Journal of Vascular Surgery
Volume 37, Issue 1 , Pages 62-71, January 2003

Lower extremity stepping-table magnetic resonance angiography with multilevel contrast timing and segmented contrast infusion☆☆☆★

Presented at the Fiftieth Annual Meeting of the American Association for Vascular Surgery, Boston, Mass, Jun 9-12, 2002.

Chicago, Ill

From the Division of Vascular Surgerya and the Department of Radiology,b Feinberg School of Medicine, Northwestern University

Received 17 June 2002; accepted 24 July 2002.

Article Outline

Abstract 

Objectives: Standard lower extremity contrast-enhanced magnetic resonance angiography (LE-CEMRA) with single injection bolus-chase methods on the basis of a single pelvis timing run can be accurate for depicting most vascular occlusive lesions but may fall short of catheter-based angiography when imaging tibial and pedal vessels. Magnetic resonance angiography techniques with a second contrast timing bolus and separate acquisitions for the calves and the pelvis greatly improve reliability and reduce venous contamination to levels that may render conventional angiography obsolete. Methods: From July to December 2001, 60 consecutive patients underwent LE-CEMRA of the calves with separate stepping-table acquisitions of the pelvis and thighs. Forty-five (75%) had complete or partial angiographic correlation during an endoluminal intervention or operative completion study. Lower extremity vessels were divided into anatomic segments (aortoiliac, femoropopliteal, tibial-pedal) for review. Three blinded observers assessed magnetic resonance source partitions, maximum-intensity projections, and volume-rendered images. Disease per segment was graded from insignificant (<20%) to occluded (100%) in 10% increments. Segments were also scored for venous contamination (scale, 0 to 3) and diagnostic quality (scale, 1 to 5). Digital subtraction angiograms were assessed similarly but separately. Results: The combination dual-timing/dual-injection technique had an overall sensitivity, specificity, and accuracy of 99%, 97%, and 98%. Venous contamination and artifact were virtually eliminated with combined technique LE-CEMRA. Diagnostic quality of calf and foot vessels was significantly superior to conventional bolus-chase magnetic resonance techniques (P < .01). Conclusion: Hybrid dual-acquisition LE-CEMRA allows complete timing specification that consistently produces high-quality, artifact-free images of the calf and pedal vessels. These techniques may be accurate enough to replace conventional digital subtraction angiogram for evaluation of lower extremity vascular occlusive disease. (J Vasc Surg 2003;37:62-71.)

 

Lower extremity contrast-enhanced three-dimensional magnetic resonance angiography (LE-CEMRA) has become a valuable noninvasive diagnostic tool in peripheral vascular occlusive disease.1, 2, 3, 4 Magnetic resonance (MR) imaging techniques still vary substantially between institutions, and local factors continue to be critically important for obtaining high-quality examinations. Some of the most important variables relate to contrast-bolus timing schemes.5, 6, 7 The wide variability in bolus transit times through diseased vessels makes it impossible to predict, on an individual basis, the optimum delay between intravenous injection and image acquisition.1 The most common approach is to measure the transit time to the aorta or iliac vessels and make assumptions about arrival times in the calf vessels. This has resulted in variable and unpredictable levels of venous filling and image degradation in the calves.

As such, complete acceptance of MR angiography (MRA) as an alternative to digital subtraction angiography (DSA) has not been universal.2, 3, 4 In fact, the use of DSA as a primary preoperative imaging method continues to be the rule rather than the exception. A number of different approaches have been applied to improve the quality and reproducibility of LE-CEMRA in an attempt to supplant invasive imaging from this primary role. Simultaneously, steps have been taken to decrease LE-CEMRA examination times in hopes of improving patient acceptance and increasing overall efficiency.

LE-CEMRA with hybrid contrast injection and acquisition schemes was created in an attempt to eliminate the “guess work” involved in imaging lower extremity vessels by measuring exact contrast arrival times separately to the pelvis and calves.8, 9 Standardization of these protocols has consistently produced high-quality images and has allowed us to almost completely eliminate DSA as the preoperative diagnostic test of choice. The purpose of this study was to review our experience with this hybrid LE-CEMRA technique and to determine how often DSA was rendered redundant.

