| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Vascular and Interventional Radiology |
1 From the Dotter Interventional Institute (C.A.B., B.D.P., J.A.K.) and Department of Diagnostic Radiology (P.D.B., J.S.), Oregon Health & Science University, Portland, Ore; Department of Radiology, Brigham and Women’s Hospital, Harvard Medical School, 75 Francis St, Boston, MA 02115 (C.A.B.); and Imaging Service, Portland VA Medical Center, Ore (P.D.B., B.D.P., J.S.). Received February 22, 2003; revision requested May 14; final revision received December 22; accepted January 13, 2004. Address correspondence to C.A.B. (e-mail: cbinkert@partners.org).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: Thirty men with symptomatic PVD were examined. MR angiography was performed at 1.5 T before conventional angiography. MR angiographic examination included 3D contrast-enhanced dedicated calf MR angiography and three-station bolus-chase MR angiography. Two radiologists blinded to conventional angiographic results evaluated the MR angiograms independently. Two angiographers evaluated the conventional angiograms in consensus. Calf artery segments were graded as having 50% or less stenosis, greater than 50% stenosis, or occlusion or as being nondiagnostic. Statistical analyses were performed with paired permutation testing.
RESULTS: Analyses of 472 calf segments and 420 pelvic and thigh segments were performed. Of the 472 calf segments, three and 75 segments (reader 1) and seven and 91 segments (reader 2) were graded as nondiagnostic at dedicated calf MR angiography and bolus-chase MR angiography, respectively. Differences in diagnostic grade between the two examinations were significant (P < .001), accounting for within-subject correlations, with a mean estimated difference of –17.1% (95% confidence interval [CI]: –25.8%, –8.4%). In the calf arteries, the dedicated and bolus-chase MR angiographic sequences had diagnostic accuracies, respectively, of 81.5% (reader 1) and 79.1% (reader 2) and of 67.8% (reader 1) and 63.4% (reader 2). The dedicated calf sequence was significantly more accurate than the bolus-chase sequence (P = .001). The point estimate of the difference was 14.7%, with estimated correct diagnosis rates of 80.3% and 65.6% for the dedicated calf and bolus-chase examinations, respectively (95% CI for difference: 4.0%, 25.4%). The diagnostic accuracy of bolus-chase MR angiography at the pelvis-thigh level was slightly higher when it was performed first: 81.9% (reader 1) and 83.8% (reader 2) versus 74.3% (reader 1) and 80.0% (reader 2) when it was performed last. The difference was not significant (P = .21).
CONCLUSION: Use of dedicated calf MR angiography led to significantly increased diagnostic accuracy in the calf arteries compared with standard bolus-chase MR angiography. Use of the dual-bolus technique did not jeopardize the diagnostic accuracy in the pelvic and thigh arteries.
© RSNA, 2004
Index terms: Angiography, comparative studies, 92.122, 92.12942, 98.122, 98.12942 • Arteries, extremities • Arteries, MR, 92.129412, 92.129416, 92.12942, 98.129412, 98.129416, 98.12942 • Arteries, stenosis or obstruction, 92.721, 98.721 • Magnetic resonance (MR), vascular studies, 92.12942, 98.12942
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
Adequate visualization of the distal runoff vessels is a requisite for accurate planning of endovascular or surgical treatment and is especially important in patients with diabetes, in whom small-caliber runoff vessels are more often affected (6).
The purpose of the present study was to compare the accuracy of three-dimensional contrast-enhanced dedicated calf MR angiography performed by using a dual-bolus technique with the accuracy of standard bolus-chase MR angiography in patients with symptomatic peripheral vascular disease, with conventional film hard-copy angiography as the reference standard.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
All patients were men aged 48–80 years (mean age, 65.8 years). All patients had symptomatic peripheral vascular disease; 10 had tissue loss, and five of these 10 men had rest pain; and 20 patients had lifestyle-limiting claudication. At the time of the angiographic examinations, 21 patients had bilateral symptoms and nine had unilateral symptoms. The average ankle-brachial index was 0.55. The patients had the following cardiovascular risk factors: 16 patients had diabetes mellitus, 20 were smokers, 21 had hypertension, and 18 had hypercholesterolemia. All patients gave informed written consent prior to participating in the study, which was approved by the institutional review board of Portland VA Medical Center.
