MRI of the Spine A Guide for Orthopedic Surgeons, 1st

Edition

Visit the link below to download the full version of this book:

https://medipdf.com/product/mri-of-the-spine-a-guide-for-orthopedic-surgeons-1st
-edition/

Click Download Now

For Binx and Boots
–William B. Morrison
Preface

MRI of the Spine is a book concept that arose from a spine imaging instruc-
tional course at the annual meeting of the American Academy of Orthopaedic
Surgeons. The course has been the only dedicated radiology-based course at
the AAOS meeting over the past decade, with radiology faculty and organiz-
ers. This unique arrangement reflects a need in the orthopaedic community,
education on MR imaging, which has become an essential component of the
diagnostic algorithm. Spine imaging incorporates degeneration, trauma,
inflammatory conditions and infection, metabolic conditions, as well as dural
and cord lesions, including neoplasia. An understanding of the MR imaging
appearances of these conditions will assist the orthopaedist in understanding
the nature and extent of the patient’s condition. Following surgery and altera-
tion of anatomy, MR imaging can be especially challenging to interpret. This
is explored, with expected post-­operative appearances and findings suggest-
ing recurrent or new pathology. An understanding of methods for acquiring
the images can also be helpful – what protocol to use and utility of contrast in
different situations. To this end, we also provide tips for optimization of the
MR imaging protocol. The editors and authors are musculoskeletal radiolo-
gists and neuroradiologists who work closely with surgeons in direct patient
care. Our goal is to create a bridge between the radiologist and the orthopae-
dist, together using cutting-­edge technology to better care for our patients.

Philadelphia, PA, USA William B. Morrison

New York, NY, USA John A. Carrino
 Adam E. Flanders

vii
Contents

1 MRI Protocol������������������������������������������������������������������������������������   1

Vishal Desai and Jehan Ghany
2 MRI in Spine Anatomy�������������������������������������������������������������������� 17
Ajit Karambelkar
3 MRI in Spine Trauma���������������������������������������������������������������������� 31
Phan Q. Duy, Ichiro Ikuta, Michele H. Johnson,
Melissa Davis, and Vahe M. Zohrabian
4 MRI in Degenerative Disease of the Spine������������������������������������ 87
Alessandra J. Sax
5 MRI in Spine Infection�������������������������������������������������������������������� 107
M. K. Jesse and Corey K. Ho
6 MRI in Non-infectious Inflammation and Arthropathies������������ 129
Sachin Dheer
7 MRI in Metabolic Disease �������������������������������������������������������������� 159
Ricardo Hernandez, Philip K. Wong, Monica Umpierrez,
and Felix M. Gonzalez
8 MRI in Neoplastic Bone Disease and Differential
Considerations���������������������������������������������������������������������������������� 169
John V. Dennison, Alexander Leyva, Andrew T. Cibulas,
Kurt F. Scherer, Jack A. Porrino, Sean C. Dodson,
Richard D. Beegle, and Laura W. Bancroft
9 MRI in Dural Lesions���������������������������������������������������������������������� 189
Mougnyan Cox
10 MRI in Cord Lesions ���������������������������������������������������������������������� 207
Kofi-Buaku Atsina
11 Principles of Postoperative Spine MRI������������������������������������������ 237
Karthik Krishnan, Sophie C. Queler, and Darryl B. Sneag
12 Identification of Complications Using Postoperative
Spine MRI���������������������������������������������������������������������������������������� 253
Prabath Kumar Mondel
Index���������������������������������������������������������������������������������������������������������� 261

