
PET/CT CT scout Standard 2
Injection rate Bolus of ⱕ30 s Standard
Imaging time 3-6 min 3
Imaging delay after injection 2-3 min (when count rate reaches 10-11 million cps) Standard 3
Imaging mode Phased/dynamic (no delay after injection) Standard 4
Gated imaging None 4
Rest attenuation correction Measured attenuation correction, before or after Standard 5
Stress testing Pharmacologic agents Standard
Stress attenuation correction
PET Measured attenuation correction with stress to Optional 5
ensure image registration
CT transmission scan Optional
PET/CT Measured attenuation correction after stress to Standard 5
ensure image registration
Reconstruction method Iterative (RAMLA) Standard 6
Reconstruction filter None Standard 7
Pixel size (reconstructed) 4 Standard
Coronary calcium scoring Gated CT acquisition separate from CTAC Optional 8

compensated for to produce accurate atten- misregistration compensation software is not avail-
uation maps.34,35 able.
7. Processing protocol: Several corrections are required
5. Stress testing: The long infusion time for Rb-82 and
for creating data sets that can be used for reconstruc-
slow uptake require some modifications to conven-
tion. Rb-82 data must be corrected for randoms,
tional stress testing. On average, the patient should
remain at peak stress for somewhat longer than scatter, dead time, attenuation, and decay before
conventional SPECT-based radionuclide stress test- reconstruction can begin. Once these corrections are
ing. The radionuclide should be injected in a manner applied, the data can be reconstructed with either
such that all of the Rb-82 is taken up in the stress filtered backprojection (FBP) or iterative algorithms.
state. See previously published guidelines for further For viability studies, it is often desirable to match the
information on pharmacologic agents for stress test- resolution of the FDG and the perfusion (rubidium)
ing.29 agent, although this is less critical when the data are
6. Stress transmission imaging: These images should be divided into 8 or fewer sectors per short-axis slice and
acquired while the patient is still at the peak of stress. comparisons made on a sector-by-sector basis. For
If the patient cannot tolerate this or if the stress testing rest/stress comparisons, the rest/stress images must
protocol will not allow this, the technologist and have matched resolution. Filtering with FBP or addi-
physician must carefully inspect the transmission and tional filtering of the ordered-subset expectation maxi-
emission data sets to ensure that they are properly mization (OSEM) (eg, Butterworth, Hanning, Gaussian)
registered in the transaxial, sagittal, and coronal is usually necessary to achieve adequate noise proper-
planes. For patients undergoing PET/CT, a separate ties. Again, care must be taken to match reconstructed
CT-based transmission scan for correction of the resolution when making pixel-by-pixel comparisons
stress rubidium images is standard. A post-stress of perfusion and metabolism.
transmission scan is preferred to minimize misregis- 8. With PET/CT systems, it is possible to acquire a
tration artifacts on the corrected Rb-82 images when separate gated CT with breath-holding for measuring
e130 Machac et al Journal of Nuclear Cardiology
PET and glucose metabolism imaging November/December 2006

coronary calcium scores. Calcium scores provide a same transport mechanism as glucose and is phosphor-
measure of atherosclerotic burden that, in combina- ylated intracellularly by a hexokinase into FDG-6-phos-
tion with the stress perfusion PET images, may be phate (FDG-6-P). In tissues with a low concentration of
useful to guide patient management. However, be- glucose-6-phosphatase, such as the myocardium,
cause the added prognostic value of this approach has FDG-6-P does not enter into further enzymatic path-
not been proven, calcium scoring remains optional. ways and accumulates intracellularly proportionally to
the glycolytic rate of the cell. The myocardium can use
various substrates according to substrate availability,
Notes for Table 4
hormonal status, and other factors. In a typical fasting
1. Given the higher sensitivity of the 3D imaging oblig- state, the myocardium primarily utilizes free fatty
atory in existing GSO systems, a lower dose of Rb-82 acids, but post-prandially or after a glucose load, it
is sufficient and preferred to reduce dead time and favors glucose.40,41
randoms. Radiation exposure is further decreased. Following injection, FDG is slowly taken up by
2. The existing GSO systems have not featured gated body tissues, including the myocardium. Imaging is
acquisition until recently. performed about 45 to 90 minutes after injection. The
3. The GSO system uses a Cs-137 rod source to acquire 110-minute physical half-life of F-18 FDG allows suffi-
a transmission scan for attenuation correction. A CT cient time for synthesis and purification, its commercial
transmission scan is an alternative for GSO PET/CT distribution in a radius of several hours from the produc-
systems. tion site, its temporary storage at the user site, the
4. The GSO system uses the RAMLA reconstruction absorption time after injection, and sufficient imaging
technique (row action maximum likelihood algo- time to yield images of high quality.
rithm). ii. Patient preparation: Introduction to Tables 5
5. For the GSO system, there should not be any recon- and 6. FDG uptake, combined with a PET or SPECT
struction filter. (The RAMLA/Blob includes a texture/ perfusion measurement, has been well validated as a
filter factor.) measure of myocardial viability (see Part 2 of these
guidelines). FDG uptake begins with facilitated diffu-
sion, followed by hexokinase-mediated phosphorylation.
C. PET FDG Metabolism Imaging
Therefore, uptake of FDG requires viable myocardial
i. Background and FDG tracer characteristics. cells. In the fasting state, normal myocardium preferen-
Tables 5, 6, and 7 summarize the recommended guide- tially utilizes free fatty acids. The uptake of glucose and
lines for performing cardiac FDG scans with dedicated, FDG is generally low, although fairly variable. In the
multicrystal PET and PET/CT cameras, as part of an fasting state, uptake inhomogeneity frequently occurs.
assessment of myocardial viability. Tables 5 and 6 Therefore, whereas uptake of FDG indicates viability,
summarize the patient preparation and method of FDG lack of uptake could either indicate nonviable tissue or
administration. Table 7 discusses the image acquisition. indicate viable tissue that was utilizing substrates other
Preserved metabolism for the production of adeno- than glucose. For this reason, every effort is made to
sine triphosphate is one of the critical features of myo- force the myocardium to utilize primarily glucose to
cardial viability. FDG is F-18 –labeled 2-deoxyglucose, meet its energy needs by stimulating a natural insulin
an analog of glucose. It is the principal workhorse in response. This is usually accomplished by having the
clinical PET viability imaging. F-18 is produced in a patient fast for at least 6 hours and then administering a
cyclotron through the (p,n) reaction, consisting of bom- standardized glucose load, either orally or intravenously.
bardment of O-18 – enriched water.36 Subsequent to glucose loading, glucose and insulin
F-18 fluorine decays by the emission of a positron plasma levels are elevated, and glucose is the preferred
with a half-life of 109.8 minutes, whereby F-18 fluorine substrate for energy metabolism.42 The approach is
is converted into O-18 oxygen. The low kinetic energy of outlined in Table 5.
the positron, 635 keV, allows the highest spatial resolu- There are several approaches to the administration
tion among all PET radionuclides. The whole body of glucose. The guiding principle seems to be that the
dosimetry from a 10-mCi dose is 0.7 rem,37 as compared blood glucose (BG) needs to be increased by glucose
to 2.0 rem from a 3.5-mCi dose of Tl-20138 and 0.5 rem administration (in order to shift myocardial metabolism
from a 30-mCi dose of Tc-99m sestamibi.39 For FDG, to glucose) and, then, to be declining (to promote FDG
the critical organ is the urinary bladder, which receives uptake) after insulin injection before FDG administra-
5.9 rem. tion. The situation is more complicated should the patient
FDG is an analog of glucose and is used as a tracer be diabetic, not achieve a sufficiently low fasting BG
of glucose metabolism. FDG enters into the cells by the level, or have too high a BG level after glucose admin-
Journal of Nuclear Cardiology Machac et al e131
Volume 13, Number 6;e121-51 PET and glucose metabolism imaging

Table 5. FDG cardiac PET: Patient preparation guidelines—An overview

For details, see
Procedure Technique note(s) in text

Fasting period Step 1: Fast patient 1

6-12 h Preferred 1
⬍6 h Suboptimal 1
Step 2: Check blood glucose and then glucose load
(choose one of the following 4 options)
Oral glucose load Option 1: Oral glucose loading
IF: fasting BG ⬍ ⬃110 mg/dL AND: No known Standard 1 and 2, as well as
diabetes THEN: (1) Oral glucose load: typically Table 6
25-100 g orally (see Table 6) (2) Monitor blood
glucose (see Table 6)
IF: fasting BG ⬎ ⬃110-130 mg/dL OR: Known Standard 1, 2, 4, and 5
diabetes THEN: See Table 6
OR
IV, protocol A Option 2: Hyperinsulinemic/euglycemic IV clamp Optional
For details, see sample protocol A 4
OR
IV, protocol B Option 3: Dextrose IV infusion Optional 5
For details, see sample protocol B
OR
Option 4: Acipimox
Acipimox Acipimox, 250 mg orally (see reference in note 3 3
for details), not available in United States
Step 3: Administer FDG
FDG injection Time: Dependent on which option was selected Standard Table 7, item 1
Administer FDG intravenously; see Table 7, item 1,
for details
Step 4: Begin imaging
Begin PET imaging Time 60-90 min post FDG injection: start imaging, Table 7
see Table 7

istration. There are a variety of methods to deal with Notes for Tables 5 and 6: Patient Preparation
these situations. Table 6 discusses several options
should BG values not reach the desired ranges. In 1. Myocardial substrate utilization. As mentioned in
addition, two sample IV protocols (protocol A and the introduction to Tables 5 and 6, for evaluation of
protocol B) are given below. These protocols illustrate myocardial viability using FDG, the substrate and
some of the various possible approaches to IV glucose hormone levels in the blood need to be pushed to
loading and BG level control. Some of the glucose- favor utilization of glucose by the myocardium.11,43
loading methodologies are easily implemented in stan- This is usually accomplished by loading the patient
dard nuclear medicine facilities, whereas others may with glucose after a fasting period of at least 6 hours
be more elaborate than some facilities feel comfort- to induce an endogenous insulin response. A shorter
able performing on a routine basis. Most laboratories fasting time may depress this physiological response.
utilize the oral glucose–loading methods, with supple- The most common method of glucose loading is with
mental insulin if needed, because of its simplicity. The an oral load of 25 to 100 g, but IV loading is also used
reader is urged to examine Tables 5 and 6 and the and has some advantages (as described in detail in the
sample protocols and use them as a guide to develop- two sample protocols below). Either can be adequate
ing an approach that will be feasible in his or her own for nondiabetic patients, if BG level falls sufficiently
setting. (see Table 6 for details) before FDG injection. The IV
e132 Machac et al Journal of Nuclear Cardiology
PET and glucose metabolism imaging November/December 2006

Table 6. Guidelines for BG maintenance (eg, after oral glucose administration) for optimal FDG cardiac
uptake, BG of approximately 100 to 140 mg/dL at FDG injection time
BG at 45-60 min after Possible restorative For details, see
administration measure Technique note(s) in text

130-140 mg/dL 1 u regular insulin Standard 1 and 2

140-160 mg/dL 2 u regular insulin
160-180 mg/dL 3 u regular insulin
180-200 mg/dL 5 u regular insulin
⬎200 mg/dL Notify physician

route avoids potential problems due to variable gas- 4.1.1. Prepare dextrose/insulin solution: 15 U of reg-
trointestinal absorption times or inability to tolerate ular insulin in 500 mL of 20% dextrose in a
oral dosage. Note that if the patient is taking medica- glass bottle. The initial 50 mL is discarded
tions that may either antagonize or potentiate the through the plastic IV tubing (no filter) to
effects of insulin, these should be taken into account decrease adsorption of the insulin to the tubing.
by the physician. 4.1.2. Prime the patient with 5 U of regular insulin
2. Diabetic patients. Diabetic patients pose a challenge, and 50 mL of 20% dextrose (10 g) IV bolus.
either because they have limited ability to produce 4.1.3. Infuse dextrose/insulin solution at a rate of 3
endogenous insulin or because their cells are less able mL · kg⫺1 · h⫺1 for 60 minutes (corresponding to
to respond to insulin stimulation. For this reason, the an insulin infusion of 1.5 mU · kg⫺1 · min⫺1 and
simple fasting/oral glucose–loading paradigm is often a glucose infusion of 10 mg · kg⫺1 · min⫺1).
not effective in diabetic patients. Fortunately, use of Monitor BG every 10 minutes (goal BG, 100-200
insulin along with close monitoring of BG (Table 6) mg/dL).
yields satisfactory results.44 Image improvement can 4.1.4. If BG at 20 minutes is 100 to 200 mg/dL
also be seen after waiting 2 to 3 hours after injection (preferably ⬍150 mg/dL), administer FDG in-
before imaging (at the expense of increased decay of travenously.
the radiopharmaceutical FDG). An alternative method 4.1.5. If BG is greater than 200 mg/dL, administer
is the euglycemic hyperinsulinemic clamp,45 a rigor- small IV boluses of 4 to 8 U of regular insulin
ous and time-consuming procedure, allowing regula- until BG decreases to less than 200 mg/dL.
tion of metabolic substrates and insulin levels and Administer FDG intravenously.
providing excellent image quality in most patients,46 4.1.6. Stop dextrose/insulin infusion at 60 minutes, and
especially in those with non–insulin-dependent diabe- start 20% dextrose at 2 to 3 mL · kg⫺1 · h⫺1.
tes mellitus (protocol 4A). A shorter IV glucose/ 4.1.7. During image acquisition, continue infusion of
insulin-loading procedure (30 minutes) has also been 20% dextrose at 2 to 3 mL · kg⫺1 · h⫺1.
used with some success47 (protocol 4B). 4.1.8. At completion of the acquisition of the images,
3. Acipimox. Acipimox is not currently available in the discontinue infusion, give a snack to the pa-
United States but has been used successfully in tient, and advise him or her regarding the risk
Europe instead of glucose loading. Acipimox is a of late hypoglycemia.
nicotinic acid derivative inhibiting peripheral lipoly- 4.1.9. ALERT: (1) If BG is greater than 400 mg/dL,
sis, reducing plasma free fatty acid levels, and indi- call the nuclear physician immediately. (2) If BG
rectly stimulating myocardial glucose utilization.48,49 is less than 55 mg/dL or if the patient develops
symptoms of hypoglycemia with BG less than
75 mg/dL, discontinue dextrose/insulin infu-
Two Sample IV Protocols
sion, administer one amp of 50% dextrose
4. Protocol A. A sample protocol for IV glucose loading intravenously, and call the nuclear physician.
is presented. This protocol is based on one in use at 4.2. IV glucose/insulin loading for diabetic patients or
Vanderbilt University Medical Center, Nashville, fasting BG greater than 110 mg/dL:
Tenn, and is adapted from Martin et al.48 4.2.1. Prepare insulin solution: 100 U of regular
4.1. IV glucose/insulin loading for nondiabetic patients insulin in 500 mL of normal saline solution in
and fasting BG is less than 110 mg/dL: a glass bottle. The initial 50 mL is discarded
Journal of Nuclear Cardiology Machac et al e133
Volume 13, Number 6;e121-51 PET and glucose metabolism imaging

through the plastic IV tubing (no filter) to constant rate for 3 minutes, and then at a
decrease adsorption of the insulin to the tubing. constant rate for the remainder of the study.
4.2.2. Prime patient with regular insulin: 4.4.3. If the BG is greater than 200 mg/dL, an
4.2.2.1. If fasting BG is greater than 140 mg/dL, prime additional bolus of insulin is given. An exoge-
the patient with 10 U of regular insulin IV nous 20% glucose infusion is started at an
bolus. initial rate of 0.25 mg · kg⫺1 · min⫺1 and ad-
4.2.2.2. If fasting BG is less than 140 mg/dL, prime justed until steady state is achieved. The BG
the patient with 6 U of regular insulin IV concentrations are measured every 5 minutes
bolus. during the insulin clamp. The glucose infusion
4.2.3. Infuse insulin solution at a rate of 1.2 is adjusted according to the plasma glucose
mL · kg⫺1 · h⫺1 for 60 minutes (corresponding over the preceding 5 minutes.
to an insulin infusion of 4 mU · kg⫺1 · min⫺1) 5. Protocol B. A sample protocol for IV glucose loading
or for the entire study (to calculate the regional is presented. Protocol B is based on the protocol in
glucose utilization rate). use at the Emory University–Crawford Long Memo-
4.2.4. After 8 to 10 minutes or when BG is less than rial Hospital (Atlanta, Ga).47 This protocol has been
140 mg/dL, start 20% dextrose infusion at 1.8 used in over 600 subjects (over one third of whom
mL · kg⫺1 · h⫺1 (corresponding to a dextrose were diabetic), resulting in good-quality images in
infusion of 6 mg · kg⫺1 · min⫺1). over 98% of studies.
4.2.5. Monitor BG every 5 to 10 minutes and adjust 5.1. If fasting BG is less than 125 mg/dL, give 50%
dextrose infusion rate to maintain BG at 80 to dextrose in water (D-50-W), 25 g, intravenously.
140 mg/dL. Hydrocortisone, 20 mg, should be added to the
4.2.6. After 20 to 30 minutes of stable BG, administer D-50-W to minimize the rather severe pain that
FDG. can occur at the injection site with D-50-W. This is
4.2.7. Maintain the IV infusion of insulin plus 20% compatible and avoids the pain that limits patient
dextrose for 30 to 40 minutes after FDG injec- cooperation. There is no negative effect on the
tion or until the end of the scan (to calculate quality of the FDG studies.
rMGU [rate of glucose utilization]). Some cen- 5.2. If fasting BG is between 125 and 225 mg/dL, give
ters confirm FDG uptake particularly in patients D-50-W, 13 g, intravenously.
with diabetes before discontinuing the clamp. 5.3. If fasting BG is greater than 225 mg/dL, adminis-
4.2.8. At completion of the acquisition of the images, ter regular aqueous insulin as per the following
discontinue infusion, give a snack to the pa- formula: Regular aqueous insulin (dose units) ⫽
tient, and advise him or her regarding the risk (BG ⫺ 50)/25.
of late hypoglycemia. 5.4. After 30 to 60 minutes, if BG is less than 150
4.3. For lean patients with type 1 juvenile-onset diabe- mg/dL, give FDG intravenously, but if BG is
tes mellitus, alter protocol 4.2 as follows: greater than 150 mg/dL, give more regular insulin,
4.3.1. If fasting BG is less than 140 mg/dL, inject 4 U of using the formula in 5.3 above, until BG is less
regular insulin and infuse insulin solution (pre- than 150 mg/dL, before giving FDG. Giving FDG
pared as in 4.2.1 above) at 0.3 mL · kg⫺1 · h⫺1 (1 when BG is 150 to 200 mg/dL resulted in many
mU · kg⫺1 · min⫺1). poor-quality studies.
4.3.2. After 8 to 10 minutes of infusion or when BG is FDG cardiac PET acquisition parameters. Acqui-
less than 140 mg/dL, start 20% dextrose at 2.4 sition parameters for PET cardiac FDG imaging are
mL · kg⫺1 · h⫺1 (8 mg · kg⫺1 · min⫺1). itemized in Table 7 and its accompanying notes.
4.4. Some centers (Munich, Ottawa, and others) have Most of the literature about viability and prediction
also applied a front-loaded infusion. of recovery after revascularization with PET is based on
4.4.1. About 6 hours after a light breakfast and their mismatch perfusion/metabolism (see Part 2 of these
usual dose of insulin or oral hypoglycemic, all guidelines). N-13 ammonia and Rb-82 provide optimal
diabetic patients have a catheter inserted in one perfusion images for comparison because the images are
arm for glucose and insulin infusion, as well as acquired with the same PET system and can be displayed
a catheter in the opposite arm for BG measure- with similar parameters as the FDG images. However, if
ment. these PET perfusion agents are not available, the FDG
4.4.2. At time 0, the insulin infusion is started. Reg- images can be interpreted in conjunction with SPECT
ular insulin is given at 4 times the final constant perfusion images (see Part 2 below).
rate50 for 4 minutes, then at 2 times the final This comparison of PET to SPECT can be difficult
e134 Machac et al Journal of Nuclear Cardiology
PET and glucose metabolism imaging November/December 2006

