MRA procedure must be explained to the patient in detail. If a patient has rest pain or skin ulcerations, appropriate care must be taken of the affected limb and analgesics provided if necessary. Patients must remain still during the scan, as movement will affect image quality. A 1.5 T MRI machine with high-performance gradient strength is adequate for peripheral MRA. Incremental gains in study quality can be obtained with higher gradient strengths, multichannel phased array coil, and a 3 T scanner. Dedicated bilateral lower extremity phased array coils are ideal for peripheral MRA image quality. Additional body coils will often be placed over the lower abdomen just above the extremity coil assembly to image the abdominal aorta and iliofemoral vessels. Patients are placed feet first with their arms by their sides and are forewarned that the table will move during the study and compression devices may be placed about the thighs to minimize venous contamination [20]. Care must be taken not to compress any regions especially the calves because this may result in pseudo stenosis. For lower extremity MRA, a 20 gauge venous access must be obtained in either antecubital fossa. Lower limb venous access is not recommended due to significant venous contamination of the region of interest. A multiphase MR-compatible injector pump is important in attaining consistency in injection timing and flow rates [21]. Extension tubing flushed with normal saline is required to reach the pump injector. Once a patient is appropriately positioned, large field of view localizers are obtained. In obese patients and crossover bypass grafts, it is important to ensure that relevant vascular anatomy is not excluded from the imaging volume. Rapid scout images in axial, coronal, and sagittal orientations are obtained. These initial images are useful to define the vascular territory to plan the eventual MRA examination . Individual 3D volumes oriented in the coronal plane are acquired sequentially in multiple stations (abdomen, pelvis, legs, and calves) chasing a gadolinium-based contrast agent as it flows down into the peripheral vessels (Fig. 18.1). Subtraction of the post-contrast images from the pre-contrast images is performed and the data is transmitted to a 3D workstation.

To maximize tissue contrast, the sequences are heavily T1 weighted, and acquisition is timed to the maximum intra-vascular concentration of gadolinium. Typically a fast 3D gradient echo sequence is used for imaging acquisition. Images are acquired first without contrast (mask images) and then same acquisition is repeated with contrast. Subtraction of mask images from the contrast-enhanced images suppresses the background tissue signal. The major limitation with this method is patient movement and breathing between the mask and contrasted portions of the scan and different table positions between stations, which may result in misregistration artifacts. Image acquisition (both mask and actual CE-MRA) must be obtained with breath holding, which significantly improves image quality [22].

In CE-MRA the synchronization of image acquisition to intra-arterial GBCA concentration is extremely important, and data acquisition must coincide with the arrival of the contrast in the vessel of interest. The GBCA dose volume is small, and since the acquisition time for MRA is long, the contrast bolus must encompass 40–60 % of the duration of acquisition for optimum image contrast. Improper timing of acquisition results in image artifacts that can be especially problematic with bolus chase MRA of extremities, in which multiple stations are acquired with one contrast bolus. On average, the contrast material requires 24 s to reach the femoral artery but only 5 more seconds to reach the popliteal artery and additional 7 s to reach the ankle artery [23]. The rapid transit of contrast makes it especially difficult for first pass CE-MRA as table movement and imaging time needs to be fast enough to keep up with the bolus transit. In recent years, parallel imaging (SENSE, SMASH, GRAPPA) has emerged as powerful method to decrease acquisition time without degrading spatial resolution. In parallel imaging, data are acquired simultaneously using multiple surface coils with different spatial sensitivities. The disadvantage of parallel imaging is reduced signal to noise ratio (SNR) in proportion to the acceleration factor. Due to intrinsically higher SNR, higher acceleration factors can be achieved with 3 T MRI scanners.

We typically use 0.1–0.2 mmol/kg dose of intravenous GBCA for CE-MRA at the infusion rate of 2 cc/s using a power injector followed by a saline flush of 20 cc to ensure that the contrast material does not pool within the tubing or peripheral veins and is delivered to the central veins. Since the time to peak arterial enhancement is dependent on multiple variables such as contrast injection volume and rate, saline flush volume and rate and cardiac output, it is not possible to predict the scan delay reliably; a “test bolus method” or “bolus-tracking method ” should be used to assess the peak-enhancement time.

