| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Vascular and Interventional Radiology |
1 From the Departments of Radiology (N.M.K., P.A.W., M.R.P., E.V., D.W.T., R.W., Y.W.) and Surgery (H.L.B.), New York Presbyterian Hospital, Weill Medical College of Cornell University, 525 E 68th St, Rm P-519, New York, NY 10021. Received February 5, 2001; revision requested March 26; final revision received November 14; accepted December 11. Address correspondence to N.M.K. (e-mail: nmkhilna@med.cornell.edu).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: Three radiologists retrospectively reviewed the pretreatment x-ray angiographic and MR angiographic studies obtained in 30 consecutive patients: 15 patients (15 limbs) evaluated for limb salvage and 15 patients (20 limbs) evaluated because of claudication. MR angiography included acquisition of 2D contrast-enhanced MR digital subtraction angiograms of the area from the adductor canal to the feet and 3D spoiled gradient-recalled-echo bolus chase MR angiograms obtained in three stations from the aorta to the middle portion of the calf. Each reader reviewed the x-ray and MR angiograms to determine the inflow and outflow segments for a hypothetical bypass graft placement.
RESULTS: The three readers selected identical segments for inflow at MR angiography and x-ray angiography in 32, 32, and 35 of the 35 limbs evaluated (mean percentages of agreement [95% CI ]: 91% [77%, 98%], 91% [77%, 98%], and 95% [90%, 100%], respectively). The readers selected identical segments for outflow in 32, 32, and 34 of the 35 limbs evaluated (mean percentages of agreement [95% CI]: 91% [77%, 98%], 91% [77%, 98%], and 97% [85%, 100%], respectively).
CONCLUSION: Preliminary data support the combining of 2D MR digital subtraction angiography with 3D bolus chase MR angiography to extend the utility of 3D MR angiography in treatment planning to include patients being evaluated for limb salvage, as well as those being evaluated for claudication.
© RSNA, 2002
Index terms: Angiography, comparative studies, 928.1221, 928.1222, 928.12942, 988.1221, 988.1222, 988.12942 • Arteries, stenosis or obstruction, 928.721, 988.721 • Magnetic resonance (MR), vascular studies, 928.129412, 928.12942, 928.12943, 988.129412, 988.12942, 988.12943
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
However, there are several limitations to relying on 3D bolus chase MR angiography as a stand-alone procedure. These limitations include unpredictable venous enhancement that can obscure arteries, motion artifacts, and the 1.5–2.0-mm spatial resolution that is inherent to the imaging matrix typically used. These limitations are most apparent when the infrapopliteal vessels are imaged, as demonstrated by Wang et al (11). In their review of 89 three-station bolus chase MR studies, diagnostic images were obtained in 100% and 96% of the abdominal-pelvic and thigh stations, respectively, but in only 43% of the calf stations. Treatment planning for patients who are being evaluated for limb salvage requires high-quality images of the calf and pedal vessels. For patients with claudication, high-quality imaging to the middle portion of the calf is required in all cases and high-quality imaging to the feet is required for those patients in whom infrageniculate bypass or endovascular procedures are being contemplated (12).
The inability to reliably and accurately image the infrapopliteal vessels with bolus chase MR angiography limits the usefulness of this examination in treatment planning, especially for patients being evaluated for limb salvage. This was demonstrated in a study (13) in which treatment planning with 3D bolus chase MR angiography was compared with that with x-ray angiography: Identical plans were generated for 10 of 11 patients who were being evaluated because of claudication but for neither of the two patients who were being evaluated for limb salvage.
To overcome the imaging limitations of 3D bolus chase MR angiography in the calf, additional acquisitions can be performed. One report (18) describes success that was achieved by using a combination of contrast-enhanced 3D MR angiography of the abdomen and pelvis with nonenhanced two-dimensional (2D) time-of-flight MR imaging of the infrainguinal vessels. However, time-of-flight MR imaging has several limitations, including inadequate depiction of arteries because of flow artifacts and long acquisition times (12,14–19).
