European Heart Journal (2013) 34, 775–781 CLINICAL RESEARCH

doi:10.1093/eurheartj/ehs022 Imaging

MR-IMPACT II: Magnetic Resonance Imaging

for Myocardial Perfusion Assessment in Coronary
artery disease Trial: perfusion-cardiac magnetic
resonance vs. single-photon emission computed
tomography for the detection of coronary
artery disease: a comparative multicentre,
multivendor trial
Juerg Schwitter 1*, Christian M. Wacker 2, Norbert Wilke3, Nidal Al-Saadi 4,
Ekkehart Sauer5, Kalman Huettle 6, Stefan O. Schönberg 7, Andreas Luchner 8,
Oliver Strohm 9, Hakan Ahlstrom 10, Thorsten Dill 11, Nadja Hoebel 12,
and Tamas Simor 13, for the MR-IMPACT Investigators
1
University Hospital Lausanne, Rue de Bugnon 46, CH-1011 Lausanne, Switzerland; 2University Hospital Wuerzburg, Wuerzburg, Germany; 3University of Florida Health Science Center,
Gainesville/Jacksonville, FL, USA; 4Franz-Volhard Clinic-Humboldt University, Berlin, Germany; 5Landstuhl Hospital, Landstuhl, Germany; 6Semmelweis University Hospital, Budapest,
Hungary; 7LMU Munich, Grosshadern, Germany; 8University Hospital Regensburg, Regensburg, Germany; 9St Gertrauden Hospital, Berlin, Germany; 10Uppsala University Hospital, Uppsala,
Sweden; 11Kerckhoff Clinics, Bad Nauheim, Germany; 12GE Healthcare Buchler GmbH & Co. KG, Munich, Germany; and 13Medical University of Science, Pecs, Hungary

Received 6 November 2011; revised 17 January 2012; accepted 23 January 2012; online publish-ahead-of-print 4 March 2012

Aims Perfusion-cardiac magnetic resonance (CMR) has emerged as a potential alternative to single-photon emission com-
puted tomography (SPECT) to assess myocardial ischaemia non-invasively. The goal was to compare the diagnostic
performance of perfusion-CMR and SPECT for the detection of coronary artery disease (CAD) using conventional
X-ray coronary angiography (CXA) as the reference standard.
.....................................................................................................................................................................................
Methods In this multivendor trial, 533 patients, eligible for CXA or SPECT, were enrolled in 33 centres (USA and Europe) with
and results 515 patients receiving MR contrast medium. Single-photon emission computed tomography and CXA were per-
formed within 4 weeks before or after CMR in all patients. The prevalence of CAD in the sample was 49%.
Drop-out rates for CMR and SPECT were 5.6 and 3.7%, respectively (P ¼ 0.21). The primary endpoint was non-in-
feriority of CMR vs. SPECT for both sensitivity and specificity for the detection of CAD. Readers were blinded vs.
clinical data, CXA, and imaging results. As a secondary endpoint, the safety profile of the CMR examination was eval-
uated. For CMR and SPECT, the sensitivity scores were 0.67 and 0.59, respectively, with the lower confidence level
for the difference of +0.02, indicating superiority of CMR over SPECT. The specificity scores for CMR and SPECT
were 0.61 and 0.72, respectively (lower confidence level for the difference: 20.17), indicating inferiority of CMR vs.
SPECT. No severe adverse events occurred in the 515 patients.
.....................................................................................................................................................................................
Conclusion In this large multicentre, multivendor study, the sensitivity of perfusion-CMR to detect CAD was superior to SPECT,
while its specificity was inferior to SPECT. Cardiac magnetic resonance is a safe alternative to SPECT to detect per-
fusion deficits in CAD.
-----------------------------------------------------------------------------------------------------------------------------------------------------------
Keywords Magnetic resonance imaging † Scintigraphy † Coronary disease † Perfusion † Ischaemia

* Corresponding author. Tel: +41 21 314 4015, Fax: +41 21 314 4013, Email: [email protected]
Published on behalf of the European Society of Cardiology. All rights reserved. & The Author 2012. For permissions please email: [email protected]
776 J. Schwitter et al.

