Dr. Ho is Director of MR Research and Ms. Hood is the MR research coordinator, Department of Radiology, Uniformed Services University of the Health Sciences, Bethesda, MD; Dr. Meaney is Chief of MRI, St. James Hospital, Dublin, Ireland; Dr. Kent is Chairman of the Department of Vascular Surgery and Dr. Watts, Dr. Wang, Dr. Winchester, and Dr. Prince are faculty in the Department of Radiology, Weill Medical College of Cornell University, New York, NY; Dr. Dong is a research associate at the University of Michigan Medical Center, Ann Arbor, MI; Dr. Choyke is Chief of MRI, Diagnostic Radiology Department, Warren G. Magnuson Clinical Center, National Institutes of Health, Bethesda, MD. The opinions or assertions contained herein are the private views of the authors and are not to be construed as official or reflecting the views of the Department of Defense or the Uniformed Services University of the Health Sciences.

Patients with a suspicion of atherosclerotic peripheral vascular disease have presented a difficult imaging challenge. Flow-limiting stenoses tend to be localized to small regions of the body. However, because of the systemic (i.e., diffuse) nature of the atherosclerotic disease process, a flow-limiting atherosclerotic stenosis or occlusion could occur anywhere from the aorta down to the calf and still cause similar ischemic symptoms in the lower limb. In addition, synchronous disease is common and a proximal lesion may mask other more distal lesions, which can also be significant. For these reasons, in patients with suspected peripheral vascular disease, it is essential to image an extensive region of arterial anatomy from the abdominal aorta down to the feet.

The time-honored, gold-standard method for the evaluation of patients with peripheral vascular disease has been x-ray angiography, sometimes referred to as the peripheral "run-off" study. With conventional x-ray angiography, the patient is subjected to the risks of arterial catheterization, iodinated contrast, sedation, and ionizing radiation. The iodinated contrast load may be large because of the extensive region of anatomy to be covered. But these patients also have a high incidence of renal artery disease as well as nephrosclerosis, which puts them at greater risk of iodinated-contrast-media­induced renal failure. These clinical considerations led to much enthusiasm with the early successes described for peripheral magnetic resonance (MR) angiography ^1-3 using two-dimensional (2D) time-of-flight (TOF). However, 2D TOF MR angiography relies on the phenomenon of "in-flow" for vascular depiction and, therefore, is prone to image artifacts related to arterial pulsation, turbulent flow, in-plane saturation, and saturation of retrograde flow--all of which can result in poor arterial signal and the overestimation of arterial stenoses. ⁴ Slice misregistration artifacts can also occur with 2D TOF in patients who cannot hold perfectly still. One additional practical consideration with peripheral 2D TOF imaging is the long scan times required for completion of the exam, which can exceed 2 hours.

Gadolinium (Gd)-enhanced three-dimensional (3D) magnetic resonance angiography (MRA) has revolutionized arterial imaging with MR and largely replaced traditional 2D TOF methods. The technique uses gadolinium-chelate contrast media to produce images that are essentially "luminograms" comparable to those of catheter angiography. Gd-enhanced 3D MRA, of course, has distinct clinical advantages over catheter angiography, which is invasive and requires the use of nephrotoxic iodinated contrast agents and ionizing radiation.

Until recently, Gd-enhanced 3D MRA has been limited to imaging of a single region of the body (i.e., a single "station") corresponding to the field of view of the MR scanner, typically 40 to 50 cm. This is not sufficient for evaluating patients with peripheral vascular disease where the distance from upper abdominal aorta to feet is typically more than 100 cm. However, by imaging rapidly and moving the patient through the isocenter of the scanner as the Gd contrast passes down the legs, it is possible to image the arteries of the entire lower half of the body using multiple overlapping fields of view. ^5-7 This technique, known as bolus chasing, is not entirely new, as it has been used for many years in conventional x-ray angiography for peripheral run-off studies using a stepping or moving table technique.

This article will discuss the practical considerations of bolus-chase MRA as they pertain to the evaluation of patients with atherosclerotic peripheral vascular disease. The authors intend to provide a foundation of understanding such that readers can build their own peripheral MRA strategy to accomplish their own specific needs as dictated by their referring surgeons, equipment, and personnel.

