| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Technical Developments |
1 From the Imaging Science Laboratories (V.M., V.V.I., M.S., J.G.S.A., D.D.S., G.M., Z.A.F.); the Zena and Michael A. Wiener Cardiovascular Institute, the Marie-Josée and Henry R. Kravis Cardiovascular Health Center, and Department of Medicine (V.M., J.G.S.A., G.M., Z.A.F.); and Department of Radiology (V.V.I., D.D.S., Z.A.F.); Mount Sinai School of Medicine, One Gustave L. Levy Place, Box 1234, New York, NY 10029-6574; and Siemens Medical Solutions, Malvern, Pa (M.S.). Received July 1, 2003; revision requested August 29; revision received October 17; accepted November 24. Supported in part by NIH/NHLBI ROI HL71021, the Howard Hughes Medical Institute Biomedical Research Support Program Grant, the Herman Goldman Foundation, the New York Community Trust, the Mount Sinai Consortium for Cardiovascular Imaging Technology, the Eva and Morris Feld Estate, the Louis B. Mayer Foundation, the Peter Jay Sharp Foundation, the Zena and Michael A. Wiener Cardiovascular Institute, the Marie-Josée and Henry R. Kravis Cardiovascular Health Center, and the Dept of Radiology, Mount Sinai School of Medicine. Address correspondence to Z.A.F. (e-mail: zahi.fayad@mssm.edu).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMaterials and Methods ResultsDiscussionREFERENCES |
|---|
© RSNA, 2004
Index terms: Aorta, diseases, 563.75 • Aorta, MR, 563.121413, 563.121416 • Magnetic resonance (MR), pulse sequences • Magnetic resonance (MR), rapid imaging • Magnetic resonance (MR), technology
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and Methods ResultsDiscussionREFERENCES |
|---|
Flow suppression (ie, black-blood imaging) improves vessel wall visibility and minimizes blood flow artifacts that can affect image quality and interpretation. Black-blood techniques include spatial presaturation (7) and double inversion-recovery (DIR) preparation pulses (8). The DIR preparation module consists of two 180° radiofrequency pulses and is applied before the image acquisition (readout). The first nonselective radiofrequency pulse inverts the magnetization of the whole volume. The second selective radiofrequency pulse restores the magnetization in the section or slab of interest. After a time delay (inversion time [TI0]), required for the magnetization of blood to reach the null point, the imaging section or slab is acquired (8).
The conventional DIR-prepared two-dimensional imaging sequence involves acquisition of multiple lines of k space from one section following each DIR block. Acquisition of multiple sections in this fashion results in long examination times. For example, to acquire 20 sections with 256 phase-encoding lines and a turbo factor of 9, the conventional DIR rapid acquisition with relaxation enhancement (RARE) readout sequence would take 1,160 R-R intervals (time intervals between two consecutive heartbeats) to complete because electrocardiographic triggering is every other heartbeat for intermediate-weighted imaging. This DIR RARE sequence has been successfully used in vivo for vessel wall imaging of different vascular structures (9,10).
Improved DIR sequences to reduce examination time have recently been developed (11–13). Song et al (11) evaluated a dual-section DIR technique. The DIR preparation was modified to include one nonselective and two section-selective inversion pulses. After the DIR preparation, k-space lines from two sections were acquired (11). Parker et al (12) improved the efficiency of the DIR RARE technique by reducing the required time delay for nulling of the blood signal. The number of selective reinversion pulses in the DIR block corresponded to the number of sections imaged within one triggering period. Yarnykh and Yuan (13) evaluated a multisection DIR technique that involved the use of several DIR blocks within one triggering period, thereby reducing the inversion time of blood signal suppression. The selective reinversion pulse of the DIR block inverted the entire imaging volume (13). With these techniques, one DIR block was followed by the acquisition of one imaging section (12,13). Single-shot RARE techniques for improved efficiency of black-blood imaging have been proposed (14,15): A half-Fourier transform performed on k-space data allowed the acquisition of a reduced number of k-space lines, thereby reducing the acquisition time involved in these techniques.