Introduction Table 1 Inclusion and exclusion criteria

Despite considerable progress in the treatment of acute and
Inclusion
chronic coronary artery disease (CAD) during the past years,
Patients scheduled for routine CXA and/or SPECT for clinical
CAD still remains the largest single killing disease in the USA1 reasons
and Europe.2 Approximately one-third of coronary attacks CXA and SPECT must be performed within 4 weeks before or after
results in death in the USA1 and similar numbers apply to CMR irrespective of findings in any of the 3 tests
Europe. These numbers stress the need for early detection of No interventions on the coronary arteries in the time period
disease. The presence of perfusion deficits has been shown to between the 3 tests
................................................................................
be one of the strongest predictors of cardiac death and non-fatal
Exclusion
myocardial infarction (MI) in large studies,3 and ischaemia
Acute MI (,2 weeks prior to study enrolment)
imaging takes a central position in guidelines for the work-up
History of coronary artery bypass grafting
of patients with known or suspected CAD.4,5 Scintigraphy is
Unstable angina pectoris
widely used for ischaemia detection and it has been shown to
Decompensated heart failure
be cost-effective.6 In recent years, several studies documented
Any interventions on the coronary arteries in the time period
a high diagnostic performance of perfusion-cardiac magnetic res- between CXA, SPECT, and CMR
onance (CMR) vs. conventional X-ray coronary angiography Contraindications for adenosine and contrast media
(CXA)7 – 13 and showed its prognostic value.14 This evidence Severe arrhythmias considered to compromise quality of CMR
triggered an increasing utilization of perfusion-CMR in clinical imaging
practice and its impact on clinical patient management was re-
cently demonstrated.15 CXA, invasive coronary angiography; SPECT, single-photon emission computed
The current MR-IMPACT II trial was designed to compare the tomography; CMR, cardiac magnetic resonance; MI, myocardial infarction.

diagnostic performance of CMR vs. single-photon emission com-

puted tomography (SPECT) for the detection of perfusion deficits
in CAD (defined as ≥75% area reduction in coronary vessels in Definition of coronary artery disease: reference standard
CXA) in a large international multicentre, multivendor design. The study was designed to determine the sensitivity and specificity of
We enrolled 533 patients in 33 study centres across Europe and CMR and SPECT to detect perfusion deficits in patients with suspected
the USA. Patients were characterized for the presence of CAD or known CAD. It was not the aim to discriminate perfusion deficits
by CXA (reference standard), perfusion-CMR, and SPECT. into ischaemia vs. scar tissue, since CMR is known for its excellent
power to differentiate viable from scar tissue,16 which could have
Cardiac magnetic resonance and SPECT were each analysed by
introduced a bias in favour of CMR. For the definition of CAD, i.e.
three blinded readers in core laboratories for the presence or
to define patients with perfusion deficits, two criteria were used:
absence of perfusion abnormalities.
first, the presence of a ≥50% diameter stenosis (i.e. ≥75% area sten-
osis) measured in two orthogonal planes as was used in previous
studies7 – 9,17 present in ≥1 coronary artery of ≥2 mm diameter
using a core laboratory (Cleveland Clinic Foundation, Cleveland,
Methods OH, USA). Thus, a ≥50% diameter reduction had to be present in
both orthogonal projections to define CAD. This criterion accounted
Study design and patient population for 94.6% of all patients in this study. Secondly, the history of a previ-
This Phase III clinical trial was conducted at 33 centres in Europe and ous MI was considered. This way, patients after MI (and thus with per-
the USA (see Supplementary material online for a list of participating fusion deficits in scar tissue on resting perfusion images) do fulfil the
sites). Eligible patients were those scheduled for a conventional CXA criteria for the presence of perfusion deficits in CAD [e.g. patients
and/or a SPECT examination for clinical reasons. Before study entry, with successful percutaneous coronary interventions (PCI)/stenting
all patients had to agree to undergo all three imaging studies. Table 1 in the setting of acute MI and consequently non-stenosed coronary ar-
shows the inclusion/exclusion criteria. The study was conducted teries]. This second criterion to define CAD (¼history of previous MI
according to the Declaration of Helsinki, the principles of Good Clin- without significant stenosis on CXA) was relevant for 5.4% of all
ical Practice, and was approved by the Health Authorities and the local patients included. Conversely, patients with a history of successful
Ethics Committee of each participating institution. All patients gave PCI/stenting (with a residual area stenosis ≤75% in the actual CXA)
written informed consent before study participation. and without a history of MI do not fulfil the definition of CAD (and
are assumed to yield normal perfusion studies at stress and at rest).
Vessels of ,2 mm diameter were not considered for definition of
CAD, since such small vessels are rarely treated (e.g. no stents avail-
Efficacy measures able for ,2 mm vessels).
Diagnostic performances of CMR and SPECT were assessed with two
efficacy measures: it was tested for non-inferiority of both sensitivity Cardiac magnetic resonance examination
and specificity for CMR vs. SPECT for the detection of CAD In 1.5 T scanners of various vendors, a breath-hold MR first-pass per-
(¼primary study endpoint). Thus, a binary approach was used, i.e. fusion examination was performed to follow a bolus of 0.075 mmol/kg
reading was assessed at one threshold. As secondary endpoints, sensi- Gd-DTPA-BMA (Omniscan, GE Healthcare, USA) injected into a
tivity, specificity, and negative and positive predictive values for the peripheral vein with power injectors at 5 mL/s (followed by a 25 mL
CMR examination were calculated as well as the safety profile of CMR. saline flush) after 3 min of adenosine infusion (0.14 mg/min/kg i.v.).
MR-IMPACT II 777