MR Angiography
MR angiography was performed just before conventional angiography was performed. The MR angiographic examination had two parts: dedicated calf MR angiography from the knee to the ankle and three-station—that is, pelvis, thigh, and calf—bolus-chase MR angiography from the middle part of the abdomen to the ankle. A dual-bolus technique was performed, with the contrast agent injections for the dedicated and three-station parts of the examination separated by approximately 15 minutes. In the first 15 patients, dedicated calf MR angiography preceded three-station bolus-chase MR angiography, and in the second group of 15 patients, three-station bolus-chase MR angiography preceded dedicated calf MR angiography. In all patients, the total MR angiographic examination time was less than 1 hour.
All MR angiographic examinations were performed by using a 1.5-T MR imaging unit (Gyroscan NT Intera, release 7; Philips Medical Systems, Andover, Mass) with an automated moving table and a five-element surface spine coil (Quadrature Synergy Spine Coil; Philips Medical Systems). Patients were placed in the supine position, feet first, on the MR imaging table, with their feet and calves positioned over spine coil elements 2–5. The spine coil was positioned with the first two elements extending over the edge of the MR imaging table. Each patient’s feet and calves were secured with sandbags—one bag on each side and one bag between the legs—to minimize motion artifacts.
The MR angiographic examination started with an automated three-station (pelvis, thigh, and calf) two-dimensional time-of-flight scout sequence in which coronal, sagittal, and transverse maximum intensity projections that guided the dedicated calf and bolus-chase MR angiographic examinations were obtained. For dedicated calf MR angiography, a three-dimensional coronal fast field-echo gradient-echo sequence was performed with 5/1.5 (repetition msec/echo time msec), a 30° flip angle, a 530 x 318-mm field of view, a 512 x 256 matrix, a 1.5-mm section thickness, 0.75-mm section spacing, 100 sections, low-high k-space ordering, and a five-element surface spine coil with elements 2–5 activated. The imaging time was 49 seconds, and the voxel resolution was 1.04 x 1.24 x 1.50 mm. This sequence was performed before (to obtain mask images) and after (to obtain dynamic images) the gadopentetate dimeglumine (Magnevist; Berlex Laboratories, Montville, NJ) injection. Between the mask and dynamic image acquisitions, an MR fluoroscopic sequence (Bolus Trak; Philips Medical Systems) was performed and yielded a subtracted coronal two-dimensional projection of the calf station every second.
The dynamic dedicated calf MR angiogram acquisition was started manually when arterial contrast enhancement was visualized in the popliteal and proximal tibial arteries; there was a 5-second preprocessing delay between the start of the acquisition and the actual acquisition of imaging data. Twenty milliliters of gadopentetate dimeglumine was injected through a forearm or antecubital vein at 1 mL/sec by using an automated injector (Spectris MR Injector; Medrad, Indianola, Pa); 20 mL of saline was then injected at 1 mL/sec.
For three-station bolus-chase MR angiography, a modified version of a commercially available moving-bed infusion-tracking (MoBI-track; Philips Medical Systems) sequence was used. The modified three-dimensional coronal fast field-echo gradient-echo moving-bed sequence was performed with 5/1.3, a 30° flip angle, a 430 x 323-mm field of view, a 384 x 192 matrix, a 3.0-mm section thickness, 1.5-mm section spacing, 60 sections acquired per station, linear k-space ordering for the pelvis station and low-high k-space ordering for the thigh and calf stations, and the previously described five-element surface spine coil for the calf station and a whole-body coil for the pelvic and thigh stations. The imaging time was 81 seconds—27 seconds per station—with a 5-second delay between stations for automated table movement. Voxel resolution was 1.12 x 1.68 x 3.00 mm.
The commercially available moving-bed sequence consists of a three-station bolus-chase examination involving two (mask and dynamic) image acquisitions separated by a delay, which is determined by using a contrast agent timing bolus. To eliminate the need for a timing bolus, we modified the moving-bed sequence to a single-acquisition sequence that was performed before (to obtain mask images) and after (to obtain dynamic images) contrast agent injection, as in the dedicated calf MR angiographic part of the examination. Between the mask and dynamic image acquisitions, an MR fluoroscopic sequence (Bolus Trak) was performed and yielded a subtracted coronal two-dimensional projection of the pelvis station every second.
The dynamic three-station MR angiogram acquisition was started manually when arterial contrast enhancement was visualized in the middle part of the abdominal aorta; there was a 5-second preprocessing delay between the start of the acquisition and the actual acquisition of imaging data. For the modified moving-bed sequence, the acquisition volume of each station was angled anteroposteriorly to cover the arteries of interest, as visualized on the two-dimensional time-of-flight scout images. Forty milliliters of gadopentetate dimeglumine was injected: 20 mL was injected at 1 mL/sec, and 20 mL was injected at 0.5 mL/sec; then, 20 mL of saline was injected at 0.5 mL/sec.