The purpose of this study was to evaluate a two-dimensional rapid extended coverage (REX) RARE pulse sequence for simultaneous multisection DIR black-blood vessel wall MR imaging.
|
Materials and Methods |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and Methods ResultsDiscussionREFERENCES |
|---|
Subjects and Patients
Aortic vessel wall MR imaging was performed in five healthy subjects (four men, one woman; mean age, 33 years ± 4) with no known history of atherosclerotic disease and five patients with atherosclerotic disease (three men, two women; mean age, 67 years ± 11.7). This study was approved by the institutional review board of Mount Sinai School of Medicine. To be included as a patient in this study, an individual had to have a known history of atherosclerotic disease, as documented with transesophageal echocardiography and/or with B-mode ultrasonography of the carotid arteries and the aortic arch (16). Healthy subjects had no known history of atherosclerosis. Informed consent was obtained from all patients and healthy subjects. The subjects were positioned headfirst and supine in the magnet bore. Three surface electrocardiographic electrodes were placed on the subjects’ chests for data acquisition triggering.
MR Pulse Sequence Design
The electrocardiographically triggered REX simultaneous multisection black-blood DIR sequence developed for vessel wall imaging in this study consisted of multiple DIR blocks and multiple RARE readouts with short TI0 times (Fig 1). The time window after the application of the DIR block, during which the blood signal is sufficiently suppressed, is used to acquire multiple section readouts (11). This allows for the simultaneous acquisition of k-space lines from more than one section within the same repetition time (TR) interval.
|
Sel180°) was calculated with the following equation:
|
Z is the thickness of each section, and Gap is the section separation. The sequence acquisition block (REX module) consisted of one DIR block followed by multiple (two to five) RARE section readouts. In this study, four to nine REX modules were acquired in two R-R intervals, yielding 16–20 closely spaced sections. An example of the pulse sequence diagram for the acquisition of 18 sections with six REX modules is shown in Figure 1. TI0 spans the time from the nonselective radiofrequency pulse (inverting the magnetization of the blood) to the middle of the section readouts to ensure as much proximity to the null point of blood as possible.
The TR for the section (TRSEC) equaled two R-R intervals (2R-R) and was different from the TR of dark blood (TRDB), which was determined by the time between two successive DIR blocks as follows:
|
A reduction in TRDB leads to a decrease in dark blood TI0 according to the equation
|
The relationship between the TRDB and the TI0 of blood when the signal from blood is nulled is illustrated in Figure 2. One dummy imaging pass was performed before data acquisition to allow for steady-state inversion recovery.
|
|
The TF is related to the echo-train length (ETL) by using the following formula:
|
Data Analysis
The quality factor (QF) was calculated for all images according to the following equation:
|
n is the SD of noise, and T is the total experiment time. QF was defined on the basis of the time it took to perform an experiment with the single-section approach for the required number of sections.
The contrast-to-noise ratio (CNR) was calculated for all images according to the following equation:
|
n is the SD of noise (ie, of a region outside the patient).
MR Imaging and Image Comparison
Initial scout images in three orthogonal planes (transverse, coronal, and sagittal) were used to locate the thoracic and abdominal portions of the descending aorta. During aortic wall imaging, the healthy subjects were asked to hold their breath at inspiration when possible. Alternatively, a saturation band covering the anterior thorax was applied to reduce breathing artifacts. All images in patients with atherosclerotic disease were obtained without breath holding. An anterior chest saturation band was used to reduce ghosting artifacts.
Table 1 shows the parameters used in the imaging protocol. To image the entire or a large portion of the descending aorta, multisection protocols for simultaneously imaging 16, 18, and 20 sections were developed. Other imaging parameters were as follows: echo spacing, 4.9 msec; echo time, 4.9 msec; acquisition matrix size, 256 x 256; section thickness, 3 mm; section separation, 0.3 mm; data acquisition bandwidth, 488 Hz per pixel; number of signals acquired, one; and field of view, 250 mm. The section excitation order descended from head to foot along the blood flow direction. The section readout time (approximately, the echo spacing times the turbo factor) ranged between 44 and 64 msec. This minimized vessel wall motion and blurring along the phase-encoding direction. Turbo factors (9 to 13) were maximized for a given number of sections to fit the readouts within the TR interval.