The patients had to refrain from coffee, tea, chocolate, or other caf- number of readers with correct diagnosis (true-positive and true-
feinated beverages and food for at least 24 h before the CMR exam. negative, respectively) divided by the number of all readers (¼3) yielding
A contrast medium (CM) dose of 0.075 mmol/kg was chosen accord- values between 0 and 1 (e.g. all three readers correct for true-positive:
ing to recommendations of the food and drug administration (FDA) to sensitivity score SensMR ¼ 1; e.g. one reader of three correct for true-
test the minimal effective dose (i.e. a dose slightly lower than the negative: specificity score SpecMR ¼ 0.33). By using scores of sensitivity
optimum dose in MR-IMPACT I9 used). During bolus arrival, three and specificity, the sensitivity and specificity results of the three readers
short-axis slices were acquired every heart beat at one-fourth, half, were combined into one value/patient, which facilitated subsequent
and three-fourth of the left ventricular (LV) long axis (non-slice select- non-inferiority testing.
ive 908-preparation, fast gradient-echo acquisition with an echo-planar
component where available; spatial resolution: 2– 3 mm × 2 – 3 mm, Single-photon emission computed tomography
slice thickness 8 – 10 mm). An example is given in Figure 1. At the examination
same locations, at 10 and 25 min after the stress imaging, a rest perfu- Stress and rest SPECT examinations were performed according to gen-
sion imaging at the same CM dose and a late enhancement study (with erally accepted guidelines18 on machines of different vendors (two- or
the inversion time nulling normal myocardium at the cumulated CM three-head cameras) with 99mTc- or 201Tl-tracer, adenosine dose as
dose of 0.15 mmol/kg) were performed, respectively. for perfusion-CMR, or physical stress, and 1 or 2 days protocols. The
Cardiac magnetic resonance data were analysed visually by three patients had to refrain from coffee, tea, chocolate, or other caffeinated
blinded readers in an independent core laboratory (Independent beverages and food for at least 24 h before the SPECT exam.
Review Center, GE Healthcare, former Nycomed Amersham Imaging, Gated-SPECT using 99mTc-tracer was strongly recommended, but
Princeton, NJ, USA). The three readers were blinded with respect to ungated acquisitions and/or 201Tl-tracers were accepted if part of the
any clinical information of the patients or results of the other examina- performing institution’s clinical routine. In the efficacy population, i.e.
tions. For the single threshold analysis, a binary assessment of the CMR all three methods completed, gated-SPECT was performed in 253
studies was performed as either showing a perfusion abnormality in any patients. 201Tl-tracer was used in 32 patients (rest and stress) and in 8
of the 16 segments of the heart at rest and/or at stress (abnormal study) additional patients for rest studies only (6.9 and 1.7%, respectively).
or not (normal study). Perfusion abnormalities were defined as myocar- Algorithms for attenuation correction or resolution recovery were
dium being black or dark grey at the peak bolus. Borderline normal per- not applied as these were not available or not identical over all sites.
fusion (myocardium being light grey) was classified together with normal Single-photon emission computed tomography data were analysed
perfusion (myocardium being bright). Additional criteria indicative for visually by three blinded readers using a core laboratory (Beacon
true hypoperfusion vs. artefacts were subendocardial signal reduction Bioscience, Inc., Doylestown, PA, USA). The three readers were
persisting longer than the CM first-pass through the LV cavity, signal re- blinded with respect to any clinical information of the patients or
duction in several slices, and neighbouring regions.9 From this binary results of the other examinations. Each reader was presented with
judgement, sensitivity and specificity scores were defined as the 10 – 12 short-axis as well as 6 – 9 vertical and horizontal long-axis

Figure 1 An example of a 67-year-old patient is shown with angina CCS II. The perfusion-cardiac magnetic resonance study during hyper-
aemia (at 0.075 mmol/kg Gd-DTPA-BMA) demonstrates a perfusion deficit in the subendocardium of the anterolateral wall (B and C; arrows). In
the apical slice, almost the entire subendocardium is hypoperfused during hyperaemia (C, arrows). Another minor perfusion deficit is detected
in the inferior wall of the basal and mid-ventricular slices (A and B, arrowheads). In the basal slice in (A), the inferior wall of the right ventricle
also shows hypoperfusion in comparison with the right ventricular free wall. Single-photon emission computed tomography in this patient was
positive with a predominant perfusion abnormality in the anteroseptal wall (F– K ). Quantitative X-ray coronary angiography (QCA) demon-
strated a severe stenosis in the left anterior descending coronary artery of 82% (LAD; D, arrow). The right coronary artery (RCA) shows
two mild stenoses of 64 and 59% diameter reduction (E, arrows).
778 J. Schwitter et al.

