Dr. Henseler is a Fellow in Interventional Radiology at The Johns Hopkins Hospital, Baltimore, MD. He recently completed his residency at the University of Wisconsin in Madison, WI. He received his MD from the Georgetown University School of Medicine, Washington, DC, in 1997

Peripheral magnetic resonance angiography (MRA) has advanced rapidly to become a less invasive alternative to catheter angiography. Peripheral MRA can be performed using numerous techniques, which can lead to confusion about how to complete the study. This report addresses the current techniques used to perform peripheral MRA, as well as appropriate patient selection and equipment. Imaging protocols currently under development will be introduced in order to familiarize the reader with technical advancements that may be available in the near future.

Magnetic resonance angiography (MRA) is a new discipline that allows for minimally invasive evaluation of the vasculature. The first applications concentrated on the carotid arteries and later the intracranial vascular tree. As the technology evolved, larger areas could be investigated, leading to the use of MRA to study the peripheral arterial system in patients with suspected peripheral vascular disease (PVD). Unique challenges are encountered when imaging the peripheral (predominantly lower extremity) arterial system. Techniques continue to evolve to maximize the evaluation of PVD. This report discusses current indications, as well as techniques and limitations of peripheral MRA. Newer directions will also be addressed. This article will serve as a foundation on which to build an understanding of peripheral MRA, as well as offer an introduction to the newer imaging techniques, which will be more widespread in the future.

Why use peripheral MRA?

Magnetic resonance angiography has become an accepted alternative to conventional angiography in many areas of the body; most notably in the head and neck, ^1,2 and more recently within the renal arteries. ^3,4 As technology has progressed, applications in the chest and abdomen have also yielded promising results for evaluating aneurysms and dissections as well as vascular disease. Peripheral MRA offers similar advantages as MRA elsewhere in the body. Peripheral MRA is minimally invasive, needing only an intravenous (IV) injection. Its major competitor, conventional angiography, remains the gold standard. Angiography yields exceptional diagnostic information as well as the ability to intervene immediately if needed. MR angiography must offer significant advantages over digital subtraction angiography (DSA) to offer an acceptable alternative. These advantages include lack of puncture site or catheter complications (hematoma, emboli, dissection, vascular injury), no significant nephrotoxicity, ^5,6 and little concern for anaphylactoid reactions. ^5,7 The test can be performed in a short period of time (approximately 1 hour), with the patient leaving immediately afterward; there is no need for observation or overnight admission.

The MRA study must be equivalent (or nearly so) to make it an acceptable alternative. Much research has been conducted to compare the accuracy of peripheral MRA with conventional angiography. Most recently Koelemay and colleagues ⁸ published a meta-analysis of MRA in lower extremity arterial disease. Thirty-four studies including 1090 patients were reviewed and the authors concluded that MRA is highly accurate for lower extremity arterial disease. ⁸ Table 1 describes the experience various investigators have had with peripheral MRA and reports the sensitivity for identifying disease compared with DSA. ^9-20

Quality peripheral MRA studies offer significant advantages to patients, clinicians, and radiologists. The patient is able to undergo a minimally invasive test that is not associated with any significant complications. The clinicians are afforded a test that can evaluate for suspected PVD with little risk to their patients. This makes the evaluation for PVD easier to justify and may allow for earlier intervention and quicker recovery. These same reasons allow the diagnosis and management to be accomplished more easily by primary care physicians, and may obviate the need for subspecialty referral. The radiologist also has the advantage of being in direct communication with the referring clinician. This yields the opportunity to discuss percutaneous interventional possibilities, and may allow for fuller participation of the interventional radiologist in the diagnosis and planning of further interventions.

Who are candidates for peripheral MRA?

According to the Legs for Life program sponsored by the Society of Interventional Radiology, as many as 8 million people in the United States may have PVD. ²¹ Obvious candidates for peripheral MRA are patients with suspected PVD, which can manifest in several ways. Symptoms may include numbness, tingling, or weakness in the leg. Claudication or rest pain is often encountered. People with PVD may also experience a cooling or color change in the skin of the legs or feet, or loss of hair on the legs. Table 2 describes risk factors for PVD. ²¹ The work-up of these patients is usually preceded by noninvasive vascular tests. If these results are abnormal, then further testing is warranted.