Conventional Angiography
Immediately after the MR angiographic examinations, the patients were brought to the conventional angiographic suite (Multistar; Siemens, Erlangen, Germany). Intravenous conscious sedation was induced with midazolam (Versed; Roche Pharmaceuticals, Nutley, NJ) and fentanyl citrate (Sublimaze; Janssen Pharmaceutica, Titusville, NJ). In all patients, a 5-F pigtail catheter was placed in the abdominal aorta by using a common femoral artery approach. A bilateral angiogram was obtained by using the film hard-copy technique. Iodixanol (Visipaque; Amersham Health, Princeton, NJ) was injected at 8–10 mL/sec, for a total injected volume of 100–140 mL. The angiograms were obtained at five levels: the pelvis, thighs, knees, calves, and feet. Additional digital subtraction angiography was performed if the attending radiologist who performed the angiographic examination did not consider the initially obtained film hard-copy angiogram to be sufficient for diagnosis.
Image Analysis
Two radiologists with three (P.D.B.) and 11 (J.A.K.) years of experience in MR angiography evaluated the MR angiograms independently. Both radiologists were blinded to the conventional angiographic findings. The calf arteries were divided into eight vessel segments: below-knee popliteal artery, tibioperoneal trunk, proximal half of the anterotibial artery, distal half of the anterotibial artery, proximal half of the posterotibial artery, distal half of the posterotibial artery, proximal half of the peroneal artery, and distal half of the peroneal artery. Each segment was assigned a grade that represented one of four categories: Grade 1 indicated normal segment or minimal disease, or stenosis of 50% or less; grade 2, significant disease, or greater than 50% stenosis but not occlusion; grade 3, occlusion; and grade 9, nondiagnostic, or assessment not possible on angiograms.
The pelvic and thigh arteries were divided into seven segments: common iliac artery, external iliac artery, common femoral artery, deep femoral artery, proximal superficial femoral artery, distal superficial femoral artery, and above-knee popliteal artery. These segments were evaluated by using the same system that was used to grade the calf artery segments.
The MR angiograms were postprocessed by subtracting the corresponding mask images from the dynamic images. From the subtracted images, a series of two-dimensional maximum intensity projection images were created by using the entire acquired volume at each station, starting with a left sagittal projection and proceeding in 15° increments through a coronal to a right sagittal projection; 13 images of each station were obtained at each acquisition. For image interpretation, the maximum intensity projection images were viewed on a computer workstation (Easy Vision, software version 4.4.1; Philips Medical Systems). Source subtracted images were also available for interpretation on the computer workstation but were not always used. Evaluation of the dedicated calf MR angiograms was separated from evaluation of the bolus-chase MR angiograms by 2 months.
The conventional angiograms were evaluated by two interventional radiologists, who had been out of fellowship for 1 (C.A.B.) and 10 (B.D.P.) years. These radiologists interpreted the conventional angiograms in a consensus reading by using the same grading system that was used to grade the MR angiographic findings. These radiologists were not involved in the MR angiogram readings. Conventional angiography served as the reference standard for evaluating the diagnostic accuracy of MR angiography. Assignments of nondiagnostic segments at MR angiography were counted as inaccurate readings.
Statistical Analyses
All segments that were considered diagnostic at conventional angiography were included in the analysis. All statistical calculations were performed by using standard statistical software (S-PLUS 2000; MathSoft, Seattle, Wash). A paired permutation test involving the use of the average difference in correct diagnosis rate between dedicated calf MR angiography and bolus-chase calf MR angiography was conducted. Possible correlations between observations were corrected by performing this test. P values were calculated on the basis of 10 000 permutations. Cohen 
statistics were used to measure the concordance of results between the two readers. Significant differences were indicated by P < .05.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
|
The diagnostic accuracies of the two MR angiographic techniques are summarized in Table 2. At imaging of the calf arteries, the diagnostic accuracy of dedicated calf MR angiography was significantly superior to that of bolus-chase MR angiography (P = .001). The point estimate of the difference was 14.7% (95% confidence interval: 4.0%, 25.4%.), with estimated correct diagnosis rates (ie, as averaged for both readers) of 80.3% for dedicated calf MR angiography and 65.6% for bolus-chase calf MR angiography. Interobserver agreement was only minimally better with dedicated calf MR angiography. Based on the results of a paired permutation test performed by using the average difference in Cohen values, a P value of .60 was calculated. The point estimate of the difference was 0.040, and the 95% confidence interval for the difference was –12.7%, 20.8%. Examples of improved diagnostic accuracy at dedicated calf MR angiography are shown in Figures 1 and 2.
