Black-Blood Steady-State Free Precession (SSFP)
Coronary Wall MRI for Cardiac Allografts:
A Feasibility Study
Kai Lin, MD, MS,1Xiaoming Bi, PhD,2Ying Liu, MD, PhD,1Kirsi Taimen, MD,1
Biao Lu, MD,1,3Debiao Li, PhD,1and James Carr, MD1*
Purpose: To assess the hypothesis that steady-state free
procession (SSFP) allows for imaging of the coronary wall
under the conditions of fast heart rate in heart transplan-
tation (HTx) patients.
Materials and Methods: With the approval of our Institu-
tional Review Board, 28 HTx patients were scanned with
a 1.5T scanner. Cross-sectional black-blood images of the
proximal portions of the left main artery, left anterior de-
scending artery, and right coronary artery were acquired
with both a 2D, double inversion recovery (DIR) prepared
turbo (fast) spin echo (TSE) sequence and a 2D DIR SSFP
sequence. Image quality (scored 0–3), vessel wall area,
thickness, signal-to-noise ratio (SNR, vessel wall), and
contrast-to-noise ratio (CNR, wall-lumen) were compared
between TSE and SSFP.
Results: The overall image quality of SSFP was higher
than TSE (1.23 6 0.95 vs. 0.88 6 0.69, P < 0.001). SSFP
had a higher coronary wall SNR (20.1 6 8.5 vs. 14.9 6
4.8, P < 0.001) and wall-lumen CNR (8.2 6 4.6 vs. 6.8 6
3.7, P ¼ 0.005) than TSE.
Conclusion: Black-blood SSFP coronary wall MRI pro-
vides higher image quality, SNR, and CNR than tradi-
tional TSE does in HTx recipients. It has the potential to
become an alternative means to noninvasive imaging of
Key Words: coronary wall MRI; steady state free preces-
sion; heart transplantation
J. Magn. Reson. Imaging 2012;35:1210–1215.
C 2012 Wiley Periodicals, Inc.
HEART TRANSPLANTATION (HTx) is the final life-sav-
ing resort to treat endstage heart failure. As a conse-
quence of this treatment, cardiac allograft vasculop-
athy (CAV) is seriously affecting long-term survival of
HTx recipients. The typical pathological change of
CAV is gradual thickening of the vessel wall through
the whole coronary tree (1).
HTx recipients require periodic follow-up exams to
monitor development of CAV. Currently, the most pop-
ular clinical examinations allowing for directly detect-
ing coronary wall are intravascular ultrasound (IVUS)
and x-ray angiography. Computed tomography (CT)
has also been used to observe coronary artery in car-
diac allografts (2). However, IVUS and x-ray angiography
are invasive and CT carries additional concerns of radi-
ation exposure. Therefore, magnetic resonance imaging
(MRI) has emerged as a promising solution for noninva-
sively detecting morphological abnormalities of the coro-
nary arteries since it is beneficial if the cumulative x-ray
dose can be avoided by using MRI (3). Currently, black-
blood 2D turbo (fast) spin echo (TSE) sequence has
been routinely used for coronary wall imaging (3). How-
ever, cardiac motion, which is strongly correlated with
heart rate, is a major threat for image quality (4,5).
Generally, heart rate is much faster in recipients of
HTx than in healthy controls because of denervation
or subnormal reinnervation of the donor heart (6).
Unfortunately, the TSE sequence is very sensitive to
cardiac motion (7); such a disadvantage therefore
impedes its clinical application in HTx recipients.
Steady-state free precession (SSFP), a fast MRI tech-
nique, is a promising technique to address such a clin-
ical need. SSFP is more time-efficient (8) and can uti-
lize a shorter time window within each cardiac cycle.
Such an advantage is valuable in coronary wall MRI
since images are acquired in a segmented fashion.
Currently, it has been proven that two-dimensional
(2D) black-blood SSFP MRI has better performance
than TSE on healthy volunteers under conditions of
fast heart rate (>80 beats/min) (9). The aim of our
study was to prospectively test the hypothesis that 2D
double inversion recovery (DIR)-prepared SSFP MRI
sequence can be applied to image coronary wall in car-
1Department of Radiology, Northwestern University, Chicago, Illinois,
2Cardiovascular MR R&D, Siemens Healthcare, Chicago, Illinois, USA.
3Department of Radiology, Beijing Anzhen Hospital, Capital Medical
University, Beijing, P.R. China.
Contract grant sponsor: National Institutes of Health (NIH); Contract
grant number: R01HL089695; Contract grant sponsor: American
Heart Association; Contract grant number: 10CRP3050051.
