Original Research
Cardiac Imaging
June 2006

MRI of Coronary Vessel Walls Using Radial k-Space Sampling and Steady-State Free Precession Imaging

Abstract

OBJECTIVE. The objective of our study was to investigate the impact of radial k-space sampling and steady-state free precession (SSFP) imaging on image quality in MRI of coronary vessel walls.
SUBJECTS AND METHODS. Eleven subjects were examined on a 1.5-T MR system using three high-resolution navigator-gated and cardiac-triggered 3D black blood sequences (cartesian gradient-echo [GRE], radial GRE, and radial SSFP) with identical spatial resolution (0.9 × 0.9 × 2.4 mm3). The signal-to-noise ratio (SNR), contrast-to-noise ratio (CNR), vessel wall sharpness, and motion artifacts were analyzed.
RESULTS. The mean SNR and CNR of the coronary vessel wall were improved using radial imaging and were best using radial k-space sampling combined with SSFP imaging. Vessel border definition was similar for all three sequences. Radial k-space sampling was found to be less sensitive to motion. Consistently good image quality was seen with the radial GRE sequence.
CONCLUSION. Radial k-space sampling in MRI of coronary vessel walls resulted in fewer motion artifacts and improved SNR and CNR. The use of SSFP imaging, however, did not result in improved coronary vessel wall visualization.

Introduction

The current diagnostic method of choice for the detection of coronary artery disease is conventional X-ray angiography. However, X-ray angiography frequently underestimates the true extent of coronary arteriosclerosis because it exclusively shows the vessel lumen and provides only limited information about the coronary plaque burden [1]. At present, intravascular sonography is considered the gold standard for visualization of coronary plaques [2]. This technique, however, is invasive and is not appropriate for screening or serial examinations for the assessment of subclinical and advanced arteriosclerosis.
Several studies have shown the use of breath-hold and navigator-gated black blood MRI for noninvasive visualization of the coronary vessel wall with good contrast between the coronary lumen and surrounding tissues [3, 4]. For routine clinical use, image quality and the robustness of the technique need to be improved. Recently, a local reinversion method in combination with spiral data acquisition that allowed 3D coronary vessel wall imaging with improved coverage was introduced [5].
In bright blood coronary MR angiography, steady-state free precession (SSFP) imaging led to an improved signal-to-noise ratio (SNR) [6] while radial k-space sampling has been shown to be less sensitive to motion artifacts [7]. Both techniques, however, have not been used for vessel wall imaging to date. The motion sensitivity of the SSFP technique may differ because of the longer effective acquisition window (constant k-space weighting). We therefore sought to investigate the impact of both radial k-space sampling and SSFP sequences on image quality with special regard to motion artifacts in free-breathing navigator-gated 3D MRI of the coronary vessel wall.

Subjects and Methods

Subjects

The right coronary artery (RCA) of 11 consecutive subjects (five men, six women; mean age ± SD, 28 ± 5 years) was scanned with the patient in the supine position using a 1.5-T whole-body MR system (Gyroscan Intera, Philips Medical Systems) equipped with a cardiac software package (R8, Philips Medical Systems) and a five-element cardiac Synergy coil (Philips Medical Systems). None of the subjects had a history of coronary artery disease, and the protocol was approved by the institutional board on clinical investigations. Informed consent was obtained from each subject participating in this study.

Scout Scanning

A transverse cardiac-triggered cine fast-field echo echo-planar imaging (EPI) sequence (TE = 9.7 msec, EPI readouts = 7, heart phases = 40) was performed to visually determine the quiescent period within the cardiac cycle. The derived trigger delay was used for all subsequent scans, including a fast navigator-gated cardiac-triggered 3D localizer SSFP sequence with axial slice orientation, covering the entire heart [8].