Bolus chase MRA: Technical considerations

Currently, there are three critical aspects for successful acquisition of diagnostic MRA examinations of the peripheral vasculature. These are accurate placement of the imaging volume, appropriate bolus administration, and image subtraction. Accurate placement of the imaging volume is critical, as the arteries of interest must be imaged with the fewest slices possible to ensure that imaging of all three data sets can be completed before onset of venous enhancement. Imaging is performed in the coronal plane with several, usually three, stations (figure 1). Typically, the first station includes the abdomen-pelvis; the second station, the thigh; and the third station, the knee-calf. Occasionally, as in the case of patients with limb-threatening ischemia or nonhealing ulcerations, illustration of a fourth station for the distal calf and feet is desired. By positioning the patient with the knees and ankles elevated, the arteries of the lower half of the body can be brought within a relatively horizontal, narrow volume of coverage. This permits use of a relatively small number of slices to cover all the arteries for rapid imaging with high resolution. However, most manufacturers now offer the possibility of independent angulation of the three imaging volumes, which improves the ease of performance, as the legs no longer need to be elevated prior to the procedure.

The 3D volume for the bolus-chase MRA is positioned by determining the anterior and posterior extent of coverage necessary to include the arteries by first acquiring axial and/or sagittal 2D TOF images that sample the anatomy at 10 to 20 cm intervals from mid-abdomen to mid-calf. Typically, the most anterior extent of the peripheral vessels is at the common femoral artery and the most posterior extent is either at the popliteal artery or the common iliac bifurcation. Failure to include segments of vascular anatomy within the coronal 3D volume may result in erroneous simulation of arterial stenosis or occlusion (figure 2). Once the anterior-posterior coverage is determined, it is then necessary to select the slice thickness, number of slices, number of phase encoding steps and bandwidth such that the acquisition time is optimized to keep up with the bolus as it passes down the legs. In general, it is necessary to reduce the resolution to be able to scan faster. If necessary, the slice thickness can be increased to 5 mm (and then interpolated or zero-filled down to 2.5 mm), the phase encoding steps reduced to 128 and the bandwidth increased to ± 62 kHz. However, typically the slice thickness is 3.6 mm with 192 phase-encoding steps and ± 32 kHz bandwidth. On some MR scanners, station-specific image prescription of the bolus chase MRA is currently available. This feature enables further optimization of speed and spatial resolution (i.e., coronal oblique imaging and imaging parameters such as partition thickness for each station). ⁸

Another significant consideration for bolus-chase MRA is the selection of contrast volume and injection rates. The image data for each station should be acquired during the arterial phase of the Gd contrast bolus that lasts at least the duration of the bolus itself (e.g., 40-mL contrast dose injected at 0.7 mL/sec has an approximately 57-sec duration). Because of the importance of central k-space data for determining image contrast, it is particularly important to acquire the data corresponding to the central half of k-space while the bolus is within the arteries for that station. In order to achieve correct bolus timing, the duration of each station must match the rate of passage of contrast down that particular station. This requires knowing in advance the rate at which contrast is flowing down the legs and also requires adjusting the resolution, bandwidth, and other imaging parameters so that the image acquisition time is correct. The speed of the image acquisition will depend on the scanner and its particular software and gradient hardware. However, the duration of the bolus should at least approximate the duration required to obtain the central k-space data for all stations (typically 40 to 60 seconds for a three-station bolus chase).

Because 3D imaging typically takes between 12 to 24 seconds per location, there is a risk of venous enhancement in some patients. This can be minimized, however, by exploiting two modifications commercially available on many state-of-the-art systems: the use of a bolus-detection technique to trigger imaging for the arterial phase and the optimization of k-space mapping. The use of a bolus-detection technique, such as an automated algorithm (e.g., MR SmartPrep, GE Medical Systems, Milwaukee, WI; figure 3) or MR fluoroscopy (e.g., BolusTrak, Philips Medical Systems, Bothell, WA; figure 4) ensures that imaging is performed during arterial enhancement (at least for the initial station), thereby minimizing the risk of venous contamination of images. In association with bolus detection, it makes sense to alter the order in which k-space is mapped for three stations. For the first station (aortoiliac), the center of k-space is acquired toward the end of the acquisition. For the second and particularly the third stations, it makes sense to acquire the data "centrically," to shorten the time from onset of injection to completion of central k-space data mapping, as this will diminish the chance of venous enhancement (figure 5). ^7,9 Additional time savings (thereby shortening bolus duration requirements) can be realized by reducing the phase-encode steps in the slice direction for the intermediate second station. ¹⁰ Exploiting a partial Fourier image acquisition scheme, the scan duration for each station might have to be as short as 15 to 20 seconds per station. With the re-mapping of k-space, a 20-second scan duration for each station can still work out well, especially for imaging the proximal 2 stations above the knee.