Table 7. FDG cardiac PET: Acquisition guidelines (for dedicated, multicrystal PET scanner)
For details, see
Feature Technique note in text

Dose 5-15 mCi (185-555 MBq) Standard 1

Image start time 45-60 min after injection (keep constant for repeat 2
studies)
Image duration 10-30 min (depending on count rate and dose) 3
Acquisition modes 2D Standard 4
3D Standard 4
Static Standard 4
Dynamic Optional 4
Total counts Knowledge of machine performance characteristics 5
(eg, noise-equivalent counts) is essential
Pixel size (reconstructed) 2-3 mm Preferred 6
4-5 mm Acceptable 6
Attenuation correction Measured attenuation correction: before or Preferred 7
immediately after scan
Segmented attenuation correction Acceptable 7
Reconstruction method FBP or iterative expectation maximization (eg, OSEM) Standard 8
Gating ECG gating of myocardium Standard 9
Patient positioning Arms out Preferred 10
Arms in (not moving) Optional 10

because of the absence or difference in the type of should be encouraged to void frequently for 3 to 4
attenuation correction for SPECT, as well as the usual hours after the study.
registration problems when comparing images on differ- 2. Wait a minimum of 45 minutes before starting the
ent instruments. Guidelines for SPECT perfusion imag- static scan. Uptake may continue to increase and
ing have been published previously51 and are being blood pool to decrease as time progresses, even after
updated concurrently. It should be noted that if Tl-201 or 45 minutes. Longer than 90 minutes may give better
Tc-99m SPECT perfusion scanning has been performed, blood pool clearance and uptake, when necessary
no waiting period is necessary (from an instrumentation (eg, diabetic or high-BG subjects), but could result
point of view) before the PET scanning is begun if the in reduced count rate. If a follow-up FDG PET study
2D acquisition mode is used. The photons from Tl-201 is envisioned, it is important to duplicate the timing
and Tc-99m do not interfere. Caution needs to be used of the scan. Because FDG uptake is time-dependent
with 3D imaging, since the Tc-99m activity can increase (ie, it is possible that uptake may continue beyond
dead time and thus decrease the “true” counts from the 60 minutes), comparing two scans acquired at dif-
FDG. On the other hand, after administration of a PET ferent post-injection acquisition times can be mis-
tracer, it is usually necessary to wait at least 15 or more leading.
half-lives (depending on dose) before a low-energy (eg, 3. Duration is typically 10 to 30 minutes. If acquired
Tl-201 or Tc-99m) scan is performed. This is because the in 3D mode (ie, septa-out), compared with 2D
511-keV photons from the PET tracers easily penetrate mode with the same machine, a smaller dose is
the collimators most commonly used for Tl-201 or typically required to achieve the same total count
Tc-99m imaging. rate, but the imaging time may or may not be
reduced, as a result of count rate limitations and
increased scatter. In some machines, beyond a
Notes for Table 7: Image Acquisition
certain dose, septa-out mode (3D) will actually
1. Dose. Typically, 5 to 15 mCi is injected in a produce poorer-quality images for the same dose
peripheral vein (see counts requirements below). and imaging time than septa-in (2D) mode. For
Injection speed is not critical (bolus to 2 minutes). this reason, it is critical to have fully characterized
To reduce patient dose to the bladder, patients the performance of the system.
Journal of Nuclear Cardiology Machac et al e135
Volume 13, Number 6;e121-51 PET and glucose metabolism imaging

4. Static versus dynamic acquisition: Static acquisition roughly 8 mm/3 ⫽ 2.7 mm/pixel. Many institutions
produces images that allow relative quantification of achieve a 3-mm sampling rate or better with a 256 ⫻
FDG uptake on a regional basis. Such images (along 256 array over the entire field of view of the camera.
with perfusion images) are the standard basis for Other institutions choose to use a 128 ⫻ 128 array over
making viability determinations.1-3 However, there a limited field of view (eg, 25-35 cm diameter) cen-
is one form of dynamic imaging that has a significant tered over the heart, in which case, 2 to 3 mm/pixel is
practical advantage. Consider what is normally the easy to achieve (cutting out extraneous structures in
10- to 30-minute duration static scan, begun around the field of view) even with a 128 ⫻ 128 array.
60 minutes after injection. It is clinically desirable to Either method is acceptable to achieve the desired 2
acquire these data as a 3-frame or greater dynamic to 3 mm/pixel. Greater than 3 mm/pixel may be
data set. If the patient should move during the end of acceptable for older PET cameras with resolution
the study, one can then utilize only those dynamic worse than 1 cm.
frames with no motion (summing them together to 7. Since attenuation correction is a far more severe
make one static image). This is easily implemented problem in coincidence imaging than in SPECT
and takes almost no additional operator time. A imaging31; it is essential that accurate attenuation
more elaborate dynamic acquisition may optionally correction be performed. Segmented attenuation cor-
be used when FDG kinetic analysis over the entire rection schemes may give errors for those slices that
uptake period is to be performed (eg, compartmental contain a mixture of lung and liver tissue adjacent to
analysis or Patlak analysis). Kinetic analysis permits the heart. Similarly, the optimal CT-based attenua-
absolute quantification of the rate of FDG utilization. tion correction to be used to image the heart for
Performing and interpreting such kinetic analyses52 viability imaging will depend on the results of future
can be complex and require experience with kinetic research. Experience at several centers has shown
modeling. that a slow CT acquisition, lasting 10 to 60 seconds
5. The counts per slice necessary to yield adequate- during free breathing, at a low CT tube current, helps
quality images will vary from institution to institu- sample the average attenuation map, corresponding
tion depending on, among other things, scatter and to the emission map.
randoms corrections, as well as the amount of 8. FBP versus iterative reconstruction method: FBP is
smoothing that is done. If one tries to achieve on the the standard method used for reconstruction. FBP
order of 7 mm full width at half maximum (FWHM) images are subject to streak artifacts, especially
in-plane resolution and has 10% to 15% scatter when too short a transmission scan is used for
(National Electrical Manufacturers Association), attenuation correction (or when the subject is obese
then a typical good-quality study in 2D mode might or large). This can affect visual analysis but usually
have on the order of 50,000 true counts per milli- does not adversely affect quantitative analysis with
meter of transaxial distance over the region of the regions of interest (the streaks tend to average out
heart (eg, for a 4.25-mm slice separation, the counts properly over typical volumes of interest). Iterative
would be 50,000 ⫻ 4.25 ⫽ 250,000 counts per methods (eg, the method of OSEM) have been
slice). These numbers are very approximate and may adopted in other FDG imaging situations (eg, oncol-
differ from one scanner type to the next. With a ogy), yielding images with better noise properties.
10-mCi injected dose, these total counts could be Although high-uptake structures, such as the heart,
achieved in 20 to 30 minutes depending on system may not improve their noise characteristics with
sensitivity. If one is willing to accept a lower OSEM, the surrounding lower-uptake structures do
resolution (eg, more smoothing) or more noise, improve, and streak artifacts are nearly eliminated,
imaging time can be reduced. Low uptake and high thus greatly improving the visual appearance of the
blood pool activity situations (eg, diabetes or high image. However, low-uptake areas (such as myocar-
glucose levels) may require longer imaging time dial defects and the LV cavity at late times) may
and/or (preferably) later imaging times. Since 3D have slightly (artificially) elevated activity levels
scanners have greater scatter, they usually require unless sufficient iterations are performed. It is rec-
more counts than a 2D scanner to achieve the same ommended that one thoroughly characterize the PET
noise level. machine and its reconstruction algorithm’s behavior
6. It is recommended that 2 to 3 mm per pixel be used. with a realistic cardiac phantom.
A “rule of thumb” in nuclear medicine physics is 9. Usually, FDG PET counts are sufficiently large to
that one needs at least 3 pixels for every FWHM of yield a high-quality ventricular motion study (typi-
resolution in the image. For example, if the data are cally 8-16 time points), in a manner similar to
reconstructed to 8-mm FWHM, then one needs SPECT gated perfusion studies (but at higher spatial
e136 Machac et al Journal of Nuclear Cardiology
PET and glucose metabolism imaging November/December 2006

resolution). Given that ventricular contraction and ammonia, at rest and during pharmacologic or exercise
thickening are often clinically useful for assessing stress. Normal myocardial perfusion obtained during
viability, gating should be performed when possible. adequate stress implies absence of significant CAD.
It is important that the gating software does not Stress-induced regional myocardial perfusion abnormal-
adversely affect the ungated images (eg, by loss of ities or inadequate augmentation in perfusion with stress
counts as a result of beat length rejection). Monitor- implies epicardial CAD or possibly small-vessel disease.
ing the length and number of the accepted beats is Impaired regional myocardial perfusion during both
highly desirable. stress and rest suggests the presence of an irreversible
10. Ideally, the patient should be positioned supine, with myocardial injury.
the arms out of the camera field of view. This can be Both N-13 ammonia and, particularly because of its
tolerated by nearly all patients, provided some care short half-life, Rb-82 can be used sequentially with
is given to support of the arms or by use of an non-traditional maneuvers as well as traditional stress
overhead bar to hold onto. There are, however, cases testing to identify and localize myocardial areas with
in which “arms-out” imaging is not possible (eg, in ischemia and coronary dysfunction. This includes hand-
patients with severe arthritis), and imaging must be grip, cold-pressor testing, mental stress, nitroglycerin
performed with the arms at the side. In this case the drip, and intra-aortic balloon pump. These applications
transmission scan time may have to be increased, are still evolving and in the process of clinical evalua-
and it is of critical importance that the arms not tion.
move between transmission and emission, or arti- Myocardial metabolism imaging with PET serves to
facts will result. In patients undergoing PET/CT identify persistent metabolic activity in dysfunctional
imaging, arms in (down) the field of view may result and hypoperfused myocardial regions. It is accomplished
in beam-hardening artifacts on the CT-based trans- with F-18 FDG as a tracer of exogenous glucose utiliza-
mission scan, which usually lead to streak artifact of tion. The regional myocardial concentrations of this
the corrected emission scans. tracer are compared with the regional distribution of
myocardial perfusion: Regional increases in FDG uptake
PART 2: INTERPRETATION AND REPORTING relative to regional myocardial blood flow (perfusion-
metabolism mismatch) signify myocardial viability—
that is, reversibility of contractile dysfunction if regional
A. General Comments
blood flow is improved. In contrast, a regional reduction
The purpose of evaluating myocardial perfusion in FDG uptake in proportion to regional reductions in
during stress testing and at rest is to determine whether myocardial perfusion (perfusion-metabolism match) sig-
there is evidence of myocardial ischemia and/or infarc- nifies irreversibility of contractile dysfunction.
tion in those patients suspected of having CAD, as well The following guidelines for interpretation and re-
as to determine the extent and severity of coronary porting of PET myocardial perfusion and metabolism
disease and any dysfunction in both patients with known images assume that the studies were performed and
coronary disease and those with suspected coronary processed in accordance with the guidelines as described
disease. The goals of diagnosing CAD are to direct in Part 1. In instances where myocardial FDG uptake was
specific therapy to alleviate symptoms of ischemia, to imaged with SPECT, as is also outlined in Part 1 of these
initiate appropriate steps to modify its risk factors, to guidelines, the interpretation of the metabolic images
retard its progression, and to improve patient outcomes. must then consider technical and methodologic aspects
The diagnosis of CAD is also important in the selection unique to positron imaging with SPECT and method-
of patients who need to be studied with invasive diag- ologic differences between PET and SPECT. This also
nostic methods (eg, coronary angiography), those with a pertains to situations where metabolism has been imaged
high cardiac risk, those with a high need to know, and with PET and perfusion with SPECT. Special consider-
those with symptoms refractory to medical therapy. ation must also be given to studies when only images of
Evaluation of the severity of disease is critically impor- myocardial metabolism but not of myocardial perfusion
tant in the selection of those patients, in whom extensive have been acquired.
and/or severe disease suggests that invasive interven-
tional therapy is warranted by the possibility of improv-
B. Display of Perfusion and Metabolism
ing clinical outcomes. Finally, myocardial perfusion
imaging can be used for assessing the efficacy of medical i. Recommended medium for display. According
or surgical treatment. to the previously established guidelines for myocardial
Myocardial perfusion imaging with PET can be SPECT imaging, it is strongly recommended for the
accomplished with the flow tracers Rb-82 or N-13 reading physician to use the computer monitor rather
Journal of Nuclear Cardiology Machac et al e137
Volume 13, Number 6;e121-51 PET and glucose metabolism imaging

than film hard copies for interpretation of myocardial obtained on the perfusion study).53,54 Because glucose
perfusion and metabolism images. A linear gray scale, a metabolism may be enhanced in regions of normal
monochromatic color scale, or a multicolor scale can be perfusion, as might occur in postischemic stunning, this
used as the type of display, depending on user experience normalization approach allows for more accurate identi-
and preference. fication and interpretation of perfusion-metabolism mis-
ii. Conventional slice display of PET perfusion match or match patterns.
and metabolism images. Recommendations for display iii. Three-dimensional display. Display of the re-
of PET perfusion rest-stress and/or perfusion-metabo- constructed image data in a 3D static or cine mode is
lism images are consistent with those listed in previous optional and may be convenient for morphologic corre-
guidelines for SPECT myocardial rest-stress perfusion lation with magnetic resonance or angiographic images,
imaging. It is necessary to examine the transaxial, if suitable software is available. Currently, however, an
coronal, and sagittal views—that is, the not-yet-reori- advantage of 3D over conventional 2D displays with
ented images for assessing the alignment of the emission regard to accuracy of PET image interpretation remains
images acquired during different conditions, rest and to be demonstrated.
stress perfusion, and metabolism, as well as the trans- iv. Polar maps. Polar maps of PET rest and stress
mission images. Fused transmission and emission images perfusion images and FDG metabolic images are based
are preferred. If the images are not aligned because of on a circumferential profile of the corresponding short-
patient or cardiac motion, this may cause serious image axis slices, in the same way as polar maps of SPECT
artifacts, especially when only one set of attenuation myocardial perfusion images (see concurrently updated
correction images has been applied to all emission guidelines on myocardial SPECT perfusion images), and
images for attenuation correction. are commonly used for visual and semiquantitative
For interpretation, the reoriented images should be assessment. These derivative polar maps, while useful,
displayed as follows: should not be considered a substitute for the examination
of the standard short-axis and long-axis cardiac tomo-
1. A short-axis view, by slicing perpendicular to the
graphic slices.
long axis of the left ventricle from apex (left) to base
(right).
2. A vertical long-axis view, by slicing vertically from C. Evaluation of Images for Technical Sources of
septum (left) to lateral wall (right). Errors
3. A horizontal long-axis view, by slicing from the
Each laboratory should perform the daily QC rec-
inferior (left) to the anterior wall (right).
ommended by the manufacturer of the PET camera to be
For interpretation and comparison of perfusion and certain that the detector blocks are working appropri-
metabolism images, slices of all data sets should be ately. Normalization of the detectors (usually monthly)
displayed aligned and adjacent to each other. should also be performed according to the manufactur-
Software routines are available for handling of er’s recommendations.
image data from different camera systems and for a When FDG PET metabolic images are evaluated
combined display of both SPECT and PET images. with SPECT perfusion images, examination of the
Combined assessment of perfusion and metabolism SPECT images for artifacts should be performed as
within a single PET session offers the advantage that, recommended in the previously developed guidelines for
provided that the patient’s position has not changed SPECT myocardial perfusion imaging.
between the two acquisition periods, the ventricular long i. Patient motion. PET images are typically gener-
axis defined during image orientation can be copied from ated with nonmoving circular arrays of scintillation
one image set to the second set so that matching of the detectors that acquire all projection data. In contrast to
perfusion with the metabolism images is optimized. SPECT imaging with rotating gamma cameras, in which
Normalization of the stress perfusion with resting patient motion leads to a typical misalignment between
perfusion images is commonly performed by using the adjacent projection images and can be identified by
maximal myocardial pixel value in each of the two or viewing a projection movie, movement during static PET
three image sets (or, for example, the average pixel value imaging affects all projections and is therefore more
with the highest 5% of activity) of the perfusion images. difficult to identify. Blurring of contours results from
Each perfusion study is then normalized to its own substantial patient motion; it is most easily confirmed by
maximum. viewing non-reoriented transaxial images or the sino-
The metabolism images are normalized to the counts grams. Algorithms for motion correction during static
in the same myocardial region on the resting perfusion PET imaging have not become available. While repeat
images (eg, with the highest count rates that were imaging with FDG is feasible, and repeat imaging with
e138 Machac et al Journal of Nuclear Cardiology
PET and glucose metabolism imaging November/December 2006

