| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Technical Developments |
1 From the Department of Radiology, Kumamoto University School of Medicine, 1-1-1 Honjo, Kumamoto 860, Japan. Received March 1, 1999; revision requested April 15; final revision received December 6; accepted December 20. Address correspondence to K.M. (e-mail: katsumit@kaiju.medic.kumamoto-u.ac.jp).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMaterials and Methods ResultsDiscussionREFERENCES |
|---|
Index terms: Angiography, comparative studies, 92.122, 93.122, 96.122, 98.122 • Magnetic resonance (MR), comparative studies, 92.129416, 93.129416, 96.129416, 98.129416, • Magnetic resonance (MR), vascular studies, 92.129416, 93.129416, 96.129416, 98.129416,
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and Methods ResultsDiscussionREFERENCES |
|---|
Three-dimensional turbo MR angiography is a fast low-angle shot, or FLASH, sequence that uses asymmetric echo encoding in the readout direction, and the number of phase-encoding steps in the readout direction can be reduced by means of asymmetric sampling. This technique permits a shorter imaging time, reducing motion artifacts. Furthermore, it may be possible to reduce intravoxel phase dispersion and signal loss with short echo times. The loss in signal-to-noise ratio that accompanies a reduction in imaging time is partially compensated by means of a noise truncation filter applied to the raw data and results in maintenance of the signal-to-noise ratio and spatial resolution. We assumed that the combination of a short echo time, zero interpolation, and reordering of phase- and section-encoding steps would enable us to reduce intravoxel phase dispersion and signal loss with short echo times and to obtain high-quality contrast material–enhanced 3D MR angiograms.
In this study, we analyzed diagnostic accuracy for evaluating stenosis of the abdominal aorta and its branches and arteries in the pelvis and lower extremity with dynamic contrast-enhanced 3D turbo MR angiography compared with conventional cut-film angiography.
|
Materials and Methods |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and Methods ResultsDiscussionREFERENCES |
|---|
All patients were referred by clinicians to detect possible arterial abnormalities. MR and conventional angiography were performed within 2 weeks in each patient. None of the patients were excluded.
Data Collection
MR angiography.—All MR images were obtained on a 1.5-T superconducting unit with a maximum gradient capability of 25 mT/m (Magnetom Vision; Siemens, Erlangen, Germany) with a body phased-array coil. Gadolinium-enhanced MR angiography was performed with a 3D fast low-angle shot sequence (repetition time msec/echo time msec of 4.4/1.4, flip angle of 30°, matrix of 224 (phase) x 320 (readout) with left-to-right phase encoding, readout bandwidth of 195 Hz per pixel, slab thickness of 90 mm, and 20 partitions). Asymmetric echo encoding in the readout direction, the number of phase-encoding steps in the section direction, and zero filling to sampling points were used for 500-point acquisition. MR angiography was performed in the coronal or coronal-oblique plane. Mean acquisition time was 19 seconds, and the mean field of view was 40 cm (range, 38–42 cm).
At 3D turbo MR angiography, only part of the k space is imaged. The remaining encoding steps in the k space are filled with zeros. The central portion of the phase-encoding steps is obtained in the beginning of acquisition. A further reduction in the number of phase-encoding steps is possible with use of a sinc interpolation technique in the section direction to create thin section partitions. In turbo MR angiography, central section-encoding steps are acquired (Fig 1). As a result, we can expect a more consistent in-plane resolution of the resultant maximum intensity projections, independent of the exact timing of bolus administration. Incorrect timing may result in a loss of spatial resolution along the section-select direction. This technique permits a short imaging time, which reduces motion artifacts.
|
|
The sequences were repeated before and after injection of gadopentetate dimeglumine (Magnevist; Schering, Osaka, Japan) with breath holding. The patients were instructed to hold their breath in a constant degree of inspiration to avoid misregistration on the subtraction images. Before gadolinium-enhanced MR angiography, a timing sequence was performed to determine the arrival time of the contrast material bolus in the imaged volume. This technique consisted of a two-dimensional turbo fast low-angle shot sequence: 5.0/2.3/300 (inversion time msec), flip angle of 15°, field of view of 300 x 400, matrix of 96 x 256, and 8-mm section thickness. With this dynamic sequence, sequential transverse images were obtained through the abdominal aorta 60 times, with a temporal resolution of 0.89 second. A 100-mm-thick caudal presaturation slab was used to suppress signal from inflowing venous blood, and an 80-mm-thick cranial presaturation slab was used to prevent inflow effects from arterial blood. A test bolus of 1 mL of gadopentetate dimeglumine with a subsequent flush of 20 mL of normal saline solution was injected by using a power injector (Spectris model SRM 230; MedRad, Pittsburgh, Pa) at the same injection rate (2 mL/sec) as was used for gadolinium-enhanced MR angiography. The dynamic sequence and the injection of 1 mL of gadopentetate dimeglumine (and subsequent 20-mL normal saline solution flush) were started simultaneously.