An important factor in achieving high diagnostic accuracy in peripheral MRA is the spatial resolution. Reliable diagnosis of arterial stenosis in MRA requires at least 3 pixels to be present across the arterial lumen. In the lower leg, this can be challenging, as the vessel diameters are 2–3 mm in this area. 3D computer workstations are used for image post-processing and viewing. The post-processing techniques used are multiplanar reformations (MPR) , curved MPR, maximum intensity projections (MIP) , and volume rendering (VR) . MPR allows viewing anatomy in all planes and the image resolution is improved if the acquisition voxels are isotropic. Curved MPR allows reformation along tortuous vessels. MIPs produce images comparable to digital subtraction angiography and can provide a quick preview of the peripheral vasculature, but diagnosis must always be confirmed by looking at thin MPR images. Volume-rendered images allow excellent display of surface-rendered 3D images but are not useful for stenosis quantitation.

There are several types of first pass CE-MRA acquisitions based on region of interest. Single-station MRA is used to acquire a single field of view when the area of interest is limited such as the abdominal aorta, carotid arteries, and renal arteries. Multi-station MRA is used to image the peripheral vasculature from the abdominal aorta to the feet by acquiring multiple overlapping fields of views. In multi-injection MRA, contiguous overlapping stations are acquired with separate injection of GBCA at each station but are limited by the total volume of GBCA that can be administered and by the effects of residual contrast in the blood pool from the previous injections. In moving table MRA, the entire region of interest is acquired by moving the scanner table through the center of the magnet in a stepwise fashion chasing a single contrast bolus through multiple discrete stations (Figs. 18.1 and 18.2). Typically three discrete stations are imaged in peripheral MRA: aortic bifurcation and the iliac arteries, femoral arteries, and the runoff vessels (Fig. 18.3). Image acquisition must be fast enough to keep up with the arterial bolus. Failure to do so will result in incomplete arterial opacification and venous contamination that can further degrade image quality. The sequential bolus chase acquisition images the lower leg region last and is most likely to suffer from venous contamination. The most common cause of venous contamination is timing problems but may also occur as a result of inflammatory hyperemia in cellulitis or spontaneous or iatrogenic formation of arteriovenous shunts in patients with diabetes and prior surgical or endovascular interventions. Faster acquisition times decrease the likelihood of venous contamination. Biphasic contrast administration with the initial rate of 1.5 mL/s followed by 0.5 mL/s prolongs the arterial phase. The latter phase equals tissue extraction and reduces venous filling [24]. Venous compression techniques using blood pressure cuffs inflated to 50–60 mmHg around the thigh or calf have proven effective in reducing venous contamination [20]. However, cuff compression should not be used in patients with superficial vascular bypass grafts due to the risk of thrombosis. In patients with bypass grafts, hybrid protocols in which the leg station is imaged first followed by abdomen and thigh result in lesser venous contamination.

Blood pool contrast agents allow us to image the peripheral vasculature beyond the short arterial first-pass phase [25]. Because there is no time constraint, MRA with very high spatial resolution can be achieved. Steady-state MRA (SS-MRA) provides images of both arteries and veins, but the excellent spatial resolution allows separation of arteries and veins. Combined first-pass and SS-MRA approach of the peripheral arteries improves results as compared to first-pass imaging alone [26].

Time-resolved MRA is based on rapid repeated acquisitions of 3D images at the same location over time. Due to the rapidity of image acquisition, there is no need for appropriate contrast timing, and it provides both arterial and venous phase information in a dynamic fashion. Time-resolved acquisition techniques when used with parallel imaging can efficiently improve temporal resolution but degrade spatial resolution. Time-resolved MRA is more reliable than first-pass MRA and digital subtraction MRA for the leg station, especially in patients with arteriovenous malformations and diabetics with small vessel disease.