rest Rb-82 is easily achieved with a second injection and careful inspection of the fused emission-transmission
acquisition, repeat imaging with stress is usually not an data sets. Fused images should be inspected in the axial
option. Therefore attention to patient motion during (lateral displacement) as well as the coronal (vertical
image acquisition is essential for avoidance of motion displacement) and sagittal (vertical displacement) slices.
artifacts. Patient positioning before and immediately Alternatively, displacement can be detected on transaxial
after image acquisition should be carefully evaluated (eg, images by counting the number of pixels by which the
by checking the alignment of the camera’s positioning cardiac image is displaced between resting and stress
laser beams with ink markers on the patient’s skin). transaxial acquisitions.
Acquisition of a brief scan or scout image after injection The degree of co-registration of transmission and
of a small dose, usually one third of the standard dose of emission images should be carefully examined using the
Rb-82 or N-13 ammonia, may facilitate accurate patient fusion software available on integrated PET-CT systems
positioning. With PET/CT systems, a CT scout scan to assess the reliability of images with attenuation
(10 mA) is routinely used for accurate patient position- correction. If there is patient motion and the cardiac
ing. In instances of patient discomfort and likely patient silhouette does not superimpose perfectly on the trans-
motion, especially during longer image acquisition times mission and emission images, the images without atten-
as with FDG, possible approaches to reducing adverse uation correction need to be examined as well. When the
effects of motion include acquisition of a series of 3 to 4 transmission maps are acquired using CT, the incidental
sequential image frames instead of a single static image findings in the portion of the chest in the field of view
of longer duration. If the quality of one of the serially should be reported when relevant to patient care.
acquired frames is compromised by motion, then that In general, vertical misalignment might be easier to
frame can be rejected and only frames that are of resolve by off-setting the alignment between the emis-
acceptable quality and are free of motion artifacts are sion and transmission scans, but this option is not
summed for the final image analysis. generally available.
ii. Attenuation correction. Correction of the emis- iii. Reconstruction artifacts. Image artifacts may
sion images for photon attenuation is critical for cardiac occur if extracardiac activity is present adjacent to the
PET imaging. Positron-emitting tracers are more sensi- myocardium. For example, intense focal activity in the
tive than single photons to attenuation artifacts. Both liver or the gastrointestinal tract may cause a reconstruc-
simultaneously emitted photons must be detected by the tion artifact with artifactually low count rates in adjacent
coincidence detection systems. As each of the two myocardium. A method to correct for such artifacts is
photons is susceptible to tissue attenuation, attenuation currently not available, but such artifacts are less prom-
artifacts are generally greater. Therefore only attenua- inent when iterative reconstruction is used instead of the
tion-corrected images should be used for clinical inter- standard FBP techniques. Streak artifacts may result
pretation. Potential sources of errors include misalign- from nonuniformity due to malfunctioning blocks of
ment between emission and transmission data resulting detector crystal. Streak artifacts result from problems
from patient motion. Misalignments of 1.5 to 2 cm, for with CT transmission images, such as insufficient x-ray
example, can lead to as much as a 30% change in the tube intensity in obese individuals, breathing mismatch,
observed regional myocardial radioactivity.55,56 truncation, and beam hardening resulting from bone
Vertical and transaxial displacement of the heart can (arms down) or metal adjacent to the heart (pacemakers
occur even in the absence of movement of the chest, and internal defibrillators).
perhaps because of a change in the pattern of breathing, iv. Image count statistics. The final count density
which may occur during pharmacologic stress. It is of PET images is influenced by additional factors such as
analogous to the “upward creep” seen in SPECT imag- body weight and build, radionuclide dose, scanner per-
ing. This can result in anterior or anterolateral defects formance, acquisition time, and in the case of metabolic
(usually resulting from under-correction due to the lower imaging, the dietary and hormonal state. Image count
attenuation coefficient of the overlapping lung tissue) or density directly affects the diagnostic quality and reli-
inferior “hot spots” (usually resulting from over-correc- ability of the study.
tion due to the higher attenuation coefficients of the
overlapping subdiaphragmatic tissues). Inspection of
D. Image Analysis and Interpretation of Perfusion
fused emission-transmission images for possible mis-
Images
alignment is essential because the resulting artifacts
would greatly affect image interpretation. The rest and stress perfusion images, if obtained,
Identification of vertical and lateral displacements and FDG images should be interpreted initially without
that result in misalignment between the emission and clinical information in order to minimize any bias in
transmission images is relatively straightforward upon study interpretation. All relevant clinical data should be
Journal of Nuclear Cardiology Machac et al e139
Volume 13, Number 6;e121-51 PET and glucose metabolism imaging

reviewed after a preliminary impression is formed. The ammonia administration or prior to 70 to 170 seconds
following sections address first the interpretation of the after Rb-82 administration, especially in patients with a
perfusion images and then the interpretation of the history of congestive heart failure and poor LV function.
metabolism images, if available. The latter section then If Rb-82 perfusion images are acquired serially, an
continues by describing the correlation of findings on the appropriate starting point, after blood pool activity has
metabolism image to those on the perfusion images. cleared sufficiently, can be chosen after the acquisition
i. Left and right ventricular size. The reader for summing the myocardial tracer uptake images. Judi-
should note whether there is an enlargement of the right cious adjustment of display threshold and contrast set-
or left ventricle at rest or whether there is transient tings can help offset this problem.
stress-related LV dilation if such images were obtained. v. Noncardiac findings. The tomographic images
Ventricular enlargement seen on the stress and rest should be carefully examined for uptake of the radio-
perfusion images, as well as on the FDG images, tracer in organs other than the myocardium, particularly
generally indicates left, right, or bi-ventricular dysfunc- in the lungs and the mediastinum. Extracardiac uptake of
tion. Transient stress-related LV dilation usually reflects a flow tracer may be of clinical significance, as it may be
extensive CAD. LV and right ventricular (RV) sizes as associated with malignancy and/or an inflammatory pro-
well as any changes associated with stress are typically cess. The 3D maximum intensity projection display, a
described qualitatively. It is important to be sure that method of displaying acquired PET images as a rotating 3D
Rb-82 blood pool activity has cleared from the LV display, can be particularly helpful in this regard. When
cavity, in order to be sure that there is not an artificial using PET/CT systems, review of the low-resolution CT-
underestimation of LV cavity at rest. A number of based transmission scans can be useful to delineate poten-
commercially developed software packages originally tially important ancillary findings (pleural and pericardial
developed for SPECT have the ability to quantitate mean effusion, coronary and/or aortic calcification, lung mass,
LV volumes and end-diastolic and end-systolic volumes
and so on).
for gated PET images, but not all such packages have
vi. Perfusion defect location. Myocardial perfusion
been validated for all PET instruments, and they should
defects should be identified through careful visual anal-
be used with caution.
ysis of the reoriented myocardial slices. Perfusion de-
ii. Lung uptake. Increased tracer activity in the
fects should be characterized by extent and location
lungs should be reported qualitatively. Increased lung
relative to the specific myocardial territory, such as the
uptake on the perfusion images, particularly when se-
anterior, lateral, inferior, septal, and/or apical walls.
vere, may reflect severe LV dysfunction with increased
Standardized nomenclature should be used, according to
LV filling pressures. Mild to moderate lung uptake may
occasionally be seen in patients with chronic pulmonary previously published guidelines relating to SPECT per-
disease and in smokers. fusion imaging. RV defects due to scarring and ischemia
iii. RV uptake. The degree of RV tracer uptake should be noted.
should also be noted. It is usually described qualitatively. vii. Perfusion defect severity and extent: Quali-
Increases may be seen in the presence of pulmonary tative. Defect extent may be qualitatively described as
hypertension with or without significant RV hypertro- small (5% to 10% of the left ventricle), medium (15% to
phy. RV hypertrophy may alter the shape of the left 20% of the left ventricle), or large (⬎20% of the left
ventricle and produce both “hot” and “cold” areas in the ventricle). Alternatively, defect extent may also be esti-
septum and adjacent LV areas. However, no definite mated by describing the segments involved, such as “mid
criteria have been established for defining abnormally and distal” or “extending from base to apex,” in a
increased RV tracer uptake. particular LV area. Defect severity is typically expressed
iv. Blood pool activity. Increased blood pool activ- qualitatively as mild, moderate, or severe. Severe defects
ity is usually a sign of relatively poor myocardial uptake may be considered as those having a tracer concentration
of the flow tracer, insufficient time for tracer uptake into equal or similar to background activity, and moderate
the myocardium, or diminished clearance from blood. A defects are considered definitely abnormal but less se-
major cause of increased blood pool activity, especially vere. Mild defects are those with a subtle but definite
for perfusion imaging with Rb-82, is impairment of reduction in regional myocardial tracer concentration
cardiac systolic function that prolongs the circulation that are, however, of uncertain clinical significance.
time. This is especially relevant when only static images Perfusion defects that remain unchanged in extent and
are acquired, because vasodilators typically increase severity between the stress and rest images are typically
cardiac output and shorten the circulation time. Increased categorized as fixed or nonreversible. In contrast, defects
blood pool activity may be seen if the acquisition of that decrease in extent, in severity, or both from stress to
images begins prior to 5 to 7 minutes after N-13 rest are categorized as reversible, in which case a
e140 Machac et al Journal of Nuclear Cardiology
PET and glucose metabolism imaging November/December 2006

Table 8. Semiquantitative scoring system of defect

severity and extent
Grade Interpretation Score

Normal counts Normal perfusion 0

Mild reduction in Not definitely abnormal 1
counts
Moderate reduction Definitely abnormal 2
in counts
Severe reduction in Definitely abnormal 3
counts
Absent counts Definitely abnormal 4

example, it is not at all uncommon to find segments 9,

Figure 1. Seventeen-segment nomenclature. Segments 1, 7,
and 13 represent the basal (1), mid (7), and apical (13) anterior 10, and 15 of the 17-segment model involved in left
segments. Segments 4, 10, and 15 represent the basal (4), mid anterior descending artery disease. Similarly, segments 5
(10), and apical (15) inferior segments. The septum contains 5 and 11 of the model may be affected by disease of the
segments: basal anteroseptal (2), basal inferoseptal (3), mid right coronary artery.
anteroseptal (8), mid inferoseptal (9), and apical septal (14).
Semiquantitative scoring system. The scoring sys-
The lateral wall is divided into the basal anterolateral (6), basal
inferolateral (5), mid anterolateral (12), mid inferolateral (11), tem provides a reproducible semiquantitative assessment
and apical lateral (16). The long-axis segment is called the apex of defect severity and extent (Table 8). A consistent
(17). (Reprinted with permission from American Society of approach to defining defect severity and extent is clini-
Nuclear Cardiology. Imaging guidelines for nuclear cardiology cally important because each of the two variables con-
procedures, part 2. J Nucl Cardiol 1999;6:G47-84.)
tains independent prognostic power. Furthermore, semi-
quantitative scoring can be used to more reproducibly
description of the degree of defect reversibility is re- and objectively designate each segment as normal or
quired (see below). abnormal. Points are assigned to each segment in direct
viii. Perfusion defect severity and extent: Semi- proportion to the observed count density of the segment.
quantitative. In addition to the qualitative assessment of In addition to individual scores, calculation of
perfusion defects, the reader may also apply semiquan- summed scores is recommended, in which the summed
titative approaches based on a validated segmental scor- stress score is the sum of the stress scores of
ing system. This approach standardizes the visual inter- all segments, the summed rest score is the sum of the
pretation of scans, reduces the likelihood of overlooking resting scores or redistribution scores of all segments,
clinically significant defects, and provides a semiquanti- and the summed difference score is the difference be-
tative index that is applicable to diagnostic and prognos- tween the summed stress and summed rest (redistribu-
tic assessments. tion) scores and serves as a measure of reversibility.
A 17-segment model for semiquantitative visual Before scoring, it is necessary for the interpreting
analysis is usually employed. The model is based on physician to be familiar with the normal regional varia-
three short-axis slices (apical, mid, and basal) to repre- tion in count distribution of myocardial perfusion PET.
sent most of the left ventricle and one vertical long-axis For example, retention of N-13 ammonia in the lateral
slice to better represent the LV apex. The basal and mid and basal inferolateral wall of the left ventricle may be
short-axis slices are divided into 6 segments. The apical diminished in normal subjects. In severity, this normal
short-axis slice is divided into 4 segments. A single variant reduction may be mild to moderate and should be
apical segment is taken from the vertical long-axis slice. considered when analyzing rest or stress N-13 ammonia
Each segment has a specific name (Figure 1). images. No such regional variation in tracer uptake has
The extent of stress and rest perfusion abnormalities, been reported for Rb-82, except for a mild reduction in
as well as an estimate of the extent of scarring and the apex and base, consistent with the known thinning of
ischemia, can be performed by counting the number of the LV myocardium in these locations.
segments. ix. Perfusion defect severity and extent: Quan-
Myocardial segments may be assigned to coronary titative. Quantitative analysis aids in supplementing
artery territories. Caution should be exercised because the visual interpretation. Most quantitative analyses
the coronary anatomy varies widely among patients. For require display of the tomographic slices in a polar map
Journal of Nuclear Cardiology Machac et al e141
Volume 13, Number 6;e121-51 PET and glucose metabolism imaging

format that provides an easily comprehensible represen- rest, it can scatter into the subendocardial defect to make
tation of the extent, severity, and reversibility of perfu- the resting scan appear artificially normal and the defect
sion abnormalities.53,57,58 The patient’s polar map is appear artificially reversible. Adjustment of display set-
compared with a reference polar map derived from tings and contrast can partly help to adjust for this
normal individuals. Because of the differences in tracer discrepancy. Complete reversibility is present when the
distributions throughout the LV myocardium, the use of tracer activity within the defect returns to a level com-
separate databases for N-13 ammonia and Rb-82 is parable to the surrounding normal myocardium. Because
recommended. Each camera system and acquisition pro- the term partial often includes segments with a wide
tocol should have its own “normal” file. The quantitative range of reversibility, it is recommended that the terms
estimates should appropriately sample all segments of moderate or significant but not complete be used instead
the left ventricle. The quantitative analysis system should to describe the degree of change in tracer activity from
be validated by appropriate studies published in peer- stress to rest. The semiquantitative scoring system may
reviewed journals. Such systems typically describe the be used to define significant reversibility as improvement
extent of a perfusion defect in terms of percentage of LV of 2 grades or more or as improvement to a score of 1.
myocardium. These analysis systems further offer an The degree of reversibility on a quantitative polar map
index of the average severity of a perfusion defect. will depend on the particular software routine in use and
Myocardial blood flow in absolute units of milli- the normal reference databases used in that program, but
liters of blood per minute per grams of myocardium. it is more objective.
A quantitative estimate of myocardial blood flow can xi. Gated myocardial perfusion PET images. ECG
complement the visual interpretation. Estimating myo- gating of the rest and peak stress images can provide
cardial blood flow in absolute units requires acquisition additional information regarding changes in LV function
of dynamic imaging data sets.7,59-61 Regions of interest
and volumes that may be useful in identifying 3-vessel
are placed on the LV myocardium and the LV blood pool
CAD with or without left main disease, which may be
and are copied to all serially acquired images for gener-
underestimated on the review of the perfusion images.
ation of myocardial tissue and blood pool time-activity
Unlike ECG gating of the post-stress SPECT images,
curves. The time-activity curves are corrected for activity
PET acquisitions take place during peak stress, espe-
spillover from the blood pool to the myocardium and for
cially when using ultra–short-lived tracers like Rb-82
radioactive decay. They are then fitted with a validated
(acquisitions are shorter than those for N-13 ammonia
tracer kinetic model, and estimates of myocardial blood
and, thus, more likely to occur while the patient is in
flow in milliliters of blood per minute per grams of
peak pharmacologic stress). Recent data suggest that
myocardium are obtained. Software programs are also
available for generating parametric polar maps that estimates of LVEF during peak stress and their change
display regional myocardial blood flows in absolute from baseline correlate with the magnitude of ischemia
units.6,62 on the perfusion images and are useful measures to
Quantitative blood flow approaches offer an objec- identify balanced ischemia resulting from extensive ob-
tive interpretation that is inherently more reproducible structive CAD.63
than visual analysis. Absolute quantitation may aid in
assessing the physiologic significance of known coro- E. Interpretation of FDG Metabolism Images
nary stenoses, especially when of intermediate severity.
Both relative quantitation and absolute quantitation are i. Noncardiopulmonary findings. The tomographic
particularly useful in describing changes between two images should be carefully examined for uptake of FDG
studies in the same patient. In addition, quantitative in organs other than the heart, particularly the lungs and
measurements of myocardial blood flow may aid in the mediastinum. Extracardiac uptake of FDG may be of
detection of balanced reductions in myocardial blood clinical significance as it may represent malignancy
flow due to multivessel CAD and, thus, may be useful for and/or inflammatory disease. FDG may be taken up in
assessing the true extent of disease. Quantitative assess- atherosclerotic plaques visible in the aorta and other
ment of myocardial blood flow at rest and stress can help larger arteries. The 3D rotating maximum intensity
detect an inadequate response to stress, resulting in projection display can be particularly helpful. If focal
decreased sensitivity for regional epicardial CAD, or can areas of FDG uptake are visualized outside of the heart,
help detect the presence of diffuse small-vessel disease. a whole-body PET acquisition may be useful for defining
x. Reversibility of flow defects. The degree of areas of abnormality more completely. One should note,
defect reversibility may be categorized as complete or however, that whole-body, non-fasting, glucose-loaded
partial. It is important to note the Rb-82 clearance from and insulin-supplemented FDG images may demonstrate
the LV blood pool, because if there is excess Rb-82 at different distribution than in the fasting state, featuring
e142 Machac et al Journal of Nuclear Cardiology
PET and glucose metabolism imaging November/December 2006

higher uptake in skeletal muscles and possibly lower Table 9. Semiquantitative scoring system of
uptake in tumors. distribution and extent of viability
ii. Blood pool activity. Increased blood pool activity
is usually a sign of diminished clearance of radiotracer Grade Interpretation* Score
from blood into tissue and into the myocardium. It is
important to note this finding while the patient is still in Highest counts Normal or ischemia 0
the scanner, while it is easiest to take corrective mea- Mildly lower counts 1
sures. This increase in blood pool activity may be caused Moderately lower counts 2
by the following: Severely reduced counts 3
Absent counts 4
1. Inadequate patient preparation with high free fatty
acid and low insulin plasma levels, as occur, for *This scoring system applies to studies performed in the glucose-
loaded state and high FDG uptake in normal (“best” perfused)
example, after fasting and without glucose loading, so myocardium. The same scoring system may also be used for
that tracer clearance from blood is delayed. studies with lower FDG uptake in normal myocardium; the descrip-
2. States of insulin resistance (insulin resistance, im- tion of the findings should then, however, indicate the segments
that are considered normal as judged from the perfusion images.
paired glucose tolerance, and type 2 diabetes mel-
litus).
3. Type 1 diabetes mellitus.
tion of insulin and thereby shifts the myocardium’s sub-
4. Elevated plasma glucose concentrations with an in-
strate utilization from free fatty acid to glucose. This shift in
crease in the distribution volume for FDG.
substrate utilization is associated with increased FDG up-
5. Insufficient time for clearance of tracer from blood
take both in normally perfused and in dysfunctional myo-
into tissue before image acquisition.
cardium. Therefore, FDG uptake may be highest in the
High blood pool activity reduces the image contrast “best”-perfused myocardium. The extent and location of
and, thus, the diagnostic quality of the images. Blood pool regional myocardial reductions in FDG uptake should be
activity frequently declines with time so that relative myo- analyzed and described analogous to the qualitative descrip-
cardial FDG uptake increases. Repeat imaging after an tion of perfusion defect severity and extent as described
additional 30 to 60 minutes may improve the myocardium– above.
to– blood pool contrast but at the expense of lower count iv. Metabolism: Distribution and extent of viabili-
densities because of radioactivity decay. The image quality ty—Semiquantitative. Analogous to the semiquantitative
can be improved and controlled by careful patient prepara- evaluation of the perfusion images, the use of a scoring
tion as detailed in Part 1 of these guidelines. It is also system provides a reproducible semiquantitative assessment
possible to accelerate the clearance of FDG from blood and of the extent and severity of regional alterations in FDG
its uptake into myocardium by IV administration of regular uptake. Such semiquantitative scoring can be used to more
short-acting insulin while carefully monitoring BG levels to reproducibly and objectively designate segments as meta-
avoid hypoglycemia.44,46 bolically viable or nonviable. Scores are assigned to each
iii. Myocardial FDG uptake: Distribution and ex- segment in direct proportion to the observed count density
tent of viability—Qualitative. After inspection of the of the segment (Table 9).
transaxial myocardial views for movement artifacts, myo- Before scoring, it is necessary for the interpreting
cardial metabolism should first be evaluated through visual physician to be familiar with the normal regional variation
analysis of the reoriented myocardial image slices. An in count distribution of myocardial metabolism, as well as
initial determination of the degree of FDG uptake in the with the dietary condition of the patient at the time of FDG
“best”-perfused myocardial territory aids in recognizing imaging. Importantly, FDG concentrations in the normal
whether images were acquired in the fasting state or after a myocardium may be increased in the lateral and posterolat-
glucose load. In the fasting state, normal (ie, “best”- eral wall.42,66 This normal variant is observed especially
perfused) myocardium will show minimal or low FDG when metabolism is imaged in the fasting state and FDG
uptake because of its preferential fatty acid uptake.42,64,65 uptake is low in the interventricular septum and the anterior
This “hot spot” imaging condition may be difficult to wall but increased in the lateral and posterolateral wall.
interpret because the image normalization process further The interpretation should include a description of the
augments the regional FDG accumulation, which may probability for the potential for improvement in regional
prompt an overestimation of the actual degree of a perfu- contractile function and in LVEF. The potential for a
sion metabolism mismatch and thus lead to an overestima- post-revascularization improvement in contractile function
tion of myocardial viability. On images obtained after is low for perfusion-metabolism matches, even if the
glucose loading, either by mouth or intravenously, the regional reductions in perfusion and in FDG uptake are only
increase in plasma glucose concentrations prompts secre- mild or moderate. Conversely, the potential for improve-
Journal of Nuclear Cardiology Machac et al e143
Volume 13, Number 6;e121-51 PET and glucose metabolism imaging