*Address reprint requests to: J.C., Department of Radiology, North-
western University, 737 N. Michigan Ave., Suite 1600, Chicago, IL
60611. E-mail: email@example.com
Received March 23, 2011; Accepted November 29, 2011.
View this article online at wileyonlinelibrary.com.
JOURNAL OF MAGNETIC RESONANCE IMAGING 35:1210–1215 (2012)
C 2012 Wiley Periodicals, Inc.
MATERIALS AND METHODS
The study was compliant with the Health Insurance Port-
ability and Accountability Act (HIPAA). Written informed
consent was obtained from all participants before scans.
With the approval of the Institutional Review Board
(IRB), 30 stable HTx patients (13 men, 17 women, age 47
6 14 [standard deviation], range 22–71 years; heart rate
96 6 11 beats/min, range 76–108 beats/min; body
weight 68 6 11 kg, range 56–89 kg) were recruited. We
excluded participants with contraindications to MR
scanning. No documented acute or chronic rejection was
reported in our participants (Table 1).
All MR scans were performed with a 1.5-T whole-body
MRI system (Magnetom Espree; Siemens Healthcare,
Erlangen, Germany) with a high-performance gradient
system (maximum gradient amplitude, 40 mT/m;
maximum slew rate, 200 mT/m/msec). A six-channel
cardiac phase array coil was used for radiofrequency
Cardiac Localization (General)
A three-plane fast localization sequence was used for
anatomic orientation for the cardiac scan. A black-
blood HASTE sequence was then run to identify the
two-chamber, four-chamber, short-axis views for ori-
entation. A four-chamber cine was used to obtain the
quiescent phase of cardiac motion.
Whole-Heart Coronary MR Angiography (for Accurate
An electrocardiogram (ECG)-triggered, T2-prepared,
fat-saturated, T2-prepared, segmented three-dimen-
sional (3D) SSFP sequence was employed for the
whole-heart coronary MRA (10). The 3D k-space data
were collected using centric order in the phase-encod-
ing direction and linear order in the partition-encoding
direction. Fifty-two transverse slices were acquired and
sinc-interpolated to 104 slices, each 0.75 mm thick.
The in-plane spatial resolution was 0.7 ? 0.7 mm2
(interpolated from 1.4 ? 1.4 mm2). Other imaging pa-
rameters included: repetition time (TR) and echo time
(TE) ¼ 3.7 and 1.7 msec; parallel acquisition factor ¼ 2
in the phase-encoding direction; flip angle ¼ 90?; lines
per heartbeat ¼ 15–25; readout bandwidth ¼ 870 Hz/
pixel. To overcome respiratory motion-induced image
artifacts, a motion adaptive navigator (NAV) technique
was used with end-expiratory imaging defined by the
dome of the right hemidiaphragm. Data were acquired
during a rest period of cardiac motion during mid-dias-
tole of the cardiac cycles.
Coronary Artery Wall Imaging: TSE
The 3D multiplanar reformations were performed on
the MR angiography images to localize the left main
coronary artery (LM), the proximal left anterior de-
scending coronary artery (LAD), and the right coronary
artery (RCA). Cross-sectional black-blood coronary
wall images were acquired using 2D, NAV-gated, ECG-
triggered, DIR prepared TSE sequence under free
breathing. A spectral-selective adiabatic inversion-re-
covery (SPAIR) pulse was used to suppress the epicar-
dial fat signal (inversion time 195–205 msec). Imaging
parameters included: echo spacing ¼ 6.9 msec; turbo
factor (k-space lines each heart beat) ¼ 5–13; apparent
TE ¼ 41 msec, trigger pulse ¼ 2, bandwidth ¼ 305
Hz/pixel, matrix ¼ 448 ? 448, field of view (FOV) ¼
420 ? 420 mm2; slice thickness ¼ 4.0 mm. Phase
directions were set perpendicular to coronary blood
flow of the imaging plane. In all cases, imaging data
were acquired during mid-diastole (rest period) of the
cardiac cycle according to 4-chamber cine (3). Based
on coronary landmarks, we acquired one cross-sec-
tional slice each in the RCA, LM, and LAD at locations
perpendicular to the long axis of each vessel and 5
mm from the origin.
Coronary Artery Wall Imaging: SSFP
A 2D, NAV gated, ECG-triggered, DIR prepared SSFP
sequence was run before or after TSE (within 15 min).