|
The QFs and CNRs of REX DIR RARE MR images were quantitatively compared with those of the images obtained with the conventional single-section DIR RARE sequence for the 16-, 18-, and 20-section protocols. The single-section DIR RARE sequence consisted of a DIR block followed by acquisition of a single section in one triggering period (two R-R intervals). The number of sections, as well as the other parameters for the single-section sequence (ie, the two-R-R-interval triggering, the echo time, the matrix size, the section thickness and separation, the bandwidth, the field of view, and the turbo factor), was chosen to be the same as those for the REX multisection sequence so that we could compare QF and CNR in the images equally.
Typical acquisition times (expressed in R-R intervals) for the REX multisection and conventional single-section DIR RARE methods are shown in Table 1. For example, if 20 sections with a 256 x 256 matrix and a turbo factor of 9 were to be imaged, acquisition of one section at intermediate- and T2-weighted imaging would take 256/9 = 29 TR intervals. If TR was two R-R intervals, it would take 58 R-R intervals to acquire one section. Therefore, to sequentially acquire 20 sections by using the conventional single-section approach, it would involve 58 x 20 = 1,160 R-R intervals. One dummy imaging pass of two R-R intervals preceded acquisition, and the total time required for the acquisition was 1,162 R-R intervals.
Qualitative analysis.—Two authors (Z.A.F., who had 12 years of cardiovascular MR imaging experience and J.G.S.A., who had 7 years of cardiovascular MR imaging experience) graded all the images (both those obtained with the single-section approach and those obtained with the REX approach) for three qualitative factors: overall quality of the image, flow suppression, and presence of artifacts by using an ordinal scale from 0–5, with 0 indicating very poor image quality, very poor flow suppression, and gross artifacts; 3 indicating average image quality, adequate flow suppression, and some artifacts; and 5 indicating excellent image quality, ideal flow suppression, and no artifacts. The concordance of the agreement between the observers’ findings was evaluated. The single-section and multisection approaches were then compared by using nonparametric statistical methods.
Statistical Analysis
Quantitative analysis.—A t test with equal variance was used to compare the ages of the patients and the healthy volunteers. A paired t test was used to compare the signal intensity values recorded by the two observers. One-way analysis of variance with a Bonferroni posttest with a statistical significance level of P < .05 was used to compare the QFs of the 16-, 18-, and 20-section DIR RARE protocols (there were two QFs for each multisection DIR RARE protocol) with the QFs of the equivalent single-section protocols. The CNRs of the multisection and single-section protocols were compared in the same manner.
Qualitative analysis.—The paired t test was used to compare the scores for overall image quality, flow suppression, and presence of artifacts recorded by the two authors. The concordance of the agreement between the observers’ findings was evaluated by using the Wilcoxon signed rank test. The single-section and multisection approaches were then compared by using Kruskal-Wallis one-way analysis of variance on ranks. Two-tailed P values of less than .05 were considered to indicate statistically significant differences.
Sigma Stat version 3.0 (SPSS, Chicago, Ill) was used to perform the statistical analyses. Power values were automatically computed by the software to verify the acceptability of the results of the statistical tests. A level of significance of .05 was used, and any power of .8 or greater was considered acceptable.
|
Results |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and Methods ResultsDiscussionREFERENCES |
|---|
|
|
The QF, CNR, and acquisition times for the rapid multisection DIR RARE sequence are quantitatively compared with those for the conventional single-section DIR RARE sequence in Table 2. The REX sequence had an improved QF compared with that of the single-section method (P < .05). The CNR values for the multisection REX DIR RARE protocols were not significantly different from those for the single-section DIR RARE protocols. The REX method demonstrated higher signal-to-noise ratios per unit time per section (16.0 ± 2.45 vs 7.5 ± 1.10, P < .05), no significant changes in contrast-to-noise ratios (15.0 ± 5.3 vs 20.1 ± 3.9, P > .05), and 17-fold improvement in acquisition time compared with those at conventional single-section DIR RARE imaging. The highest values of QF (17.2 ± 4.2) and CNR (15.6 ± 5.7) for the REX sequence were observed for the protocol involving 16 sections and four DIR blocks (Table 2). Time improvement factors (ie, ratios between the acquisition times for the single-section sequences and the acquisition times for the corresponding REX sequences) are shown in Table 2. The time improvement factors ranged from 13.0 (for the 16-section REX protocols) to 17.3 (for the 20-section REX protocols).