ments in regional contractile dysfunction is high if perfu- tion of the count difference between regional FDG and flow
sion is normal, if both perfusion and FDG uptake are tracer uptake (usually expressed as counts in defect divided
normal, or if FDG uptake is significantly greater than by counts in normal or maximal count zone) is preferable to
regional perfusion (mismatch). Finally, the potential of a determinations of ratios of regional FDG and flow tracer
post-revascularization improvement in LVEF by at least 5 uptake (ie, FDG/Rb-82 ratios for each pixel). Such ratios
or more EF units is high if the mismatch affects 20% or are susceptible to regional statistical variations in count
more of the LV myocardium.44,67,68 densities and to image noise.
NOTE: Comments by the Working Group regarding Myocardial glucose utilization in absolute units of
interpretation of metabolism images in early post-infarction micromoles of glucose per minute per grams of myo-
patients: It has been the experience of some members of the cardium. Quantitative estimates of myocardial glucose
Working Group that in early post-infarction patients (1-2 utilization have not been found to aid in the assessment or
weeks), overestimation of FDG uptake may occur. This characterization of myocardial viability due to the variabil-
may be related to intense glycolytic activity of leukocytes in ity in substrate utilization by the myocardium, even when
the necrotic zones of the myocardial infarction. Alterna- FDG images are acquired during a hyperinsulinemic-eugly-
tively, FDG uptake may be lower than perfusion (reverse cemic clamp.42,45,46 Methods for deriving quantitative esti-
mismatch) especially in early post-infarction patients after mates of myocardial metabolism require acquisition of
thrombolysis. This may lower the predictive value of PET serial images.36,41 Regions of interest are placed on the
perfusion-metabolism imaging after an acute myocardial myocardium and the LV blood pool and are copied to all
infarction and should be stated accordingly in the report. serially acquired images in order to generate myocardial
RV uptake. When tracer activity concentrations are tissue and blood pool time-activity curves. The time-
increased in the RV myocardium, they should be reported. activity curves are corrected for spillover of activity from
The normal RV myocardial activity concentrations of FDG the blood pool into the myocardium and for radioactive
images are about 50% of those of the LV myocardium. decay. The time-activity curves are then fitted with a
Increased RV uptake is indicative of RV hypertrophy or validated tracer kinetic model, and estimates of regional
overload, usually as a result of pulmonary hypertension. On myocardial glucose utilization are obtained in micromoles
the other hand, an apparent increase in RV activity may of glucose per minute per grams of myocardium. Measure-
only be relative because of diminished LV uptake of FDG. ments of glucose metabolic rates further require determina-
However, little if any information on the significance of tion of glucose concentrations in arterial or arterialized
increased RV FDG uptake either associated with a concor- venous blood. Similar to myocardial perfusion, parametric
dant increase in perfusion or in excess of RV perfusion is images and polar maps are also available for display of rates
available, so the clinical relevance of such uptake patterns of regional myocardial glucose utilization. Regional meta-
remains uncertain at present. bolic rates on such parametric images are coded by a color
v. Metabolism and extent of viability: Quantitative. scale and can be determined noninvasively for any myo-
Quantitative analysis can supplement the visual interpreta- cardial region through regions of interest assigned to the
tion. As with the perfusion images, quantitative analysis of polar map.69
myocardial metabolism requires display of the tomographic vi. Comparison of myocardial metabolism to per-
short-axis slices in a polar map format that provides an fusion. The comparison of perfusion and metabolism im-
easily comprehensible representation of the extent and ages obtained with PET is relatively straightforward be-
severity of perfusion and metabolic abnormalities, as well cause both image sets are attenuation-corrected. Thus,
as their relationship.54,55 The “raw” perfusion and FDG relative increases or decreases in myocardial metabolism
polar maps are normalized to myocardial regions with the relative to perfusion generally reflect the presence or ab-
highest perfusion. To identify myocardial regions with sence of metabolic viability.
increases in FDG uptake relative to perfusion, the FDG Special considerations for combining SPECT per-
polar maps are then compared with the perfusion map, fusion with PET metabolism images. In current clinical
resulting in a “difference” polar map. The patient’s polar practice, FDG PET images are often read in combination
maps (perfusion, metabolism, and their difference) are then with SPECT myocardial perfusion images. The interpreting
compared with corresponding reference polar maps of physician should be careful when comparing the non–
normal subjects. Depending on the degree of FDG uptake, attenuation-corrected SPECT images with attenuation-cor-
hypoperfused regions are then categorized into normal, rected FDG PET images. Myocardial regions showing an
perfusion-metabolism mismatch, or perfusion-metabolism excessive reduction in tracer concentration as a result of
match. The quantitative measurements should appropriately attenuation artifacts, such as the inferior wall or the anterior
sample all segments of the left ventricle. The quantitative wall in female subjects, may be interpreted as perfusion-
analysis system should be validated by appropriate studies metabolism mismatches, resulting in falsely positive perfu-
published in peer-reviewed journals. Generally, determina- sion-metabolism mismatches. Two approaches have proved
e144 Machac et al Journal of Nuclear Cardiology
PET and glucose metabolism imaging November/December 2006

useful for overcoming this limitation. First, because assess- FDG images. The separate adjustment of threshold and
ment of viability is relevant only in myocardium with contrast settings can help compensate for these discrepan-
regional contractile dysfunction, gated SPECT or PET cies.
images offer means for determining whether apparent Four distinct resting perfusion-metabolism patterns
perfusion defects are associated with abnormal regional may be observed in dysfunctional myocardium44,64,68,73-90:
wall motion. Second, quantitative analysis with polar map
1. Normal blood flow associated with normal FDG
displays that are compared with tracer- and gender-specific
uptake.
databases (for SPECT images) may be a useful aid to the
2. Reduced blood flow associated with preserved or en-
visual interpretation. Neither approach is foolproof, so the
hanced FDG uptake (perfusion-metabolism mismatch).
use of attenuation-corrected PET images acquired on the
3. Proportionally reduced blood flow and FDG uptake
same instrument to minimize these problems is clearly
(perfusion-metabolism match).
preferable.
4. Normal or near-normal blood flow with reduced FDG
For myocardial FDG images acquired with SPECT
uptake (reversed perfusion-metabolism mismatch).
equipped with ultrahigh-energy collimators or with
SPECT-like coincidence detection systems, additional Some laboratories have added a fifth pattern, a mild
problems may be encountered especially when the images perfusion-metabolism match in which the regional uptake
are not corrected for photon attenuation.70-72 Myocardial of both the tracer of blood flow and of FDG is mildly to
regions with severely reduced tracer activity concentrations moderately reduced.80,84 Because contractile function in
due to attenuation artifacts on both perfusion and metabo- such “mild” matches generally does not improve after
lism imaging, such as the inferior wall in men or the revascularization, the pattern is subsequently included in the
anterior wall in women, may be interpreted erroneously as general category of perfusion-metabolism matches. The
perfusion-metabolism matches. Attenuation of the high- patterns of abnormal perfusion-metabolism and of perfu-
energy 511-keV photons is less than that for the 140-keV sion-metabolism mismatch identify potentially reversible
photons of Tc-99m or the 60- to 80-keV photons of Tl-201 myocardial dysfunction, whereas the pattern of a perfusion-
so that attenuation artifacts are less prominent for FDG metabolism match identifies irreversible myocardial dys-
images and may result in an apparent mismatch. Further- function.
more, the lower spatial resolution of SPECT imaging If stress and rest perfusion imaging information is
systems for FDG imaging, especially when using high- available, it is useful to add an estimate of the extent of
energy photon collimation and then comparing with Tc- stress-inducible ischemia in regions of normal resting per-
99m or Tl-201 images, causes apparent mismatches for fusion and FDG uptake, in regions with matched resting
small defects, at the base of the left ventricle, or at the edges perfusion-FDG defects, or in regions with resting perfusion
or borders of large perfusion defects. Such artifacts result- FDG-metabolic mismatch. The simultaneous display of
ing from the use of different photon energies can be avoided stress and rest perfusion and FDG metabolic images is most
by using dedicated PET systems for both perfusion and helpful but not available on all display workstations. In
metabolism imaging. Again, quantitative analysis through circumstances where only resting perfusion imaging is
polar map displays with comparison to radiotracer- and performed alongside FDG metabolic imaging, besides re-
gender-specific databases of normal may aid in the visual porting on the extent of scar and extent of hibernating
interpretation. myocardium, it is useful to indicate that one cannot rule out
vii. Integration of perfusion and metabolism re- stress-inducible ischemia.
sults. The combined evaluation of regional myocardial In circumstances where only stress perfusion imag-
perfusion and FDG metabolism images allows identifica- ing is available in combination with FDG metabolic
tion of specific flow-metabolism patterns that are useful to imaging, the following patterns can be found in segments
differentiate viable from nonviable myocardium. It is useful with contractile dysfunction:
to start with a functional assessment, ideally from gated
1. Stress perfusion defect with preserved FDG uptake
PET or SPECT imaging, as dysfunctional segments are
indicates ischemically jeopardized tissue (but distinction
those suitable for evaluation of myocardial viability. If
between ischemia, stunning, and hibernation is no
stress perfusion images as well as resting perfusion images
longer possible), but revascularization is needed any-
are available, jeopardized myocardium can be distinguished
way.
from normal myocardium, and myocardium perfused nor-
2. Stress perfusion defects without preserved FDG uptake
mally at rest but dysfunctional as a result of repetitive
indicate scar tissue, and revascularization is not needed.
stunning can be distinguished from myopathic or remod-
eled myocardium. The interpretation should include a description of the
Differences in blood pool concentration of tracers can probability for the potential for improvement in regional
impact on the apparent match or mismatch of perfusion- contractile function and in LVEF after revascularization.
Journal of Nuclear Cardiology Machac et al e145
Volume 13, Number 6;e121-51 PET and glucose metabolism imaging

The potential for a post-revascularization improvement in a mismatch or the difference score serves as a predictor of
contractile function is low for perfusion-metabolism the degree of functional improvement.
matches, even if the regional reductions in perfusion and in Reverse mismatch. A pattern of normal perfusion
FDG uptake are only mild or moderate. Conversely, the associated with reduced FDG uptake may be observed.90-93
potential for improvements in regional contractile dysfunc- In most circumstances this results from a lack of or
tion is high if perfusion is normal, if both perfusion and incomplete normalization of the myocardial FDG uptake. It
FDG uptake are normal, or if FDG uptake is significantly has also been observed in patients with multivessel CAD, in
greater than regional perfusion (mismatch). Finally, the patients with left bundle branch block, in patients after
potential of a post-revascularization improvement in the coronary angioplasty and thrombolysis, and in patients with
LVEF by at least 5 or more EF units is high if the mismatch diabetes. When observed, it is important to consider possi-
affects 20% or more of the LV myocardium.44,68 ble technical explanations for this pattern, such as inade-
Qualitative or semiquantitative approaches can be ap- quate or insufficient normalization of the image data. An
plied to the interpretation of perfusion-metabolism patterns. important consideration for the interpretation of the “true”
Consistent with the qualitative analysis of myocardial per- reverse mismatch is that perfusion is generally normal,
fusion defects, as described under paragraphs 14 and 15, suggesting that such regions, when dysfunctional, are via-
concordant regional reductions in perfusion tracer and in ble. Although comments on the presence of such reverse
FDG uptake should be defined as matches. Their extent mismatch may be made in the report, it is generally not
may be small (5%-10% of the left ventricle), moderate advisable to include its relevance in the final impression, as
(15%-20% of the left ventricle), or large (⬎20% of the left it might lead to confusion.
ventricle). The severity of a match is similarly expressed as Interpretation of FDG images when perfusion im-
mild, moderate, or severe. Regional uptake of FDG in ages have not been obtained. Interpretation of FDG
excess of the reduced perfusion in the same myocardial images without perfusion images and/or angiographic in-
region is interpreted as a mismatch. Depending on the formation and/or without information on regional wall
difference between the uptakes of both tracers, the magni- motion can be difficult and unreliable. This approach is not
tude of a mismatch can be defined qualitatively as low, recommended for several reasons:
moderate, and high. Important for relating the FDG to the
perfusion images is to define the myocardial region or 1. The presence of relatively well-preserved FDG up-
segment that is normal. It is identified best on the perfusion take in dysfunctional myocardium does not differen-
images as the myocardium with the highest perfusion tracer tiate an ischemic from a nonischemic process.
uptake. Depending on the study conditions and the dietary 2. Regional absence and, conversely, regional increases
state, the uptake of FDG in this region may be highest but in FDG uptake alone do not distinguish between
also may be diminished or absent. normal and irreversibly injured myocardium. For
Semiquantitative assessments of perfusion and metab- example, uptake would be low in normal myocardium
olism patterns may also be helpful. Similar to the scoring of but increased in dysfunctional myocardium when
perfusion defects and stress-rest perfusion differences, nu- studies are performed after more than 6 hours of
merical scores can be applied to matches and mismatches. fasting. Conversely, in the glucose-loaded state, up-
Identical scores for perfusion and FDG uptakes are consis- take is highest in normal myocardium but frequently
tent with matches. In contrast, at least one score difference reduced in dysfunctional myocardium.
in scores between the regional FDG uptake and perfusion 3. The degree of FDG accumulation over and above
tracer uptake should be considered as a mismatch. Again, it regional perfusion helps to assess the relative amount
is important for the scoring approach to use the perfusion of scar and metabolically viable myocardium. The
images for defining the normal myocardium. The segmental latter information may significantly influence the
scores on the FDG images should be normalized to the power of the test for predicting functional recovery. It
perfusion scores (ie, the scores need to be adjusted so that is therefore recommended that FDG metabolic images
scores in normal myocardium will be 0). The normalization be analyzed in combination with perfusion images
approach and calculation of difference scores are depicted obtained either with SPECT or, preferably, with PET.
in Figure 2.
Critical for prediction of the potential for an improve-
F. Gated Myocardial FDG PET Images
ment in global LV function is the number of segments with
scores of mismatches. On the basis of the 17-segment Different from gated myocardial perfusion imaging
model, four or more mismatch segments (or about 20% of with SPECT, the clinical usefulness of gated FDG PET is
the left ventricle) might be considered as predictive for a less firmly established, although parameters of global and
potential improvement in LV function after revasculariza- regional function derived from gated PET have been found
tion. Current evidence has not verified that the magnitude of to correlate closely with those obtained through standard
e146 Machac et al Journal of Nuclear Cardiology
PET and glucose metabolism imaging November/December 2006

Figure 2. Semiquantitative scoring of metabolism images relative to perfusion images. The

segmental perfusion and metabolism tracer uptake is shown schematically on a short-axis slice.
Examples of two conditions are schematically depicted on short-axis slices, one with FDG uptake
being highest in the normally perfused myocardium and one with FDG uptake being highest in the
hypoperfused myocardium. As seen in the upper panel, FDG uptake on the metabolism No. 1 image
in segments with normal perfusion scores (scores of 0) in the top left panel is scored with a 0,
reflecting the highest uptake. The perfusion scores in the inferior wall of 1 and 2 (top left panel)
reflect the mildly to moderately diminished perfusion. FDG uptake in these segments is normal and,
hence, scored as 0. Metabolism perfusion difference scores are obtained by subtracting the perfusion
scores from the metabolism scores; as depicted below the metabolism No. 1 images, the difference
scores are ⫺1 and ⫺2 and reflect a perfusion-metabolism mismatch. In contrast, the metabolism
No. 2 image schematically depicts an example of the highest FDG uptake in the hypoperfused
inferior wall (again in comparison to the perfusion scores shown in the top left panel); the scores are
therefore 0, whereas the scores of 1 in the normally perfused myocardium on the metabolism No.
2 image indicate the lower FDG uptake. Before deriving difference scores, the scores for the FDG
uptake in the normal myocardium need to be normalized to those of perfusion (ie, they need to be
set to 0). This is accomplished by subtracting a value of ⫺1 from all segments on the FDG images
so that the scores in the normally perfused myocardium become 0. The normalized scores for FDG
uptake in the hypoperfused segments then become 0 and ⫺1, and furthermore, the difference scores,
as depicted in the right lower panel, are ⫺1, ⫺2, and ⫺3. A difference score of 0 reflects normal
or matched myocardial regions, and negative scores reflect a perfusion-metabolism mismatch.