|
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
The quality of any contrast-enhanced imaging modality is dependent on the timing of the image acquisition in accordance with the maximal contrast enhancement to optimize the signal-to-noise ratio. Until recently, the long image acquisition times associated with peripheral MR angiography had led to the inability to follow the contrast agent bolus down the leg and therefore to poor signal-to-noise ratios and venous contamination (7). This phenomenon has been especially true for the small arteries in the distal calf. Until these issues are resolved, peripheral MR angiography will not replace conventional catheter angiography as the sole pretreatment planning modality for patients with critical lower extremity ischemia.
In this study of a dual-contrast technique used to perform bolus-chase MR angiography combined with dedicated calf MR angiography, we were able to adequately image the tibial arteries. All examinations were performed in patients with symptomatic peripheral vascular disease, and the MR angiographic examination was performed in less than 1 hour. Most important, both techniques were directly compared with conventional angiography that was performed on the same day.
The concept of contrast-enhanced bolus-chase MR angiography of the peripheral arteries is based on the goal of following the arterial contrast agent bolus from the pelvis to the calves. With a peripheral venous injection, the average arterial contrast agent bolus arrives in the common femoral artery in 24 seconds and in the middle-calf tibial arteries in 30 seconds (8). The contrast agent arrival time in the veins was measured, and mean arterial phase windows of 49, 45, and 35 seconds were calculated for the pelvic, thigh, and calf stations (8). To achieve a sufficient signal-to-noise ratio and spatial resolution, the acquisition time for each station during bolus-chase MR angiography is as long as 20 seconds (5) to 28 seconds (9). In this study, the image acquisition time for each of the three stations examined at bolus-chase MR angiography was 27 seconds. An arterial contrast agent bolus will generally travel from the common femoral artery to the tibial arteries of the middle part of the calf in approximately 6 seconds (8).
With use of a multistation bolus-tracking contrast-enhanced MR angiographic technique with long acquisition times, the further the bolus travels down the legs, the longer the acquisition will lag behind the contrast agent bolus, decreasing the signal-to-noise ratio and increasing the probability of venous overlap. In addition, the arterial phase window becomes increasingly shorter down the leg, especially in patients with cellulite, the presence of which increases the chances of venous contamination (8,10). We found that owing to long acquisition times, a decreasing arterial phase window, and increasing venous contamination, the diagnostic accuracy of bolus-chase angiography was lower for the calf station (67.8% for reader 1, 63.4% for reader 2) than for the pelvic and thigh stations (78.1% for reader 1, 81.9% for reader 2). Similar findings have been demonstrated in other studies: diagnostic quality values of 100% in the pelvis, 96% in the thigh, and only 43% in the calf (5).
To overcome these limitations, the use of a time-of-flight sequence has been recommended for imaging the distal calf arteries (11). However, this technique is limited by in-plane saturation and prolonged acquisition times and has been shown to be inferior to two-dimensional contrast-enhanced MR angiography for imaging the distal tibial arteries (12). Two-dimensional contrast-enhanced MR angiography yields excellent, diagnostic-quality images of the calf arteries, but owing to the lack of dimensionality, additional sequences have been required in up to 32% of cases (5).
In this series, the dedicated three-dimensional contrast-enhanced MR angiographic examination of the calf began with a manual timing of the arterial contrast agent bolus to synchronize the acquisition to the arterial phase window and to avoid venous contamination. Use of this technique led to a reduced percentage of nondiagnostic calf segments: to 0.6%–1.5%. These results are similar to the 0.7% of nondiagnostic segments seen in the pelvic and thigh stations and represent a significant (P < .001) reduction from the 15.9%–19.3% of nondiagnostic segments seen in the calf at standard bolus-chase MR angiography.
To increase spatial resolution, the voxel size was reduced from 1.12 x 1.68 x 3 mm (with bolus-chase sequence) to 1.04 x 1.24 x 1.5 mm (with dedicated calf sequence). The increased spatial resolution led to an increased imaging time: from 27 to 49 seconds. However, because of the accurate timing of the dedicated calf sequence, severe venous overlap did not occur. With these technique alterations, the diagnostic accuracy of dedicated calf MR angiography was comparable to that of bolus-chase MR angiography of the pelvic and thigh arteries. The highest diagnostic accuracy for imaging the calf arteries was achieved when the dedicated calf sequence was performed before the standard bolus-chase sequence.