|
|
|
Discussion |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and Methods ResultsDiscussionREFERENCES |
|---|
The REX sequence separates the TR of the imaging section from the TR of blood by introducing multiple DIR blocks within two R-R intervals (the TR of a section). The time interval between two successive DIR blocks is the TR of blood. This shortened TRDB therefore reduces TI0 according to Equation (3) (12). The advantage of this TI0 shortening is the ability to interleave (acquire k-space lines from many sections within one TR) multiple inversion-recovery experiments within two R-R intervals.
Song et al (11) have found that the blood signal remains dark for a certain period of time after the application of a DIR block. By using this property, it was possible to acquire several section readouts (two to five) following each DIR block. The combination of the shortening of the TI0 with use of the acceptable blood nulling time window allowed rapid simultaneous acquisition of up to 20 sections. The time improvement factor yielded with use of the REX DIR RARE sequence (as compared with use of the single-section DIR RARE method with the same turbo factor), measured as a ratio of R-R intervals, was slightly less than the number of sections acquired with the REX protocol (16–20 sections) owing to the time used to perform the dummy imaging pass (two R-R intervals). Practically, owing to differences in the heart rate of a given subject during the experiment, time improvement factor measured as a ratio of the actual experiment times was different from time improvement factor measured as a ratio of theoretical values.
In the DIR technique, the blood flowing into the imaging plane after the TI0 interval has zero longitudinal magnetization because of the prior application of the nonselective inversion pulse. In our study, the sections were acquired along the blood flow direction (from head to foot) to augment outflow effects and hence improve blood suppression (ie, CNR). The second slab-selective radiofrequency pulse in the DIR block reinverted the magnetization of the entire slab of interest (16–20 sections), not just the sections imaged after the respective DIR block. This prevented the incomplete recovery of the longitudinal magnetization from the rest of the sections in the imaging slab during the time between two successive DIR blocks (200–450 msec, Table 1) and the resulting loss of muscle signal (ie, signal-to-noise ratio).
The time window during which the signal from blood was sufficiently suppressed (within 10% of the null point) was approximately 250 msec. For the multisection protocols, the total readout times ranged from 113 to 265 msec, enabling their acquisitions to fit into this time window.
The limitations of the present method included the non-ideal blood nulling in some sections caused by the different TI0 times for the sections in the same DIR block. Due to the time constraints of fitting readouts of multiple sections (>16) into two R-R intervals, the REX sequence for the protocols in this study can only be used with turbo factors of up to 13. Higher turbo factors could be used with the REX sequences if three R-R intervals are used. However, since image quality is compromised with longer echo trains (ie, images are blurred) and because we wanted to compare QFs and CNRs between the REX protocol and the single-section protocol, both protocols were performed with the same turbo factors (ie, factors of 9–13).
The images obtained by using the REX sequences were not compared with images obtained by using the half-Fourier DIR single-shot fast spin-echo method. The spatial resolution of the images was 0.97 x 0.97 x 3 mm. Higher-spatial-resolution imaging may need to be used for differentiation and quantification of plaque components. This technique can easily be adapted to a matrix size of 512 x 512, thereby increasing spatial resolution to 0.5 x 0.5 x 3 mm. However, more signal averages must be acquired to maintain sufficient signal-to-noise ratio and image quality. Signal-to-noise ratio may also be improved by using higher-field-strength magnets.