techniques such as echocardiography, left ventriculography, Cardiology guidelines. These segments are scored for
and radionuclide angiography.94-100 However, regional regional wall motion (scoring system: normokinesis, hypo-
function can be more difficult to assess with gated FDG kinesis, akinesis, and dyskinesis). In the segments, the
PET because of the prevalence of severe global LV dys- information on perfusion (stress-rest) and metabolism is
function in patients referred for PET viability studies. In integrated.
these patients, overall wall motion and thickening are likely Findings on gated PET images if available may be
to be poor and regional differences may therefore be only included in the final report and may be interpreted within
subtle. Nevertheless, because gated image acquisition can the context of the analysis of the images of perfusion and
be performed at minimal expense, some incremental diag- metabolism. LV volumes at end diastole and end systole
nostic information may be obtained at little if any additional and the LVEF, if available, should be included, as they
cost. Display, analysis, and interpretation of gated PET contain independent diagnostic and prognostic information.
myocardial perfusion and metabolism images should follow In addition, evidence of contractile function can be helpful
a structured and standardized approach. This approach to identify stunning in dysfunctional myocardium with
should follow the guidelines published previously for gated preserved perfusion and metabolism. Residual contractile
SPECT perfusion imaging, as the methodologic aspects of function in areas of a moderate matched perfusion-metab-
gated PET as well as its diagnostic content are comparable olism defect may also aid in assessing the degree of
to those of gated SPECT. The LV short-axis slices are viability and in predicting the magnitude of functional
divided into 17 segments (basal, mid, and distal) as pro- recovery after revascularization in such regions. Impaired
posed by American Heart Association/American College of function in areas of preserved resting perfusion and pre-
Journal of Nuclear Cardiology Machac et al e147
Volume 13, Number 6;e121-51 PET and glucose metabolism imaging

served FDG uptake without evidence of stress-induced iii. History and key clinical findings. A brief descrip-
perfusion defects suggests myopathic or remodeled myo- tion of the patient’s clinical history and findings can
cardium. Nonetheless, further studies are needed to estab- contribute to a more appropriate and comprehensive inter-
lish the incremental diagnostic value of gated PET. pretation of the rest (and stress) perfusion and of the
metabolism images. This information may include past
myocardial infarctions and their location, revascularization
G. Modification of the Interpretation by Relevant
procedures, the patient’s angina-related and congestive
Clinical Information
heart failure–related symptoms, presence of diabetes or
With regard to perfusion imaging, the initial interpre- hypertension, and other coronary risk factors. Information
tation should be made without knowledge of the clinical on regional and global LV function can similarly be
characteristics and stress ECG findings. However, the final important for the interpretation of regional perfusion and
interpretation should be adjusted to some limited degree metabolism patterns.
according to clinically relevant information. Age, pretest A description of the ECG findings may serve as an aid
likelihood, history of known CAD, findings on the stress in the study interpretation, such as the presence of Q waves
and rest electrocardiograms, and chest pain (of very limited and their location or conduction abnormalities (eg, left
significance after pharmacologic stress) all contain diagnos- bundle branch block) for exploring septal perfusion and/or
tic and prognostic information that may impact the inter- metabolic abnormalities.
pretation of the perfusion and viability study. iv. Type of study. The imaging protocols should be
Similar to the guidelines for SPECT perfusion imag- stated concisely. This should include the type of imaging of
ing, the adjusted interpretation should not vary from the myocardial perfusion and of myocardial metabolism with,
initial assessment by more than one category of probability. for example, a PET or PET/CT system or SPECT perfusion
For example, if the initial conclusion was “probably nor- and FDG PET metabolism imaging. If stress perfusion
mal” (a 5-category scale as described above), the final imaging was performed, this should be stated accordingly,
interpretation could be changed to “equivocal” or “defi- indicating the type of stress, such as treadmill, dipyridam-
nitely normal,” but it should not be re-categorized as ole, or adenosine. Radiopharmaceuticals used for the per-
“probably abnormal” or “definitely abnormal.” The modi- fusion and the metabolism imaging studies should be stated,
fication has the effect of improving the overall concordance and their radioactivity doses should be given. The acquisi-
of information sent to clinicians without substantially alter- tion modes and image sequences should be described, such
ing the information contained in the scintigraphic examina- as static or dynamic image acquisition, for stress and rest
tion. perfusion imaging, perfusion and metabolism imaging on
With regard to metabolism imaging, for the interpre- different days, and the use of gating.
tation of the metabolic images, the integration of clinical The main body of the report following this introduc-
information may be even more relevant. Angiographic data tory descriptive information should then be tailored to the
(if available) and rest and stress imaging results are rele- specific clinical question asked by the referring clinician
vant. Regional wall motion data from radionuclide angiog- and the procedural approach chosen for answering this
raphy or 2D echocardiography can be critical for under- question. For example, the report for a stress-rest myocardial
standing findings on the metabolism images. In many cases PET perfusion study will be different from a report describing
it may be necessary to seek additional information from the and interpreting a myocardial perfusion and myocardial me-
referring clinician. tabolism study. Furthermore, the report will again be different
for a study that includes a stress and rest myocardial perfusion
study and a myocardial metabolism study.
H. Reporting of PET Myocardial
v. Summary of stress data. If myocardial perfusion
Perfusion-Metabolism Study Results
has been evaluated during stress, the type of stress should
One is referred to a concurrently updated guideline be indicated (eg, physical or pharmacologic). In instances of
on reporting. pharmacologic stress, the agent should be described, as
i. Patient information. The report should start with should the dose and route of administration. Side effects
the date of the study and then state the patient’s age, sex, and symptoms experienced during stress should be indi-
height, and weight or body surface area, as well as the cated. If pharmacologic stress was discontinued prema-
patient’s medical identification number. turely, the reasons should be indicated.
ii. Indication for study. Stating the reasons for The report should contain additional details on the
performing the study aids in focusing the study interpre- stress study including changes in heart rate and blood
tation on the clinical question asked by the referring pressure, occurrence of arrhythmias or conduction ab-
clinician. In addition, a clear statement of the reasons for normalities, and development of ST-segment changes
the study has become critical for billing purposes. and their location. Type of chest pain during pharmaco-
e148 Machac et al Journal of Nuclear Cardiology
PET and glucose metabolism imaging November/December 2006

logic stress has limited predictive value for myocardial estimates of regional myocardial flow reserve, as there is
ischemia, but whether it resembles chest pain from prior greater familiarity with this particular index. Normal values
history should be noted, and its severity (mild, moderate, for myocardial perfusion reserves and for resting and
or severe) should be reported. hyperemic myocardial blood flows should be given for
vi. Summary of clinical laboratory data and dietary comparison. Estimates of myocardial blood flow at rest
state. Information on the dietary state (eg, fasting or should also be related to the rate-pressure product as an
post-prandial) and on interventions for manipulating plasma index of cardiac work, an important determinant of resting
glucose levels through, for example, oral or IV administra- myocardial blood flow, which could influence the apparent
tion of glucose or use of the euglycemic hyperinsulinemic coronary flow reserve capacity (ability to augment flow
clamp should be given. If pharmacologic measures, such as during stress).
nicotinic acid derivatives, have been used, this should be viii. Image description and interpretation: Metab-
described. Furthermore, BG levels, if obtained at baseline olism. The report should describe the relative distribution of
or after intervention, should be listed, as they are useful for myocardial blood flow at rest and the location, extent, and
the interpretation of the metabolic images. severity of regional perfusion defects. The report should
vii. Image description and interpretation: Perfu- continue with a description of the FDG uptake in the
sion images. A statement regarding image quality is im- myocardium and indicate the tracer activity concentrations
portant. Reduced quality may affect interpretation or the in normally perfused and in hypoperfused myocardium.
confidence the interpreter has in the accuracy of the find- The glucose utilization state as evident from the radiotracer
ings. If the cause of the reduced quality is known or uptake in normally perfused myocardium and also from
suspected, then this should be stated accordingly. This blood pool activity should then be reported and be related to
information may prove useful when repeat images are the dietary state or the presence of insulin resistance
obtained in the same patient. (including impaired glucose tolerance and type 2 diabetes).
The pretest likelihood of CAD may be determined with This should be related to the residual blood pool activity as
available algorithms that rely on parameters including age, additional evidence for inadequate clearance of FDG from
sex, character of chest pain, number of coronary risk blood into tissue and provide the information for low tracer
factors, and findings of stress electrocardiography. The uptake in normally perfused myocardium. The regional
findings on the stress/rest myocardial perfusion images are FDG concentrations in hypoperfused myocardial regions
then included in order to derive a probability of CAD. An should be described and categorized into perfusion-metab-
estimate of the probability may be reported based on the olism mismatches or matches. An estimate of the degree of
definite presence or absence of a perfusion defect, its extent mismatch (or the difference between FDG uptake and
and severity, and its change from stress to rest. Other findings perfusion) and, more importantly, of the extent of a mis-
such as a stress-related transient LV dilation should be in- match or match, or both, when present, should be given.
cluded. When the presence of CAD has already been estab- Findings on semiquantitative or quantitative image analysis
lished, then the likelihood of stress-induced ischemia should be approaches may be added. Location and, in particular,
reported instead of the likelihood of significant CAD. extent of matches and mismatches are important because
NOTE: It is acknowledged that in contrast to the they contain prognostic information on future cardiac
extensive experience with SPECT myocardial perfusion events and predictive information on potential outcomes in
imaging and the accordingly large body of evidence re- regional and global LV function, congestive heart failure–
ported in the scientific literature, the experience with PET related symptoms, and long-term survival after revascular-
myocardial perfusion imaging is more limited. It seems ization. Finally, the description of the perfusion-metabolism
acceptable, however, to extrapolate predictive information findings may include a correlation to regional wall motion
from the SPECT literature for assessing the probability of abnormalities and should indicate the potential for a post-
CAD or of stress-induced myocardial ischemia as well as revascularization improvement in the regional and global
assessment of cardiac risk through myocardial perfusion LV function.
imaging. Nevertheless, such extrapolations should be made ix. Final interpretation. Results should be succinctly
with caution because of methodologic differences in PET summarized first to address the clinical question being
and SPECT for assessing the relative distributions of asked, such as whether there is evidence of CAD or
myocardial perfusion and its disease-related alterations. evidence of myocardial viability. Potential confounding
When the report includes a quantitative estimate of effects of image artifacts or other quality concerns should
regional myocardial blood flows, it should be given in be stressed in the summary. It is useful to conclude with a
milliliters per minute per grams of myocardium, describe summary of the percentage of the LV muscle mass that is
changes in flow from rest to pharmacologically induced viable, as well as the percentage that shows the potential for
hyperemic stress, and be related to the specific clinical ischemia.
question to be answered. It is also useful to include The results should be compared with available prior
Journal of Nuclear Cardiology Machac et al e149
Volume 13, Number 6;e121-51 PET and glucose metabolism imaging

study results and any diagnostic changes in the interim emission tomographs. Washington, DC: National Electrical
highlighted. If possible, an estimate of the likelihood of Manufacturers Association; 2001.
16. Schelbert HR, Phelps ME, Huang SC, MacDonald NS, Hansen H,
CAD in instances of perfusion images or of the predicted Selin C, et al. N-13 ammonia as an indicator of myocardial blood
outcome in regional or global LV function after revascu- flow. Circulation 1981;63:1259-72.
larization should be provided. If additional diagnostic clar- 17. DeGrado TR, Bergmann SR, Ng CK, Raffel DM. Tracer kinetic
ification seems needed, the physician may recommend an modeling in nuclear cardiology. J Nucl Cardiol 2000;7:686-700.
alternative modality. 18. International Commission on Radiological Protection. Radiation
dose to patients from radiopharmaceuticals. ICRP Publication 80.
Ann ICRP 2000;28:113.
19. International Commission on Radiological Protection. Radiation
References dose to patients from radiopharmaceuticals. ICRP Publication 53.
Ann ICRP 1988;18:62.
20. Kim SH, Machac J, Almeida O, et al. Optimization of rubidi-
1. Bacharach SL, Bax JJ, Case J, et al. PET myocardial glucose
um-82 generator performance [abstract]. Clin Nucl Med 2004;29:
metabolism and perfusion imaging: part 1— guidelines for patient
135P.
preparation and data acquisition. J Nucl Cardiol 2003;10:543-54.
21. Meerdink DJ, Leppo JA. Experimental studies of the physiologic
2. Schelbert HR, Beanlands R, Bengel F, et al. PET myocardial
properties of technetium-99m agents: myocardial transport of
perfusion and glucose metabolism imaging: part 2— guidelines
perfusion imaging agents. Am J Cardiol 1990;66:9E-15E.
for interpretation and reporting. J Nucl Cardiol 2003;10:557-71.
3. Beanlands RSB, Muzik O, Melon P, et al. Noninvasive quantifi- 22. Leppo JA, Meerdink DJ. Comparison of the myocardial uptake of
cation of regional myocardial flow reserve in patients with a technetium-labeled isonitrile analogue and thallium. Circ Res
coronary atherosclerosis using nitrogen-13 ammonia positron 1989;65:632-9.
emission tomography. Determination of extent of altered vascular 23. Mack RE, Nolting DD, Hogancamp CE, et al. Myocardial
reactivity. J Am Coll Cardiol 1995;26:1465-75. extraction of Rb-86 in the rabbit. Am J Physiol 1959;197:1175-7.
4. Bergmann SR. Quantification of myocardial perfusion with 24. Becker L, Ferreira R, Thomas M. Comparison of Rb-86 and
positron emission tomography. In: Sobel BE, editor. Positron microsphere estimates of left ventricular blood flow distribution.
emission tomography of the heart. Mount Kisco (NY): Futura; J Nucl Med 1977;15:969-73.
1992. p. 97-127. 25. Selwyn AP, Allan RM, L’Abbate A, et al. Relation between regional
5. Cardiogen-82 rubidium Rb-82. Princeton (NJ): Bracco Diagnos- myocardial uptake of rubidium-82 and perfusion: absolute reduction
tics; 2000.[package insert] of cation uptake in ischemia. Am J Cardiol 1982;50:112-21.
6. Choi Y, Huang SC, Hawkins RA, et al. A simplified method for 26. Goldstein RA, Mullani NA, Marani SK, et al. Myocardial
quantification of myocardial blood flow using nitrogen-13-ammo- perfusion with rubidium-82. II. Effects of metabolic and pharma-
nia and dynamic PET. J Nucl Med 1993;34:488-97. cological interventions. J Nucl Med 1983;24:907-15.
7. Hutchins GD, Schwaiger M, Rosenspire KC, et al. Noninvasive 27. Schelbert HR, Ashburn WL, Chauncey DM, et al. Comparative
quantification of regional blood-flow in the human heart using myocardial uptake of intravenously administered radionuclides.
N-13 ammonia and dynamic positron emission tomographic J Nucl Med 1977;15:1092-1100.
imaging. J Am Coll Cardiol 1990;15:1032-42. 28. Chow BJ, Ruddy TD, Dalipaj MM, et al. Feasibility of exercise
8. Krivokapich J, Smith GT, Huang SC, et al. 13N ammonia rubidium-82 positron emission tomography myocardial perfusion
myocardial imaging at rest and with exercise in normal volun- imaging [abstract]. J Am Coll Cardiol 2003;41:428A.
teers. Quantification of absolute myocardial perfusion with dy- 29. Garcia EV, Bacharach SL, Mahmarian JJ, et al. Imaging
namic positron emission tomography. Circulation 1989;80:1328- guidelines for nuclear cardiology procedures: part 1. J Nucl
37. Cardiol 1996;3:G1-46.
9. Schwaiger M, Ziegler SI, Bengel FM. Assessment of myocardial 30. Hickey KT, Sciacca RR, Bokhari S, et al. Assessment of cardiac
blood flow with positron emission tomography. In: Shaw PM, wall motion and ejection fraction with gated PET using N-13
editor. Imaging in cardiovascular disease. Philadelphia: Lippin- ammonia. Clin Nucl Med 2004;29:243-8.
cott Williams and Wilkins; 2000. p. 195-212. 31. Schwaiger M, editor. Cardiac positron emission tomography.
10. Schoder H, Schelbert HR. Positron emission tomography for the Boston: Kluwer Academic Publishers; 1996. p. 366.
assessment of myocardial viability: noninvasive approach to cardiac 32. Moser KW, Hsu BL, Cullom SJ, et al. Count-based attenuation
pathophysiology. In: Dilsizian V, editor. Myocardial viability: a clinical maps reconstructed with a Bayesian algorithm for more efficient
and scientific treatise. Armonk (NY): Futura; 2000. p. 391-418. cardiac PET imaging [abstract]. J Nucl Med 2005;46:262p.
11. Schelbert HR. F-18-deoxyglucose and the assessment of myocar- 33. Loghin C, Sdringola S, Gould KL. Common artifacts in PET
dial viability. Semin Nucl Med 2002;32:60-9. myocardial perfusion images due to attenuation-emission misreg-
12. Shoder H, Schelbert HR. Positron emission tomography for the istration: clinical significance, causes, and solution. J Nucl Med
assessment of myocardial viability. In: Dilsizian V, editor. Myo- 2004;45:1029-39.
cardial viability: a clinical and scientific treatise. Armonk (NY): 34. Hsu BL, Moser KW, Cullom SJ, et al. Correction of imaging
Futura; 2000. p. 391-418. artifacts from implanted leads in cardiac PET/CT: a phantom
13. Di Carli M. Advances in positron emission tomography. J Nucl evaluation [abstract]. J Nucl Med 2005;46:174p.
Cardiol 2004;11:719-32. 35. Difilippo FP, Brunken RC, Kaczur T, Jaber W. Artifacts from
14. Kamel E, Hany TF, Burger, et al. CT vs 68GE attenuation implanted leads in cardiac PET using CT-based attenuation
correction in a combined PET/CT system: evaluation of the effect correction [abstract]. J Nucl Card 2004;11:S23.
of lowering the CT tube current. Eur J Nucl Med Mol Imaging 36. Hamacher K, Coenen HH, Stauocklin G. Efficient sterospecific
2002;29:346-50. synthesis of no-carrier-added 2-[F-18]-fluoro-2-deoxy-D-glucose
15. National Electrical Manufacturers Association. NEMA standards using aminopolyether supported nucleophilic substitution. J Nucl
publication NU 2-2001: performance measurements of positron Med 1986;27:235-8.
e150 Machac et al Journal of Nuclear Cardiology
PET and glucose metabolism imaging November/December 2006