Two-dimensional contrast-enhanced MR angiography with use of complex vector digital subtraction depicts significant stenoses in infrapopliteal vessels better than time-of-flight MR imaging (20) or x-ray angiography (4). The 2D contrast-enhanced digital subtraction angiographic (DSA) sequence yields time-resolved projectional angiograms with a 1.5-second frame rate (21). These images generally are obtained as coronal or sagittal 2D slabs that have an appearance that is similar to that of digital subtraction x-ray angiograms. The in-plane spatial resolution of 2D contrast-enhanced MR DSA of the infrageniculate arteries with a head transmit-receive coil is 1 mm.
In our clinical practice, MR evaluation of peripheral vascular disease involves the use of 2D MR DSA of the area from the adductor canal to the middle portion of the foot combined with three-station 3D bolus chase MR angiography in a single examination. This combined 2D-3D MR angiographic protocol can be performed easily within 45 minutes without patient preparation, sedation, or recovery. At our institution, this imaging strategy is increasingly being used without x-ray angiographic correlation to make treatment decisions for patients with peripheral vascular disease. However, to our knowledge, the accuracy of this protocol in treatment planning has never been objectively evaluated. The purpose of this study was to evaluate the utility of 2D MR DSA combined with 3D bolus chase MR angiography, as compared with the utility of x-ray angiography, for treatment planning.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
To be a part of this review, each patient had to have undergone complete MR angiographic and complete x-ray angiographic examinations of the symptomatic extremity. Complete MR angiography consisted of 3D bolus chase MR angiography of the abdomen, pelvis, and lower extremities, as well as 2D MR DSA evaluation of the area from the adductor canal to the feet. Complete x-ray angiography consisted of DSA of the area from the infrarenal aorta to the foot for patients being evaluated for limb salvage and DSA of the area extending to at least the middle part of the calf for patients with claudication. The infrapopliteal arteries of the limb or limbs in question had to be examined with selective angiography; no nonselective infrageniculate x-ray angiograms were included in this study. No interventions or changes in the clinical status of the limb ischemia could have occurred between the two angiographic studies, which had to have been performed within 20 weeks of each other.
Only clinically relevant extremities that were being imaged for revascularization were included in this analysis, even if an asymptomatic contralateral extremity was completely imaged. The reports on each of the potential candidates for the study were reviewed, and the patients with metal arterial stents (two patients), with gross motion artifact that made the images uninterpretable to the examining physician reader (three patient’s MR angiograms and one patient’s x-ray angiogram), in whom x-ray angiography was performed with carbon dioxide (two patients), and in whom imaging was unsuccessful (one patient because of coil failure) were excluded from analysis.
The resulting analysis involved 30 patients, 19 men and 11 women, aged 38–92 years (mean age, 67.9 years). In these 30 patients, 15 limbs (in 15 patients) were evaluated for limb salvage and 20 limbs (in 15 additional patients) were evaluated because of severe lifestyle-limiting claudication. All of the patients evaluated in this study had chronic symptoms. The mean interval between the MR angiographic and x-ray angiographic studies was 32.5 days. The patterns of vascular disease and the vascular interventions previously performed in these patients are summarized in Tables 1 and 2, respectively.
|
|
Two-dimensional MR DSA.—The patients were placed in the supine position with their feet positioned to enter the magnet first. The lower part of the legs was immobilized with foam cushions and stabilizing straps and placed into the head transmit-receive coil (GE Medical Systems). Multiphase 2D fast gradient-recalled-echo DSA acquisitions (21) were performed with the following parameters: 9.0/1.9 (repetition time msec/echo time msec), 50 ° flip angle, 40-cm field of view, 40–100-mm slab thickness, 256 x 192 matrix, and 15.6-kHz receiver bandwidth. The slab location and thickness were selected to include all of the arteries and bypass grafts of interest in the region imaged. Each 2D MR DSA acquisition began concurrently with the initiation of a 5–7-mL bolus of gadodiamide (Omniscan; Nycomed Amersham, Princeton, NJ) administered with a power injector at a rate of 2 mL/sec through a 20–22-gauge intravenous catheter placed in an antecubital vein. All contrast material injections were followed by the injection of 20 mL of normal saline at the same rate to facilitate contrast material passage into the central circulation. Thirty to 40 images were typically acquired in 1 minute at about 1–2 seconds per image.