Coronary MR Angiography

Subsequently, a free-breathing navigator-gated and cardiac-triggered bright blood 3D SSFP coronary angiogram was acquired (TR/TE = 6.1/3.0, excitation angle = 120°, field of view = 360 mm2) [9]. For accurate and reproducible targeting of the RCA, double oblique orientations were prescribed using a three-point-plan scan tool. Data were acquired using radial k-space sampling with a 384 × 384 matrix (i.e., 384 radials and 384 samplings). Before the imaging sequence, five dummy radiofrequency pulses were performed to allow a smooth and rapid approach toward the steady-state condition. In addition, a T2-preparation pulse and a fat-saturation prepulse were used to suppress signal from myocardium and epicardial fat.

MRI of Coronary Vessel Walls

Imaging sequence—Coronary vessel wall imaging was performed using 3D cartesian gradient-echo (GRE), radial GRE, and radial SSFP imaging sequences (Fig. 1). Spatial resolution was identical for all three sequences (field of view = 360 × 360 mm, matrix= 384 × 384, in-plane resolution = 0.9 × 0.9 mm reconstructed to 0.35 × 0.35 mm using a 1024 × 1024 matrix, twelve 2.4-mm-thick slices reconstructed to twenty-four 1.2-mm-thick slices). The data acquired with radial k-space sampling were reconstructed by a making a grid (i.e., the data in k-space were interpolated onto a rectangular grid by means of a convolution, and a 2D Fourier transform was performed). The imaging parameters for the sequences were as follows: cartesian GRE sequence, TR/TE of 8.0/2.3, excitation angle of 30°, acquisition window of 64 msec; radial GRE sequence, 7.1/2.0, excitation angle of 30°, acquisition window of 57 msec; and radial SSFP sequence, 5.3/2.7, excitation angle of 55°, acquisition window of 42 msec, and five dummy radiofrequency excitations. To minimize cardiac motion artifacts, only eight profiles were sampled per R-R interval.
Radial imaging was performed using a “stack-of-stars” approach with radial sampling along the x-y direction and centric phase encoding (Fourier encoding) in the z-direction [7, 9]. Centric phase encoding in the z-direction within each R-R interval was chosen to optimize prepulse efficiency and, thus, contrast, which is essential for the stack-of-stars approach [7]. Cartesian imaging was performed with centric phase encoding along the y-direction.
For all sequences, the scanning time was 11 min for a heart rate of 70 beats per minute (bpm).
Black blood prepulse—Signal from blood was suppressed using a modified double-inversion prepulse consisting of a nonselective inversion pulse immediately followed by a 2D selective reinversion pulse. The 2D pencil beam was positioned along the RCA. By imaging every other heartbeat, the inversion time was approximately 600 msec [10], thereby allowing sufficient blood exchange and thus good black blood properties. To suppress signal from the epicardial fat, an additional spectral inversion recovery (IR) prepulse was applied.
MR navigator—All sequences were equipped with a right hemidiaphragmatic prospective real-time navigator for respiratory motion artifact suppression during free breathing [11]. Data were accepted if the lung-liver interface was within the end-expiratory gating window (5 mm) or otherwise were rejected and reacquired during the subsequent R-R interval.