37. ACRP-53. CDE Dosimetry Services. 2001. Available from: URL: severity and extent assessment in myocardial perfusion imaging
http://www.internaldosimetry.com/freedosestimates/adult/index. using positron emission tomography. Eur J Nucl Med
html. Accessed October 19, 2006. 1998;25:1313-21.
38. Stabin MG, Stubbs JB, Toohey RE. Radiation dose estimates for 55. McCord ME, Bacharach SL, Bonow RO, et al. Misalignment
radiopharmaceuticals. Radiation Internal Dose Information Cen- between PET transmission and emission scans: its effect on
ter. Oak Ridge Institute of Science and Education, Oak Ridge, myocardial imaging. J Nucl Med 1992;33:1209-14, discussion
Tenn. Available from: URL:http://www.nrc.gov/reading-rm/doc- 1214-5.
collections/nuregs/contract/cr6345/cr6345.pdf. 56. Bettinardi V, Gilardi MC, Lucignani G, et al. A procedure for
39. Package insert for Tc-99m sestamibi kit (Cardiolite). Oak Ridge patient repositioning and compensation for misalignment be-
(TN): Radiation Internal Dose Information Center, Oak Ridge tween transmission and emission data in PET heart studies. J Nucl
Institute for Science and Education. Med 1993;34:137-42.
40. Gallagher BM, Ansari A, Atkins H, et al. Radiopharmaceuticals 57. Laubenbacher C, Rothley J, Sitomer J, et al. An automated
XXVII. F-18-labeled 2-deoxy-2-fluoro-D-glucose as a radiophar- analysis program for the evaluation of cardiac PET studies: initial
maceutical for measuring regional myocardial glucose metabo- results in the detection and localization of coronary artery disease
lism in vivo: tissue distribution and imaging studies in animals. using nitrogen-13-ammonia. J Nucl Med 1993;34:968-78.
J Nucl Med 1977;18:990-6. 58. Gould KL, Martucci JP, Goldberg DI, et al. Short-term choles-
41. Ratib O, Phelps ME, Huang SC, et al. Positron tomography with terol lowering decreases size and severity of perfusion abnormal-
deoxyglucose for estimating local myocardial glucose metabo- ities by positron emission tomography after dipyridamole in
lism. J Nucl Med 1982;23:577-86. patients with coronary artery disease. A potential noninvasive
42. Choi Y, Brunken RC, Hawkins RA, et al. Factors affecting marker of healing coronary endothelium. Circulation 1994;89:
myocardial 2-[F-18] fluoro-2-deoxy-D-glucose uptake in positron 1530-8.
emission tomography studies of normal humans. Eur J Nucl Med 59. Krivokapich J, Stevenson LW, Kobashigawa J, Huang SC,
1993;20:308-18. Schelbert HR. Quantification of absolute myocardial perfusion at
43. Ohtake T, Yokoyama I, Watanabe T, et al. Myocardial glucose- rest and during exercise with positron emission tomography after
metabolism in noninsulin-dependent diabetes-mellitus patients human cardiac transplantation. J Am Coll Cardiol 1991;18:512-7.
evaluated by FDG-PET. J Nucl Med 1995;36:456-63. 60. Kuhle WG, Porenta G, Huang SC, et al. Quantification of regional
44. Schöder H, Campisi R, Ohtake T, et al. Blood flow-metabolism myocardial blood flow using 13N-ammonia and reoriented dy-
imaging with positron emission tomography in patients with namic positron emission tomographic imaging. Circulation 1992;
diabetes mellitus for the assessment of reversible left ventricular 86:1004-17.
contractile dysfunction. J Am Coll Cardiol 1999;33:1328-37. 61. Muzik O, Beanlands RS, Hutchins GD, et al. Validation of
45. Knuuti MJ, Nuutila P, Ruotsalainen U, et al. Euglycemic hyper- nitrogen-13-ammonia tracer kinetic model for quantification of
insulinemic clamp and oral glucose load in stimulating myocar- myocardial blood flow using PET. J Nucl Med 1993;34:83-91.
dial glucose utilization during positron emission tomography. 62. Blanksma PK, Willemsen AT, Meeder JG, et al. Quantitative
J Nucl Med 1992;33:1255-62. myocardial mapping of perfusion and metabolism using paramet-
46. Vitale GD, deKemp RA, Ruddy TD, Williams K, Beanlands RS. ric polar map displays in cardiac PET. J Nucl Med 1995;36:
Myocardial glucose utilization and optimization of (18)F-FDG 153-8.
PET imaging in patients with non-insulin-dependent diabetes 63. Dorbala S, Limaye A, Samson U. Normal and abnormal re-
mellitus, coronary artery disease, and left ventricular dysfunction. sponses of left ventricular ejection fraction during vasodilator
J Nucl Med 2001;42:1730-6. stress Rb-82 positron emission tomography (PET/CT) [abstract].
47. Streeter J, Churchwell K, Sigman S, et al. Clinical glucose J Nucl Med 2005;46:P268.
loading protocol for F-18 FDG myocardial viability imaging 64. Tamaki N, Yonekura Y, Yamashita K, et al. Prediction of
[abstract]. Mol Imaging Biol 2002;4:192. reversible ischemia after coronary artery bypass grafting by
48. Martin WH, Jones RC, Delbeke D, Sandler MP. A simplified positron emission tomography. J Cardiol 1991;21:193-201.
intravenous glucose loading protocol for fluorine-18 fluorodeoxy- 65. Mäki M, Luotolahti M, Nuutila P, et al. Glucose uptake in the
glucose cardiac single-photon emission tomography. Eur J Nucl chronically dysfunctional but viable myocardium. Circulation
Med 1997;24:1291-7. 1996;93:1658-66.
49. Bax JJ, Veening MA, Visser FC, et al. Optimal metabolic 66. Gropler RJ, Siegel BA, Lee KJ, et al. Nonuniformity in myocar-
conditions during fluorine-18 fluorodeoxyglucose imaging; a dial accumulation of fluorine-18-fluorodeoxyglucose in normal
comparative study using different protocols. Eur J Nucl Med fasted humans [see comments]. J Nucl Med 1990;31:1749-56.
1997;24:35-41. 67. Pagano D, Townend JN, Littler WA, et al. Coronary artery bypass
50. DeFronzo RA, Tobin JD, Andres R. Glucose clamping technique: surgery as treatment for ischemic heart failure: the predictive
a method for quantifying insulin secretion and resistance. Am J value of viability assessment with quantitative positron emission
Physiol 1979;237:E214-23. tomography for symptomatic and functional outcome. J Thorac
51. Port SC, Berman DS, Garcia EV, et al. Imaging guidelines for Cardiovasc Surg 1998;115:791-9.
nuclear cardiology procedures. J Nucl Cardiol 1999;6:G47-84. 68. Gerber BL, Ordoubadi FF, Wijns W, et al. Positron emission
52. Gambhir SS, Schwaiger M, Huang SC, et al. Simple noninvasive tomography using (18)F-fluoro-deoxyglucose and euglycaemic
quantification method for measuring myocardial glucose utiliza- hyperinsulinaemic glucose clamp: optimal criteria for the predic-
tion in humans employing positron emission tomography and tion of recovery of post-ischaemic left ventricular dysfunction.
fluorine-18 deoxyglucose. J Nucl Med 1989;30:359-66. Results from the European Community Concerted Action Multi-
53. Porenta G, Kuhle W, Czernin J, et al. Semiquantitative assess- center study on use of (18)F-fluoro-deoxyglucose positron emis-
ment of myocardial blood flow and viability using polar map sion tomography for the detection of myocardial viability. Eur
displays of cardiac PET images. J Nucl Med 1992;33:1628-36. Heart J 2001;22:1691-701.
54. Nekolla SG, Miethaner C, Nguyen N, Ziegler SI, Schwaiger M. 69. Choi Y, Hawkins RA, Huang SC, et al. Parametric images of
Reproducibility of polar map generation and assessment of defect myocardial metabolic rate of glucose generated from dynamic
Journal of Nuclear Cardiology Machac et al e151
Volume 13, Number 6;e121-51 PET and glucose metabolism imaging

cardiac PET and 2-[18F]fluoro-2-deoxy-d-glucose studies. J Nucl 85. Flameng WJ, Shivalkar B, Spiessens B, et al. PET scan predicts
Med 1991;32:733-8. recovery of left ventricular function after coronary artery bypass
70. Bax JJ, Visser FC, van Lingen A, Visser CA, Teule GJ. operation. Ann Thorac Surg 1997;64:1694-701.
Myocardial F-18 fluorodeoxyglucose imaging by SPECT. Clin 86. Bax JJ, Visser FC, van Lingen A, et al. Metabolic imaging using
Nucl Med 1995;20:486-90. F18-fluorodeoxyglucose to assess myocardial viability. Int J Card
71. Bax JJ, Visser FC, Cornel JH, et al. Improved detection of viable Imaging 1997;13:145-55, discussion 157-60.
myocardium with fluorodeoxyglucose-labeled single-photon 87. Haas F, Haehnel CJ, Picker W, et al. Preoperative positron
emission computed tomography in a patient with hibernating emission tomographic viability assessment and perioperative and
myocardium: comparison with rest-redistribution thallium 201- postoperative risk in patients with advanced ischemic heart
labeled single-photon emission computed tomography. J Nucl disease [see comments]. J Am Coll Cardiol 1997;30:1693-700.
Cardiol 1997;4:178-9. 88. Beanlands RS, Hendry PJ, Masters RG, et al. Delay in revascu-
72. Bax JJ, Wijns W. Fluorodeoxyglucose imaging to assess myo- larization is associated with increased mortality rate in patients
cardial viability: PET, SPECT or gamma camera coincidence with severe left ventricular dysfunction and viable myocardium
imaging? J Nucl Med 1999;40:1893-5. on fluorine 18-fluorodeoxyglucose positron emission tomography
73. Tillisch J, Brunken R, Marshall R, et al. Reversibility of cardiac imaging. Circulation 1998;98:II51-6.
wall-motion abnormalities predicted by positron tomography. 89. Fath-Ordoubadi F, Pagano D, Marinho NV, et al. Coronary
N Engl J Med 1986;314:884-8. revascularization in the treatment of moderate and severe
74. Marwick TH, Hollman J. Acute myocardial infarction associated postischemic left ventricular dysfunction. Am J Cardiol 1998;
with intravenous dipyridamole for rubidium-82 PET imaging. 82:26-31.
Clin Cardiol 1990;13:230-1. 90. Bax JJ, Poldermans D, Elhendy A, et al. Improvement of left
75. Carrel T, Jenni R, Haubold-Reuter S, et al. Improvement of ventricular ejection fraction, heart failure symptoms and prog-
severely reduced left ventricular function after surgical revascu- nosis after revascularization in patients with chronic coronary
larization in patients with preoperative myocardial infarction. Eur artery disease and viable myocardium detected by dobutamine
J Cardiothorac Surg 1992;6:479-84. stress echocardiography. J Am Coll Cardiol 1999;34:163-9.
76. Lucignani G, Paolini G, Landoni C, et al. Presurgical identifica-
91. Maes A, Van de Werf F, Nuyts J, et al. Impaired myocardial
tion of hibernating myocardium by combined use of technetium-
tissue perfusion early after successful thrombolysis. Impact on
99m hexakis 2-methoxyisobutylisonitrile single photon emission
myocardial flow, metabolism, and function at late follow-up.
tomography and fluorine-18 fluoro-2-deoxy-D-glucose positron
Circulation 1995;92:2072-8.
emission tomography in patients with coronary artery disease.
92. Maes A, Mortelmans L, Nuyts J, et al. Importance of flow/
Eur J Nucl Med 1992;19:874-81.
metabolism studies in predicting late recovery of function fol-
77. Tamaki N, Kawamoto M, Takahashi N, et al. Prognostic value of
lowing reperfusion in patients with acute myocardial infarction.
an increase in fluorine-18 deoxyglucose uptake in patients with
Eur Heart J 1997;18:954-62.
myocardial infarction: comparison with stress thallium imaging.
93. Yamagishi H, Akioka K, Hirata K, et al. A reverse flow-
J Am Coll Cardiol 1993;22:1621-7.
metabolism mismatch pattern on PET is related to multivessel
78. Maes A, Flameng W, Nuyts J, et al. Histological alterations in
disease in patients with acute myocardial infarction. J Nucl Med
chronically hypoperfused myocardium. Correlation with PET
1999;40:1492-8.
findings. Circulation 1994;90:735-45.
79. Paolini G, Lucignani G, Zuccari M, et al. Identification and 94. Yamashita K, Tamaki N, Yonekura Y, et al. Quantitative analysis
revascularization of hibernating myocardium in angina-free pa- of regional wall motion by gated myocardial positron emission
tients with left ventricular dysfunction. Eur J Cardiothorac Surg tomography: validation and comparison with left ventriculogra-
1994;8:139-44. phy. J Nucl Med 1989;30:1775-86.
80. Schwarz E, Schaper J, vom Dahl J, et al. Myocardial hibernation 95. Yamashita K, Tamaki N, Yonekura Y, et al. Regional wall
is not sufficient to prevent morphological disarrangements with thickening of left ventricle evaluated by gated positron emission
ischemic cell alterations and increased fibrosis. Circulation 1994; tomography in relation to myocardial perfusion and glucose
90:I-378. metabolism. J Nucl Med 1991;32:679-85.
81. vom Dahl J, Eitzman D, Al-Aouar A, et al. Relation of regional 96. Boyd HL, Gunn RN, Marinho NV, et al. Non-invasive measure-
function, perfusion, and metabolism in patients with advanced ment of left ventricular volumes and function by gated positron
coronary artery disease undergoing surgical revascularization. emission tomography. Eur J Nucl Med 1996;23:1594-602.
Circulation 1994;90:2356-66. 97. Boyd HL, Rosen SD, Rimoldi O, Cunningham VJ, Camici PG.
82. Depré C, Vanoverschelde JL, Melin JA, et al. Structural and Normal values for left ventricular volumes obtained using gated
metabolic correlates of the reversibility of chronic left ventricular PET. G Ital Cardiol 1998;28:1207-14.
ischemic dysfunction in humans. Am J Physiol 1995;268: 98. Hor G, Kranert WT, Maul FD, et al. Gated metabolic positron
H1265-7. emission tomography (GAPET) of the myocardium: 18F-FDG-
83. Maes A, Flameng W, Borgers M, et al. Regional myocardial blood PET to optimize recognition of myocardial hibernation. Nucl
flow, glucose utilization and contractile function before and after Med Commun 1998;19:535-45.
revascularization and ultrastructural findings in patients with chronic 99. Willemsen AT, Siebelink HJ, Blanksma PK, Paans AM. Auto-
coronary artery disease. Eur J Nucl Med 1995;22:1299-305. mated ejection fraction determination from gated myocardial
84. vom Dahl J, Altehoefer C, Sheehan F, et al. Recovery of regional FDG-PET data. J Nucl Cardiol 1999;6:577-82.
left ventricular dysfunction after coronary revascularization: im- 100. Hattori N, Bengel FM, Mehilli J, et al. Global and regional
pact of myocardial viability assessed by nuclear imaging and functional measurements with gated FDG PET in compar-
vessel patency at follow-up angiography. J Am Coll Cardiol ison with left ventriculography. Eur J Nucl Med 2001;28:
1996;28:948-58. 221-9.
 Standardized reporting of myocardial perfusion
images
Peter L. Tilkemeier, MD

INTRODUCTION In the future, structured reporting will be an impor-

tant mechanism for exchange of patient data across
The myocardial perfusion scan report is the final multiple users and the different computer platforms. It is
product of a complicated process outlined in the other the portion of the examination that will make the results
sections of this document. As such, it needs to contain portable as our patients move through the health care
sufficient detail to portray the complexity of the proce- system while allowing the practitioner rapid and efficient
dure while simultaneously being succinct and providing access to the appropriate parts of the patient’s medical
a basic “bottom line” result to the referring physician. data. Reporting will therefore evolve from the present
The American Society of Nuclear Cardiology open state to one in which the elements noted above will
(ASNC) has published two guidelines and numerous be required as part of all myocardial perfusion imaging
articles and editorials that address the importance of reports. This will ensure that a report currently contains
reporting.1-5 The two documents that appear below all of the necessary data for completeness while allowing
address the required/recommended elements that should us to build toward a future of report portability.
be contained in a report, as well as the possible response
to each of the elements. Hendel et al describe a subjec-
tive structure for a myocardial perfusion report, while References
Tilkemeier et al carried this further with a detailed
1. Wackers FJ. Intersocietal Commission for the Accreditation of
objective description of the elements contained in a
Nuclear Cardiology Laboratories (ICANL) position statement on
myocardial perfusion imaging report and the possible standardization and optimization of nuclear cardiology reports.
content in each of those elements. These elements will be J Nucl Cardiol 2000;7:397-400.
the future organization for a structured report utilizing 2. Gonzalez P, Canessa J, Massardo T. Formal aspects of the user-
the Digital Imaging and Communications in Medicine friendly nuclear cardiology report. J Nucl Cardiol 1998;5:365-6.
3. Cerqueira MD. The user-friendly nuclear cardiology report: what
(DICOM) and Integrating the Healthcare Enterprise
needs to be considered and what is included. J Nucl Cardiol
(IHE) standards. 1996;3:350-5.
4. Hendel RC, Wackers FJT, Berman DS, Ficaro E, DePuey EG, Klein
Quality Assurance Committee. L, et al. Reporting of radionuclide myocardial perfusion imaging
J Nucl Cardiol 2006;13:e152-71. studies. J Nucl Cardiol 2003;10:705-8.
1071-3581/$32.00 5. Tilkemeier PL, Cooke CD, Ficaro EP, Glover DK, Hansen CL,
Copyright © 2006 by the American Society of Nuclear Cardiology. McCallister BD. Standardized reporting matrix for radionuclide
doi:10.1016/j.nuclcard.2006.08.010 myocardial perfusion imaging. J Nucl Cardiol. In press 2006.