An image obtained prior to contrast material enhancement was used as a mask and subtracted from the subsequent images. Complex subtraction was performed on the raw k-space data prior to Fourier transformation, and the magnitude images generated after Fourier transformation formed the final angiograms. This data processing typically took 5–10 minutes to perform after the data acquisition. A coronal slab was acquired first from the adductor canal to the middle portion of the calf after the first contrast material injection. A coronal slab was acquired also from the middle portion of the calf to the feet after the second injection. If it was necessary to image the tibial or pedal arteries in a perpendicular projection, an additional sagittal slab was then acquired at the same level in the extremity of interest following a third injection. The physician (D.W.T., M.R.P., N.M.K., P.A.W.) performing the procedure obtained five subtracted images and one nonsubtracted image from each acquisition and displayed them as six images on one sheet of film.
Three-dimensional bolus chase MR angiography.—Three-station bolus chase MR angiography was performed with multiphase 3D fast spoiled gradient-recalled-echo sequences by using dynamic k-space sampling, a body coil, manual table translation, and complex subtraction (3,6,11). The 3D MR imaging parameters were as follows: 4.8/1.1, 30° flip angle, 40-cm field of view, 2–3-mm section thickness, 256 x 192 matrix; and 62.5-kHz receiver bandwidth. This protocol resulted in the three contiguous stations being imaged in 20–25 seconds each. The dynamic k-space sampling orders were edge then center for the first station; edge, center, edge for the second station; and center then edge for the third station (6,22). Prior to bolus chase MR imaging, a 2D MR DSA timing sequence was performed in the second station by using the body coil, the 2D imaging parameters described herein earlier, and 3–5 mL of the gadolinium-based contrast material injected at a rate of 1.5 mL/sec. The 3D contrast material bolus chase was timed such that imaging in the second station began when the leading edge of the contrast material bolus arrived to the knee.
The three 35-cm-long imaging stations were (a) the abdomen and pelvis, (b) the upper thighs to about the knees, and (c) the knees and calves. Table stepping was performed manually (2 seconds per stepping) by using 35-cm wooden blocks to control table translation (3). A 3D acquisition that served as a mask for subsequent subtraction was performed in each station prior to contrast material injection. The patient was repositioned to the first station, and 30–40 mL of the gadolinium-based contrast material was injected at a rate of 1.5 mL/sec. After the calculated delay, the imaging sequence began with breath holding for imaging in the first station. After imaging in each station was completed, the table was manually translated and imaging in the next station was initiated. On the MR imaging unit computer, complex subtraction was performed on the raw k-space data acquired during the mask sequence and the contrast-enhanced sequence. Zero filling was used to interpolate the section spacing to 1.5 mm in image reconstruction. The images were recorded in the database of the imaging unit after postprocessing (approximately 6 minutes after imaging completion). The physician performing the procedure reconstructed the data into maximum intensity projection images in multiple oblique projections and recorded all pertinent images on film.
X-ray Angiography
The patients were instructed not to eat for at least 6 hours prior to the procedure. Conscious sedation was induced as necessary by administering midazolam (Versed; Roche, Nutley, NJ) and morphine sulfate with standard monitoring. Leg immobilization was used only when needed. The common femoral artery contralateral to the extremity with the most severe symptoms was accessed with the Seldinger technique, and a 4-F catheter (Omni-flush; AngioDynamics, Queensbury, NY) was positioned at the T11–12 level. An aortogram was obtained by using a 25–38-cm image intensifier with a 1,024 x 1,024 display matrix (Integris V-3000; Philips, Best, the Netherlands). The catheter was then withdrawn to just above the aortic bifurcation, where DSA was performed, also by using a 25–38-cm image intensifier, in several overlapping fields to the level of the popliteal artery, including oblique projections of the pelvis and femoral bifurcations.