Data Analysis

Multiplanar reformatting and subsequent subjective and objective image analyses were performed using a previously described postprocessing tool [12]. Subjective image quality was determined on a consensus basis by two observers blinded to sequence parameters. The observers used two scales from 1 to 3. One scale was for evaluation of the level of motion artifacts: 1, no visible motion artifacts; 2, moderate artifacts; and 3, pronounced motion artifacts. The second scale was used to assess image quality: 1, good image quality with sharply defined vessel wall borders; 2, reduced image quality but vessel wall still visible; and 3, poor image quality with poor visualization of the coronary vessel wall. Although motion artifact level was examined in the entire image, image quality was judged only by the appearance of the right coronary vessel wall.
Signal intensities (SIs) were measured in the aorta (blood) and in epicardial fat (fat) using user-specified regions of interest (ROIs). In addition, the signal of the coronary vessel wall (vw) was determined as the average SI along the centerline of the coronary vessel wall. The SNR of the coronary vessel wall (SNRvw) was calculated as follows:
\[ \[SI_{vw}{/}SD_{noise}\] \]
where SDnoise refers to the SD of the signal in a region of air ventral to the chest. The CNR between the coronary vessel wall and blood signal and the epicardial fat (CNRvw-blood / fat) was calculated as follows:
\[ \[(SI_{vw}-SI_{blood{/}fat}){/}SD_{noise}.\] \]
Furthermore, the maximal length of the visible vessel wall and the maximal vessel wall sharpness were assessed on multiplanar reformatted images [12].
Fig. 1 —Diagram of free-breathing navigator-gated 3D black blood sequences using cartesian and radial gradient-echo or radial steady-state free precession (SSFP) imaging. Black blood properties are maintained using double inversion prepulse consisting of nonselective and selective inversion pulses. Real-time navigator precedes five preparatory pulses, which are used to approach steady-state conditions. Epicardial fat saturation is achieved using spectral inversion (SPIR) pulse. Imaging was performed during mid-diastole, which is quiescent period in cardiac cycle.
Fig. 2A —Bright blood MR angiography shows right coronary artery of two healthy volunteers. Radial steady-state free precession (SSFP) image of 31-year-old healthy man shows right coronary artery.
Fig. 2B —Bright blood MR angiography shows right coronary artery of two healthy volunteers. Scans corresponding to A that were acquired using cartesian gradient-echo (GRE) (B), radial GRE (C), or radial SSFP (D) sequences show coronary vessel walls (arrowheads). Note pronounced motion artifacts (arrows, B) in phase-encoding direction originating from cardiac motion in cartesian GRE images.
Fig. 2C —Bright blood MR angiography shows right coronary artery of two healthy volunteers. Scans corresponding to A that were acquired using cartesian gradient-echo (GRE) (B), radial GRE (C), or radial SSFP (D) sequences show coronary vessel walls (arrowheads). Note pronounced motion artifacts (arrows, B) in phase-encoding direction originating from cardiac motion in cartesian GRE images.
Fig. 2D —Bright blood MR angiography shows right coronary artery of two healthy volunteers. Scans corresponding to A that were acquired using cartesian gradient-echo (GRE) (B), radial GRE (C), or radial SSFP (D) sequences show coronary vessel walls (arrowheads). Note pronounced motion artifacts (arrows, B) in phase-encoding direction originating from cardiac motion in cartesian GRE images.
Fig. 2E —Bright blood MR angiography shows right coronary artery of two healthy volunteers. Radial SSFP image of 38-year-old healthy woman shows right coronary artery.
Fig. 2F —Bright blood MR angiography shows right coronary artery of two healthy volunteers. Scans corresponding to E that were acquired using cartesian GRE (B), radial GRE (C), or radial SSFP (D) sequences show coronary vessel walls (arrowheads). Note pronounced motion artifacts (arrows, F) in phase-encoding direction originating from cardiac motion in cartesian GRE images.
Fig. 2G —Bright blood MR angiography shows right coronary artery of two healthy volunteers. Scans corresponding to E that were acquired using cartesian GRE (B), radial GRE (C), or radial SSFP (D) sequences show coronary vessel walls (arrowheads). Note pronounced motion artifacts (arrows, F) in phase-encoding direction originating from cardiac motion in cartesian GRE images.
Fig. 2H —Bright blood MR angiography shows right coronary artery of two healthy volunteers. Scans corresponding to E that were acquired using cartesian GRE (B), radial GRE (C), or radial SSFP (D) sequences show coronary vessel walls (arrowheads). Note pronounced motion artifacts (arrows, F) in phase-encoding direction originating from cardiac motion in cartesian GRE images.

Statistical Analysis

The results were expressed as geometric means ± 1 SD. For statistical comparisons, nonparametric methods using statistics software (JMP Discovery software [version 5.0.1.2], SAS Institute) were applied. The objective image quality parameters were analyzed with the Wilcoxon's rank sum test, and for the analysis of the ordinal data, the Pearson's chi-square test was applied. Differences between groups with a p value of less than 0.05 were considered to be statistically significant.