American Society of Nuclear Cardiology consensus statement: Reporting of radionuclide

myocardial perfusion imaging studies
Robert C. Hendel, MD, Frans J. Th. Wackers, MD, PhD, Daniel S. Berman, MD,
Edward Ficaro, PhD, E. Gordon DePuey, MD, Larry Klein, MD,
Manuel Cerqueira, MD

PREAMBLE ing and provide this information so that it may serve as

a standard for all nuclear cardiology laboratories. Infor-
Although clear and effective guidelines have been mation regarding reporting has been collated from a
published regarding the optimal practice for nuclear number of sources, some of which provide extensive and
cardiology procedures, including reporting, the Writing comprehensive guidelines for many aspects of nuclear
Group believed that it was important to specifically cardiology procedures. It is the intent of this Consensus
identify the critical factors involved in effective report- Statement to specifically focus on reporting and provide

e152
Journal of Nuclear Cardiology Tilkemeier e153
Volume 13, Number 6;e152-71 Standardized reporting of myocardial perfusion imaging

a consolidation of recommendations with regard to this Table 1. Clinical information

critically important task.
Demographics (age, sex, race)
Body habitus (height, weight)
INTRODUCTION Symptoms
Medications
The final task in performing a myocardial perfusion
Cardiac risk factors
imaging study and often the only portion that is directly
Prior cardiac events
examined by the referring physician is the written report.
Prior diagnostic tests
The purpose of this report is to communicate the findings
Therapeutic cardiac procedures
and clinical implications of the myocardial perfusion
imaging procedure. The report should also document the
technical aspects of the procedure for purposes of ac-
Table 2. Procedure
countability and reimbursement. In addition, the written
report provides appropriate documentation necessary for Type and protocol of stress procedure
certification of physicians and accreditation of the nu- Pharmacologic agents used, with total dose
clear cardiology laboratory. The report also functions as Adequacy of stress
part of ongoing quality assurance measures and allows Symptoms during stress
comparison with other modalities, such as coronary Hemodynamic response (heart rate, blood pressure)
angiography. Templates for exercise (Appendix 1) and Electrocardiographic changes
pharmacologic stress imaging (Appendix 2) are included Radiopharmaceuticals used (with dose)
in this Consensus Statement. Imaging protocol
Functional data
COMPONENTS OF REPORT Use of attenuation/scatter correction

The individual components of the report include the

indications for the procedure, clinical history, procedure, in the interpretation of perfusion studies. The type of
findings, and impression. Although individual laboratory symptoms for which the study is being performed should
and practitioner variability may be present with regard to be identified, as well as the current medications that the
how these fields are constructed, each of the specific patient is receiving and whether the patient is under the
areas contains vital information that should be part of the influence of specific cardioactive medications at the time
final report. Each section of the report will be examined of stress testing. The latter may have a dramatic impact
individually. on the overall interpretation of the study. A reasonably
detailed cardiac history should also be included includ-
ing all past procedures and the major cardiovascular risk
Indication
factors. Finally, previous diagnostic tests or therapeutic
The specific purpose for which the test is being procedures may also be included so as to add to the
performed should be clearly identified. This provides the clinical relevance of the report.
required documentation for the medical necessity of the
study and focuses the report on the question asked by the
Procedure
referring physician. Although many indications are pos-
sible, they may be broken down into 5 general fields: (1) All aspects of the stress and imaging procedures
diagnosis of coronary artery disease, (2) delineation of should be well documented within the report (Table 2).
extent and severity of disease, (3) risk stratification, (4) The type of stress test (ie, exercise or pharmacologic)
determination of myocardial viability, and (5) assess- should be identified. For a pharmacologic stress exami-
ment of acute chest pain syndromes. nation, the total doses of the stress agent and timing of
administration should be noted, as well as whether
adjunctive exercise was performed. If an exercise test is
Clinical History
performed, the type of protocol (Bruce, modified Bruce,
Selective clinical information should be included Naughton) should be noted, as well as the adequacy of
within the report so as to clarify the image findings the stress results (peak heart rate, percent of the maxi-
(Table 1). These may include patient demographics such mum predicted heart rate, duration of exercise, and
as age and sex. Body habitus, height, weight, chest estimated metabolic equivalents [optional]). The pres-
circumference, and bra size (optional) may also be ence of symptoms during the test, as well as whether they
included but should definitely be collected for assistance provided an indication for termination, should be noted.
e154 Tilkemeier Journal of Nuclear Cardiology
Standardized reporting of myocardial perfusion imaging November/December 2006

Table 3. Results ratio should also be described. Finally, abnormal right

ventricular activity and size should be noted.
Study quality
Assessment of left ventricular function should be
Size of left and right ventricle at stress and rest
performed with stress and/or rest gated techniques. The
Defect description (size, reversibility, severity, location)
report should describe regional wall motion abnormali-
Extensiveness (transient cavity dilation/transient ischemic
ties in terms of both severity (hypokinesis, akinesis,
dilation, lung activity, right ventricular activity)
dyskinesis) and location. Differentiation between global
Left ventricular function (global, regional)
and focal abnormalities should be noted. A comment
Extracardiac activity
regarding the overall left ventricular function should be
made, such as normal or moderately depressed left
ventricular systolic function. Finally, the quantitative left
ventricular ejection fraction should be included within
Angina should be classified as typical or atypical, and the the report. For ejection fractions greater than 60%, the
location of chest pain may be described. Hemodynamic actual calculated number should be included in the report
information, such as the heart rate and blood pressure and mention made of overestimation in patients with
response, should be noted. The baseline and peak heart small hearts. Ventricular volumes may also be reported
rate and blood pressure may also be included. Finally, (optional).
the presence or absence of significant electrocardio- The vascular location may be commented on by the
graphic changes should be mentioned for both exercise interpreter. Each report should attempt to differentiate
and pharmacologic stress, including the amount of ST- between the presence of single-vessel or multivessel
segment deviation and whether resting abnormalities coronary artery disease, and markedly abnormal studies
were present. The imaging procedure should also be should be clearly identified. The specific assignment of
delineated including the protocol used, the radiopharma- vascular territories, especially to the inferolateral regions
ceutical and its dose, the timing of injection (optional), and the apex, may be difficult. Significant extracardiac
and the time between injection and imaging (optional). activity (eg, abnormal focal tracer uptake, which may
The use of gating or attenuation correction should be represent malignancy) should be described.
noted.
Impression
Findings
The most critical portion of the report is the impres-
The first portion of the “Results” section (Table 3) sion. The final interpretation must possess clarity and
should comment about image quality if it is technically must state whether the study is “normal” or “abnormal.”
inadequate, as this has important ramifications regarding The reader is encouraged to use those categories that
the accuracy and interpretation of the result. This assess- provide the greatest clarity for the report. The categories
ment of study quality should be based on the degree of of “probably normal,” “probably abnormal,” and “equiv-
patient motion, subdiaphragmatic/hepatic activity inter- ocal” should be used as infrequently as possible but may
fering with interpretation, insufficient myocardial activ- allow for communication of interpretive uncertainty. The
ity, and other technical features. cause of the uncertainty, whether it is technical or related
Perfusion defect(s) should be well described in to specific patient issues, should be stated. A small
terms of their size (small, medium, or large), type percentage of patients may fall into an “equivocal”
(reversible, persistent, or mixed), and severity (mild, category, but this should be used in less than 10% of all
moderate, or severe). Summed stress, rest, and difference studies interpreted.
scores may be reported, providing a more quantitative Of special importance is the finding of normal
assessment of defect size and severity, and the true perfusion in the setting of other abnormal findings, such
quantitative measures, such as percent of left ventricle, as left ventricular dysfunction. Such studies should be
may also be used (optional). The location should be described as “abnormal,” with clarification that the
noted and described in concordance with the position perfusion data are homogeneous or normal but other
paper on standardized myocardial segmentation. aspects of the imaging study led to this conclusion. If the
Evidence of extensive abnormalities should be men- patient is unable to achieve an adequate level of stress or
tioned, including abnormal radiopharmaceutical lung the images are of inadequate quality, the term “nondiag-
activity, and the presence of cavity enlargement, either as nostic” should be used to describe the overall impression
stress-only cavity enlargement or as persistent cavity of the study. In the former circumstance, repeat testing
dilation, should be mentioned. The transient cavity dila- with pharmacologic stress should be recommended.
tion (alternatively known as transient ischemic dilation) After the final impression, which is stated in the first
Journal of Nuclear Cardiology Tilkemeier e155
Volume 13, Number 6;e152-71 Standardized reporting of myocardial perfusion imaging

sentence, the presence of apparent perfusion abnormali- 3. Cerqueira MD. The user-friendly nuclear cardiology report: what
ties may be noted and coordinated with the final impres- needs to be considered and what is included. J Nucl Cardiol
1996;4:350-5.
sion by acknowledging that artifacts—such as those 4. Port SC. Imaging guidelines for nuclear cardiology procedures, part
resulting from soft-tissue attenuation, patient motion, or II. J Nucl Cardiol 1999;6:G48-84.
left bundle branch block—may result in apparent perfu- 5. Wackers FJTh. Intersocietal Commission for the Accreditation of
sion abnormalities in the presence of coronary artery Nuclear Medicine Laboratories (ICANL) position statement on
disease or infarction. A report may still be interpreted as standardization and optimization of nuclear cardiology reports.
J Nucl Cardiol 2000;7:397-400.
normal even if it possesses an artifactually produced
perfusion abnormality. The functional information
should be included in the final section by describing the APPENDIX 1: TEMPLATE FOR EXERCISE
presence or absence of regional and/or global abnormal- MYOCARDIAL PERFUSION IMAGING
ities. Correlation with clinical stress testing and angio-
graphic data, as well as any comparisons to prior studies, Stress/Rest (or Rest/Stress) Single-/Dual-Isotope Sin-
should also be included in the impression section. Fi- gle Photon Emission Computed Tomography
nally, the report should address the clinical question that (SPECT) Imaging With Exercise Stress and Gated
was posed as the indication for the procedure. SPECT Imaging

STRUCTURED REPORTING Indication

ASNC supports the development of structured re- (select one)

porting for myocardial perfusion imaging. It is antici- Diagnosis of coronary disease
pated that many of the components of a final report as Evaluation of extent and severity of coronary artery
outlined in this Consensus Statement will be included as disease
part of the data elements for structured reporting. Evaluation of myocardial viability
A DICOM subcommittee specifically dealing with Risk stratification—post–myocardial infarction (MI)/
this issue is presently delineating the critical fields. All preoperative/general
key organizations and societies are working with instru- Assessment of acute chest pain
mentation manufacturers in the development of struc-
tured reporting. Clinical History
____ year old man/woman with (no) known coronary
CONCLUSIONS artery disease
In summary, ASNC is strongly in favor of the Cardiac risk factors include: ____
standardization of myocardial perfusion imaging reports. Previous cardiac procedures include: ____
Furthermore, we encourage clarity and clinically relevant Current symptomatology includes: ____
conclusions. Finally, the report should contain adequate
information to support the medical necessity of the Procedure
procedure.1-5
The patient performed treadmill exercise/bicycle exer-
cise using a modified Bruce/Bruce/Naughton/____ pro-
Acknowledgment tocol, completing ____ minutes and completing an esti-
The authors have indicated they have no financial conflicts mated workload of ____ metabolic equivalents (METS).
of interest. The heart rate was ____ beats per minute at baseline and
increased to ____ beats at peak exercise, which was
____% of the maximum predicted heart rate. The blood
References pressure response to exercise was normal/hypotensive/
1. Berman DS, Germano G. An approach to the interpretation and
hypertensive. The patient did/did not develop any symp-
reporting of gated myocardial perfusion SPECT. In: DS, Berman G, toms other than fatigue during the procedure; specific
editors. Germano Clinical gated cardiac SPECT. Armonk (NY): symptoms include ____. The resting electrocardiogram
Futura Publishing; 1999. demonstrated ____ and did/did not show ST-segment
2. Cerqueira MD, Weissman NJ, Dilsizian V, et al. Standardized changes consistent with myocardial ischemia.
myocardial segmentation and nomenclature for tomographic imag-
ing of the heart: a statement for healthcare professionals from the
Myocardial perfusion imaging was performed at rest
Cardiac Imaging Committee of the Council on Clinical Cardiology (____ minutes following the injection of ____ mCi of
of the American Heart Association. J Nucl Cardiol 2002;9:240-5. ____). At peak exercise, the patient was injected with
e156 Tilkemeier Journal of Nuclear Cardiology
Standardized reporting of myocardial perfusion imaging November/December 2006

____ mCi of ____ and exercise was continued for ____ Cardiac risk factors include: ____
minute(s). Gating post-stress tomographic imaging was Previous cardiac procedures include: ____
performed ____ minutes after stress (and rest). Current symptomatology includes: ____

Findings Procedure
The overall quality of the study is poor/fair/good/excel- Pharmacologic stress testing was performed with aden-
lent. osine/dipyridamole/dobutamine with a dose ____. Addi-
Left ventricular cavity is noted to be normal/enlarged tionally, low-level exercise was performed along with
on the rest (and/or stress) studies. There is evidence of the vasodilator infusion (specify: ____). The heart rate
abnormal lung activity. Additionally, the right ventricle was ____ at baseline and rose to ____ beats per minute
is normal/abnormal (specify: ____). during the adenosine/dipyridamole/dobutamine infusion.
SPECT images demonstrate homogeneous tracer dis- This corresponds with ____% of the maximum predicted
tribution throughout the myocardium OR a small/mod- heart rate. Blood pressure response was normal/hyper-
erate/large perfusion abnormality of mild/moderate/se- tensive/hypotensive during the stress procedure. The
vere intensity is present in the ____ (location) region on patient developed significant symptoms, which included
the stress images. The rest images reveal ____. Gated ____. The resting electrocardiogram demonstrated ____
SPECT imaging reveals normal myocardial thickening and did/did not show ST-segment changes consistent
and wall motion. OR Gated SPECT imaging demon- with myocardial ischemia.
strates hypokinesis/dyskinesis/akinesis of the ____ (lo- Myocardial perfusion imaging was performed at rest
cation). The left ventricular ejection fraction was calcu- (____ minutes following the injection of ____ mCi of
lated to be ____% OR the left ventricular ejection ____). At peak pharmacologic effect, the patient was
fraction was normal (⬎60%). injected with ____ mCi of ____. Gating post-stress
tomographic imaging was performed ____ minutes after
stress (and rest).
Impression
Myocardial perfusion imaging is normal/abnormal.
Findings
There is a small/moderate/large area of ischemia/infarc-
tion in the ____ location. Overall left ventricular systolic The overall quality of the study is poor/fair/good/excellent.
function was normal/abnormal with/without regional Left ventricular cavity is noted to be normal/enlarged
wall motion abnormalities (as noted above). Compared on the rest (and/or stress) studies. There is evidence of
to the prior study from ____ (date), the current study abnormal lung activity. Additionally, the right ventricle
reveals ____. is normal/abnormal (specify: ____).
SPECT images demonstrate homogeneous tracer dis-
APPENDIX 2: TEMPLATE FOR PHARMACOLOGIC tribution throughout the myocardium OR a small/mod-
MYOCARDIAL PERFUSION IMAGING erate/large perfusion abnormality of mild/moderate/se-
vere intensity is present in the ____ (location) region on
Stress/Rest (or Rest/Stress) Single-/Dual-Isotope the stress images. The rest images reveal ____. Gated
SPECT Imaging With Pharmacologic Stress and SPECT imaging reveals normal myocardial thickening
Gated SPECT Imaging and wall motion. OR Gated SPECT imaging demon-
strates hypokinesis/dyskinesis/akinesis of the ____ (lo-
cation). The left ventricular ejection fraction was calcu-
Indication
lated to be ____% OR the left ventricular ejection
(select one) Diagnosis of coronary disease fraction was normal (⬎60%).
Evaluation of extent and severity of coronary artery
disease
Impression
Evaluation of myocardial viability
Risk stratification—post-MI/preoperative/general Myocardial perfusion imaging is normal/abnormal.
Assessment of acute chest pain There is a small/moderate/large area of ischemia/infarc-
tion in the ____ location. Overall left ventricular systolic
function was normal/abnormal with/without regional
Clinical History
wall motion abnormalities (as noted above). Compared
____ year old man/woman with (no) known coronary to the prior study from ____ (date), the current study
artery disease reveals ____.
Journal of Nuclear Cardiology Tilkemeier e157
Volume 13, Number 6;e152-71 Standardized reporting of myocardial perfusion imaging

American Society of Nuclear Cardiology information statement: Standardized reporting matrix

for radionuclide myocardial perfusion imaging
Peter L. Tilkemeier, MD,a C. David Cooke, MSEE,b Edward P. Ficaro, PhD,b
David K. Glover, MD, PhD,b Christopher L. Hansen, MD,b
Benjamin D. McCallister, Jr, MDb

INTRODUCTION Acknowledgment
The standardization of reporting is one of the pre- The authors have indicated they have no financial conflicts
requisites for the development of electronic records. of interest.
Incorporating the report in the patient’s electronic record
allows the report to become part of a saved/transmitted References
image set. This image set is independent of the platform
1. Imaging guidelines for nuclear cardiology procedures, part 2.
used to acquire, process, and interpret the data as well as American Society of Nuclear Cardiology. J Nucl Cardiol 1999;6:
generate a report that complies with current guidelines. G47– 84.
Standardized reporting is the method that defines the 2. Hendel RC, Wackers FJ, Berman DS, Ficaro E, DePuey EG, Klein
necessary fields that are part of the report and the L, et al. American Society of Nuclear Cardiology consensus
statement: reporting of radionuclide myocardial perfusion imaging
potentially allowable responses in those fields. As part of studies. J Nucl Cardiol 2003;10:705-8.
the process, each field is also assigned a priority level for
inclusion in the report. The matrix that follows divides
the myocardial perfusion imaging report into each of its TABLES
parts (demographics, perfusion image results, and so on)
and defines allowable responses as well as assigning Tables
priorities to each of the fields. This becomes a dynamic The following list outlines the tables that have been
document as the field of nuclear cardiology evolves. created containing the data elements in a standardized
report.

HISTORY Description
ASNC has previously published two important doc-
Study demographics
uments with regard to standardized reporting, “Imaging
Patient demographics Table 1
guidelines for nuclear cardiology procedures, part 2,”1
Clinical information Table 2
and “Reporting of radionuclide myocardial perfusion
Stress testing data Table 3
imaging studies.”2 These two documents are the histor-
ECG data
ical foundation on which this matrix is based. The former
Resting ECG Table 4
provides many of the responses for the data fields, and
Stress ECG Table 5
the latter gives structure and organization to the report. Imaging data
This information statement was created to fuse the work Imaging parameters Table 6
of the prior authors and assign priorities to the fields. It LV perfusion Table 7
also gives groups, such as DICOM, a reference guide as LV perfusion quantitation Table 8
they create standards for industry to use in the develop- Stress LV function parameters Table 9
ment of standardized reporting tools and templates for Rest LV function parameters, if performed Table 10
radionuclide myocardial perfusion images. Miscellaneous data Table 11
RV parameters Table 12
Overall impression Table 13

ECG, Electrocardiographic; LV, left ventricular; RV, right ventricular.

e158 Tilkemeier Journal of Nuclear Cardiology
Standardized reporting of myocardial perfusion imaging November/December 2006

STUDY DEMOGRAPHICS

Table 1. Patient demographics

Variable Description Datatype Priority Response

Last name Patient last name Text Required Variable

First name Patient first name Text Required Variable
Patient ID Patient unique identifier Text/numerical Required Variable
Patient DOB Date of birth Numerical Required Variable
Gender Patient gender Text Required M/F
Date of study Date of study Numerical Required 00/00/0000
Weight Patient weight Numerical Recommended Value in pounds
Height Patient height Numerical Recommended Value in inches
Chest circumference Chest circumference Numerical Optional Value in inches
Bra cup size Bra cup size Text Optional Variable
Ethnicity Ethnic origin Text Optional Variable
Inpt/outpt In or outpatient Text Optional I/O
Testing site Location Text Optional Variable text
Ordering MD MD name Text Required Text
Interpreting MD MD name Text Required Text
Date of interpretation Date of interpretation Date Required 00/00/0000
Date of transcription Date of transcription Date Required 00/00/0000

Table 2. Clinical information

Variable Description Datatype Priority Response

Indication Study indication Text Required Chest pain

Preoperative evaluation
CAD
Heart failure
Coronary risk factors
Dyspnea
History of PTCA
History of CABG
Abnormal stress test
Abnormal electrocardiogram
Arrhythmia
Angina pectoris
Hypertension
Palpitations
SVT
Syncope
Other

Pretest chest pain Type of chest pain Text Required Typical angina
Atypical angina
Nonanginal chest pain
Anginal equivalent
No chest pain
Journal of Nuclear Cardiology Tilkemeier e159
Volume 13, Number 6;e152-71 Standardized reporting of myocardial perfusion imaging

Table 2. Continued.
Variable Description Datatype Priority Response

Medications Medications Text Recommended ␤-Blocker

Ca⫹⫹ blocker
Nitrates
Digoxin
ACE/ARB
Diuretics

Test medications Medications on day of Text Recommended ␤-Blocker

test Ca⫹⫹ blocker
Nitrates
Digoxin
ACE/ARB
Diuretics

Cardiac risk factors Risk factors Text Recommended Hypertension

Diabetes
Hypercholesterolemia
Family history
Smoking
Obesity
Metabolic syndrome
Peripheral vascular disease
Other

Cardiac history Cardiac history Text Recommended s/p PTCA/stent

s/p CABG
s/p MI
History of peripheral vascular disease
Other (text)

Prior testing Prior cardiac testing Text Optional ETT

Perfusion imaging
Stress echo
Catheterization
MRI
CT

Comments Text Optional Free text comments

CAD, Coronary artery disease; PTCA, percutaneous transluminal coronary angioplasty; CABG, coronary artery bypass grafting; SVT,
supraventricular tachycardia; ACE, angiotensin-converting enzyme; ARB, angiotensin receptor blocker; MI, myocardial infarction; ETT,
exercise tolerance test; MRI, magnetic resonance imaging; CT, computed tomography.

