MRA has the advantage of being noninvasive, does not require iodinated contrast or ionizing radiation, and provides an unlimited number of projections of the carotid lumen from a single acquisition. Two techniques are used: time-of-flight imaging (TOF), a flow-dependent technique, and contrast-enhanced MRA (CE MRA), a filling-dependent technique, comparable to the technique of CTA. TOF is a widely used technique to establish the diagnosis of carotid stenosis. “This technique is optimized to minimize the signal from stationary tissue, thereby increasing relative signal from the fresh spins delivered to the volume by blood flow from outside the imaging volume” [17]. There are two modes of TOF, two dimensional and three dimensional. 2D TOF is more sensitive to slower flow, while 3D TOF depicts a wide range of flow velocities and has greater accuracy than 2D for defining internal and external carotid artery morphology [17]. Because this imaging is flow dependent, there is some distortion of the carotid anatomy with high-grade lesions or in lesions with turbulent flow. “TOF spins may remain in imaging volume long enough to see numerous pulses and become saturated, thereby causing loss of signal within vessel lumen and inability to depict the vessel contiguous with the lesion” [17]. While this can lead to overestimation of the degree of stenosis and a high false-positive rate, in contrast, the false-negative rate is quite low. Demonstration of minimal disease at the carotid bifurcation on MRA is a highly accurate diagnosis [22].

Contrast-enhanced MRA (CE MRA) uses MRA technique to provide flow-independent anatomic information [17]. This technique is somewhat similar to CTA with first-pass MRA. Because these images are not dependent on flow, they provide a more accurate assessment of stenosis and visualization of ulcerated plaques. There may be some technical difficulty in capturing the timing bolus; however once this is overcome, the shorter imaging time increases accuracy secondary to a decreased risk of motion artifact.

Unfortunately, the use of MRA as a diagnostic tool for carotid stenosis is often dependent on local expertise and familiarity with the test. In centers where the test is widely used, it can provide valuable additional data from that obtained by a carotid duplex. MRA has no ionizing radiation or ionic contrast and is quite safe for most patients. Additionally, information about the cerebral circulation can be obtained simultaneously, including patency of the carotid siphon and middle cerebral artery.

MRA can display vessel anatomy as a rotating three-dimensional angiogram that can be readily interpreted by those who did not perform the study (Fig. 18.3). Severe or tight stenoses (≥70%) are usually seen as a flow gap (Fig. 18.4).

More advanced techniques are now being evaluated for visualization of the atherosclerotic plaque and vessel wall . These methods include standard CE MRA and TOF obtained with specialized surface carotid coils to increase the signal to noise ratio. Additional techniques include 3D bright blood MRA, 2D spin echo, and fast spin-echo methods. Data obtained from these techniques demonstrates good correlation with ex vivo plaque morphology. However, the tests are time consuming and technology laden and, as yet, merely experimental [23, 24].

The sensitivity and specificity for diagnosing 70–99% stenosis with TOF MRA are identical to duplex ultrasound (88% and 84%, respectively); however, MRA is less accurate in diagnosing 50–69% stenosis but is quite accurate in diagnosing carotid occlusion [25, 26]. MRI can be used to analyze plaque morphology, specifically the structure of the atherosclerotic plaque. It can identify the lipid-rich necrotic core and the fibrous capsule with high sensitivity and specificity [27] and can distinguish between intact thick, thin, or ruptured fibrous cap [28]. Using dedicated protocols, MRA can also demonstrate specific plaque components, e.g., calcium, lipid, fibrocellular element, or thrombus within the plaques.