Back to Article Outline

Materials and methods 

Between July and December 2001, 60 consecutive patients (34 men [57%] and 26 women [43%]; mean age, 66.4 years; range, 37 to 88 years; standard deviation, 12.5 years) underwent clinically indicated hybrid LE-CEMRA for nonaneurysmal atherosclerotic limb ischemia. For more than 2 years before the study period, patients had been imaged with standard bolus-chase methods. The 33 most recent consecutive bolus-chase scans on 22 men (67%) and 11 women (33%; mean age, 62.8 years; range, 16 to 90 years; standard deviation, 18.2 years) from a 4-month period just before July 2001 were selected for retrospective comparison with the hybrid studies. (Because all images acquired before this were available only in a hardcopy format, we chose not to perform more than 33 analyses.) All examinations were performed by experienced MR technologists on a standard, commercially available, 1.5-T clinical MR imaging scanner (Magnetom Symphony, Siemens Medical Systems, Iselin, NJ) at Northwestern Memorial Hospital. In addition, 45 of the 60 patients (75%) who underwent hybrid scans and 19 of the 33 patients (58%) who had bolus-chase studies had complete or partial angiographic correlation available for review. Patient demographics and indications for LE-CEMRA are listed in Table I.

Table I. Patient demographics
Bolus-chaseHybrid
Average age (y)62.866.4
Male:female ratio2:11.3:1
Correlation studies19 (58%)45 (75%)
Claudication14 (23%)12 (36%)
Critical ischemia43 (72%)19 (58%)
Other indication3 (5%)2 (6%)
The Northwestern University Medical School Institutional Review Board approved the retrospective review of all patients' images and records.

Image analysis 

A single attending vascular surgeon (MDM) and two attending radiologists (FSP and JCC), blinded to patient identity and clinical history, evaluated original three-dimensional MRA source partition data, maximum intensity projections, and volume-rendered three-dimensional MR images derived from standard bolus-chase techniques and, after July 2001, with the hybrid MRA technique. The MRAs were read twice, once by the observers independently and then by consensus to eliminate significant discrepancies in the individual interpretations. All MRA observations were carried out on a picture archiving and communications system workstation (PACS, Pathspeed 8.1, General Electric, Milwaukee, Wis). The areas of stenosis or occlusion identified on each study were recorded on separate schematic diagrams for later review (Fig 1).

For both bolus-chase and hybrid LE-CEMRA techniques, the presence of disease at each of 29 anatomic segments was graded and recorded. Analyzed vascular segments are labeled and numbered in Fig 1. The most severe focus of disease was graded in each vascular segment. Disease was graded into 10 separate categories from an insignificant stenosis (<20%) to a complete occlusion by 10% increments. A total of 2648 (941 bolus-chase and 1707 hybrid) segments of disease were evaluated (total does not equal 2697 because four patients were amputees). Renal and pedal vessels (dorsalis pedis and ankle posterior tibial arteries), because they are typically at the extreme of the receiver coil volumes, were graded as either visualized or not visualized. When the vessels were included in the field of view, nonvisualization was interpreted as occlusion.

For both bolus-chase and hybrid LE-CEMRA, venous contamination was graded for each of the three imaging stations (pelvis, thighs, and calves) on a 4-point scale: 0, not visible; 1, barely visible but did not affect diagnostic quality; 2, visible and may have affected diagnostic quality; and 3, present and definitely compromised diagnostic quality. Image diagnostic quality was also graded at each of the three imaging stations on a scale of 1 to 5: 1, nondiagnostic; 2, poor quality and not confident; 3, fair quality and marginally confident; 4, good quality and confident; and 5, excellent quality and highly confident.

In all 93 patients, a search was undertaken to find correlating angiographic studies that had been performed either at Northwestern Memorial Hospital or at an outside institution within a 3-month period before or after the LE-CEMRA. These angiographic correlates were sought retrospectively from different sources, including diagnostic or therapeutic digital subtraction interventional radiology studies and therapeutic or completion arteriography performed in the operating room angiography suite. Forty-five hybrid studies (75%) and 19 bolus-chase examinations (58%) were found to have useful complete or partial angiographic correlates. Because LE-CEMRA has been considered our primary diagnostic test, most of the angiographic correlates (96%) were from the interventional radiology suite (Integris, Phillips Medical Systems, Bautzen, Germany) or from the fixed (Integris) or portable (Series 9800, O.E.C. Medical Systems, Salt Lake City, Utah) angiography equipment in the operating room. Correlates were usually found after a therapeutic endoluminal intervention or an operative completion study. Angiographic images were obtained in the operating room with the same biplane digital subtraction techniques that are used in the interventional radiology suite and were usually completed with a contralateral femoral artery puncture. Twenty-one of the 64 patients (33%) who had a correlative angiogram had a complete correlation study showing the aortoiliac segments and both limbs to include the feet. The remaining 43 patients (67%) had partial correlation studies, usually showing the pelvis and one limb entirely.