images for both stress and rest conditions. Gated-SPECT data were
also presented to the readers, if they had been acquired. For the
Results
primary endpoint, a binary assessment of the SPECT studies was
performed as either being normal or not according to generally
Patient characteristics
accepted guidelines.18 Specifically, SPECT studies were categorized From the 533 patients enrolled, 515 entered the safety analysis
as either showing a perfusion abnormality in any segment of a (¼patients received MR CM; Figure 2). Of the 465 patients with
17-segment model at rest and/or at stress (¼abnormal study) or data of all three modalities complete (¼efficacy population;
not (¼normal study). Also, patients with a transient ischaemic LV Table 2), 227 (48.8%) had coronary artery stenoses with ≥75%
dilation were categorized as abnormal. From this binary judgement, area reduction, 73 had occlusions (15.7%), 129 (27.7%) had infarc-
sensitivity and specificity scores (SensSPECT and SpecSPECT) were cal- tions, and 25 patients (5.4%) of those with infarctions showed no
culated as described for the CMR data. significant stenoses (,75% area reduction) on CXA. The preva-
lence of CAD in the population without a history of infarction
was 29%. No patients of the previous MR-IMPACT I were included
Safety parameters in the analyses of MR-IMPACT II.
In all patients dosed with the CM, the following was recorded: adverse
events, findings of physical examinations (1 – 24 h before CM adminis-
Non-inferiority analysis: binary sensitivity
tration and 24 h thereafter), vital signs (systolic/diastolic blood pres-
sure, heart rate, and respiratory rate at 10 and 6 min before CM and specificity score: primary endpoint
administration, during adenosine infusion, and at the end of the CMR For this evaluation, 26 (5.6% of 465) CMR and 17 (3.7% of 465,
study, as well as 15, 60 –90 min, and 24 h thereafter), laboratory para- P ¼ 0.21 vs. CMR) SPECT studies were deemed non-evaluable
meters (serum biochemistry, hematology, and coagulation parameters by the MR and SPECT readers, respectively. The prevalence of
within 36 h before and 24 h after first CM administration), and 12-lead CAD on CXA was similar in the studies excluded and included
electrocardiographic (ECG) tracings (immediately before the CMR in the efficacy analysis (21 of 40 vs. 206 of 425, respectively,
study, and 60– 90 min and 24 h thereafter). Core laboratories for
blood sample analyses and 12-lead ECG analyses were CRL-Medinet
Europe, Breda, The Netherlands, and Biomedical Systems (BMS)
Europe, Brussels, Belgium, respectively.

Statistical analysis
Sample size calculation for non-inferiority for sensitivity
and specificity
To meet the efficacy measure for non-inferiority of sensitivity, this cri-
terion had to be met in each of two parallel substudies. Accordingly,
for the primary endpoint of sensitivity, 139 subjects were calculated
to yield a 90% power to show non-inferiority of CMR vs. SPECT at
an equivalence limit difference of 20.1 with a target significance
level of 0.025 (nQuery Advisor 5.0) using an SD of the difference of
0.36 for CMR (derived from the Phase II clinical study: MR-IMPACT1).
Hence, non-inferiority is inferred if the lower bound of the confidence
interval (CI) falls within the equivalence margin of 20.10 (¼10%
non-inferiority margin: H0: SensMR – SensSPECT, 20.1). The criterion
of non-inferiority of specificity had to be met in the entire study
(¼substudies 1 and 2). Accordingly, for the primary endpoint assess-
ment of specificity, the two identical Phase III substudies combined
had to achieve a 90% power to show non-inferiority of CMR vs.
SPECT at an equivalence limit difference of 20.1 with a target signifi-
cance level of 0.025 (nQuery 5.0) expecting 30 – 40% of negative sub-
jects from each of the two Phase III substudies. The substudies 1 and 2
included 238 and 227 patients, respectively (efficacy population ¼
dosed patients with complete data sets).
For the primary endpoint of non-inferiority of CMR vs. SPECT,
non-inferiority was inferred, if the lower bounds of the CIs for the
sensitivity and specificity scores fall within the equivalence margin
of 20.10 (¼10% non-inferiority margin: H0: SensMR – SensSPECT, Figure 2 Flow chart demonstrating the number of eligible
20.1). In the case of superiority, i.e. if the lower bounds of the patients and drop-outs. CMR, cardiac magnetic resonance; CM,
CIs for sensitivity or specificity fall above 0, superiority is reported.19 contrast medium (Gd-DTPA-BMA); CXA, coronary X-ray angi-
All tests were two-sided and a P-value of ,0.05 was considered ography; Pats, patients; SPECT, single-photon emission computed
statistically significant. Statistical analyses were performed using tomography.
SASw software (Version 8.2).
MR-IMPACT II 779