A bolus of gadopentetate dimeglumine (0.2 mmol per kilogram of body weight) was intravenously infused with a power injector for 10–12 seconds before imaging; this injection was immediately followed with 20 mL of normal saline solution to complete delivery of the entire dose of contrast material and to flush the vein. Because the 3D turbo MR angiographic sequences used in this study fill in k space from the center, the central portion of k space was acquired at the beginning of the data acquisition. The injection was started with an imaging delay determined on the basis of the timing of the test injection data. The imaging protocol was completed with two contiguous data acquisitions (measurements), and no time elapsed between acquisitions.
The magnitude data of the measurement before administration of gadopentetate dimeglumine was subtracted on a section-by-section basis from the magnitude data of the measurement after contrast enhancement, and then a maximum intensity projection algorithm was used for image reconstruction. The position of the coil was moved in imaging of the pelvic or femoral arteries, and additional images were obtained with the same 3D turbo MR angiographic sequence.
Conventional angiography.—Conventional angiography (model DR-2400; Shimadzu, Kyoto, Japan) was performed after MR angiography. The standard pelvic angiographic protocol consisted of puncture of the common femoral artery and placement of a 5-F pigtail catheter in the distal aorta just above the bifurcation (if examination of two legs was necessary). Nonionic contrast matertial (Iopamiron 350; Nihon Schering, Osaka, Japan) was injected at a rate of 18 mL/sec with a dose of 35 mL for conventional angiography or at a rate of 18 mL/sec with a dose of 25 mL for digital subtraction angiography. The acquisition rate for conventional angiography and frame rate for digital subtraction angiography were three images per second for 3 seconds and one image every 11 seconds, respectively. All digital subtraction angiograms consisted of two frames per second for 3 seconds and one frame every 10 seconds. The matrix size of the digital image intensifier was 1,024 x 1,024.
For unilateral studies, selective angiography was performed after insertion of a 5-F cobra type catheter. Contrast material was injected at a rate of 5 mL/sec with a dose of 20 mL for conventional angiography or at a rate of 5 mL/sec with a dose of 15 mL for digital subtraction angiography. The acquisition rate and frame rate were the same as those used in the standard pelvic angiographic protocol. Standard selective renal angiography consisted of placement of a 5-F cobra type or a duck-head type catheter in the proximal main renal artery and injection of contrast material at a rate of 5 mL/sec with a dose of 15 mL for conventional angiography or at a rate of 5 mL/sec with a dose of 12 mL for digital subtraction angiography. Either conventional or intraarterial digital subtraction angiograms were obtained in the posteroanterior projection. Oblique views were obtained when severe tortuosity of the occlusive arteries was seen.
Data Analysis
In each patient with atherosclerotic obstructive disease, the arterial tree on the maximum intensity projections and conventional angiograms was divided into the following segments: lower abdominal aorta, common iliac artery, external iliac artery, common femoral artery, superficial femoral artery, and deep femoral artery. Because selective angiography was not performed in some patients due to complete occlusion, the number of arterial segments differed from segment to segment: 18 lower abdominal aorta, 18 right common and 18 external iliac arteries, 17 left common and 17 external iliac arteries, 18 right common and 18 superficial femoral arteries, 17 left common and 17 superficial femoral arteries, and 18 right and 17 left deep femoral arteries. With the addition of eight renal arteries in four patients with renal artery stenosis, a total of 201 segments were evaluated.
Quantitative Analysis
All MR angiograms were evaluated independently by two observers (K.M., Y.Y.), vascular radiologists who were unaware of the results of conventional angiography. The observers evaluated all imaged vessel segments with regard to the degree and length of stenoses. They correlated the degree of stenosis for each segment on MR and conventional angiograms side by side. To measure the percentage of stenosis on conventional and digital subtraction angiograms, calipers were used to measure the diameter of the most severely stenotic portion of the segment relative to the diameter of a nonstenotic portion of the artery. To maximize measurement accuracy, a ruler with 0.5-mm markings was used for measurement. The length of stenosis on MR angiograms was defined as the length of a vessel with an abnormal diameter compared with that of a normal segment. For conventional angiography and digital subtraction angiography, the length of the stenosis was measured by means of calibration with an angiographic catheter.