Results

All three high-resolution MR vessel wall sequences were successfully completed in all subjects. Representative double oblique MR images of the RCA vessel wall using all three sequences are shown in Figures 2A, 2B, 2C, 2D, 2E, 2F, 2G, and 2H. The bright coronary vessel wall is well delineated from the dark-appearing blood and epicardial fat. In some volunteers, visualization (i.e., image quality score = 3) of the coronary vessel wall was poor on images obtained using the cartesian GRE (n = 2) and radial SSFP (n = 3) sequences, while sufficient image quality (i.e., image quality score = 1 or 2) was observed in all volunteers when the radial GRE sequence was performed (Table 1).
TABLE 1: Subjective Image Quality Scores
No. of Subjects (n = 11)
ScoreCartesian GRE ImagesRadial GRE ImagesRadial SSFP Images
Image qualitya   
   1244
   2774
   3203
Artifact levelb   
   101111
   21000
   3
1
0
0
Note—GRE = gradient-echo, SSFP = steady-state free precession
a
1, good image quality with sharply defined vessel wall borders; 2, reduced image quality but vessel wall still visible; 3, poor image quality with poor visualization of the coronary vessel wall
b
1, no visible motion artifacts; 2, moderate artifacts; 3, pronounced motion artifacts
With regard to motion artifacts, radial k-space sampling was superior to cartesian k-space sampling (Table 1). Cartesian imaging was found to be more susceptible to cardiac motion, thus resulting in artifacts along the phase-encoding direction (Figs. 2A, 2B, 2C, 2D, 2E, 2F, 2G, and 2H). These visually apparent artifacts were translated into a higher level of subjective motion artifact when compared with radial scanning with or without SSFP (p < 0.05).
The SNRvw was superior using the radial SSFP sequence (16.9 ± 4.3) compared with the cartesian GRE sequence (12.2 ± 3.7; p <0.05) (Fig. 3A). The radial GRE images showed intermediate SNRvw (14.2 ± 4.4; p, not significant [NS]). The best CNRvw-blood was also seen with radial SSFP imaging (cartesian GRE, 9.6 ± 3.0; radial GRE, 12.0 ± 4.0; radial SSFP, 13.6 ± 4.4), whereas CNRvw-fat was comparable for all three sequences (cartesian GRE, 9.1 ± 3.9; radial GRE, 8.7 ± 4.0; radial SSFP, 9.1 ± 5.4; p, NS).
Higher SNR and higher CNR values in radial scanning, however, did not result in better vessel border definition (Fig. 3B). Objective vessel sharpness analyses were similar for all three sequences: cartesian GRE, 34.5 ± 8.9; radial GRE, 37.6 ± 6.4; radial SSFP, 34.2 ± 4.9 (p, NS). The maximal visible vessel wall length was slightly longer on images obtained using the radial SSFP sequence (p, NS).

Discussion

In this study, we show the successful use of a navigator-gated cardiac-triggered 3D GRE sequence with radial k-space sampling and a double-inversion black blood prepulse for imaging of the RCA vessel wall.