Cross-sectional images were acquired at the same
position as for the TSE images. Imaging parameters
included: TR/TE ¼ 4.8/2.4 msec, trigger pulse ¼ 2,
FOV ¼ 420 ? 420 mm2; slice thickness ¼ 4.0 mm;
readout matrix ¼ 448 ? 448. The same number of k-
space lines was acquired in each cardiac cycle as in
the TSE sequence. Bandwidth: 531 Hz/pixel. NAV and
the shimming box were set at the same position as
TSE. The same SPAIR pulse as in the TSE sequence
was used for fat saturation. The acquisition windows
was also optimized for individual patients to avoid
exceeding the coronary rest period duration and to
begin after the onset of the coronary rest period, as
defined on the 4-chamber cine images (3).
Image Evaluation and Data Processing
Images were transferred to an imaging workstation
(Dell, Studio XPS 435T). Images were ranked with a
4-point system and required consensus from two
authors. The scoring system can be described thus: in
Grade 0, no structure could be seen; in Grade 1, the
vessel wall (lumen) could barely be identified with ref-
erence images; in Grade 2, the vessel wall (lumen)
was seen clearly with some signal loss; lastly, in
Grade 3, the vessel wall (lumen) was distinct from the
surrounding tissue and was free of signal loss. A
Body weight (kg)
Heart rates (beats/minute)
Unless indicated, data are numbers of patients (percentages).
Coronary MRI for Cardiac Allografts1211
grade of 1, 2, or 3 was considered eligible for quanti-
zoomed to 1000%. The eligible coronary borders (with
image grades 1, 2, 3) were manually traced by one
experienced radiologist (with 5 years experience of
cardiovascular imaging) using VesselMASS software
(Leiden University, the Netherlands). The area of the
vessel wall was defined as the area between the outer
wall and the lumen. The mean thickness of the vessel
wall and lumen area was calculated automatically
according to vessel contours. The signal-to-noise ratio
(SNR, SI wall / SD noise) and the contrast-to-noise ra-
tio (CNR [SI wall?SI lumen]/SD noise) of the coronary
artery wall vs. lumen were also calculated (11–13).
The noise was measured in the pictures where there
were no anatomic structures or artifacts.
the coronarywall were
Data are presented as mean 6 one standard deviation
(SD). General image quality was compared between
SSFP and TSE using the Mann–Whitney U-test. The
SNR and CNR of the two imaging techniques (SSFP vs.
TSE) were compared with the t-test. Pearson correla-
tion efficient and Bland–Altman plots were applied to
evaluate agreements of vessel wall area and wall thick-
ness measured on matched SSFP and TSE images. All
statistical analysis was performed with SPSS software
(v. 13.0, Chicago, IL). For all calculations and results,
P < 0.05 was considered statistically significant.
General Image Quality
Two patients were excluded because they had metal
residues (pacing wires) in the body. In total, 28 scans
were completed. The length of cardiac rest periods
was 53 6 27 msec. In total, 19 coronary segments
(with TSE) and 40 coronary segments (with SSFP)
were carried out for quantitative analysis (grades 1, 2,
3). Seventeen pairs of coronary images (from 15
patients) were matched for the same anatomy.
Coronary Indices Comparison
The image quality score for SSFP was higher than for
TSE (1.23 6 0.95 vs. 0.88 6 0.69, P < 0.001). Com-
pared with TSE, SSFP had a higher mean coronary
wall SNR (20.1 6 8.5 vs. 14.9 6 4.8, P < 0.001) and
wall-lumen CNR (8.2 6 4.6 vs. 6.8 6 3.7, P ¼ 0.005)
(see Figs. 1–3 for images of three typical cases; see Ta-
ble 2 for data summary). Good agreements of wall
area and thickness measurements between SSFP and
TSE were observed on matched coronary images, with
Pearson correlation coefficients of 0.783 and 0.624,
respectively (P < 0.001) (Figs. 4, 5).
In this study we report the feasibility of imaging coro-
nary wall in cardiac allografts using black-blood SSFP
MRI. Compared with traditional TSE, SSFP offers bet-
ter image quality, higher SNR and CNR, with compa-
rable coronary measurements under conditions of fast
heart rate in HTx patients.
According to data from the registry of the Interna-
tional Society of Heart and Lung Transplantation
(ISHLT), ?3300 heart transplants were performed in
North America in 2008 as a final solution to irreversi-
ble heart failure (13). However, many serious compli-
cations follow this high-cost treatment. Due to con-
tinuing humoral and cell-mediated responses by the
recipient to human leukocyte antigen (HLA) present in
Figure 1. A 55-year-old male who had a heart transplant 2 years ago. Heart rate: 76 bpm. The LM could be clearly seen on
both SSFP (Grade 3) and TSE (Grade 3) images. a: SSFP. b: TSE.