Because the diagnostic accuracy of dedicated calf MR angiography remained higher, even when it was performed after three-station bolus-chase MR angiography, a dedicated calf acquisition could possibly be performed after the bolus-chase examination in the event that a calf-station segment was judged to be nondiagnostic. Injection of the contrast agent before the bolus-chase MR angiographic examination did not significantly alter the diagnostic accuracy of the angiograms of the pelvic and thigh arteries that were acquired afterward.
Two techniques to improve contrast-enhanced MR angiography of the calf arteries have recently been described (13,14). Maki et al (13) used sensitivity encoding and elliptical centric phase encoding to shorten the acquisition time in the pelvic and thigh stations and thus decrease the delay before the calf-station acquisition. Despite the shorter upper-station image acquisition time, there was still a 34-second delay between contrast agent arrival in the aorta and the beginning of the acquisition at the calf level. The 34-second delay combined with a prolonged imaging time at the calf station (71 seconds) seems to increase the chance of venous contamination compared with the technique described in the present study.
Swan et al (14) described a time-resolved three-dimensional contrast-enhanced MR angiographic technique in which images of each station (ie, pelvis, thigh, and calf) are acquired separately so that optimal contrast timing can be achieved by acquiring a "snapshot" every 6–7 seconds. The described voxel size of this technique at the calf station is 1.54 x 1.19 x 2.00 mm. The poorer spatial resolution, as compared with that of the dedicated calf sequence described herein, potentially decreases the diagnostic accuracy. Neither sensitivity-encoding MR angiography nor time-resolved MR angiography are widely available yet, whereas the sequences and hardware described in the present study are readily available.
There were some limitations to the present study. The 60 mL of gadopentetate dimeglumine administered for MR angiography added considerable cost to the examination. Smaller amounts of contrast agent may be sufficient, and the lower limit has yet to be established. The pedal arteries, which were not evaluated during this study, can be important for distal bypass surgery. In selected patients who were not included in this study, we performed a similar dedicated MR angiographic examination, the results of which indicated that the principles presented herein seem to apply for the feet.
In conclusion, the use of dedicated three-dimensional contrast-enhanced MR angiography increases the diagnostic accuracy of imaging of the calf arteries as compared with the diagnostic accuracy of standard bolus-chase MR angiography. Contrast agent administration for dedicated calf MR angiography before the bolus-chase part of the examination did not jeopardize the diagnostic accuracy of bolus-chase angiography of the pelvic and thigh arteries.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Author contributions: Guarantors of integrity of entire study, C.A.B., P.D.B.; study concepts and design, C.A.B., P.D.B., B.D.P.; literature research, C.A.B., P.D.B.; clinical studies, C.A.B., P.D.B., B.D.P., J.A.K.; data acquisition, all authors; data analysis/interpretation, C.A.B., P.D.B., B.D.P., J.A.K.; manuscript preparation, C.A.B.; manuscript definition of intellectual content, C.A.B., P.D.B., B.D.P.; manuscript editing, P.D.B., B.D.P.; manuscript revision/review, P.D.B., B.D.P., J.S., J.A.K.; manuscript final version approval, all authors
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  H. Ersoy and F. J. Rybicki MR Angiography of the Lower Extremities Am. J. Roentgenol., June 1, 2008; 190(6): 1675 - 1684. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Andreisek, T. Pfammatter, K. Goepfert, D. Nanz, P. Hervo, R. Koppensteiner, and D. Weishaupt Peripheral Arteries in Diabetic Patients: Standard Bolus-Chase and Time-resolved MR Angiography Radiology, December 19, 2006; (2006) 2422051111. [Abstract] [Full Text]
|
|
|
|
|
|
M. Lapeyre, H. Kobeiter, P. Desgranges, A. Rahmouni, J.-P. Becquemin, and A. Luciani Assessment of Critical Limb Ischemia in Patients with Diabetes: Comparison of MR Angiography and Digital Subtraction Angiography Am. J. Roentgenol., December 1, 2005; 185(6): 1641 - 1650. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Ota, K. Takase, H. Rikimaru, M. Tsuboi, T. Yamada, A. Sato, S. Higano, T. Ishibashi, and S. Takahashi Quantitative Vascular Measurements in Arterial Occlusive Disease RadioGraphics, September 1, 2005; 25(5): 1141 - 1158. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