Motion Artifacts

Data acquisition using radial k-space sampling showed reduced motion artifacts and thus improved robustness toward cardiac and respiratory motion. Using cartesian k-space sampling, ghosting artifacts were observed in the anteroposterior direction (phase-encoding direction) and originated mainly from cardiac motion. The anteroposterior phase-encoding direction was chosen to avoid “back-folding” artifacts. However, the craniocaudal direction could have also been considered because of the limited sensitivity of the receiver coils in the foot-head direction. It is conceivable that scanning in this orientation may have led to an improved sharpness of the coronary vessel wall because the course of the RCA was nearly perpendicular to the frequency-encoding direction.
In contrast to cartesian imaging, radial k-space sampling showed only minor artifacts, which were distributed homogeneously over the entire image, resulting in an overall lower artifact level. Because of the short TR, radial k-space sampling is a good nonrectilinear data acquisition candidate for SSFP imaging. Spiral imaging seems to be less favorable due to the relatively long TR of one spiral interleave and therefore was not evaluated in this study.
Another reason for the reduced sensitivity of radial imaging to motion may be the choice of the encoding scheme. Intrashot encoding (Fourier encoding) was performed in the z-direction, whereas intershot encoding used radial encoding in the x-y plane. Respiratory motion and arrhythmias are expected to primarily influence intershot encoding, whereas cardiac motion is expected to influence intrashot encoding. Surprisingly, improved motion artifact suppression using radial scanning did not translate into better objective vessel border definition in our study. In radial scanning, each k-space line covers the entire k-space, which may result in different motion sensitivity. Similar vessel sharpness values were found for all three sequences.
Fig. 3A —Analysis of objective image quality parameters for cartesian gradient-echo (cartGRE, gray bars), radial gradient-echo (radGRE, black bars), and radial steady-state free precession (radSSFP, white bars) sequences. Bar graph shows signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR). vw = vessel wall.
Fig. 3B —Analysis of objective image quality parameters for cartesian gradient-echo (cartGRE, gray bars), radial gradient-echo (radGRE, black bars), and radial steady-state free precession (radSSFP, white bars) sequences. Bar graph shows vessel length and sharpness.

SNR and CNR

The SNR and CNR between the coronary vessel wall and blood were improved for both radial GRE and radial SSFP coronary vessel wall sequences when compared with cartesian k-space sampling. To calculate the SNR and CNR, the noise levels were determined as the SD in user-defined ROIs in the background. As we mentioned earlier, motion artifacts were distributed homogeneously over the entire image in radial imaging, resulting in a lower mean artifact level than with cartesian data sampling, which in part may have contributed to the increased SNR and CNR in radial imaging.
The use of a nonselective inversion in concert with a pencil-shaped reinversion black blood prepulse resulted in good signal suppression of the blood, thereby allowing good delineation of the vessel wall from coronary blood. In addition, the signal of epicardial fat was suppressed using a spectral IR prepulse. The potential of SSFP imaging has already been shown for coronary MR angiography [6] but has not been investigated for coronary vessel wall imaging. As expected, SSFP imaging resulted in a higher SNRvw and CNRvw-fat when compared with spoiled GRE approaches. However, due to the higher spatial resolution in vessel wall imaging, longer TRs and TEs (TR/TE, 5.3/2.7 compared with 3.9/1.8 in previous SSFP coronary MR angiography studies) [6] are required. In addition, the effective acquisition window setting is prolonged in SSFP imaging due to the rephrasing of the transverse magnetization and subsequent use during the next radiofrequency excitation (i.e., constant k-space weighting, resulting in a narrow point-spread function). This may lead to a higher sensitivity to off-resonances and motion, thereby potentially degrading the visualization of the coronary vessel walls.
In conclusion, radial k-space sampling in free-breathing navigator-gated cardiac-triggered MRI of the coronary vessel wall leads to a reduced motion artifact level and improved SNR and CNR compared with cartesian k-space sampling. Although SSFP imaging led to further increase in SNR and CNR, vessel wall border definition was not improved. Studies in patients with known coronary artery disease are now warranted to investigate the impact of coronary vessel wall imaging with radial k-space sampling in a clinical setting.

Footnotes

Supported in part by a Biomedical Engineering Grant from the Whitaker Foundation (RG-02-0745).
M. Stuber is compensated as a consultant by Philips Medical Systems NL, the manufacturer of equipment described in this presentation. The terms of this arrangement have been approved by the Johns Hopkins University in accordance with its conflict of interest policies.
Address correspondence to M. Katoh ([email protected]).