Atherosclerotic disease is a systemic process, and often patients with symptomatic vascular disease in one area may have disease elsewhere. Patients with symptomatic carotid or coronary artery disease are at high risk for PVD, and these patients are often studied to help evaluate the extent of disease and determine a more global approach to treatment.

Another category of patients for whom peripheral MRA can be helpful are patients with diabetes who have nonhealing ulcers. The microvascular disease caused by longstanding diabetes is often the culprit of the local ischemia that results in a nonhealing ulcer. Although neuropathy and microvascular disease conspire to cause the nonhealing ulcer in the majority of patients, PVD can also have a profound effect on the prognosis and treatment of these patients. The presence of concomitant PVD can exacerbate the tissue ischemia and reduce ulcer healing. Peripheral MRA can help to diagnose PVD elsewhere in the extremity and can guide further treatment, either surgical (bypass) or interventional (angioplasty). ²²

Peripheral MRA can be used to assess patients with previous surgical or percutaneous interventions. The patency of a graft or prior angioplasty site can be assessed, as well as the possibility of new disease to account for new symptoms.

As with all patients who are to enter the magnet, contraindications include cardiac pacing and defibrillator wires, certain neurovascular aneurysm clips, and cochlear implants. Problems unique to peripheral MRA include susceptibility to artifact from a hip or knee prosthesis. This can be impossible to overcome in many cases, but information about the anatomy of the vascular tree distant from the prosthesis can be determined and can be of diagnostic value. Discussion with the vascular surgeon and interventionalist in these cases should be undertaken to assure the clinical question can be answered.

What equipment is needed
to perform peripheral MRA?

The majority of studies on peripheral vascular MRA utilize >=1 T magnets, with most performed at 1.5 T. Significant studies below 1 T have not been done; therefore, it is recommended that a >=1 T magnet be utilized for optimal quality peripheral MRA. Because of the large field of view (FOV) needed for peripheral studies and the high resolution required, a dedicated peripheral vascular coil is also recommended. The multistation protocols are not suitable for either general body coils or adapted coils from elsewhere. The high-performance gradients are needed for peripheral MRA in order to obtain the low repetition time (TR)/echo time (TE) needed for the ultra-fast scanning protocols used in peripheral MRA.

Peripheral MRA requires optimal contrast opacification, and often multiple injections. For this reason a power injector is needed. A steady injection rate is needed to keep the gadolinium concentration uniform and sufficient during the entire scan time. Multi-station injection protocols would be difficult to time and administer without a power injector. To aid in timing the arrival of contrast with the scan acquisition, further software is required. This can entail a "fluoro"-triggered option. Automated threshold functions are also available, which trigger the scan when a certain contrast enhancement threshold has been reached. Further upgrades, including time-resolved acquisitions, are also available.

Once image acquisition has been achieved, postprocessing is required to render the data set manageable, to provide images that project the anatomy in familiar planes, and to maximize anatomic detail. This is achieved by first creating a mask of soft tissue (Figure 1A), which is then subtracted by the software on all further sequences, thereby maximizing visualization of the contrast-enhanced vasculature (Figure 1B). Further processing aids in the diagnosis of disease and offers the clinician a familiar image with which to plan further possible interventions. The numerous axial images are combined and rendered into a single image representative of the arterial system (Figure 2). The standard for this is the maximum intensity projection (MIP) where the most conspicuous pixels (the contrast-enhanced vessels) in each single image are shown preferentially (Figures 1B and 2). Ideally, this leaves an image of only the opacified arterial system. More advanced image presentation is available using multiplanar volume reconstructions (MPVR). Multiplaner volume reconstructions offer the advantage of 3D representations, and, therefore, are less susceptible to reconstruction artifacts seen in the 2D MIP reconstructions.

What techniques are currently in practice?

The initial foray into vascular imaging was conducted using time-of-flight (TOF) imaging. These acquisitions are performed by saturating the spinning protons in an imaging slice (2D) or volume (3D). This leaves little signal from the protons residing in the tissues, which are then imaged before the signal can regrow. The inflowing blood that has not received the saturation pulse has a signal that can be sampled, and the resulting image reveals bright signal in the flowing blood with lack of signal (dark) in the surrounding tissues. Several limitations of this technique make it suboptimal for peripheral MRA. Flow within the plane is also suppressed so vessels that do not have a perpendicular course in the image can have signal suppressed and yield false-positive studies. Slow flow signal drop out from blood residing in the volume too long is also encountered. Lastly, the time needed to cover the entire lower extremity and the resultant high probability of patient motion make TOF imaging ill-suited for the peripheral vascular system.