In all extremities studied, selective images of the symptomatic extremity were also obtained with the catheter repositioned in the external iliac or common femoral artery of the extremity. The selective images were obtained in multiple stations and with many injections on a 17–31-cm field of view image intensifier with the 1,024 x 1,024 display matrix. Diatrizoate meglumine 30% (Reno-Dip; Bracco Diagnostics, Princeton, NJ) or iohexol 300 (Omnipaque; Nycomed Amersham) was administered in all studies at the discretion of the performing radiologist (D.W.T., N.M.K., P.A.W.); 12–30 mL was injected into the aorta for multistation nonselective imaging, and 5–30 mL was injected for selective imaging. The catheters were removed with manual compression, and standard supine recovery for 4–6 hours in a monitored setting followed. Several images from each diagnostic series, as selected by the radiologic technologist and angiographer performing the procedure, were displayed as six images on one sheet of film. Standard postprocessing software (part of Integris V-3000) with pixel shifting and remasking capability was used for image creation.
Image Analysis
On the MR angiographic and x-ray angiographic films, the names and medical record numbers of the patients were obscured by tape. Each MR angiographic and x-ray angiographic case was assigned a study number. This study number, the extremity or extremities of clinical interest, and the status as to whether the patient had claudication or limb-threatening ischemia were posted on the films for the readers. The MR and x-ray angiographic studies were randomly organized, and the images were placed on a film alternator so that the MR and x-ray angiogram interpretations for a given patient were performed in different orders and on separate days.
For each extremity for which therapy was being considered, the vascular tree was divided into 16 potential segments: aorta, common iliac artery, external iliac artery, common and deep femoral arteries, superficial femoral artery, above-knee popliteal artery, below-knee popliteal artery, tibioperoneal trunk and upper half of peroneal artery, lower half of peroneal artery, upper half of anterior tibial artery, lower half of anterior tibial artery, upper half of posterior tibial artery, lower half of posterior tibial artery, dorsal artery of the foot, either plantar artery, and bypass graft (if present). The images were interpreted by three fellowship-trained cardiovascular and interventional radiologists (N.M.K., P.A.W., M.R.P.) who had experience with the x-ray angiographic and MR angiographic examinations performed in this study.
The readers were provided with a grading sheet for each MR angiogram and each x-ray angiogram on which they recorded the vascular segment that would serve as the inflow vessel for a revascularization procedure. This vessel was defined as the segment that was free of significant (
50%) stenosis and one segment above the most cephalic hemodynamically significant lesion–containing segment. In a similar fashion, the readers then recorded the vascular segment that would serve as the outflow vessel for revascularization. For patients being evaluated for limb salvage, outflow segments were chosen to provide direct flow to the pedal circulation or to the terminal peroneal artery (if there were good collateral vessels to the foot from this artery) without an intervening significant stenosis. The outflow artery would be one segment below the most caudal significant stenosis. For patients with claudication, outflow segments were chosen to provide direct flow to at least the middle portion of the calf without an intervening significant stenosis or direct flow to the foot if an infrageniculate outflow target was selected.
At each inflow and outflow segment interpretation, the readers also graded the depiction of the arterial anatomy as (a) inadequate—that is, it was impossible to confidently determine a treatment plan; (b) intermediate—that is, image quality was intermediate but sufficient for treatment planning; or (c) excellent—that is, depiction of the anatomy was excellent and more than sufficient for treatment planning. The reasons for any inadequate or intermediate grades were recorded. Finally, a section for other comments was provided beneath each recorded inflow and outflow segment on the grading sheet.
Data Analysis
For each reader, the fraction of the cases in which the inflow segments determined with MR angiography and x-ray angiography were concordant was recorded. This process was repeated for the outflow segment determination of each reader. CI analysis of these data was performed to evaluate the equivalence of the two angiographic studies in enabling each reader to predict the proximal and distal target segments (23).
The three-level image quality grades recorded by the readers for each MR angiographic and x-ray angiographic inflow and outflow segment determination were compared by computing the percentages of agreement (with 95% CIs) with a 3 x 3 contingency table (23).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
|
|
|
|
|
|
Outflow Segment Determination
The three readers selected the identical segment for outflow with MR angiography and x-ray angiography in 32 (reader 1), 32 (reader 2), and 34 (reader 3) of the 35 evaluated limbs (mean percentages of agreement [95% CI]: 91% [ 77%, 98%] for readers 1 and 2 and 97% [ 85%, 100%] for reader 3) (Table 3). Infrapopliteal segments were selected for outflow in 37% (36 of 98) of the concordant interpretations.