Table 2 Demographics of study population Table 3 Safety profile of the cardiac magnetic
resonance examination
Characteristics n (%)
................................................................................
Safety population (patients received MR contrast 515
All patients enrolled: 515 medium)
Age, years (mean + SD) 60 + 10.3 Serious adverse eventsa 6 (1.2)
Range 26–85 Angina pectoris 1 (0.2)
Body mass index, kg/m2 (mean + SD) 28.2 + 4.3 Prolonged hospital stays 5 (1.0)
Range 16.0–50.0 Death 0 (0)
Male sex 377 (73.2) ................................................................................
Adverse events (in 74 patients) 114 (22.1)
Race: Caucasian 488 (94.8)
................................................................................ Requiring treatment 12 (2.3)
Angina pectoris 414 (80.4) Angina pectoris 4 (0.8)
CCS I 87 (16.9) Headache 4 (0.8)
CCS II 227 (44.1) Chest pain 3 (0.6)
CCS III 46 (8.9) Injection site bruising 1 (0.2)
CCS IV 21 (4.1) Mild 91 (17.7)
................................................................................
Risk factors Moderate 23 (4.5)
Hypertension 358 (69.5) Severe 0 (0)
Hypercholesterolaemia 354 (68.8) Subject withdrawal due to adverse events 0 (0)
Diabetes 92 (17.8)
a
Safety: all six serious adverse events were classified by the treating physician as
Myocardial infarctions 139 (27.0)
not drug-related. Prolonged hospital stays were due to treatment by PCI/CABG of
PCI 170 (33.0) severe CAD during the same hospitalization.
History of heart failure 106 (20.6)
................................................................................
Efficacy population (patients with all 3 test data sets 465
complete): n
Coronary artery disease 227 (48.8) for the difference of 20.17, indicating inferiority of CMR vs.
Left main 14 (3.0) SPECT for specificity. For CMR and SPECT, sensitivities
LAD 134 (28.8) (mean + SD of all readers) were 75 + 7 and 59 + 10%, respect-
LCX 104 (22.4) ively (P ¼ 0.03) and specificities were 59 + 8 and 72 + 14%, re-
RCA 112 (24.1) spectively (P ¼ 0.03). Positive and negative predictive values and
Multivessel disease 113 (24.3) accuracies (mean + SD of all readers) for CMR were 70 + 5,
Myocardial infarction 129 (27.7) 65 + 5, and 68 + 5%, respectively, and for SPECT 73 + 8, 60 +
................................................................................ 3, and 65 + 3%, respectively (no significant differences).
Medication
Any drugs 496 (96.4) Safety profile of the cardiac magnetic
b-Blockers 367 (71.3)
Lipid lowering 354 (68.8)
resonance examinations
Angiotensin-converting enzyme inhibitors 306 (59.4)
In all 515 patients, who received the MR CM, no severe adverse
Diuretics 131 (25.5)
events and no deaths occurred. Table 3 shows the moderate and
Calcium channel blockers 99 (19.2)
mild adverse events. There were no trends for clinically significant
................................................................................ changes in vital signs or ECG changes following MR CM
Antithrombotic 425 (82.6) administration.
MR—not evaluable 26 (5.6)
SPECT—not evaluable 17 (3.7)
Discussion
CCS, Canadian Cardiovascular Society; PCI, percutaneous coronary intervention;
LAD, left anterior descending coronary artery; LCX, left circumflex coronary The main results of the trial can be summarized as follows: (i) the
artery; RCA, right coronary artery primary endpoint of non-inferiority of CMR vs. SPECT for the de-
tection of CAD was met for sensitivity, but not for specificity. (ii)
No severe adverse effects occurred in the 515 patients who
received the MR CM during pharmacological stress CMR.
P ¼ 0.74). When applying a single, i.e. binary threshold, to the
CMR and SPECT images, the sensitivity scores were 0.67 and
0.59, respectively (P ¼ 0.024, paired t-test), with the lower confi-
Perfusion-cardiac magnetic resonance
dence level for the difference of +0.02, indicating superiority of and single-photon emission computed
CMR over SPECT for sensitivity (efficacy population: n ¼ 465). tomography comparison
The specificity score for CMR and SPECT was 0.61 and 0.72, re- This large multicentre perfusion trial demonstrates a higher sensi-
spectively (P ¼ 0.038, paired t-test) with a lower confidence level tivity of perfusion-CMR to detect perfusion deficits in CAD than
780 J. Schwitter et al.