ix
Contributors

Kofi-Buaku Atsina, MD Department of Radiology, Thomas Jefferson

University Hospital, Philadelphia, PA, USA
Laura W. Bancroft, MD, FACR University of Central Florida School
of Medicine, Orlando, FL, USA
Florida State University School of Medicine, AdventHealth, Orlando, FL,
USA
Richard D. Beegle, MD Department of Diagnostic Radiology, AdventHealth,
Orlando, FL, USA
Andrew T. Cibulas, MD AdventHealth GME Radiology, Orlando, FL, USA
Mougnyan Cox, MD Hospital of the University of Pennsylvania,
Philadelphia, PA, USA
Melissa Davis, MD, MBA Department of Radiology & Biomedical Imaging,
Yale School of Medicine, New Haven, CT, USA
John V. Dennison, MD AdventHealth GME Radiology, Orlando, FL, USA
Vishal Desai, MD Thomas Jefferson University, Philadelphia, PA, USA
Sachin Dheer, MD Department of Radiology, Thomas Jefferson University
Hospital, Jefferson Health, Philadelphia, PA, USA
Sean C. Dodson, MD Department of Diagnostic Radiology, AdventHealth,
Orlando, FL, USA
Phan Q. Duy, BS Department of Radiology & Biomedical Imaging, Yale
School of Medicine, New Haven, CT, USA
Jehan Ghany, MD Thomas Jefferson University, Philadelphia, PA, USA
Felix M. Gonzalez, MD Department of Radiology and Imaging Sciences,
Emory University Hospital, Atlanta, GA, USA
Emory University Orthopaedics & Spine Center, Atlanta, GA, USA
Ricardo Hernandez, MS-III Philadelphia College of Osteopathic Medicine,
Suwanee, GA, USA
Corey K. Ho, MD University of Colorado Anschutz Campus, Aurora,
CO, USA

xi
xii Contributors

Ichiro Ikuta, MD Department of Radiology & Biomedical Imaging,

Yale School of Medicine, New Haven, CT, USA
M. K. Jesse, MD University of Colorado Anschutz Campus, Aurora, CO,
USA
Michele H. Johnson, MD Department of Radiology & Biomedical Imaging,
Yale School of Medicine, New Haven, CT, USA
Ajit Karambelkar, MBBS, MD Thomas Jefferson University Hospital
Philadelphia, Philadelphia, PA, USA
Karthik Krishnan, MS Weill Medical College of Cornell University,
New York, NY, USA
Department of Radiology and Imaging, Hospital for Special Surgery,
New York, NY, USA
Alexander Leyva, MD AdventHealth GME Radiology, Orlando, FL, USA
Prabath Kumar Mondel, MBBS, MD Philadelphia, PA, USA
Jack A. Porrino, MD Yale School of Medicine – New Haven Hospital,
New Haven, CT, USA
Sophie C. Queler, BA Department of Radiology and Imaging, Hospital for
Special Surgery, New York, NY, USA
Alessandra J. Sax, MD Department of Radiology, Thomas Jefferson
University Hospital, Philadelphia, PA, USA
Kurt F. Scherer, MD University of Central Florida School of Medicine,
Orlando, FL, USA
Florida State University School of Medicine, AdventHealth, Orlando,
FL, USA
Darryl B. Sneag, MD Weill Medical College of Cornell University,
New York, NY, USA
Department of Radiology and Imaging, Hospital for Special Surgery,
New York, NY, USA
Monica Umpierrez, MD Department of Radiology and Imaging Sciences,
Emory University Hospital, Atlanta, GA, USA
Philip K. Wong, MD Department of Radiology and Imaging Sciences,
Emory University Hospital, Atlanta, GA, USA
Vahe M. Zohrabian, MD Department of Radiology & Biomedical Imaging,
Yale School of Medicine, New Haven, CT, USA
 MRI Protocol
1
Vishal Desai and Jehan Ghany