Seven discordant outflow segment determinations occurred in four separate limbs (Table 5). At these discordant interpretations, the outflow segment selected was one level lower in the same vessel in four cases (peroneal artery in all cases) and a more distal segment of a different tibial artery in two cases. In all of these cases, the segment selected for outflow at contrast-enhanced MR angiography would have been an acceptable target vessel for revascularization according to the x-ray angiographic determination.
|
|
|
|
|
|
|
|
Image Quality Evaluation
The image quality grades given by the readers are displayed in six 3 x 3 contingency tables in Table 6. The percentages of agreement, defined as the percentages of cases in which identical image quality grades for the two angiographic examinations were given (in situations in which the data are found on the diagonal of agreement on the contingency table), with 95% CIs, also are presented in Table 6.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
Some of the minor differences in the selection of an inflow segment with x-ray angiography versus MR angiography are related to the difficulty in determining the significance of intermediate stenoses, particularly those in the iliac arteries. Determining the hemodynamic significance of intermediate stenosis solely on the basis of anatomic criteria is difficult, regardless of the imaging study used. This fact is well recognized with regard to x-ray angiography, and most radiologists liberally use pressure measurements to overcome this problem when performing x-ray angiography. Three-dimensional MR angiograms allow viewing of stenosis on many oblique projections from a single acquisition, but they still leave the reader to make an interpretation without physiologic information.
When differences existed between the MR angiogram and x-ray angiogram interpretations in this study, the degree of stenosis tended to be read as more severe with MR angiography. Such a problem is likely to persist, even with high-spatial-resolution imaging with blood pool agents. The hope is that in the future, an additional MR angiographic sequence, such as phase-contrast or cine-phase contrast MR angiography, will help determine the hemodynamic significance of each lesion identified.
In this study, susceptibility artifacts from metallic surgical clips led to some differences between the treatment planning determined with MR angiography and that determined with x-ray angiography. As reliance on MR angiography increases in the vascular medicine community, increased awareness of this imaging pitfall will likely lead to decreased use of metallic clips and possibly a manufacturing change to an alternative nonferromagnetic material. Artifacts from metal clips and stents will also diminish as improvements in imaging unit gradient performance lead to shorter echo times. Similar artifact-related difficulties occur with intravascular stents; however, patients with stents were excluded from this study.
Motion artifact at 2D MR DSA of the infrapopliteal vessels contributed to the few remaining discrepancies between the MR angiogram and x-ray angiogram interpretations. During the study period, the 2D MR images were not subtracted prior to the bolus chase because of software and networking issues. Since that time, improvements have enabled online subtraction of 2D acquisitions. We now liberally move masks and repeat sequences in an attempt to optimize imaging. These advancements should result in decreased numbers of studies with suboptimal infrapopliteal 2D MR images.
In one case, geniculate branches may have obscured a significant lesion in the popliteal artery examined with frontal x-ray angiography and coronal 2D MR DSA. Crossing vessels can limit the interpretation of 2D projectional images such as x-ray angiograms and 2D MR DSA images. This is not a problem with 3D acquisitions because the images can be reformatted in an infinite number of oblique projections. We recently refined our bolus chase technique such that we can now achieve reliable 3D arterial depiction in the third of four stations that include the popliteal artery and the proximal tibial arteries (22). We expect the inclusion of these 3D data to improve the interpretation of lesions at these levels and ultimately lead to an elimination of the need for 2D acquisition in the knee and upper part of the calf.
In designing a study to compare treatment plans for peripheral vascular disease, we recognized that specific bypass or endovascular procedure recommendations are based on the interpretation of images, as well as on an understanding of the pathophysiologic features of the disease, the durability of specific reconstructions, and specific technical biases. To eliminate these biases from the interpretations, we chose to assess treatment plans based on inflow and outflow levels by using the segmental approach. This type of analysis also enabled us to relate the differences in treatment plans to specific judgments about the inflow or outflow vessels and to examine trends based on vascular levels. The rules by which inflow and outflow segments were selected were clearly defined, and each reader analyzed the MR angiographic and x-ray angiographic data on every patient to maintain consistency. The type of decision making that the readers were required to perform is intuitive and a vital part of every analysis for planning treatment of peripheral vascular disease. The results of such analysis are categorical and easy to analyze.