SPECT. Thus, this trial confirms trends and results of earlier multi- In addition, when assessing the current performances of both
centre9 and single-centre studies,11 respectively, and objectivates CMR and SPECT, it should be kept in mind that these results
an important feature of CMR, i.e. to correct for both cardiac were obtained in a fully blinded fashion and, thus, are expected
motion and respiratory motion during the few seconds of CM first- to potentially underestimate test performance in the clinical
pass during hyperaemia. This approach preserves the nominally setting where imaging information is generally interpreted in the
high spatial resolution of perfusion-CMR and thereby allows context of additional patient information.
detecting small even subendocardial perfusion deficits. However,
at a single threshold reading of the CMR and SPECT data, this Safety of perfusion-cardiac magnetic
level of sensitivity was associated with a specificity for CMR resonance examinations
lower than that for SPECT. One may also speculate that this rela-
The results of this MR-IMPACT II confirm the high safety profile
tively low specificity of perfusion-CMR is related to the fact that
reported in earlier CMR perfusion trials8 – 10,12 as no severe
perfusion was compared with the macroscopic coronary artery
adverse events occurred among the 515 patients in MR-IMPACT
anatomy, which does not assess, for example, collateral flow on
II. This finding is in good agreement with the safety profile
the microvascular level.
observed in the European CMR registry with mild adverse reac-
The MR-IMPACT II results are in accord with a previous smaller
tions reported in 0.1% of 7285 stress CMR studies and where
multicentre single-vendor perfusion-CMR study,8 which yielded a
no severe reactions occurred.21 Unlike SPECT, CMR does not
sensitivity and specificity of 93% (95% CI of 77 –99%) and 75%
expose patients to ionizing radiation.22
(95% CI of 48 –92%), respectively, vs. MR-IMPACT II with 75%
(69 –80%) and 59% (52 –65%), respectively. The 95% CIs for sen-
sitivity are overlapping between the current MR-IMPACT II and the
Limitations of the study
previous MR-IMPACT I (69 –80 vs. 69 –93%, respectively); The area stenosis on CXA was used as the reference standard for
however, there is a trend towards slightly lower sensitivity in the the definition of CAD. Coronary artery disease was also present
MR-IMPACT II vs. MR-IMPACT I with 85 vs. 75%, respectively. by definition in a small portion of 5.4% of the study patients
Also specificity showed a trend towards better performance in with a history of infarction where the infarct-related artery was
MR-IMPACT I with 67% vs. MR-IMPACT II with 59%, while the successfully treated by PCI, and thus, the treated vessel was no
95% CIs are overlapping (35–89 vs. 52 –65%, respectively). This longer stenosed. Accordingly, this definition of CAD is primarily
might be related to the larger number of participating sites in dependent on coronary anatomy and it is well known that the
MR-IMPACT II, by which less experienced centres could have con- presence of perfusion deficits is not only dependent on stenoses
tributed to the database. Also, in MR-IMPACT II, a slightly lower of epicardial coronary arteries, but also on collateral flow and
CM dose was used than the most effective dose in MR-IMPACT microcirculatory alterations. Nevertheless, this definition was
I. The current MR-IMPACT II results are also in agreement with deemed best as it is relatively easy to measure, is frequently
the large CE-MARC single-centre CMR study performed in 628 used in such comparative studies, and often sets the basis for
patients and published recently.20 In this CE-MARC trial, a sensitiv- patient management in clinical routine. For future comparative
ity of 80% on the receiver-operator characteristic (ROC) curve studies, however, invasive perfusion assessment by fractional
corresponds to a specificity of 70%.20 flow reserve measurements would be desirable. Importantly, an
The SPECT results of the MR-IMPACT II with a sensitivity optimal patient management should always consider the patient
and specificity of 59 and 72%, respectively, are also in close prognosis. Perfusion techniques are very powerful prognostic
match which those of MR-IMPACT I. In MR-IMPACT I, a sensi- tools, and in this regard, evidence is particularly well established
tivity of 60% corresponded to a specificity of 75% on the for SPECT. The current study results apply for pharmacological
ROC curve, which is also in line with the single-centre stress testing only.
CE-MARC trial, where a SPECT sensitivity of 60% corre- In this trial, patients with decompensated heart failure, after
sponded to a specificity of 70% on the ROC curve.20 Never- bypass surgery, and with relevant arrhythmias were excluded,
theless, one might have expected somewhat better results for and thus, the findings of this study cannot be applied to these
either, SPECT and/or CMR for the detection of CAD in the patient groups. The frequency of CAD in this study was 48.8%,
MR-IMPACT II. The aim of this large MR-IMPACT II was to and therefore, the study results are applicable to patient popula-
assess test performances in a realistic clinical environment tions with an intermediate CAD prevalence, but the trial results
across a substantial number of countries and not to repeat cannot be extrapolated to other populations with lower disease
the results of a few leading high-performance centres. To this prevalence, e.g. to asymptomatic screening populations.
end, it was crucial to accept in the protocol a certain range For this evaluation, 26 (5.6% of 465) CMR and 17 (3.7% of 465,
of imaging parameters as long as they were in agreement P ¼ 0.21 vs. CMR) SPECT studies were deemed non-evaluable by
with key features of the technique (such as spatial and temporal the MR and SPECT readers, respectively, which limit the applicabil-
resolution for CMR and concordance with established guidelines ity of the results to patients with evaluable studies.
for SPECT). Of note, gated SPECT was performed in 54% of
the patients and attenuation correction algorithms were not
applied (due to a lack of standardized criteria). These circum-
Conclusions
stances should be taken into account when considering the This large international, multicentre, multivendor, prospective trial
overall SPECT performance. performed at 33 centres demonstrates a high performance of
MR-IMPACT II 781