Angiograms were reviewed by a single observer (MDM), and the results were recorded on the schematic vessel diagrams. Angiographic evaluation was completed in a blinded fashion, separate from the MR interpretations, so that other readers were not aware of the results. DSA images were graded for disease on the basis of the same 10-category scale described previously for grading LE-CEMRA. For statistical analysis, segment-by-segment comparison of the LE-CEMRA images with the corresponding angiographic correlates was performed.

Bolus-chase MRA technique 

Before July 2001, patients were imaged with a standard three-station bolus-chase technique with three-dimensional gradient echo pulse sequences (repetition time/echo time, 3.5/1.2; flip angle, 25 degrees) and a dedicated peripheral vascular phased array surface coil. A single pelvic timing run with 2 mL gadopentetate dimeglumine (Magnevist, Berlex Laboratories, Wayne, NJ) injected at 2 mL/s was performed first. After a successful timing run, precontrast three-dimensional mask acquisitions were performed at each of the three stations—pelvis, thighs, and calves. Next, a single injection, stepping-table, bolus-chase LE-CEMRA was performed with start time for the pelvic acquisition derived from the timing run. A total 58 mL of contrast was injected at a divided rate with the first 20 mL injected at 2 mL/s and the remainder at 0.8 mL/s (total contrast infusion time, 57.5 seconds). The parameters (field of view, matrix size, number of partitions, etc) at each station were optimized to permit rapid scanning while maximizing inplane and through-plane resolution. Imaging parameters were dependent on height and body habitus (Table II). Average total acquisition time was 1 minute, including automated table movement. Total time spent in the scanner averaged 45 minutes for bolus-chase studies.

Table II. MR imaging parameters
Matrix, phaseMatrix, readTHKTR (ms)TE (ms)αPart.FOV, phaseFOV, readoutPixel, phasePixel, read
Hybrid pMRA
Pelvis2505122.43.541.1922403504501.390.87
Thigh26051223.2120403204301.240.85
Calves2285121.43.22120643004401.330.86
Bolus-chase pMRA
Pelvis1655122.13.971.3725443004501.850.98
Thigh1885122.13.54120403104501.70.92
Calves2005122.24.03120403104501.590.87

THK, Slice thickness; TR, repetition time; TE, echo time; Part, number of partitions; FOV, field of view; pMRA, peripheral MRA; α, flip angle.

Hybrid MRA technique 

Between July and December 2001, 60 consecutive patients were imaged with a hybrid technique with the same three-dimensional fast low-angle shot gradient echo pulse sequence and the same dedicated peripheral vascular coil. With this hybrid technique, two independent timing measurements were performed, one for the pelvis at the level of the aortic bifurcation and one at the calves at the level of the tibial trifurcation. Timing runs were performed with a flow-insensitive, T1-weighted gradient-echo sequence with an automated image-subtraction algorithm.8

After the precontrast mask acquisitions were obtained in the calves/feet, distal LE-CEMRA acquisitions were performed on the basis of the calf timing run. All patients were asked to actively plantar flex the foot so that the pedal vessels were within the imaging volume. With the measured contrast arrival time, a 20-mL contrast bolus administered at 2 mL/s was performed. Two consecutive three-dimensional acquisitions then were obtained.

After complete imaging of the calves, mask acquisitions for the pelvis and thigh stations were obtained. Pelvis and thigh stepping-table LE-CEMRA was performed with a second infusion of 30 to 35 mL of gadolinium contrast. The start of the pelvic acquisition was determined from the pelvic timing run in a manner similar to the bolus-chase technique. Imaging parameters at each station were optimized to permit rapid scanning while maximizing inplane and through-plane resolution (Table II). Average total acquisition time was 33 seconds for the pelvis and thigh acquisition and 25 seconds for the calf acquisition. Total MR examination time averaged 45 minutes.

Back to Article Outline

Results 

Compared with the angiographic correlates as gold standards, the overall sensitivity, specificity, and accuracy for bolus-chase technique were 95%, 88%, and 90%, respectively, and for the hybrid technique were 99%, 97%, and 98%, respectively. Sensitivity and specificity for the calf stations alone were 94% and 80% for bolus-chase methods, and sensitivity and specificity figures were significantly improved to 100% and 91% with the hybrid technique (Table III).

Table III. Sensitivity and specificity for bolus-chase versus hybrid techniques
Bolus-chaseHybrid
SensitivitySpecificitySensitivitySpecificity
Overall95%88%99%97%
Pelvis100%83%94%100%
Thighs100%100%100%95%
Calves94%80%100%91%
With hybrid LE-CEMRA, venous contamination was marginally reduced in the proximal stations and was virtually eliminated (P < .01) in the calf station when compared with bolus-chase LE-CEMRA (Fig 2). No significant difference was seen in the diagnostic quality of the pelvis and thigh arterial segments between bolus-chase and hybrid techniques, but, importantly, the diagnostic quality of the calf vessel images was found to be significantly (P < .01) and consistently higher with the hybrid technique (Fig 3). Differences in imaging quality must be attributed to the techniques because the imaging parameters for each station were similar (Table II).