References

1.
Ward MR, Pasterkamp G, Yeung AC, Borst C. Arterial remodeling: mechanisms and clinical implications. Circulation 2000; 102:1186-1191
2.
Abizaid AS, Mintz GS, Mehran R, et al. Long-term follow-up after percutaneous transluminal coronary angioplasty was not performed based on intravascular ultrasound findings: importance of lumen dimensions. Circulation 1999; 100:256-261
3.
Fayad ZA, Fuster V, Fallon JT, et al. Noninvasive in vivo human coronary artery lumen and wall imaging using black-blood magnetic resonance imaging. Circulation 2000; 102:506-510
4.
Botnar RM, Stuber M, Kissinger KV, Kim WY, Spuentrup E, Manning WJ. Noninvasive coronary vessel wall and plaque imaging with magnetic resonance imaging. Circulation 2000; 102:2582-2587
5.
Botnar RM, Kim WY, Bornert P, Stuber M, Spuentrup E, Manning WJ. 3D coronary vessel wall imaging utilizing a local inversion technique with spiral image acquisition. Magn Reson Med 2001; 46:848-854
6.
Spuentrup E, Bornert P, Botnar RM, Groen JP, Manning WJ, Stuber M. Navigator-gated free-breathing three-dimensional balanced fast field echo (True-FISP) coronary magnetic resonance angiography. Invest Radiol 2002; 37:637-642
7.
Spuentrup E, Katoh M, Buecker A, et al. Free-breathing 3D steady-state free-precession coronary MR angiography with radial k-space sampling: comparison with cartesian k-space sampling and cartesian gradient-echo coronary MR angiography—pilot study. Radiology 2004; 231:581-586
8.
Spuentrup E, Ruebben A, Schaeffter T, Manning WJ, Gunther RW, Buecker A. Magnetic resonance-guided coronary artery stent placement in a swine model. Circulation 2002; 105:874-879
9.
Spuentrup E, Katoh M, Stuber M, et al. Coronary MR imaging using free-breathing 3D steady-state free precession with radial k-space sampling. Rofo 2003; 175:1330-1334
10.
Fleckenstein JL, Archer BT, Barker BA, Vaughan JT, Parkey RW, Peshock RM. Fast short-tau inversion-recovery MR imaging. Radiology 1991; 179:499-504
11.
Stuber M, Botnar RM, Danias PG, Kissinger KV, Manning WJ. Submillimeter three-dimensional coronary MR angiography with real-time navigator correction: comparison of navigator locations. Radiology 1999; 212:579-587
12.
Etienne A, Botnar RM, Van Muiswinkel AM, Boesiger P, Manning WJ, Stuber M. “Soap-bubble” visualization and quantitative analysis of 3D coronary magnetic resonance angiograms. Magn Reson Med 2002; 48:658-666

Information & Authors

Information

Published In

American Journal of Roentgenology
Pages: S401 - S406
PubMed: 16714616

History

Submitted: December 7, 2004
Accepted: February 15, 2005

Keywords

  1. angiography
  2. arteriosclerosis
  3. cardiovascular imaging
  4. coronary artery disease
  5. hemodynamics
  6. MR angiography
  7. MRI
  8. radial k-space sampling

Authors

Affiliations

Marcus Katoh
Department of Diagnostic Radiology, RWTH Aachen University Hospital, Pauwelsstrasse 30, 52057 Aachen, Germany.
Elmar Spuentrup
Department of Diagnostic Radiology, RWTH Aachen University Hospital, Pauwelsstrasse 30, 52057 Aachen, Germany.
Arno Buecker
Department of Diagnostic Radiology, RWTH Aachen University Hospital, Pauwelsstrasse 30, 52057 Aachen, Germany.
Tobias Schaeffter
Philips Research Laboratories, Hamburg, Germany.
Matthias Stuber
Department of Radiology, Division of MRI Research, Johns Hopkins University Medical School, Baltimore, MD.
Rolf W. Günther
Department of Diagnostic Radiology, RWTH Aachen University Hospital, Pauwelsstrasse 30, 52057 Aachen, Germany.
Rene M. Botnar
Department of Medicine, Cardiovascular Division, Beth Israel Deaconess Medical Center and Harvard Medical School, Boston, MA.

Metrics & Citations

Metrics

Citations

Export Citations

To download the citation to this article, select your reference manager software.

Articles citing this article

View Options

View options

PDF

View PDF

PDF Download

Download PDF

Media

Figures

Other

Tables

Share

Share

Copy the content Link

Share on social media