1212 Lin et al.
the donor tissue, CAV, also known as chronic rejec-
tion, is a leading cause of graft loss and death in
patients who survive the first year after transplanta-
tion (1). In some cases, undetected CAV may silently
cause graft failure with global ischemia without docu-
mented signs of infarction or pathology on ECG (1).
Sometimes, CAV may develop quickly and serial fol-
low-up imaging of cardiac allografts is absolutely
essential. Therefore, MR coronary wall imaging, fea-
tured as ‘‘noninvasive’’ and ‘‘no radiation,’’ is expected
to be an ideal candidate for examining HTx recipients.
However, cardiac motion has always been considered
to be a major technical impediment to clinical use of
coronary wall MRI. Rapid heart rate, which is posi-
tively related to severe cardiac motion, has become a
main barrier restricting HTx patients from such
In various research studies, TSE has already been
accepted as the conventional technique for black-
blood coronary wall imaging (16–18). It requires short
echo trains to limit signal decay (T2 decay) caused by
transverse relaxation and to minimize cardiac motion
during signal readout (19,20), resulting in a relatively
long imaging time. For TSE coronary wall MRI, the
Figure 2. A 73-year-old female who had a heart transplant 5 years ago. Heart rate: 96 bpm. The LAD could be depicted on
both SSFP (Grade 3) and TSE (Grade 1). Visually, the image quality of SSFP is better. a: SSFP. b: TSE.
Figure 3. A 61-year-old female who had a heart transplant 7 years ago. Heart rate: 108 bpm. The RCA can only been seen
on SSFP (Grade 3), but not on TSE. a: SSFP. b: TSE.
Coronary MRI for Cardiac Allografts 1213
reduced number of k-space lines (lower acquisition
time during signal cardiac cycle) acquired per cardiac
cycle may be a possible solution for the short rest pe-
riod caused by a high heart rate. However, a direct
consequence of this strategy is the longer gross imag-
ing time, which may result in more unexpected
motion artifacts and irregular breathing modes. All of
them may significantly reduce image quality. Beta
blockers have been routinely used to improve image
quality in coronary CT angiography by reducing heart
rates. However, the long-term effects of beta blockers
on HTx patients have not been thoroughly studied.
Therefore, we did not apply them in our study.
Balanced SSFP sequences have a shorter imaging
time than TSE for vessel wall imaging. As in high heart
rate conditions, the ‘‘rest period’’ of cardiac motion is
usually short (10). When balanced SSFP is applied to
coronary wall MRI, a shorter data acquisition window
available in cardiac cycles much more easily fulfills the
needs of SSFP than that of TSE. Hence, time efficiency
becomes a valuable characteristic of SSFP for coronary
wall imaging in HTx patients.
Furthermore, with a gradient structure that is bal-
anced in all directions (readout, phase-encoding, and
partition-encoding), balanced SSFP allows for a high
flip angle without substantial signal decay. SSFP
acquires images at a faster speed, using a higher
readout bandwidth and shorter TR as compared to
TSE. Hence, both a balanced gradient structure and a
shorter acquisition time may together contribute to
less signal loss and intravoxel dephasing caused by
cardiac motion (21).
Our study had several limitations. First, the sample
size of eligible coronary segments was relatively small
for quantitative comparison. Under conditions of fast
heart rate, the success rate of coronary MRI is really
low, especially for TSE. However, it is well known that
elevated heart rate is a main adverse factor for coro-
nary imaging. In this feasibility study we demonstrated
the advantage of SSFP in coronary imaging and make
some progress in solving this problem. Second, we
were unable to identify any specific lesions of the ves-
sel wall, such as CAV. This might be due to our stable
and asymptomatic HTx patients and limited coverage
of 2D coronary wall MRI. However, our results affirm
the capability of SSFP coronary wall MRI for acquiring
important indices for monitoring CAV, such as coro-
nary wall area and wall thickness. As proven by
another study, the spatial resolution should be appro-
priate to find coronary wall thickening (3).
In conclusion, 2D DIR-prepared SSFP coronary wall
MRI has higher image quality as compared with 2D
DIR TSE in HTx patients. It has the potential to serve
as an eligible option of coronary wall MRI for cardiac
allografts under fast heart rate conditions.
1. Schmauss D, Weis M. Cardiac allograft vasculopathy: recent
developments. Circulation 2008;117:2131–2141.