The renal vessels were visualized in 23 of the patients (70%) undergoing bolus-chase LE-CEMRA and 44 of the patients (73%) undergoing hybrid LE-CEMRA. In 35 of 64 patient limbs (54%) from the bolus-chase group and 96 of the 118 patient limbs (81%) from the hybrid group, the ankle posterior tibial and the dorsalis pedis arteries were visualized or could be clearly diagnosed as occluded. In seven hybrid studies, the acquisition data identified patent target vessels in the foot that DSA labeled as occluded. Conversely, DSA identified patent vessels eight times when hybrid MRA missed them. MRA missed the patent foot vessels because of poor positioning, not because of contrast mistiming, in three of the eight. Metallic suspectibility artifacts from arterial stents, knee prostheses, or hip prostheses degraded images in two of the 33 bolus-chase studies and in seven of the 60 hybrid examinations.

We were able to determine disposition in all 93 patients. We were successful in formulating a sound treatment plan, without further diagnostic imaging, on the basis of the dataset provided with MR images alone, in 80 patients (86%), 21 (64%) from the bolus-chase group and 58 (97%) from the hybrid group. The treatment plan included an open surgical intervention in 51 patients (55%), 12 patients (13%) who went on to have an interventional procedure (angioplasty ± stent), four patients (4%) who had both, 22 patients (24%) who had no intervention, and four patients (4%) who had a primary major amputation. A preoperative diagnostic angiogram was required to formulate a treatment plan in 12 of the bolus-chase LE-MRA patients (36%) and in two of the hybrid LE-MRA patients (3%).

Back to Article Outline

Discussion 

The results indicate that, with the hybrid protocol used in this study, the image quality, reliability, and diagnostic accuracy of MRA is sufficient to render DSA unnecessary in the vast majority of patients. Early MRA studies with contrast enhancement and signal reception from surface coils were completed by moving a phased array coil from station to station for separate acquisitions of the calves, thighs, and pelvis. These techniques, devised for stationary MR tables, required separate paramagnetic contrast injections for each of the three imaging stations, the calf being the last. Progressive accumulation of contrast in the soft tissues hampered optimal visualization of calf vessels.5, 6, 7, 9

Alternative MRA techniques, such as two-dimensional time-of-flight acquisitions, on the basis of thin axial slices, are sensitive to slow flow and when performed properly can be better for identifying patent tibial and pedal runoff vessels than conventional DSA.10, 11, 12 Unacceptably long acquisition times limit the clinical practicality of time-of-flight techniques, however.13, 14, 15, 16 LE-CEMRA permits rapid imaging along the major vessel axis and does not rely on thin transaxial imaging like time-of-flight MRA.17 The development of automated stepping-table techniques1, 18, 19, 20 and dedicated peripheral array coils21 have greatly improved the speed and quality of LE-CEMRA by capitalizing on bolus-chase methods of contrast infusion. The implementation of bolus-chase techniques has extended the quality of the images to all three regions (pelvis, thighs, and calves), albeit inconsistently (Fig 4).

For the two proximal arterial regions, sensitivity and specificity values exceed 90% with bolus-chase techniques.18, 20 Chasing the arterial bolus all the way to the feet while attempting to provide quality images of the pelvis, thighs, and calves before venous contamination occurs remains a challenge, however.1 Although many techniques, including cardiac gating,22 fluoroscopic real-time bolus monitoring,1 elliptic centric k-space acquisition,23 projection reconstruction, and k-space undersampling, have improved LE-CEMRA, no single technique has been universally accepted.24, 25 Despite significant improvements, the quality of standard bolus-chase techniques remains inconsistent. 26, 27 It is because of these limitations and because of increasingly rigorous demands by interventionalists and surgeons that, in all but a few very specialized centers, the role of lower extremity MRA has remained subjugated to a position that is secondary to invasive DSA for preoperative and preinterventional diagnosis and planning.2, 4

Digital subtraction angiography has its own limitations; it involves an arterial puncture, radiation exposure, and nephrotoxic iodinated contrast. Contrast angiography uses radiograph projection techniques, which are known to overestimate or underestimate nonconcentric stenoses unless multiple projections are used. Also, as seen with bolus-chase LE-CEMRA, timing differences exist between legs, and this can lead to poor imaging of distal segments with catheter angiography when mistiming occurs. In fact, when compared with certain MR techniques, DSA may provide inferior pedal vessel images unless a concerted effort is made to image with vasodilators. Previous studies that compared MRA and DSA underscored these limitations when time-of-flight or bolus-chase techniques identified varying numbers of runoff segments as patent after they were missed or interpreted as occluded with DSA.2, 12, 26 Similarly, computed tomographic angiography is limited by the requirements for ionizing radiation and nephrotoxic iodinated contrast.