There were several limitations to our study. First, this was a retrospective analysis of only the images archived in the film library. Greater accuracy may have been achieved if the readers had had access to the raw MR angiographic and x-ray angiographic data at the respective workstations. Second, the criteria used to select the patients who underwent either or both of the angiographic examinations were not uniform. Earlier in our experience with contrast-enhanced MR angiography, the referring surgeons obtained the MR images first. After reviewing the findings of that study, they would then arrange for an x-ray angiographic examination to be performed just prior to surgery or referral for endovascular treatment. Later during the study period, as the surgeons became more comfortable with the validity of MR angiographic findings, many patients underwent surgery without having undergone x-ray angiography. The patients who underwent x-ray angiography in the later part of the study period were those who were being referred either for endovascular intervention or because of questions that persisted after the MR angiographic examination. These later cases might have negatively biased our conclusions about the utility of MR angiography. Finally, given the small number of patients evaluated with both angiographic examinations, this study lacks the power to enable a conclusive determination of the equivalence of combined 2D-3D MR angiography to x-ray angiography.
On the basis of our clinical experience, it is clear that 3D bolus chase MR angiography is becoming an important tool for the preoperative evaluation of patients with peripheral vascular disease worldwide. Future work with this technique and improvements in imaging unit performance will lead to improved infrapopliteal arterial depiction. Currently, the supplemental 2D sequences yield reliable high-spatial-resolution infrapopliteal angiograms and provide valuable timing information that is used to optimize the 3D depiction of the iliac and femoral vessels in all patients.
In conclusion, the results of this preliminary study demonstrate that combining contrast-enhanced 2D MR DSA of the area below the adductor canal with 3D bolus chase MR angiography of the abdomen, pelvis, and lower extremities extends the utility of MR angiography to provide reliable treatment planning information for patients with claudication, as well as patients who are being evaluated for limb salvage.
|
FOOTNOTES |
|---|
Author contributions: Guarantor of integrity of entire study, N.M.K.; study concepts, N.M.K.; study design, N.M.K., H.L.B.; literature research, N.M.K., Y.W., E.V.; clinical studies, N.M.K., P.A.W., D.W.T., M.R.P., R.W.; data acquisition, N.M.K., E.V.; data analysis/interpretation, N.M.K.; statistical analysis, Y.W.; manuscript preparation, N.M.K., Y.W., D.W.T.; manuscript definition of intellectual content, N.M.K.; manuscript editing, N.M.K., Y.W., M.R.P.; manuscript revision/review, N.M.K., Y.W., M.R.P., H.L.B., P.A.W.; manuscript final version approval, N.M.K.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  O. A. Meissner, J. Rieger, C. Weber, U. Siebert, B. Steckmeier, M. F. Reiser, and S. O. Schoenberg Critical Limb Ischemia: Hybrid MR Angiography Compared with DSA Radiology, April 1, 2005; 235(1): 308 - 318. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. Janka, C. Fellner, E. Wenkel, W. Lang, W. Bautz, and F. A. Fellner Contrast-enhanced MR Angiography of Peripheral Arteries including Pedal Vessels at 1.0 T: Feasibility Study with Dedicated Peripheral Angiography Coil Radiology, April 1, 2005; 235(1): 319 - 326. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. L. Zhang, N. M. Khilnani, M. R. Prince, P. A. Winchester, P. Golia, P. Veit, R. Watts, and Y. Wang Diagnostic Accuracy of Time-Resolved 2D Projection MR Angiography for Symptomatic Infrapopliteal Arterial Occlusive Disease Am. J. Roentgenol., March 1, 2005; 184(3): 938 - 947. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. L. Zhang, B. Y. Ho, M. Chao, K. C. Kent, H. L. Bush, P. L. Faries, A. I. Benvenisty, and M. R. Prince Decreased Venous Contamination on 3D Gadolinium-Enhanced Bolus Chase Peripheral MR Angiography Using Thigh Compression Am. J. Roentgenol., October 1, 2004; 183(4): 1041 - 1047. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