STRESS TESTING DATA

Table 3. Stress testing data

Variable Description Datatype Priority Response

Test type Type of test Text Required Exercise

Pharmacologic
Pharmacologic/exercise
Other (free text)
e160 Tilkemeier Journal of Nuclear Cardiology
Standardized reporting of myocardial perfusion imaging November/December 2006

Table 3. Continued.
Variable Description Datatype Priority Response

Pharmacologic stress Pharmacologic Text Required Dipyridamole

agent stress agent Dobutamine
Adenosine
Atropine
Other (free text)

Indication for Reason exercise Text Recommended LBBB or pacemaker

pharmacologic stress only is not Other
appropriate

Pharmacologic stress Pharmacologic Text Required mg/kg or ␮g · kg⫺1 · min⫺1

dose stress dose

Pharmacologic stress time Time to deliver Numerical Required Minutes

pharmacologic
stress dose

Pharmacologic stress Adjunctive Text Required Yes

exercise exercise use No

Exercise protocol Exercise protocol Text Required Bruce

used Modified Bruce
Naughton
Ramp
Modified Naughton
Bicycle ergometer
Other

Stress HR Maximum HR Numerical Required Beats/min

achieved

HR response HR response to Text Recommended Normal

exercise Blunted
Accentuated

% MPHR % of MPHR Numerical Required %

Stress BP Peak BP achieved Numerical Required mm Hg

during test

BP response BP response to Text Recommended Normal

exercise Hypotensive
Hypertensive
Blunted

Exercise duration Time on treadmill Numerical Required Minutes (0.0 format)

Functional capacity Exercise functional Text Recommended Average

capacity Below average
Above average

METS Peak estimated Numerical Recommended METS

METS level

Anginal stress symptoms Chest pain Text Required Typical angina

symptoms Atypical angina
during stress Nonanginal chest pain
Anginal equivalent
No chest pain
Journal of Nuclear Cardiology Tilkemeier e161
Volume 13, Number 6;e152-71 Standardized reporting of myocardial perfusion imaging

Table 3. Continued.
Variable Description Datatype Priority Response

Duration of symptoms Duration of Numerical Required if anginal Seconds

anginal stress stress symptom
symptoms is present

Severity of anginal Severity of anginal Numerical Required if anginal Numerical value on 1-10
symptoms symptoms stress symptom scale (1, mild; 10, severe)
is present

Other stress symptoms Other symptoms Text Recommended Dyspnea

during stress Claudication
Syncope
Flushing
Nausea
Dizziness
Fatigue

Reason for termination Reason for Text Required Chest pain

termination ECG changes
Fatigue
Dyspnea
Patient request
Achievement of target HR
Hypotension
Hypertension
Arrhythmia
Claudication
End of protocol
Other (free text)

LBBB, Left bundle branch block; HR, heart rate; MPHR, maximal predicted heart rate; BP, blood pressure; METS, metabolic equivalents; ECG,
electrocardiographic.

ELECTROCARDIOGRAPHIC DATA

Table 4. Resting ECG data

Variable Description Datatype Priority Response

Resting HR Resting HR Numerical Required Beats per minute

Resting BP Resting BP Numerical Required mm Hg

Rest rhythm Resting ECG rhythm Text Required Sinus rhythm

Sinus bradycardia
Sinus tachycardia
Junctional rhythm
SVT
Ectopic atrial rhythm
Atrial fibrillation
Atrial paced
Ventricular paced
AV sequential paced
Other (free text)
e162 Tilkemeier Journal of Nuclear Cardiology
Standardized reporting of myocardial perfusion imaging November/December 2006

Table 4. Resting ECG data

Variable Description Datatype Priority Response

Resting conduction Resting IV conduction Text Required Normal

IVCD
LBBB
RBBB
Incomplete RBBB
Incomplete LBBB
RBBB ⫹ LAHB
RBBB ⫹ LPHB
First-degree AV block
Second-degree AV block
Third-degree AV block
Pre-excitation
Other (free text)
Resting arrhythmias Resting ECG arrhythmias Text Required None
APC
VPC
Atrial fibrillation
SVT
Nonsustained ventricular
tachycardia

Repolarization Resting ECG repolarization Text Required Normal

Early repolarization
Nonspecific ST-T
abnormality
ST depression
ST elevation
Secondary ST-T
abnormality

ECG, Electrocardiographic; HR, heart rate; BP, blood pressure; SVT, supraventricular tachycardia; LBBB, left bundle branch block; RBBB, right
bundle branch block; IV, intraventricular; IVCD, intraventricular conduction delay; LAHB, left anterior hemiblock; LPHB, left posterior
hemiblock; AV, atrioventricular; APC, atrial premature contraction; VPC, ventricular premature contraction.
Journal of Nuclear Cardiology Tilkemeier e163
Volume 13, Number 6;e152-71 Standardized reporting of myocardial perfusion imaging

Table 5. Stress ECG data

Variable Description Datatype Priority Response

Stress rhythm Stress ECG rhythm Text Required Sinus rhythm

Sinus bradycardia
Sinus tachycardia
Junctional rhythm
SVT
Ectopic atrial rhythm
Atrial fibrillation
Atrial paced
Ventricular paced
AV sequential paced
Other (Free text)

Stress conduction Stress ECG IV Text Recommended Normal

conduction IVCD
LBBB
RBBB
Incomplete RBBB
Incomplete LBBB
Bifascicular block
RBBB ⫹ LAHB
RBBB ⫹ LPHB
First degree AV block
Second degree AV block
Third degree AV block

Stress arrhythmias Stress-induced ECG Text Required None

arrhythmias APC
VPC
Atrial fibrillation
SVT
Non-sustained ventricular
tachycardia
Ventricular tachycardia
Ventricular fibrillation

Stress repolarization Stress-induced ECG Text Required Normal

repolarization Early repolarization
changes Nonspecific ST-T changes
ST depression
ST elevation
Secondary ST-T changes

ST-segment change Stress-induced ST- Text Required Normal

segment change Nondiagnostic low heart rate
Nondiagnostic resting ST
abnormalities
Nondiagnostic V-pacing or
LBBB
Mildly positive
Moderately positive
Strongly positive
Strongly positive–ST
elevation

ST-segment location Location of ST- Text Required if Anterior

segment change ST-segment Inferior
change is Lateral
not normal Inferolateral
e164 Tilkemeier Journal of Nuclear Cardiology
Standardized reporting of myocardial perfusion imaging November/December 2006

Table 5. Continued.
Variable Description Datatype Priority Response

Timing of ST-segment Time when ST Text Required if ST-segment Stress only

depression segment change is not normal Stress and recovery
depression Recovery only
occurs
ST-segment Configuration of Text Required if ST-segment Horizontal
configuration ST-segment change is not normal Upsloping
change Downsloping
Elevation

ST-segment Millimeters of ST- Numerical Required if ST-segment mm

depression amount segment change change is not normal

Maximum ST- Maximum Numerical Required if ST segment mm

segment change millimeters of change is not normal
ST-segment
change
Number of leads with
Number of leads Numerical Required if ST segment Numerical
ST-segment change with ST-segment change is not normal
change

ETT compare to prior– Comparison to Text Recommended Same

exercise tolerance prior ETT Lower
Higher

ETT compare to prior– Comparison of ST Text Recommended No change

ST segment segment to prior New ischemia
test Resolution of ischemia
More ischemia than prior

Duke treadmill score Duke score Numerical Recommended XXX

Duke treadmill score Duke prognosis Text Recommended Low

prognosis Moderate
High

Heart rate recovery Heart rate recovery Text Recommended Normal

Abnormal
Markedly abnormal

ECG, Electrocardiographic; SVT, supraventricular tachycardia; LBBB, left bundle branch block; RBBB, right bundle branch block; ETT, exercise
tolerance test; IV, intraventricular; IVCD, intraventricular conduction delay; LAHB, left anterior hemiblock; LPHB, left posterior hemiblock; AV,
atrioventricular; APC, atrial premature contraction; VPC, ventricular premature contraction.
Journal of Nuclear Cardiology Tilkemeier e165
Volume 13, Number 6;e152-71 Standardized reporting of myocardial perfusion imaging

IMAGING DATA

Table 6. Imaging parameters

Variable Description Datatype Priority Response

Imaging protocol Describes protocol used to Text Required Stress thallium

acquire images Rest thallium/stress
technetium
Rest technetium/stress
technetium 1 day
Rest technetium/stress
technetium 2 day
Stress technetium/rest
technetium 1 day
Viability only

Imaging position Describes patient Test Recommended Supine

positioning Prone

Stress radiopharmaceutical Stress imaging agent used Text Required Thallium

Tc-99m sestamibi
Tc-99m tetrofosmin

Stress dose Dose of radioactivity (mCi) Numerical Required Numerical value XX.X

Stress date Date of stress study Numerical Required XX/XX/XXXX

Stress injection time Time of stress injection Numerical Recommended XXXX hours (military
time)

Stress imaging time Time of stress imaging Numerical Required XXXX hours (military
time)

Exercise time after Exercise time after Numerical Optional XX:XX min:sec
injection injection

Rest radiopharmaceutical Rest imaging agent used Text Required Thallium

Tc-99m sestamibi
Tc-99m tetrofosmin

Rest dose Dose of radioactivity (mCi) Numerical Required Numerical value XX.X

Rest date Date of rest study Numerical Required XX/XX/XXXX

Rest injection time Time of rest injection Numerical Recommended XXXX hours (military
time)

Rest imaging time Time of rest imaging Numerical Required XXXX hours (military
time)

Attenuation correction Use of attenuation Text Required Yes

correction No

Attenuation correction Type of attenuation Text Recommended Transmission

type correction Computed
tomography– based
Other
e166 Tilkemeier Journal of Nuclear Cardiology
Standardized reporting of myocardial perfusion imaging November/December 2006

Table 7. LV perfusion
Variable Description Datatype Priority Response

Perfusion defect location Location of perfusion Text Required Basal inferior

defect Basal anteroseptal
Basal inferoseptal
Basal inferior
Basal inferolateral
Basal anterolateral
Mid anterior
Mid anteroseptal
Mid inferoseptal
Mid inferior
Mid inferolateral
Mid anterolateral
Apical anterior
Apical septal
Apical inferior
Apical lateral
Apex
None

Perfusion defect size Size of perfusion defect Text Required Small

Medium
Large

Perfusion defect severity Intensity of perfusion Text Required Mild

defect Moderate
Severe
Background

Type of perfusion defect Perfusion defect type Text Required Reversible

Persistent
Mixed

TCD/TID Transient cavitary dilatation Text Required Yes

No

TCD/TID value Transient cavitary dilatation Numerical Optional Ratio

ratio
Comparison to prior images Prior image comparison Text Recommended No change
LV perfusion New ischemia
New infarction
Resolution of
prior ischemia

Prior study date Date of prior study Date Recommended XX/XX/XXXX

Lung uptake, stress and Lung uptake stress and Text Recommended Yes
rest rest No

Lung uptake, stress only Stress lung uptake Text Recommended Yes
No

RV myocardial uptake, RV uptake stress/rest Text Recommended Normal

stress and rest Increased

RV myocardial uptake, RV uptake stress only Text Recommended Normal

stress Increased

Increased LV uptake Subjective LV uptake Text Optional Normal

Hypertrophied

The information in this table may be repeated as required to describe multiple perfusion defects.
LV, Left ventricular; TCD, transient cavity dilation; TID, transient ischemic dilation; RV, right ventricular.
Journal of Nuclear Cardiology Tilkemeier e167
Volume 13, Number 6;e152-71 Standardized reporting of myocardial perfusion imaging

Table 8. Left ventricular perfusion quantitation

Variable Description Datatype Priority Response

SSS 17-Segment SSS Numerical Optional XX

SRS 17-Segment SRS Numerical Optional XX

SDS 17-Segment SDS Numerical Optional XX

Stress perfusion extent Regional stress defect extent (% Numerical Optional XX%
myocardium involved)

Rest perfusion extent Regional rest defect extent (% Numerical Optional XX%
myocardium involved)

Stress ischemia extent Regional reversibility defect extent Numerical Optional XX%
(% myocardium involved)

SSS, Summed stress score; SRS, summed rest score; SDS, summed difference score.

Table 9. Stress LV function parameters

Variable Description Datatype Priority Response

Timing of stress function Timing of LV function Text Recommended During exercise (ie,
images assessment first pass)
Post-exercise
Global LV function Subjective LV function Text Recommended Normal
Mildly reduced
Moderately reduced
Severely reduced
LV EF Calculated EF Numerical Recommended XX%
LV volume subjective Subjective LV volume Text Optional Normal
Mildly enlarged
Moderately enlarged
Markedly enlarged
EDV EDV Numerical Optional XX mL
ESV ESV Numerical Optional XX mL
Regional wall motion Subjective regional wall Text Optional Normal
motion Mild hypokinesis
Moderate hypokinesis
Severe hypokinesis
Akinesis
Dyskinesis
e168 Tilkemeier Journal of Nuclear Cardiology
Standardized reporting of myocardial perfusion imaging November/December 2006

Table 9. Continued
Variable Description Datatype Priority Response

Regional wall motion Subjective regional wall Text Optional Basal inferior
location motion Basal anteroseptal
Basal inferoseptal
Basal inferior
Basal inferolateral
Basal anterolateral
Mid anterior
Mid anteroseptal
Mid inferoseptal
Mid inferior
Mid inferolateral
Mid anterolateral
Apical anterior
Apical septal
Apical inferior
Apical lateral
Apex
None
Diffuse
Regional wall motion Comparison of function to Text Optional Consistent with
summary imaging perfusion
Inconsistent with
perfusion
Comparison to prior Prior image comparison Text Recommended No change
images New wall motion
abnormality
Improvement of wall
motion

The information in this table may be repeated as required to describe multiple perfusion defects.
LV, Left ventricular; EF, ejection fraction; EDV, end-diastolic volume; ESV, end-systolic volume.
Journal of Nuclear Cardiology Tilkemeier e169
Volume 13, Number 6;e152-71 Standardized reporting of myocardial perfusion imaging

Table 10. Rest LV function parameters if performed

Variable Description Datatype Priority Response

Rest global LV function Subjective LV function Text Optional Normal

Abnormal
Mildly reduced
Moderately reduced
Severely reduced
Rest LV EF Calculated EF Numerical Optional XX%
Rest LV volume Subjective LV volume Text Optional Normal
subjective Mildly enlarged
Moderately enlarged
Severely enlarged
Rest EDV EDV Numerical Optional XX mL
Rest ESV ESV Numerical Optional XX mL
Rest regional wall motion Subjective regional wall Text Optional Normal
motion Mild hypokinesis
Moderate hypokinesis
Severe hypokinesis
Akinesis
Dyskinesis
Rest regional wall motion Subjective regional wall Text Optional Basal inferior
location motion Basal anteroseptal
Basal inferoseptal
Basal inferior
Basal inferolateral
Basal anterolateral
Mid anterior
Mid anteroseptal
Mid inferoseptal
Mid inferior
Mid inferolateral
Mid anterolateral
Apical anterior
Apical septal
Apical inferior
Apical lateral
Apex
None
Diffuse
Rest regional wall motion Comparison of function to Text Optional Consistent with perfusion
summary imaging Inconsistent with perfusion
Rest regional wall motion Comparison of rest to stress Text Optional No significant change from
comparison wall motion rest to stress
Increased from rest to
stress
Decreased from rest to
stress
Comparison to prior Prior image comparison Text Recommended No change
images New wall motion
abnormality
Improvement of wall
motion

The information in this table may be repeated as required to describe multiple perfusion defects.
LV, Left ventricular; EF, ejection fraction; EDV, end-diastolic volume; ESV, end-systolic volume.
e170 Tilkemeier Journal of Nuclear Cardiology
Standardized reporting of myocardial perfusion imaging November/December 2006

Table 11. Miscellaneous data

Variable Description Datatype Priority Response

Overall study quality Overall quality of study Text Required Excellent

Good
Poor
Uninterpretable
Other
Study quality/artifacts Specific problems Text Recommended Breast attenuation
Diaphragmatic attenuation
Motion artifact
Subdiaphragmatic activity
Other
Extracardiac activity Describe extracardiac activity Text Recommended Normal
Increased lung uptake
Other (free text)

Table 12. RV parameters

Variable Description Datatype Priority Response

RV perfusion Subjective RV perfusion Text Optional Normal

Abnormal
Other
RV uptake Subjective Text Optional Normal
Increased
Global RV function Subjective RV function Text Optional Normal
Abnormal
Mildly reduced
Moderately reduced
Severely reduced
RV EF Calculated EF Numerical Optional XX%
RV volume subjective Subjective LV volume Text Optional Normal
Mildly enlarged
Moderately enlarged
Severely enlarged
RV regional wall motion Subjective regional wall Text Optional Normal
motion Mild hypokinesis
Moderate hypokinesis
Severe hypokinesis
Akinesis
Dyskinesis
RV regional wall motion Comparison of function to Text Optional Consistent with perfusion
summary imaging Inconsistent with perfusion

RV, Right ventricular; EF, ejection fraction.

Journal of Nuclear Cardiology Tilkemeier e171
Volume 13, Number 6;e152-71 Standardized reporting of myocardial perfusion imaging

OVERALL IMPRESSION

Table 13. Overall impression

Variable Description Datatype Priority Response

Number of diseased vessels Number of diseased vessel Numerical Optional None

territories One
Two
Three
Diseased vessels Diseased vessels Text Optional Left anterior descending
Right coronary
Left circumflex
Left main
Stress test summary Summary of stress results Text Recommended Normal
Abnormal
Nondiagnostic
LV perfusion summary Summary of LV perfusion Text Required Normal
Equivocal
Abnormal
Ischemia
Infarction
Negative for ischemia but
____________
RV perfusion summary Summary of RV perfusion Text Optional Normal
Abnormal
Prior imaging studies Integration of prior imaging study Text Optional Agreement
results Disagreement
Scan significance Significance of perfusion results Text Recommended Normal
Low risk
Low to moderate risk
Moderate risk
Moderate to high risk
High risk
Uncertain risk

LV, Left ventricular; RV, right ventricular.