Introduction MRI Physics

Magnetic resonance imaging (MRI) is the main- To get a better understanding of the commonly
stay for noninvasive evaluation of the spine, pro- performed MRI sequences and what information
viding detailed anatomical assessment and can be extracted from each, an overview of MRI
excellent sensitivity for pathology, including physics is helpful [1]. MRI utilizes the body’s
degenerative disc disease, tumors, infection, natural magnetic properties for imaging, specifi-
bone marrow processes, spinal cord abnormali- cally the hydrogen nucleus due to its prevalence
ties, traumatic injuries, and compression frac- throughout the body in water and fat.
tures. Unlike other imaging modalities, MRI can
evaluate the spinal cord, meninges, cerebrospi-
nal fluid, marrow, and supporting structures in Magnetic Field
one routine study.
Advanced MR imaging can be obtained with Hydrogen protons contain a net positive charge,
additional sequences and/or with intravenous providing them with their own magnetic spins
contrast to gather more information, help with and a local magnetic field. With the patient in
troubleshooting, or assist in evaluating patients the MRI scanner, a uniform magnetic field is
with prior spinal surgery with or without hard- applied to the protons in the slice or slab of
ware. The high yield of MRI and the lack of ion- interest, causing the randomly oriented protons
izing radiation make it the imaging modality of to now align parallel to the external magnetic
choice for the spine in nearly all populations and field, precess at a certain frequency, and contain
for most indications. a net magnetization in the longitudinal direc-
tion (Fig. 1.1).

RF Pulse

Next, a radiofrequency (RF) pulse or series of RF

pulses is applied to the protons, dependent on the
sequence and information desired. The energy
V. Desai (*) · J. Ghany from the RF pulse is absorbed by the protons,
Thomas Jefferson University, Philadelphia, PA, USA
e-mail: [email protected]; causing the net magnetization to tilt away
[email protected] from the longitudinal direction (Fig. 1.2).

© Springer Nature Switzerland AG 2020 1

W. B. Morrison et al. (eds.), MRI of the Spine, https://doi.org/10.1007/978-3-030-43627-8_1
2 V. Desai and J. Ghany

+ +
+ +

+
+

External magnetic field

+

+ +
+
+

Fig. 1.1 With the application of an external magnetic field, the previously randomly oriented protons are now aligned
and have a net magnetization

a z
b z
RF Energy
Longitudinal
direction

External External
magnetic magnetic
field field

y y
Transverse plane

x x
c z d z

External External
magnetic magnetic
field field

y y

x x

Fig. 1.2 After the protons are aligned by the external decay in the transverse plane and the time to relax to the
magnetic field (a), RF energy is applied by the scanner to longitudinal direction are related to the property of the tis-
tilt the protons into the transverse plane (b). As the pro- sues – T2 and T1, respectively (c, d). This provides impor-
tons move back to their lower energy state, the time to tant imaging characteristics
1 MRI Protocol 3

Again, the extent of this “tilt” depends on the RF Image Quality

pulse and can be manipulated to obtain different
tissue characteristics. An optimal imaging protocol considers the clini-
cal question and tailors the examination to answer
it, with high yield sequences acquired in the min-
Relaxation imum amount of time and with excellent quality
to ensure diagnostic adequacy. Quality is deter-
As with any structure with high energy, there is mined by image resolution, image contrast,
a natural tendency toward a lower energy state. signal-­to-noise ratio, and lack of artifacts. Fine-­
With MRI, this means the protons which were tuning imaging protocols are required to find the
tilted and rotated by the RF pulse (high energy appropriate balance between resolution, contrast,
state) will realign with the magnetic field (low signal, noise, and overall study time.
energy state). The time taken for a particular
tissue to relax in the longitudinal direction is
referred as the T1 relaxation rate. The time Resolution
taken for transverse magnetization to decay to
37% of its initial value is referred to as the T2 Image resolution determines the ability to evalu-
relaxation rate (Fig. 1.2). Both T1 and T2 relax- ate small structures or pathologies and can be
ation occur simultaneously based on the intrin- regarded as the level of detail in the scan. A high-­
sic properties of the excited tissue and its local resolution image is able to distinguish between
environment. adjacent structures, whereas a low-resolution
As the tissues relax toward the lower energy image would blur them together (Fig. 1.3). This
states, RF energy is emitted and can be mea- depends on the size of the image pixel (or voxel
sured. The time from the delivery of the RF for 3D sequences), which in turn depends on the
pulse and peak of the signal emitted (also matrix size, field of view, and slice thickness.
known as an echo) is referred to as the “time to Increasing the matrix size increases the total
echo” (TE). The time interval between each number of pixels/voxels, which means a higher-­
excitation pulse is referred to as the “repetition resolution image but at the cost of more time and
time” (TR). These parameters (time, duration, with less signal in each voxel. A larger field of view
and sequence of RF pulses) can be manipulated may be required to image the entire region of inter-
to create different types of images to character- est (such as the thoracic spine), but that means that
ize different structures and help answer the each voxel now contains more area, decreasing spa-
clinical question. tial resolution. Similarly, increasing slice thickness
covers more area, decreasing spatial resolution.