2. Bastarrika G, Broncano J, Arraiza M, et al. Systolic prospectively
ECG-triggered dual-source CT angiography for evaluation of the
coronary arteries in heart transplant recipients. Eur Radiol 2011;
3. Miao C, Chen S, Macedo R, et al. Positive remodeling of the coro-
nary arteries detected by magnetic resonance imaging in an
asymptomatic population: MESA (Multi-Ethnic Study of Athero-
sclerosis). J Am Coll Cardiol 2009;53:1708–1715.
4. Hofman MB, Wickline SA, Lorenz CH. Quantification of in-plane
motionof thecoronaryarteries duringthe cardiac cycle:
Image Quality and Comparisons
Wall thickness (mm)
Lumen area (mm2)
Imaging scores were compared on 84 coronary positions from 28
participants. SNR and CNR were compared on SSFP (40 coronary
segments) and TSE (19 coronary segments). Wall thickness and
lumen area were compared on SSFP (17 segments) and TSE (17
Figure 4. Bland–Altman plot of measurements of coronary
wall thickness with SSFP and TSE. r ¼ 0.783.
Figure 5. Bland–Altman plot of measurements of coronary
lumen area with SSFP and TSE. r ¼ 0.624.
1214Lin et al.
implications for acquisition window duration for MR flow quanti- Download full-text
fication. J Magn Reson Imaging 1998;8:568–576.
5. Ustun A, Desai M, Abd-Elmoniem KZ, Schar M, Stuber M. Auto-
mated identification of minimal myocardial motion for improved
image quality on MR angiography at 3 T. AJR Am J Roentgenol
6. Keeley EC, Toth ZK, Goldberg AD. Long-term assessment of heart
rate variability in cardiac transplant recipients. J Heart Lung
7. Berkowitz SJ, Macedo R, Malayeri AA, et al. Axial black blood
turbo spin echo imaging of the right ventricle. Magn Reson Med
8. Koktzoglou I, Chung YC, Carroll TJ, et al. Three-dimensional black-
blood MR imaging of carotid arteries with segmented steady-state
free precession: initial experience. Radiology 2007;243:220–228.
9. Lin K, Bi X, Zuehlsdorff S, et al. Coronary wall MR imaging in
patients with rapid heart rates: a feasibility study of black-blood
steady-state free precession (SSFP). Int J Cardiovasc Imaging
2011 [Epub ahead of print].
10. Brittain JH, Hu BS, Wright GA, Meyer CH, Macovski A, Nishi-
mura DG. Coronary angiography with magnetization-prepared T2
contrast. Magn Reson Med 1995;33:689–696.
11. Bi X, Deshpande V, Simonetti O, Laub G, Li D. Three-dimen-
sional breathhold SSFP coronary MRA: a comparison between
1.5T and 3.0T. J Magn Reson Imaging 2005;22:206–212.
12. Stuber M, Botnar RM, Fischer SE, et al. Preliminary report on in
vivo coronary MRA at 3 Tesla in humans. Magn Reson Med 2002;
13. Taylor DO, Stehlik J, Edwards LB, et al. Registry of the Interna-
tional Society for Heart and Lung Transplantation: twenty-sixth
official adult heart transplant report-2009. J Heart Lung Trans-
14. Wang Y, Vidan E, Bergman GW, et al. Cardiac motion of coronary
arteries: variability in the rest period and implications for coro-
nary MR angiography. Radiology 1999;213:751–758.
15. Jahnke C, Paetsch I, Achenbach S, et al. Coronary MR imaging:
breath-hold capability and patterns, coronary artery rest periods,
and beta-blocker use. Radiology 2006;239:71–78.
16. Fayad ZA, Fuster V, Fallon JT, et al. Noninvasive in vivo human
coronary artery lumen and wall imaging using black-blood mag-
netic resonance imaging. Circulation 2000;102:506–510.
17. Spuentrup E, Ruebben A, Mahnken A, et al. Artifact-free coro-
nary magnetic resonance angiography and coronary vessel wall
imaging in the presence of a new, metallic, coronary magnetic
resonance imaging stent. Circulation 2005;111:1019–1026.
18. Botnar RM, Stuber M, Kissinger KV, et al. Noninvasive coronary
vessel wall and plaque imaging with magnetic resonance imaging.
19. Constable RT, Gore JC. The loss of small objects in variable TE
imaging: implications for FSE, RARE, and EPI. Magn Reson Med
20. Listerud J, Einstein S, Outwater E, Kressel HY. First principles of
fast spin echo. Magn Reson Q 1992;8:199–244.
21. Deshpande VS, Shea SM, Laub G, et al. 3D magnetization-pre-
pared true-FISP: a new technique for imaging coronary arteries.
Magn Reson Med 2001;46:494–502.
Coronary MRI for Cardiac Allografts1215