perfusion-CMR to detect CAD. Sensitivity of perfusion-CMR was 8. Giang T, Nanz D, Coulden R, Friedrich M, Graves M, Al-Saadi N, Lüscher T, von
Schulthess G, Schwitter J. Detection of coronary artery disease by magnetic res-
superior to perfusion SPECT, while specificity of perfusion-CMR
onance myocardial perfusion imaging with various contrast medium doses: first
was inferior in comparison to SPECT. In selected patients (no European multicenter experience. Eur Heart J 2004;25:1657 – 1665.
severe arrhythmias), CMR is a safe approach and an alternative 9. Schwitter J, Wacker C, van Rossum A, Lombardi M, Al-Saadi N, Ahlstrom H,
Dill T, Larsson HB, Flamm S, Marquardt M, Johansson L. MR-IMPACT: compari-
to SPECT to detect perfusion abnormalities in CAD.
son of perfusion-cardiac magnetic resonance with single-photon emission com-
puted tomography for the detection of coronary artery disease in a
multicentre, multivendor, randomized trial. Eur Heart J 2008;29:480 – 489.
Supplementary material 10. Plein S, Kozerke S, Suerder D, Luescher TF, Greenwood JP, Boesiger P,
Schwitter J. High spatial resolution myocardial perfusion cardiac magnetic reson-
Supplementary material is available at European Heart Journal ance for the detection of coronary artery disease. Eur Heart J 2008;29:
online. 2148 –2155.
11. Ishida N, Sakuma H, Motoyasu M, Okinaka T, Isaka N, Nakano T, Takeda K. Non-
infarcted myocardium: correlation between dynamic first-pass contrast-enhanced
Funding myocardial MR imaging and quantitative coronary angiography. Radiology 2003;
229:209–216.
This work was supported by GE Healthcare (formerly Amersham 12. Nagel E, Klein C, Paetsch I, Hettwer S, Schnackenburg B, Wegscheider K, Fleck E.
Health). Magnetic resonance perfusion measurements for the noninvasive detection of
coronary artery disease. Circulation 2003;108:432 – 437.
Conflict of interest: J.S. and N.A.-S. served as consultants for GE 13. Al-Saadi N, Nagel E, Gross M, Bornstedt A, Schnackenburg B, Klein C, Klimek W,
Healthcare (former Amersham Health) and received honoraria. A.L. Oswald H, Fleck E. Noninvasive detection of myocardial ischemia from perfusion
reserve based on cardiovascular magnetic resonance. Circulation 2000;101:
received consultancy honorarium from GE Healthcare, and O.S. is a
1379 –1383.
consultant of Circle (a software company not involved in data analysis 14. Jahnke C, Nagel E, Gebker R, Kokocinski T, Kelle S, Manka R, Fleck E, Paetsch I.
in this trial). N.H. was an employee of GE Healthcare and was respon- Prognostic value of cardiac magnetic resonance stress tests: adenosine stress per-
sible for the statistical analyses (current employer: INC Research fusion and dobutamine stress wall motion imaging. Circulation 2007;115:
GmbH, Munich, Germany). 1769 –1776.
15. Bruder O, Schneider S, Nothnagel D, Dill T, Hombach V, Schulz-Menger J,
Nagel E, Lombardi M, van Rossum A, Wagner A, Schwitter J, Senges J, Sabin G,
References Sechtem U, Mahrholdt H. EuroCMR (European Cardiovascular Magnetic Reson-
1. Heart disease and stroke statistics: update 2009. Circulation 2009;119:e21 –e181. ance) Registry. J Am Coll Cardiol 2009;54:1457 –1466.
2. Allender S, Scarborough P, Peto V, Rayner M, Leal J, Luengo-Fernandez R, Gray A. 16. Wagner A, Mahrholdt H, Holly T, Kim RJ. Contrast-enhanced MRI and routine