Image Formation
Signal-to-Noise Ratio
The last step is to translate the received signal
into an image. The signal is received as frequency For any imaging modality, the goal is to obtain the
information, encoded with location information highest amount of signal with the least amount of
and intensity determined by the tissue character- noise. For MRI, this often comes at the expense of
istics and sequence parameters. Through Fourier spatial resolution, as more signal is obtained when
transformation, this frequency information is the voxel size is larger (increased slice thickness,
converted into shades of gray in a matrix of pix- smaller matrix, and larger field of view). Additional
els, which forms an image. strategies to increase signal and decrease noise
4 V. Desai and J. Ghany

a b

Fig. 1.3 An example of a high-resolution MRI image (a) compared to a low-resolution image (b), in which discerning
small structures is difficult

without affecting spatial resolution include increas-

ing the number of ­excitations (NEX) and utilizing
RF coils. By increasing the NEX, more signal is
received per acquisition and allows for averaging
of the signal for a higher-quality, less noisy image.
Proper coil selection is a crucial but often
overlooked concept in MR imaging. There are a
wide variety and multiple configurations of coils
available, including several optimized for spinal
imaging. The coil should sit closely against the
area of interest and should be the smallest size
possible. RF coils serve as the “antenna” for the
MR scanner. By optimizing the shape, size, and
location of this antenna, the signal received back
from the body part being imaged is greater,
improving image quality.

Contrast

In order to detect pathology, the image must be

able to display a difference, or contrast, in signal
intensity between normal and abnormal tissue.
MRI inherently has high contrast sensitivity and Fig. 1.4 An example of T2-weighted sequence demon-
is adept at demonstrating differences in tissues in strating the excellent contrast resolution of MRI, with a
the body, both anatomy and pathology (Fig. 1.4). clear distinction between CSF, fat, and soft tissues
1 MRI Protocol 5

Even though MRI has high contrast at baseline, MRI scanners are available in both open and
the parameters must be optimized for each exam- closed configurations. Open MRI scanners often
ination. By adjusting the TR and TE, the images have a lower field strength and several limita-
can highlight tissue differences on T1-weighted, tions, including a longer scan time which can
T2-weighted, and proton density sequences. predispose to motion artifact, poorer fat suppres-
More advanced sequences can evaluate and dis- sion, and a wider field of view to collect more
play other tissue characteristics if desired. signal. These limitations can contribute to lower-­
image quality compared to traditional scanners.
As a result, open scanners should be reserved for
Field Strength claustrophobic and obese patients who may not
fit in or cannot tolerate a closed system.
Besides intrinsic tissue contrast, several external
factors can impact imaging quality, with one of
the major ones being the primary property of the Artifacts
scanner – magnetic field strength. Magnetic field
strength varies from 0.2 Tesla to 3 Tesla for clini- There are numerous artifacts that can impact image
cal applications, with 1.5 Tesla the most com- quality on MRI, which can be attributable to MR
monly available and most commonly used. The hardware and software, shielding of the MRI scan-
higher the field strength, the better the contrast, ner room and magnetic field inhomogeneities, tis-
the higher the resolution, and the higher the sue heterogeneity, foreign bodies, patient motion,
signal-­to-noise ratio. This means that a stronger and physiologic motion [2] (Fig. 1.5).
magnet improves overall image quality and thus Based on the type of artifact, solutions exist
diagnostic ability. to remove or at least partially correct for the