Hybrid contrast-injection and acquisition LE-CEMRA schemes were developed to limit the variability involved in imaging the vessels of the lower extremities (Fig 5).

The hybrid technique precludes venous contamination in the calf while enabling near isotropic resolution where identification of distal bypass targets is critical (Fig 6). With the hybrid technique, venous contamination in the pelvis and thigh stations is minimal because of calf contrast equilibration throughout the extracellular fluid and precontrast mask volume subtraction. As with bolus-chase techniques, hybrid sequences preserve both inplane and through-plane spatial resolution in the pelvis and thighs to allow retrospective image reconstruction in an infinite number of projections. With DSA, this information can be derived only from additional contrast boluses and additional projections. When tibial contrast arrival times are discrepant, as seen in patients with asymmetric occlusive disease or in patients with a unilateral bypass graft, the hybrid technique, with prior knowledge of the timing difference, allows operators to tailor the examination to allow optimal imaging of both extremities. Also, because of a reduction in venous signal and tissue enhancement, pedal vessel visualization is better than with bolus-chase techniques. However, poor patient positioning still contributed to missing a patent foot vessel in at least three patients and in as many as 19% of the patients who underwent a hybrid study. We realize that this continues to be a limitation to success, and we are making efforts to further improve pedal imaging. We have found that pedal vessel visualization can be improved significantly by stressing the importance of maintaining plantar flexion throughout the calf/foot acquisition. We are in the process of developing custom orthotic foot braces to assist patients who are unable to hold a plantar flexed position for several minutes.

This MR imaging protocol has allowed us to successfully replace invasive angiography in most cases. From clinical experience, we recommended that, when reviewing any LE-CEMRA examination that uses image-mask subtraction, the contrast-enhanced original partitions should always be viewed to avoid erroneous conclusions from misregistration artifacts that could be translated onto the subtracted datasets. After this protocol, only 3% of the hybrid MR study patients who went on to an intervention needed any further diagnostic imaging. Such diagnostic confidence in MR echoes that described by prior authors.2, 4, 27

Although significantly safer than other accepted imaging methods, there are a few well-recognized contraindications to MR imaging. Patients with pacemakers, ocular metallic foreign bodies, or ferromagnetic intracranial aneurysm clips should not be imaged. MR imaging may also suffer from artifacts around intravascular stents and near joint prostheses, limiting visualization in the vicinity of these ferromagnetic objects. We do not, however, consider intravascular stents or other prosthetic materials to be contraindications to MR imaging. If useful noninvasive ultrasound scan data are available and we interpret the obscured vessels to be patent, we will proceed to intervention with these data alone.

Several limitations may have affected the results and the conclusions of this paper. Specific well-known limitations arise from the retrospective nature of our review. Because the two MR techniques were not developed simultaneously and because bolus-chase scans were not performed after hybrid sequences were implemented, a prospective analysis would have been difficult to perform. It is also important to note the possibility of the introduction of verification bias. This exists when the decision to perform what may be considered the gold standard examination (DSA) depends on the results of the examination under investigation (MRA). Because most of the gold standard images were generated during invasive therapy, these correlate DSA studies were performed more often in patients and on arterial segments with LE-CEMRA results that showed disease than on normal vessels. This type of bias tends to result in overestimation of sensitivity and underestimation of specificity. In addition, unlike the MR acquisitions, the DSA images that were used for correlation were not produced with uniform techniques. These studies were performed for varying indications, were obtained from differing sources, were performed by numerous operators, and were often limited in scope. When the MR studies were read, interobserver variation was reduced by reaching agreement through consensus when observers graded stenoses with differences of greater than 30%. The fact that the gold-standard angiograms were read by a single observer does introduce risk for observer biases, but because review was blinded and results ultimately showed close correlation, this would not appear to have unduly contaminated the results. Finally, DSA was considered our gold-standard test against which both bolus-chase and hybrid LE-CEMRA studies were compared. As discussed previously, DSA may not be suitable as a gold-standard test because this imaging method has its own significant limitations.3

Back to Article Outline

Conclusion 

Hybrid LE-CEMRA is robust, accurate, and reproducible in clinical practice. This technique consistently provides high-quality images that satisfy the rigorous preoperative demands of vascular surgeons and interventional radiologists by providing data sufficient to formulate sound treatment plans, in most cases rendering conventional DSA unnecessary.