European Cardiovascular Disease Statistics—2008 Edition. European Heart single photon emission computed tomography (SPECT) perfusion imaging for de-
Network, www.ehnheart.org/cdv-statistics.html: 1– 112. tection of subendocardial infarcts: an imaging study. Lancet 2003;361:374 –379.
3. Iskander S, Iskandrian AE. Risk assessment using single-photon emission com- 17. Hendel RC, Berman DS, Cullom SJ, Follansbee W, Heller GV, Kiat H, Groch MW,
puted tomographic technetium-99m sestamibi imaging. J Am Coll Cardiol 1998; Mahmarian JJ. Multicenter clinical trial to evaluate the efficacy of correction for
32:57–62. photon attenuation and scatter in SPECT myocardial perfusion imaging. Circulation
4. Smith SJ, Feldman T, Hirshfeld JJ, Jacobs A, Kern M, King S III, Morrison D, 1999;99:2742 – 2749.
O’Neill W, Schaff H, Whitlow P, Williams D. ACC/AHA/SCAI 2005 guidelines 18. Imaging guidelines for nuclear cardiology procedures, part 1: American Society of
update for percutaneous coronary intervention: a report of the American Nuclear Cardiology. J Nucl Cardiol 1996;3:G1 –G46.
College of Cardiology/American Heart Association Task Force on Practice 19. Piaggio G, Elbourne D, Altman D, Pocock S, Evans S, for the CONSORT Group.
Guidelines (ACC/AHA/SCAI Writing Committee to Update 2001 Guidelines Reporting of noninferiority and equivalence randomized trials. J Am Med Assoc
for Percutaneous Coronary Intervention). Circulation 2006;113:e166– e286. 2006;295:1152 – 1160.
5. Silber S, Albertsson P, Aviles F, Camici P, Colombo A, Hamm C, Jorgensen E, 20. Greenwood J, Maredia N, Radjenovic A, Brown J, Nixon J, Farrin A, Dickinson C,
Marco J, Nordrehaug J-E, Ruzyllo W, Urban P, Stone G, Wijns W. Guidelines Younger J, Ridgway J, Sculpher M, Ball S, Plein S. Clinical evaluation of magnetic
for percutaneous coronary interventions. Eur Heart J 2005;26:804 –847. resonance imaging in coronary heart disease: the CE MARC study. Lancet 2011;
6. Shaw LJ, Hachamovitch R, Berman D, Marwick T, Lauer M, Heller G, Iskandrian A, doi:10.1016/S0140 –6736(11)61335 – 4.
Kesler K, Travin M, Lewin H, Hendel R, Borges-Neto S, Miller D. The economic 21. Bruder O, Schneider S, Nothnagel D, Pilz G, Lombardi M, Sinha A, Wagner A,
consequence of available diagnostic and prognostic strategies for the evaluation of Dill T, Frank H, van Rossum A, Schwitter J, Nagel E, Senges J, Sabin G,
stable angina patients: an observational assessment of the value of precatheteriza- Sechtem U, Mahrholdt H. Acute adverse reactions to gadolinium based contrast
tion ischemia. J Am Coll Cardiol 1999;33:661 –669. agents in cardiovascular MR. Multicenter experience with 17767 patients from the
7. Schwitter J, Nanz D, Kneifel S, Bertschinger K, Buchi M, Knusel PR, Marincek B, EuroCMR registry. J Am Coll Cardiol Cardiovascular Imaging 2011 (in press).
Luscher TF, von Schulthess GK. Assessment of myocardial perfusion in coronary 22. Gerber T, Carr J, Arai A, Dixon R, Ferrari V, Gomes A, Heller G, McCollough C,
artery disease by magnetic resonance: a comparison with positron emission tom- McNitt-Gray M, Mettler F, Mieres J, Morin R, Yester M. Ionizing radiation in
ography and coronary angiography. Circulation 2001;103:2230 –2235. cardiac imaging. Circulation 2009;119:1056 –1065.