Qualitative Analysis
To evaluate the diagnostic accuracy for hemodynamically significant stenosis depicted at MR angiography, MR angiograms—including source images and maximum intensity projections—were evaluated separately by same two blinded vascular radiologists. A focal stenosis was considered hemodynamically significant when there was a reduction in diameter of more than 50%. Vessels were classified as patent or occluded. Patent segments were graded as normal or minimally diseased, focally stenotic, or diffusely diseased (multiple [more than three] hemodynamically significant stenoses). The presence of collateral vessels and any artifacts or limitations was also noted. Finally, vascular segments that were interpreted differently by the observers were reevaluated for consensus.
Conventional angiograms were evaluated with regard to degree of stenosis and the diagnostic accuracy for hemodynamically significant stenosis by the two observers and categorized in the same way as were the MR angiograms. The degree of stenosis on the conventional angiograms was established by consensus during sessions 4 weeks before interpretation of the MR angiograms to avoid recall. Disagreements about the degree of stenosis occurred for two lesions, but the observers reached consensus after conference. The validity of MR angiography, as judged on the basis of the degree of stenosis and the diagnostic accuracy for hemodynamically significant stenosis depicted on MR angiograms, was determined by comparing the results with those at conventional angiography as the standard of reference.
Statistical Analyses
The correlations of degree and length of stenoses measured on MR and conventional angiograms were assessed with the Pearson test. A P value of less than .05 was used as a threshold for statistical significance. The agreement between interobserver variability for detection of stenotic lesions on MR and conventional angiograms was determined with an unweighted 
statistic. The values greater than 0 were considered to indicate positive correlation; 0.40 or less, positive but poor correlation; 0.41–0.75, good correlation; and greater than 0.75, excellent correlation. Standard statistical techniques were used to determine diagnostic accuracy (sensitivity, specificity) for hemodynamically significant stenosis at MR angiography compared with conventinal angiography as the reference method.
|
Results |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and Methods ResultsDiscussionREFERENCES |
|---|
|
|
|
Quantitative Analysis
The degree of stenosis depicted at MR angiography correlated well with that depicted at conventional angiography (P < .001) (Fig 3). The Pearson correlation coefficient between MR and conventional angiography was 0.77, but the degree of stenosis tended to be overestimated at MR angiography in comparison with conventional angiography.
|
|
|
|
|
|
|
|
= 0.85) was better than that at MR angiography ( = 0.79). For the 201 vascular segments, the results at MR angiography for distinguishing vessels with more than 50% stenosis from those that were normal or mildly stenotic also showed good correlation between the two observers, as well as consensus reading, as shown in Table 2. At consensus reading, 148 results were true-negative; 32, true-positive; 18, false-positive; and three, false-negative (Table 2). These results yielded a sensitivity of 91%, specificity of 89%, and accuracy of 90%. Although the sensitivity of MR angiography was excellent, the overestimation of stenotic vessels resulted in somewhat limited specificity.
|
|
Discussion |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and Methods ResultsDiscussionREFERENCES |
|---|
Mulligan et al (9) reported an overall accuracy of 71% with two-dimensional time-of-flight technique for peripheral arteries. Yucel et al (11) reported a sensitivity of 92% and a specificity of 88% for distinction of vessels with greater than 50% stenosis. In other reports (14,16), direct comparison between two methods indicated that results obtained with contrast-enhanced 3D MR imaging were significantly better than those obtained with time-of-flight MR technique. Adamis et al (6) and Douek et al (7) initially reported dynamic contrast-enhanced subtraction MR angiography. When a phased-array coil is used, increased signal from subcutaneous fat may obscure small vessels, especially when the maximum intensity projection algorithm is employed. Subtraction is an effective means of suppressing the effect of background tissue. With the 3D subtraction MR imaging technique, high-spatial-resolution images can be obtained with a relatively small amount of gadopentetate dimeglumine in different regions. Venous enhancement with the previous gadopentetate dimeglumine injection can be effectively canceled, and images can be obtained with a relatively small amount of gadopentetate dimeglumine (5 mL for the lower extremities) (8).
The concentration of paramagnetic contrast agent during the time when the data are acquired is an important factor in obtaining MR angiograms with good quality. Since cardiac output is inconsistent, especially in patients with significant atherosclerotic disease, the arrival time and peak concentration of contrast material within the aorta may vary considerably. Ideally, maximal arterial enhancement should occur during the midpoint of the data acquisition, when the central portion of the k space is being filled (17,18). If imaging is improperly timed, data may be acquired prior to or after peak arterial enhancement, which results in a study with suboptimal contrast and in image blurring and nonuniform vascular enhancement as a result of amplitude modulation in the MR signal (19).