a b

Fig. 1.5 (a) Motion artifact from the patient moving during image acquisition introduces blur and artifactual lines on
the image. (b) Susceptibility artifact from dental hardware limits evaluation of the upper cervical spine
6 V. Desai and J. Ghany

artifact to improve overall image quality. No

contrast
Artifacts could contribute to confusing imaging FAT
Max
contrast

Longitudinal Magnetization
findings (pseudo-­masses) or preclude diagnostic
evaluation (extensive susceptibility artifact from
hardware). While a full discussion on MRI arti-
facts is beyond the scope of this chapter, recog- CSF
nizing that they exist and factors contributing to
them can be helpful in image optimization and
study interpretation.

Routine Sequences Time

Short TR Long TR
The routine imaging protocol for a spine MRI
Fig. 1.6 A short TR and short TE maximize T1 tissue
should include a T1-weighted sequence (excel- contrast
lent for evaluating anatomy and bone marrow), a
T2-weighted sequence (for anatomy and pathol-
ogy), and a fluid-sensitive sequence (excellent for Practical Application
detecting pathology). At the minimum, a spine This sequence is universally utilized in nearly all
protocol should contain sequences in both the MR imaging protocols, in part due to its very high
sagittal and axial plane. signal-to-noise ratio, and is primarily used for
evaluation of anatomy [3]. Due to the bright
appearance of fat on this sequence, T1-weighted
T1-Weighted Sequence imaging is especially useful for evaluation of nor-
mal bone marrow in the spine. Normal marrow in
Sequences with T1 weighting evaluate differ- adults is predominantly fatty (yellow marrow)
ences in the T1 contrast of tissues. After excita- and should appear hyperintense. Pathologic mar-
tion, tissues relax back to equilibrium at different row, which can be seen with lymphoma, leuke-
rates based on its intrinsic characteristics. mia, metastases, infection, and other infiltrative
Relaxation in the longitudinal direction provides process, results in hypointense marrow. A rule of
the T1 property of the tissue. thumb for spinal imaging is that normal bone
marrow on T1-weighted imaging should be
Physics brighter than the disc space; otherwise this should
In general, T1-weighted imaging utilizes a short raise a red flag for an infiltrative process (Fig. 1.7).
TR and short TE to maximize tissue contrast Besides fat, other structures that appear hyperin-
(Fig. 1.6). Fat, for instance, rapidly realigns from tense on T1-weighted imaging include methemoglo-
the longitudinal direction back to equilibrium, bin, melanin, slow-flowing blood, and proteinaceous
yielding a higher signal and thus appearing fluid. T1-weighted imaging is also utilized after
bright, or “hyperintense,” on T1-weighted imag- administration of IV contrast agents (such as gado-
ing. Conversely, water realigns slowly from the linium), which will be discussed separately.
longitudinal direction, yielding less signal and
thus appearing dark, or “hypointense,” on
T1-weighted imaging. Without a short TR, all of T2-Weighted Sequence
the protons would relax back to equilibrium
yielding an image with equal intensity for all tis- T2-weighted imaging is another mandatory
sues; thus, a short TR is required for T1-weighted sequence for nearly all MR imaging and is useful
imaging to achieve tissue contrast. for evaluating both anatomy and pathology. This
1 MRI Protocol 7

a b

Fig. 1.7 (a) T1-weighted image demonstrating normal The vertebral body appears darker than the adjacent disk,
fatty bone marrow. (b) T1-weighted image in a different a sign of a pathologic marrow process, which in this case
patient shows replacement of the normal fatty marrow. was lymphoma

sequence evaluates differences in the T2 contrast

of tissues, another inherent property (although it
Transverse Magnetization

can be somewhat impacted by magnetic field

inhomogeneities). Whereas T1-weighted imag-
ing relies on relaxation in the longitudinal direc- CSF
tion, T2-weighted imaging relies on decay in the
transverse direction [3]. FAT

Max
Physics contrast
For T2-weighted imaging, long TR and TE times
are required to accentuate differences in tissue Time
contrast (Fig. 1.8). Water has a long T2 relaxation Long TE
time, resulting in a hyperintense appearance on a Fig. 1.8 A long TR and long TE maximize T2 tissue
T2-weighted image. Fat also appears bright on contrast
T2-weighted imaging.