Phase-contrast MRA has also been used to evaluate the vascular system. This technique exploits the signal change found in the movement of spinning protons in the vessel in relation to the stationary protons in the tissues to acquire an image that is sensitive to a prescribed velocity of flow. Protons that are moving in that velocity range are subtracted from a flow-compensated mask image, and this results in images that can be tailored to desired flow properties. The flow profile must be known before imaging because the velocities that are encoded are preselected for a particular area. Velocities outside this range do not appropriately register on the image. This technique also has not found wide application in peripheral MRA.

Three-dimensional contrast-enhanced magnetic resonance angiography (CEMRA) has shown itself to be the most practical technique to use for peripheral MRA. The properties of gadolinium chelates, which lower the T1 of blood significantly, allow one to design imaging protocols that use a very fast TR (<10 ms). The protocols use a low flip-angle gradient-recalled echo technique to take advantage of the short T1 of gadolinium in blood (<50 ms) to obtain images with high intravascular signal. Knowing that fat also has a short T1 (270 ms at 1.5 T) that would confound the image, a mask image is obtained to minimize the signal from the soft tissues and maximize the vessel-to-background ratio. Several different strategies have been developed to exploit the contrast properties and obtain high-resolution images in the shortest time possible.

The first hurdle in acquiring an image during a CEMRA is bolus timing. The IV injection of gadolinium must pass through the venous system into the heart and travel to the portion of the arterial system to be imaged. Because the time required for this transit is variable, several strategies were developed to allow the image acquisition during peak arterial enhancement. The three major systems use either a test bolus, a fluoroscopic trigger, or empiric triggering. The test bolus begins with a 1-mL gadolinium injection with a saline chaser. A 2D fast image over the region of interest is scanned consecutively approximately every second. The bolus arrival time is obtained by these scans and the imaging protocol is triggered in a similar interval after the gadolinium bolus is initiated. Fluoroscopic triggering uses a similar fast (approximately 1 image per second) 2D acquisition that is running during the main bolus injection. The technologist then triggers the previously prepared 3D MRA sequence when the contrast bolus has arrived. Thirdly, software can be used to trigger the 3D examination by beginning scanning when a threshold enhancement has been reached over the region of interest, again using a fast 2D acquisition for the bolus timing.

At the time of bolus arrival there are two major techniques that are currently used to image during the peak of arterial contrast enhancement. The first uses a moving table in an attempt to follow the gadolinium bolus through the arterial phase by imaging very quickly at each station of the lower extremities, while scanning early enough to minimize the venous enhancement. The other technique involves performing separate contrast bolus and acquisition at different stations in the lower extremity. Because the FOV is usually limited to approximately 40 cm and the examination may need to cover 120 cm or more, imaging is performed separately at multiple stations, typically the pelvis, thigh, and calf (Figure 3). An abdominal MRA is often also performed in conjunction with the pelvic MRA to evaluate for concomitant aortic or renal artery disease (Figure 3A). An additional run-off station is occasionally necessary with tall patients.

The moving table approach attempts to follow a bolus of contrast through the 3 to 4 stations needed to image the entire lower extremity. The goal is a very short image time that would correspond to the contrast arrival at each station. This goal is difficult to achieve with the standard imaging protocol using a 20- to 25-second acquisition. This technique can be performed with automated table movements (slower table movement speeds can be a limiting factor) or with technologists moving the table manually to the prescribed positions. The bolus is prolonged to allow arterial enhancement during the entire study. This is usually achieved by a rapid bolus with a continuing injection at a slower rate that attempts to prolong the time the gadolinium resides in the arteries. A common injection might start with 1.5 mL/s for 15 mL, then decrease to 0.7 mL/s for 25 to 35 mL. The advantage of this type of examination is that a larger bolus is used, which leads to better vascular opacification and therefore better vessel-to-background ratio. There is also an advantage in that the middle stations have little venous con-tamination. Problems are encountered in the later (calf) stations, as the contrast injection yields peak arterial opacification during the first two stations but poor arterial signal (poor enhancement) later in the study. The later stations also suffer from problems of venous contamination resulting in imaging artifacts. Another obstacle is that the fast scans necessarily yield less signal-to-noise (ie, less time is spent sampling the arterial signal) and have difficulty in achieving the resolution needed to confidently diagnose vascular disease at the level of the ankle and foot where vessels are only 2 to 3 mm. Another limitation encountered is the presence of asymmetric disease. This leads to different arrival times. Because of the quick imaging and table movement, it can be difficult to image the bolus in the delayed (diseased) extremity. The amount of contrast used may limit the ability to reinject if this is not immediately recognized.