One method of obtaining optimal contrast is to use an MR-compatible mechanical injector to standardize the rate of administration of contrast material and improve performance of timing studies with a small amount of contrast material (20). Furthermore, reordering of phase- and section-encoding steps as in this study is another approach to obtain maximal enhancement at desirable timing. If data are acquired with centric ordering, maximal arterial enhancement is obtained at the beginning, when data acquisition is initiated on the basis of imaging delay time in a timing study. We used a power injector and performed a timing study before the practical study in all our patients, and none had studies with suboptimal contrast. The combination of a short echo time, zero interpolation, and reordering of phase- and section-encoding steps enabled us to obtain high-quality contrast-enhanced 3D MR angiograms. Our technique had a sensitivity of 91%, a specificity of 89%, and diagnostic accuracy for distinction of vessels with greater than 50% stenosis from those that were normal or mildly stenotic.
Although we used a short echo time, spatial resolution by means of zero interpolation, timing on the basis of a timing study, and reordering of encoding steps, several regions of disease were still overestimated with MR imaging. This overestimation is probably a result of spin dephasing from complex high-velocity flow (21,22). The turbulent flow could not be compensated for despite the use of gadopentetate dimeglumine and a sequence with very short echo time. Furthermore, partial volume averaging due to the lower spatial resolution of MR angiography would be another possible cause of overestimation at 3D turbo MR angiography. Although the interpolation technique may increase the apparent image resolution, it does not affect physical spatial resolution (23,24). Calcification in the stenotic portion may also cause spin dephasing. Image subtraction can be imperfect owing to patient motion or peristalsis of the intestine. Although these were not a clinically important problem in the present study, the results may be affected when larger numbers of patients are examined. Furthermore, development of the technique is necessary to improve its specificity with shorter echo time to reduce flow-related dephasing and to increase imaging resolution. Accurate measurement of the length of occlusive arteries may be difficult in a single projection with either conventional or MR angiography. Therefore, we compared images with the same projection for both conventional and MR angiography.
In conclusion, the turbo MR angiographic technique is a modification of contrast-enhanced 3D MR angiography that combines a short echo time, zero interpolation, and reordering of phase- and section-encoding steps. Although turbo MR angiography has a number of inherent pitfalls, such as overestimation of stenosis and inability to depict calcification, it allows excellent visualization of normal and diseased vessels within a short imaging time. Therefore, we believe this technique is useful for screening of patients with arterial occlusive diseases. However, the definition of the role of this imaging technique in place of conventional angiography awaits further investigation.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Author contributions: Guarantors of integrity of entire study, K.M., Y.Y.; study concepts, K.M., Y.Y.; study design, K.M., T.S.; definition of intellectual content, K.M., M.T.; literature research, K.M.; clinical studies, Y.H., K.M., Y.Y.; data acquisition, Y.H., K.M., Y.Y.; data analysis, K.M., Y.Y.; statistical analysis, K.M., I.O.; manuscript preparation and editing, K.M., Y.Y.; manuscript review, Y.Y., M.T.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and Methods ResultsDiscussionREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  R. Ouwendijk, M. C. J. M. Kock, K. Visser, P. M. T. Pattynama, M. W. de Haan, and M. G. M. Hunink Interobserver Agreement for the Interpretation of Contrast-Enhanced 3D MR Angiography and MDCT Angiography in Peripheral Arterial Disease Am. J. Roentgenol., November 1, 2005; 185(5): 1261 - 1267. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Sakuma, Y. Ichikawa, N. Suzawa, T. Hirano, K. Makino, N. Koyama, M. Van Cauteren, and K. Takeda Assessment of Coronary Arteries with Total Study Time of Less than 30 Minutes by Using Whole-Heart Coronary MR Angiography Radiology, October 1, 2005; 237(1): 316 - 321. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. K. Willmann, S. Wildermuth, T. Pfammatter, J. E. Roos, B. Seifert, P. R. Hilfiker, B. Marincek, and D. Weishaupt Aortoiliac and Renal Arteries: Prospective Intraindividual Comparison of Contrast-enhanced Three-dimensional MR Angiography and Multi-Detector Row CT Angiography Radiology, March 1, 2003; 226(3): 798 - 811. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E A Dick, R Aviv, I Francis, G Hamilton, D Baker, C Black, A Platts, and A Watkinson Catheter angiography and angioplasty in patients with scleroderma Br. J. Radiol., December 1, 2001; 74(888): 1091 - 1096. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