Back to Article Outline

Acknowledgements 

We thank Mrs Jan Goldstein, administrative assistant, for her expertise and her assistance in the preparation of this manuscript.

Back to Article Outline

Discussion 

Dr David C. Brewster (Boston, Mass). Congratulations to Dr Morasch and his associates. I think this is a real advance and another example of the almost astonishing evolution of different imaging techniques to facilitate and enhance our ability to provide optimal care to our patients.

Like other groups, we have had a long interest with use of MR. We have found it very useful in the renal and carotid areas in terms of screening and diagnosis. But, as you implied, there have been some limitations in regard to lower extremity occlusive disease, both in acquisition time required and the quality of images and information obtained. So, your presentation is really quite exciting.

I have several questions. First, I did not quite follow whether your method employs standard contrast or gadolinium?

Secondly, I believe one of your tables mentioned acquisition time, but perhaps you could elaborate a little bit on that.

Although cost was not your focus, can you give us a sense of a comparison of conventional angiography and this method?

And finally, are these simply modifications of the software program that any MR facility can do, or does this require special equipment that is not yet available in most communities?

Dr Mark D. Morasch. All of these studies were performed with gadolinium, not with any type of ionated contrast.

With regard to acquisition times, a segment of the presentation suggested that patients were on the scanner for a total of 45 minutes. This would be a fairly long time if the patient actually had to hold still. But the acquisition times are really more on the order of seconds to minutes. The actual period of time that the patients have to hold still and hold their feet in plantar flexion is fairly brief. And so these studies, we have found, tend to be fairly well tolerated. We have been able to coach even the most claustrophobic patients through the exam.

We did not look at cost specifically. This technique is clearly less expensive than sending the patient down to the interventional radiology suite for an angiogram. Most of the disposable cost with contrast MR is the expense of the gadolinium. Although in this study the volumes of gadolinium were pretty much the same as with traditional approaches, we found, over the course of the last few months, that we have really been able to decrease the amount of gadolinium that we need to use.

Lastly, are these just simple modifications that anybody can do? In fact, they are. The actual imaging parameters, the numbers that you plug into the computer when you run patients through the machine, are all the same as with the traditional approach. The only thing that we have really changed is the sequence in which we perform the acquisitions.

Dr Lloyd M. Taylor, Jr (Portland, Ore). I have two questions. You have told us about the acquisition times and the length of time that the patients spend. But can you give us an idea of how much workstation time is required to produce these beautiful images?

The second question is, who is sitting at the workstation? Are you doing these studies, are the radiologists doing it, or do you do it together and produce the images you need to perform the operation?

Finally, your numbers seem rather small to me, because I am familiar with your busy practice. And so my question is, has this totally replaced contrast arteriography in your lower extremity practice, or are there some patients in whom you still obtain conventional arteriograms?

Dr Morasch. The patient is on the scanner for a total of about 45 minutes, but, after that, there is still a substantial amount of time involved in postprocessing the images. This is generally done by radiologists (or by technicians or research medical students). Someone has to sit at the computer afterwards and postprocess these images so that they look like the pictures that I am showing to you. So, certainly, there is more time involved with the total process.

We do these together and then look at them together when we can. One of the things we found in looking back was that the radiologists were not always showing us exactly what we wanted to see. By sitting down at the computer with the MR radiologists, we were able to get across what is really important to vascular surgeons. Surgeons want to actually see pedal vessels. We are looking for target vessels for bypass. We would encourage anyone out there who is trying to develop MR to go down and spend some time with their MR radiologists so that this interaction can occur.

Some patients do still end up down in the interventional radiology suite for an angiogram, but these cases have nearly disappeared. Patients with ferromagnetic implants, like pacemakers, or patients with ocular metal fragments still have contraindications to MR. It has gotten to the point, though, since we have a fixed C-arm in our operating room, that if we cannot get an MR, we will just take our patient into the operating room with a duplex result and perform our procedure, whatever it may be, without ever sending them down to IR.

Dr William D. Turnipseed (Madison, Wis). There are some concerns that I have about conduct of the study. Number one, it is retrospective. It sounds like you had to work very hard to find true angiographic correlates. Just looking at the numbers, only 16 out of the 93 patients actually had DVI correlates where you had vascular imaging done in a unit where you could utilize multiview technology. Over half of your angiograms were done in the operating room. Not everybody uses OEC or fluoroscopy. If you are only doing single-plane imaging, comparisons are much more difficult. It is important for you to clarify whether you use fluoroscopy or whether you are using single-plane arteriography in the operating room. I am concerned over the fact that a large number of these people either were not surgically treated or had amputations. What did you do with those data, or could you do anything with them?