MRA does not visualize the surrounding soft tissue structures, unless additional magnetic resonance imaging (MRI) is performed and calcium within the plaque is not defined. It cannot be used in patients with implanted ferromagnetic devices (e.g., implantable defibrillators, pacemakers) and is of limited use in uncooperative patients and those with claustrophobia. Additionally, small carotid lumens and tortuous vessels can be seen as occlusion or severe stenosis. A major concern in the use of gadolinium-enhanced MRI is nephrogenic systemic fibrosis (NSF); this is a debilitating and largely untreatable disease. Several risk factors have been associated with NSF, including type of gadolinium, dose administered, renal failure, acute inflammatory state, and elevated phosphate levels. The incidence of NSF is 1–6% for dialysis patients and patients with a glomerular filtration rate less than 30 [29–31]. This fact has led to wide screening of MRI patients for renal dysfunction and considering a GFR less than 30 a relative contraindication for administration of gadolinium. The Food and Drug Administration issued a black box warning to avoid the use of gadolinium for patients with end-stage renal disease or a GFR less than 15 mL/min per 1.73 m², unless necessary. Prior to performing MRA, physicians must insure that the patient is well hydrated and the acute inflammatory state has been resolved. Low-dose or non-contrast MRA should be considered first and patients on dialysis should be dialyzed immediately after the MRA. Finally, the test is quite costly and, therefore, is a less desirable test.

Many authorities still consider carotid conventional DSA to be the gold standard of all other imaging modalities in patients with extracranial cerebrovascular disease. In one test, evaluation of the aortic arch, extracranial, and intracranial cerebrovascular system can be performed. Additionally, DSA is the only test that can definitively diagnose lesions of flow dynamics, such as subclavian steal, with clear demonstration of delayed and collateral filling. Flow-dependent lesions are not well demonstrated by CTA or MRA.

Measurement of carotid stenosis using this method is generally based on the NASCET method [3]. Conventional angiography is usually reserved for patients with conflicting imaging studies prior to carotid endarterectomy or in patients considered for carotid stenting. DSA provides high-quality imaging, which is accurate, objective, and easy to interpret. It can identify lesions from the aortic arch to the intracranial vessels. Major limitations of angiography, that make it inappropriate as a screening modality, include its cost and associated risks, specifically stroke [32–34]. Although the stroke risk has been quoted as high as 1% in the NASCET trial, in practice, the stroke/transient ischemic attack (TIA) risk from carotid angiography is closer to 0.5% [35]. Additionally, the risk of femoral sheath hematoma, while low with a five-French sheath required for diagnostic angiography, is still 0.1–0.5% [36]. Additionally, in comparison to MRA, where there is no ionic contrast or ionizing radiation, DSA holds some risk of renal insufficiency and ionizing radiation.

Finally, DSA provides accurate information regarding the lumen of the carotid artery but fails to provide any information about plaque composition, the vessel wall, or surrounding cervical structures.

Overall, DSA is most useful in patients when less invasive imaging studies produce conflicting results. When duplex is equivocal, DSA is preferred over CT and MRA in evaluating patients with renal dysfunction (by minimizing contrast load), obesity, or in-dwelling ferromagnetic material, which renders CTA or MRA technically inadequate or difficult.

The UK Health Technology Assessment (HTA) concluded that although contrast-enhanced MRA was the most accurate imaging modality overall, it was limited by inaccessibility, unavailability, and delays. They concluded that color duplex ultrasound remained the preferred imaging modality for identifying patients with 70–99% stenosis [37]. As such, CDUS is the preferred imaging modality for identification of asymptomatic stenosis.

This recommendation was based on several factors, including low cost, a much higher number of strokes likely to be prevented in the long-term by the rapid availability of carotid duplex ultrasound in contrast to other imaging, and the good sensitivity of imaging in detecting significant stenosis. However, the HTA highlighted the concern of the accuracy of duplex ultrasound in diagnosing 50–69%, which carries a sensitivity of only 36%, with a specificity of 91% [37]. The utility of CDUS depends on the clinical presentation of the patient. In neurologically symptomatic patients, a diagnosis of stenosis between 50 and 69% by CDUS is sufficient to proceed with surgery, based on its specificity. However, the low sensitivity of CDUS in this setting would mandate another imaging study if the CDUS was negative. In neurologically asymptomatic patients, a moderate stenosis (50–69%) diagnosed by CDUS should be confirmed by another imaging study before intervention is undertaken.