Other techniques for improving the efficiency of black-blood imaging include the use of single-shot RARE sequences (14,15). The image blurring caused by the long echo train length is the major limitation of the single-shot RARE technique. Results of one study involving use of the single-shot RARE sequence showed that image spatial resolution was improved by using such field of view reduction techniques as selective presaturation pulses (14). The REX technique could also be enhanced by implementing such field of view reduction techniques. This would improve the spatial resolution of black-blood imaging to a level that would facilitate vessel wall segmentation (17). Another technique for suppressing flow after administration of a T1-shortening contrast agent is quadruple inversion recovery (18). This technique is nonsensitive to variations in the T1 of blood and was used to image the carotid arterial wall in atherosclerotic patients and after contrast enhancement with gadolinium chelates.
Three-dimensional black-blood acquisitions have been used for vessel wall imaging (19–21). Advantages of three-dimensional techniques include a better section profile and signal-to-noise ratio. However, any motion that is not compensated for by gating will corrupt all sections in the imaging pass. A wrapping artifact is also present in Fourier-encoded three-dimensional imaging.
Multicontrast-weighted imaging (with T1, T2, and intermediate weighting) is possible with the REX technique. For T1-weighted images it is still possible to image up to 10 sections per TR interval. If the REX technique is used in combination with field of view reduction techniques, more efficient k-space coverage techniques such as the use of spiral readouts (22), and parallel imaging techniques such as generalized autocalibrating partially parallel acquisitions (GRAPPA) (23), sensitivity encoding (SENSE) (24), or simultaneous acquisition of spatial harmonics (SMASH) (25), it may be possible to image the entire length of the coronary arteries in a single breath hold.
In this report, a REX simultaneous multisection DIR RARE technique for black-blood vessel wall MR imaging has been described. The results were quantitatively compared with those obtained by using the conventional single-section DIR RARE technique. The technique developed in this study enables black-blood imaging of the aortic wall in healthy subjects and patients with atherosclerosis. The time improvement factor ranged from 13.0 to 17.3. With the technique described in this study, vessel wall imaging, especially in the aorta, can be performed in a reasonably short examination time to allow progression (26) and regression (27) studies of atherosclerotic plaques.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
V.M. and V.V.I. contributed equally to this work.
Author contributions: Guarantors of integrity of entire study, Z.A.F., V.M., V.V.I. The complete list of author contributions appears at the end of this article.
Author contributions: Guarantors of integrity of entire study, Z.A.F., V.M., V.V.I.; study concepts, M.S., V.M., V.V.I., Z.A.F.; study design, V.M., V.V.I., Z.A.F.; literature research, V.M., V.V.I.; clinical and experimental studies, V.M., V.V.I., J.G.S.A., D.D.S., G.M., Z.A.F.; data acquisition, V.M., V.V.I., G.M., M.S.; data analysis/interpretation, V.M., V.V.I., J.G.S.A., D.D.S., Z.A.F.; statistical analysis, V.M., V.V.I.; manuscript preparation and definition of intellectual content, V.M., V.V.I., Z.A.F.; manuscript editing, revision/review, and final version approval, all authors
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and Methods ResultsDiscussionREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  R. Bitar, A. R. Moody, G. Leung, S. Symons, S. Crisp, J. Butany, C. Rowsell, A. Kiss, A. Nelson, and R. Maggisano In Vivo 3D High-Spatial-Resolution MR Imaging of Intraplaque Hemorrhage Radiology, October 1, 2008; 249(1): 259 - 267. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
I. Koktzoglou, Y.-C. Chung, T. J. Carroll, O. P. Simonetti, M. D. Morasch, and D. Li Three-dimensional Black-Blood MR Imaging of Carotid Arteries with Segmented Steady-State Free Precession: Initial Experience Radiology, April 1, 2007; 243(1): 220 - 228. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. A. Wasserman, R. J. Wityk, H. H. Trout III, and R. Virmani Low-Grade Carotid Stenosis: Looking Beyond the Lumen With MRI Stroke, November 1, 2005; 36(11): 2504 - 2513. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 V. Fuster, Z. A. Fayad, P. R. Moreno, M. Poon, R. Corti, and J. J. Badimon Atherothrombosis and High-Risk Plaque: Part II: Approaches by Noninvasive Computed Tomographic/Magnetic Resonance Imaging J. Am. Coll. Cardiol., October 4, 2005; 46(7): 1209 - 1218. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
ns = 180° nonselective inversion pulse, 