I have a couple more questions. In the bolus-chase technique, you said that 12 patients required further arteriography before you could make a management plan. Was that a failure of MR to identify the location, the extent, or the severity of disease?

In the group of 93 patients, how many of them were distal tibial management problems to begin with? Was this the reason that the arteriograms were being performed, or was this an incidental data-gathering opportunity?

Finally, what are your selection criteria? Of the total group that you studied, how many intent-to-image failures did you have because of motion, claustrophobia, or artifact?

Dr Morasch. I fully acknowledge the shortcomings of the retrospective design. All I can say is that we have begun now to look at these patients prospectively. We are particularly interested in finding out how this modality has altered our practice patterns. I can tell you that at this point in time almost no patients go down to the interventional radiology suite. I do not have the exact numbers of patients who did not have an MR because of the various contraindications, but I can tell you those numbers are small.

With regards to the question about what we do in the operating room, how we got our correlations, I do understand that in many institutions a flat plate of the distal anastomosis is all that is performed. Fortunately, in our operating room, we have fixed biplane C-arm equipment. As such, most of the correlates that were obtained at the time of a completion angiogram were complete. Oftentimes, these were done with biplanar images of the pelvic vessels and with runoff all the way down to both feet.

Dr Victor M. Bernhard (Palisade, Colo). I have two questions about the correlations with, number one, ultrasound, if you did it; and secondly, what did it look like when you got to the target vessel? How often did you see something that was either obscured or not clearly defined by your technique when you looked at the vessel with the open leg?

Dr Morasch. We did not look specifically at duplex correlates.

With regards to what the vessels looked like, I was impressed, as we improved our techniques, at how well the MR images correlated, even all the way down in the foot. True, you cannot identify heavy calcification, like you can with contrast angiography. But, if a vessel showed up on MRA, in virtually every case, it was a vessel that could accept a bypass.