Practical Applications loss of normal CSF space, such as in the case of a

For spine imaging, T2-weighted sequences allow disc herniation narrowing the central canal or
excellent visualization of the cerebrospinal fluid neural foramen, will be readily apparent on this
(CSF) due to its high water content. As a result, sequence (Fig. 1.9).
8 V. Desai and J. Ghany

Most pathologic conditions – tumors, infections,

a
inflammatory process – have a component of edema
and as a result appear hyperintense on T2-weighted
imaging. For instance, an epidural collection will
generally appear bright or intermediate in signal on
this sequence, perhaps with heterogeneity in the
case of infection or blood products.

Fluid-Sensitive Sequence

While routine T2-weighted image will demon-

strate fluid (and, as a result, most pathology) as
hyperintense, fat also appears hyperintense which
can make identifying and characterizing the
abnormality difficult. Thus to get a truly fluid-­
sensitive sequence, different techniques can be
applied to suppress the fat and highlight the pres-
ence of edema and pathologic tissues (Fig. 1.10).

Fat-Saturated T2-Weighted Sequence

The most commonly used technique to achieve a
fluid-sensitive image is frequency-selective fat
suppression. This method applies an RF pulse to
b
the slice at the same resonance frequency of lipids
to saturate all tissues with fat followed by a gradi-
ent pulse to nullify any signal from the lipid [4].
The advantage of this technique is that it can
be applied to any sequence to suppress fat,
including post-contrast imaging. However, it is
highly susceptible to inhomogeneities in the
magnetic field, which can lead to failure in fat
suppression. Fat saturation techniques should not
be used when metallic hardware is present in the
area of interest (such as fixation devices in the
spine) due to the resulting artifact.

STIR Sequence
An alternative and also commonly used sequence for
fluid-sensitive imaging of the spine is short tau inver-
Fig. 1.9 T2-weighted imaging is excellent for evaluation
of degenerative disc disease due to contrast in disc mate-
sion recovery, or STIR. Since fat has such a short T1
rial, CSF, and nerve roots. There is a large disc extrusion relaxation time, shorter than most other tissues, its
seen on sagittal (a) and axial (b) T2-weighted sequences signal can be selectively nullified without impacting
in this patient at the L3-L4 level other tissues through the use of RF pulses [4].
1 MRI Protocol 9

a b

Fig. 1.10 Fluid-sensitive imaging (STIR, in this exam- fracture readily apparent (a) compared to non-fat-­
ple) is helpful to highlight pathology, particularly edema, suppressed imaging (b)
which in this case makes the pars interarticularis stress

STIR imaging is significantly less susceptible to Imaging Planes

inhomogeneities in the magnetic field, leading to Routine imaging of the spine should include
more uniform fat suppression. This is the preferred T1-weighted, T2-weighted, and fluid sensitive
method for fluid-sensitive imaging in the presence (fat-saturated or STIR) sequences in the sagittal
of hardware. The only disadvantage is that the fat imaging plane. This combination allows for eval-
suppression is not selective for lipids but applies to uation of the bone marrow, CSF, and for most
any tissues with short T1, such as melanin, mucus, pathologies. Axial imaging with at least
and, of particular interest in spine pathology, met- T2-weighting should also be performed to allow
hemoglobin. For the same reason, STIR should not for evaluation in more than one plane. Often, a
be used for post-contrast imaging, as the T1 relax- T2-weighted sequence in an axial plane relative
ation properties of c­ ontrast are similar to fat and to the disc space is included to improve evalua-
would result in signal loss of both. tion of degenerative disc disease (Fig. 1.11).