In general, the rate of bolus administration is predicated on the time required for image acquisition (i.e., scan time for the MRA). One successful bolus infusion scheme for a 60-second image duration requirement would be to inject 40 mL of Gd-chelate contrast at approximately 0.7 to 1 mL/sec. ^5,6 Because of the added time required to complete bolus chase imaging, the contrast media injection rates are much slower than the 2 mL/sec recommended for single-station Gd-enhanced 3D MRA. ¹¹ However, with the development of faster gradients, substantially shorter scan times could be used in association with faster injection rates--the resultant lower blood T1 would offset a lower signal-to-noise ratio (SNR) predicted by use of a shorter repetition time (TR), because SNR varies as a function of 1/T1 and a function of TR. The use of a fixed 30- to 40-mL volume for bolus-chase MRA simplifies the procedure but may result in inconsistencies between patients because larger individuals may require more contrast media, and smaller patients may require less. To offset these concerns, an alternative weight-based dose schedule can be used. In this alternative scheme, 0.2 mmol/kg dose of Gd-chelate contrast media is diluted to a fixed volume (such as 45 mL for a 45-second bolus duration; or 60 mL for a 60-second bolus duration) with sterile saline and injected at a fixed rate of 1 mL/sec. ⁷ This also achieves the desired goal of matching bolus duration with imaging duration. The benefit of this strategy is that it affords the opportunity for the operator to repeat a Gd-enhanced 3D MRA with 0.1 mmol/kg should one region require re-evaluation (figure 6).

Finally, a significant improvement in vascular image contrast is obtained by subtracting a precontrast "mask" data set from the arterial phase data (figures 3 and 4). Subtracting background tissue signal improves visualization of smaller vessels that might be obscured by volume averaging. Ideally, this background, mask subtraction is performed in Fourier space as a complex subtraction to maximize visualization of the smallest arteries. Acquiring a precontrast mask requires a table-positioning mechanism that allows the table to be re-positioned precisely at the abdomen-pelvis, thigh, and calf stations before and during Gd infusion with minimal variation in position. Image subtraction is particularly helpful for imaging the calf vessels (figures 3 and 4), which can be difficult to visualize against the background of high signal in overlaying marrow fat within the tibia and fibula as well as in subcutaneous fat.

Discussion of literature

There have been numerous investigations on the accuracy of MRA techniques for evaluating patients suspected of having peripheral vascular disease (figure 7). Two-dimensional TOF has been shown to be accurate for evaluating infra-popliteal arteries; however, the examination is cumbersome to perform because of the requirement of multiple coil placements for each leg. More proximally in the pelvis, 2D TOF has been more limited due to artifactual signal loss from vessel tortuosity, pulsatile flow, and susceptibility from bowel gas. Gd-enhanced 3D MRA, on the other hand, is highly accurate in the pelvis and thigh. Comparative studies have established a high level of accuracy in this region (Table 1).

However, visualization of the smaller vessels below the knee can often be less reliable using the bolus-chase technique. For this reason, some centers have advocated performing 2D TOF initially on the symptomatic leg below the knee prior to a three-station bolus-chase 3D MRA. Alternatively, the infrapopliteal vessels can be evaluated with the 2D projection MRA technique (figure 8), which if performed first can also provide crucial bolus timing information for subsequent multi-station bolus-chase 3D MRA.

Future advances

The 3D bolus-chase technique is just beginning to be developed and introduced. As MR scanner gradient performance and speed continue to increase, it will be possible to obtain higher and higher resolution bolus-chase studies fast enough to keep up with the bolus passing down the legs. There is also considerable effort to develop hybrid techniques, such as interleaving a faster 2D MRA for imaging the relatively straight superficial femoral arteries of the thigh (station 2) between the 3D MRAs for the tortuous vessels of the pelvis (station 1) and calf (station 3). ²³ Spiral k-space trajectories also offer the potential for greater sampling efficiency.

Alternative k-space acquisition schemes, such as segmented volume acquisitions (a.k.a. "shoot and scoot" ²⁴ ), may further improve bolus-chase MRA. With this method, only the central k-space data for each station is acquired during the initial arterial-phase "pass" and the remaining k-space data is acquired later during a second delayed-phase "pass." Because the central k-space data is responsible for the majority of image contrast, this method maximizes the resultant image contrast data for the arterial phase of the bolus and enables the reduction of the imaging time requirements for each station and a faster bolus chase.

New contrast agents with higher T1 relaxivity (R1), better R1/R2 ratios, or those approved for use at higher doses will allow greater T1 shortening of the blood to enhance image SNR. Increased SNR will also be possible with improvements in coil design that allow better matching of the coil to body to increase the coil-filling factor.

Summary

Bolus-chase MRA is revolutionizing the management of peripheral vascular disease. It is now possible to rapidly image the arterial anatomy of the entire lower half of the body in a simple, safe, fast examination with no ionizing radiation, no arterial catheterization, and no nephrotoxicity. Preliminary reports suggest that "whole-body MRA" using five fields of view and extremely rapid scanning techniques will allow depiction of whole body atherosclerotic burden (excluding the coronary arteries) using a hybrid of the technique presented here. ²⁵ This will increasingly allow general practitioners and internists to participate more actively in the management of these patients. AR