Back to Article Outline

References 

  1. Wang Y, Lee H, Khilnani N, Trost D, Jagust M, Winchester P, et al.  Bolus-chase MR digital subtraction angiography in the lower extremity. Radiology. 1998;207:263–269
  2. Cambria R, Kaufman J, L'Italien G, Gertler J, LaMuraglia G, Brewster D, et al.  Magnetic resonance angiography in the management of lower extremity arterial occlusive disease: a prospective study. J Vasc Surg. 1997;25:380–389
  3. Nelemans P, Leiner T, de Vet H, van Engelshoven J. Peripheral arterial disease: meta-analysis of the diagnostic performance of MR angiography. Radiology. 2000;217:105–114
  4. Huber T, Back M, Ballinger R, Culp W, Flynn T, Kubilis P, et al.  Utility of magnetic resonance arteriography for distal lower extremity revascularization. J Vasc Surg. 1997;26:415–424
  5. Aksit P, Frigo F, Polzin J, Choyke P, Ho V, Hood M, et al.  “Shoot and scoot”: a segmented volume acquisition method of high-resolution multi-station imaging of peripheral vasculature. In: Proceedings of the 10th Annual Scientific Meeting and Exhibition of the ISMRM Honolulu. 2002 May 18–24;
  6. Leiner T, Kessels A, Van Engelshoven J. Total runoff peripheral MRA in patients with critical ischemia and tissue loss can detect more patient arteries than IA-DSA. In: Proceedings of the 10th Annual Scientific Meeting and Exhibition of the ISMRM Honolulu. 2002 May 18-24;
  7. Sabati M, Lauzon M, Frayne R. Interactive large field-of-view peripheral MRA. In: Proceedings of the 10th Annual Scientific Meeting and Exhibition of the ISMRM Honolulu. 2002 May 18-24;
  8. Finn J, Francois C, Moore J, Niemczura R. Lower extremity MRA with full prior specification of bolus transit times. In: Proceedings of the 10th Annual Scientific Meeting and Exhibition of the ISMRM Honolulu. 2002 May 18-24;
  9. Carr JC, Pereles FS, Collins JC, Morasch MD, McCarthy RM, Niemczura R, et al.  Stepping-table lower extremity MR angiography with separate calf acquisition and dual-level bolus timing. In: Proceedings of the 10th Annual Scientific Meeting and Exhibition of the ISMRM Honolulu. 2002 May 18-24;
  10. Owen RS. Magnetic resonance imaging of angiographically occult runoff vessels in peripheral arterial occlusive disease. N Engl J Med. 1992;326:1577–1581
  11. Cortell ED, Kaufman JA, Geller SC, Cambria RP, Rivitz SM, Waltman AC. MR angiography of tibial runoff vessels: imaging with the head coil compared with conventional arteriography. AJR Am J Roentgenol. 1996;167:147–151
  12. Baum RA, Rutter CM, Sunshine JH, Blebea JS, Blebea J, Carpenter JP, et al.  Multicenter trial to evaluate vascular magnetic resonance angiography of the lower extremity. JAMA. 1995;274:875–880
  13. McCauley TR, Monib A, Dickey KW, Clemett J, Meier GH, Egglin TK, et al.  Peripheral vascular occlusive disease: accuracy and reliability of time-of-flight MR angiography. Radiology. 1994;192:351–357
  14. Schmiedl UP, Yuan C, Nghiem HV, Winter TC, Freeny PC. MR angiography of the peripheral vasculature. Semin Ultrasound CT MR. 1996;17:404–411
  15. Snidow J, Harris V, Trerotola S, Cikrit D, Lalka S, Buckwalter K, et al.  Interpretations and treatment decisions based on MR angiography versus conventional arteriography in symptomatic lower extremity ischemia. J Vasc Interv Radiol. 1995;6:595–603
  16. Yucel E, Silver M, Carter A. MR angiography of normal pelvic arteries: comparison of signal intensity and contrast-to-noise ratio for three different inflow techniques. AJR Am J Roentgenol. 1994;136:197–201
  17. Prince M, Yucel E, Kaufmann J, Harrison D, Geller S. Dynamic gadolinium-enhanced three-dimensional abdominal MR angiography. J Magn Reson Imaging. 1993;3:877–881
  18. Ho K, Leiner T, de Haan M, Kessels A, Kitslaar P, van Engelshoven J. Peripheral vascular tree stenoses: evaluation with moving-bed infusion-tracing MR angiography. Radiology. 1998;206:683–692
  19. Leiner T, Ho K, Nelemans P, de Haan M, van Engelshoven J. Three-dimensional contrast-enhanced moving-bed infusion-tracking (MoBI-Track) peripheral MR angiography with flexible choice of imaging parameters for each field of view. J Magn Reson Imaging. 2000;11:368–377
  20. Meaney J, Ridgway J, Chakraverty S, Robertson I, Kessel D, Radjenovic A, et al.  Stepping-table gadolinium-enhanced digital subtraction MR angiography of the aorta and lower extremity arteries: preliminary experience. Radiology. 1999;211:59–67
  21. Ruehm S, Hany T, Pfammatter T, Schneider E, Ladd M, Debatin J. Pelvic and lower extremity arterial imaging: diagnostic performance of three-dimensional contrast-enhanced MR angiography. Am J Roentgenol. 2000;174:1127–1135
  22. Glickerman D, Obregon R, Schmiedl U, Harrison S, Macaulay S, Simon H, et al.  Cardiac-gated MR angiography of the entire lower extremity: a prospective comparison with conventional angiography. Am J Roentgenol. 1996;167:445–451
  23. Snidow J, Johnson MS, Harris V, Margosian P, Aisen A, Lalka S, et al.  Three-dimensional gadolinium-enhanced MR angiography for aortoiliac inflow assessment plus renal artery screening in a single breath hold. Radiology. 1996;198:725–732
  24. Du J, Block W, Carroll T, Lu A, Grist T, Mistretta C. Peripheral angiography with a time-resolved VIPR sequence. In: Proceedings of the 10th Annual Scientific Meeting and Exhibition of the ISMRM Honolulu. 2002 May 18-24;
  25. Fain S, Du J, Browning F, Polzin J, Block W, Mistretta C. Floating table isotropic projection imaging (FLIPR): a technique for fast, extended FOV, contrast-enhanced MRA. In: Proceedings of the 10th Annual Scientific Meeting and Exhibition of the ISMRM Honolulu. 2002 May 18-24;
  26. Kreitner K-F, Kalden P, Neufang A, Duber C, Krummenauer F, Kustner E, et al.  Diabetes and peripheral arterial occlusive disease. Am J Roentgenol. 2000;174:171–179
  27. Levy M, Baum R, Carpenter J. Endovascular surgery based solely on noninvasive preprocedural imaging. J Vasc Surg. 1998;28:995–1005

☆ Competition of interest: nil.

☆☆ Reprint requests: Mark D. Morasch, MD, Division of Vascular Surgery, Ste 10-105, 201 E Huron St, Chicago, IL 60611 (e-mail: mmorasch@nmh.org).

★ 0741-5214/2003/$30.00 + 0

PII: S0741-5214(02)75146-X

doi:10.1067/mva.2003.5

Journal of Vascular Surgery
Volume 37, Issue 1 , Pages 62-71, January 2003

Article Tools

* Image Usage & Resolution