The other major technique uses smaller divided gadolinium boluses with dedicated scanning at each station. This technique attempts to optimize imaging at each station, but suffers from the smaller contrast doses and concern for possible venous contamination inherent in the longer examination time. The advantage of this technique is that scan times at each station do not need to be shortened to the extent that the image resolution could be jeopardized. The study takes slightly longer, as a mask must be obtained at each station immediately before the contrast arrival to minimize signal from the veins. This technique also has the advantage that if the technologist perceives asymmetric contrast arrival, the scan can be repeated immediately to improve the likelihood of obtaining peak contrast images of the delayed extremity.

A novel technique using the multistation method has been developed that involves continued imaging at a particular station using very fast scan times resulting in many sequences obtained during the contrast injection. Each of these ultra-short scans suffers from the limitations inherent in short scan times, however, the sequences are summed to obtain a final image that uses k-space data from the constituent images that enable the creation of a high-resolution image at that station. This time-resolved sequence allows improved imaging with asymmetric arrival time as well as giving physiologic information about the blood flow timing in the extremity (Figure 4).

What sequencing protocols are now being developed that may be seen in the future?

Using the moving table approach, several centers have developed further refinements including whole body MRA. The angiographic system for unlimited rolling field-of-view, or AngioSURF, technique uses the torso/body coil as a stationary coil through which the patient is moved during the examination. ^23,24 This technique allows for ultra-fast imaging at a fairly low resolution to image the entire arterial tree (carotids to calf). The technique currently is seen as a screening examination because of the resolution; however, diagnostic quality examinations of the entire arterial tree are achievable. Another refinement of CEMRA technique includes the "Shoot and Scoot" protocol. ²⁵ This acquisition involves initial sampling of central k-space at each station, yielding the bulk of image contrast, but postpones acquisition of peripheral k-space until after all stations have been imaged. Because most of the information lies in central k-space, the delayed sampling of peripheral k-space adds to image resolution but does not add any significant image contrast, therefore the venous signal present during the sampling of peripheral k-space does not contribute to image contrast and is not seen in the final image. The result is ultra-fast scanning that can image nearly all stations at peak contrast enhancement. WakiTrack uses sensitivity encoding, or SENSE, technology to image the pelvic stations more rapidly and, therefore, increases the likelihood of imaging the lower stations during peak contrast enhancement. ²⁶ This allows increased signal-to-noise ratio because of increased intravascular signal during imaging, diminished venous signal, especially in the lower station(s), and ability to perform higher resolution scans at the calf/ankle where resolution is most often suboptimal. A device has been developed to more rapidly move the patient between stations in a multistation bolus chase protocol. ²⁷ Stepping kinematic imaging platform, or SKIP, is a separate moving table with stationary coils attached, which is placed on the gantry table and allows the patient to be moved between stations in 3 to 4 seconds instead of 8 to 10 seconds if the standard table hardware is used to reposition the patient. This improves imaging in the calf station because the contrast bolus arrival more closely corresponds to imaging time, yielding better arterial enhancement and less venous contrast superimposition. Finally, research continues into blood pool agents that will remain in the arterial system longer, allowing the opportunity for longer imaging and thereby improving spatial resolution and limiting the venous enhancement, which can obscure anatomy.

Conclusion

Magnetic resonance angiography is a technique that has been studied extensively in the peripheral vascular system. It offers significant advantages over conventional DSA and has been shown to have similar diagnostic accuracy. No one technique has been shown to be the best at imaging the peripheral arterial system; therefore, much research continues into refinements of current techniques and advancement of newer techniques. One should now be familiar with the current protocols and their limitations, as well as the newer techniques that might be available in clinical practice in the near future. The radiologist who does not perform peripheral MRA, as well as those in training, should now be equipped with a knowledge base that will allow them to understand the issues and advantages of peripheral MRA, and those who are more familiar with it may have become acquainted with new imaging protocols that are being developed.