Intracranial atherosclerotic disease (ICAD) is defined as stenosis or occlusion of the intracranial ICA, middle cerebral artery (MCA) M1 and M2 segments, or the vertebrobasilar system. As such, hypoperfusion and strokes in these regions tend to come to clinical attention earlier than those in other intracranial vessels. ICA lesions classically present with amaurosis fugax, or the shade-like obscuration of vision in a single eye, but can also present with motor weakness when lesions occur in the vascular watershed regions. MCA lesions are more commonly associated with motor weakness and speech changes. So-called limb shaking episodes can be seen in patients with severe hemodynamic compromise. Vertebrobasilar lesions often manifest with lightheadedness, syncope, and autonomic nervous system changes. Atherosclerotic vascular disease is usually seen in a more elderly population but can be present in younger patients with predispositions, such as hyperlipidemia, smoking, and diabetes mellitus.

The gold standard for the diagnosis of ICAD is digital subtraction angiography (Fig. 48.1). However, several
characteristic findings on conventional tomographic imaging can suggest the diagnosis. Low attenuation lesions on computed tomography (CT) or high signal intensity on fluid-attenuated inversion recovery or diffusion-weighted imaging magnetic resonance imaging (MRI) in the watershed regions or in the posterior fossa can often be a clue to this diagnosis. Even without the use of angiographic sequences, absent vascular flow voids on MRI (particularly well seen on T₂-weighted spin-echo sequences), can suggest this diagnosis. Finally, CT and MR angiography (MRA) also depict the vascular lesions of ICAD. Prior studies suggest that CT angiogram has a higher concordance with digital subtraction angiography than does MRA.^12,13

The typical perfusion pattern in patients with atherosclerotic cerebrovascular disease of the intracranial vessels is variable. Mild to moderate disease may only be visible on high-resolution angiographic images, with regions of stenosis and irregularity. In more severe cases, perfusion abnormalities within the affected vascular territory can be seen on dynamic susceptibility contrast (DSC) or arterial spin label (ASL) imaging. on DSC, typically the most sensitive indication of intracranial disease is delayed contrast arrival, either in the vascular border zones or a specific vascular territory (Fig. 48.2). Both arterial input function (AIF)-corrected (time to the maximum of the residue function [Tmax]) or uncorrected (time-to-peak [TTP] maps) can be used for this purpose. If the stenosis or occlusion is particularly hemodynamically significant, increases in other parameters, such as the relative cerebral blood volume (rCBV) or mean transit time (MTT), can be present, although in the authors’ experience, this is less common. Kim et al.¹⁴ demonstrated that relative MTT lesions (normalized by the contralateral hemisphere) showed abnormalities in patients with symptomatic MCA stenosis. Finally, most patients do not have significant CBF changes without evidence of ischemic stroke, suggesting that they have not exceeded the limits of autoregulation.

Standard bolus DSC is challenging in these patients for several reasons. The first is the question of where to choose the AIF. This was addressed by Mukherjee et al.¹⁵ in patients with unilateral carotid occlusion. They showed that in six of seven patients, it had no effect on the accuracy of the CBF measurements when compared with positron emission tomography. Nevertheless, very little information on the effects of AIF placement exists for
patients with intracranial stenosis or occlusion. Calamante et al.¹⁶ discussed many of these issues for patients with moyamoya disease, who may have bilateral occlusion of the MCA and anterior cerebral arteries (ACAs) with extensive collaterals. Although bolus DSC is usually performed with echo-planar imaging, which typically shows distortion at the lower slices, the authors recommend using the basilar artery in ICAD patients (assuming the disease is affecting the anterior circulation). An example of bolus DSC studies in a patient with symptomatic unilateral MCA occlusion is shown in Figure 48.2, demonstrating that there was a reduction in Tmax and possible normalization of MTT within the MCA territory following hypertensive therapy.

ASL imaging is particularly sensitive to mild delays in arterial transit times, which is a disadvantage for measuring quantitative CBF. Nevertheless, this sensitivity, which arises because of the dependence of the label on the T₁ of the blood (1.2 to 1.6 seconds at typical field strengths), can be useful for identifying a region with a mild transit delay (Fig. 48.3). The extreme case of this is likely the so-called border-zone sign,¹⁷ which represents severe transit delay so that the ASL signal is only present in the feeding arteries and has yet to enter the (capillary network of the) tissue at the time of imaging (Fig. 48.4). It should be noted that the precise etiology of the border-zone sign is unclear, and transit delays can occur due to poor cardiac output or small vessel ischemic disease in addition to stenosis of large and medium-sized intracranial arteries that is depicted on angiographic sequences.

Moyamoya disease is named after the Japanese term for “puff of smoke.” This is because of the difficulty in spatially resolving the multiple tiny collateral vessels that occur at the base of the brain, in an attempt to revascularize stenosis and occlusion of the region of the internal carotid artery terminus and proximal M1 and A1 segments. The cause of moyamoya disease is still currently unknown, although there is some evidence for a small genetic component. Most cases occur either in the pediatric population or in young adults. Although the disease was first identified in the Japanese and is thought to be more common in Asians, it affects all ethnic groups. Furthermore, with the more widespread availability of noninvasive imaging with MRI and CT angiography, it is being identified more frequently. Most patients present with transient ischemic attacks or small strokes, typically in border-zone regions. Less commonly, but still well recognized, is its presentation with hemorrhage, and moyamoya disease factors high in the considerations that must be made when confronted with a young patient with intracranial hemorrhage. Less common presentations include headaches, seizures, and developmental delay.

Moyamoya disease can be assessed using conventional MRI, where the absence of T₂ flow voids and small
serpiginous collaterals at the base of the brain are frequently seen. Another typical finding is the “ivy sign,” defined as high signal on fluid-attenuated inversion recovery imaging in the supratentorial sulci, and representative of slow flow in large pial leptomeningeal collaterals. MRA is also characteristic, with the mildest cases showing stenosis of the ICA terminus region and more severe cases demonstrating occlusion of the proximal segments of the MCA and ACA. Rarely, diffusion abnormalities in the watershed regions can be seen and, if present in a young patient, should raise suspicion for moyamoya disease.

Perfusion patterns in moyamoya are similar to those seen in atherosclerotic disease, with the exception that abnormalities are frequently bilateral and affect primarily the anterior circulation. The most common findings on DSC are prolonged contrast arrival times,^16,18 particularly if delay-insensitive circular deconvolution is used.¹⁹ If severe, changes in MTT that represent an increased ratio of CBV to CBF can also be seen. An example of this in a patient with bilateral moyamoya is shown as Figure 48.5.

ASL imaging has been applied to moyamoya as well.^20,21,22,23 One important caveat is that delay times are quite long in moyamoya, and underestimation of CBF or artifacts related to slow flow will be seen if ASL sequences with typically used inflow times (1 to 2 seconds) are used (Fig. 48.5).²⁰ One recent study has shown that the use of longer inflow times (i.e., longer postlabel delays [PLD]) for pseudo-continuous ASL) demonstrates an increase in the apparent CBF at longer PLD in moyamoya patients.²¹ In fact, it can be that conventional ASL at long PLD overestimates CBF because the label spends more time in the blood rather than in the tissue, where T₁ is longer, and therefore experiences less decay than expected. ASL approaches that use multiple time points to simultaneously acquire CBF and arrival time can be useful in this population, although it is important to remember that very long inflow times must be acquired.^24,25 Finally, velocity-selective ASL,²⁶ in which the label is applied not spatially in the neck but rather within the voxel itself based on blood velocity, can be particularly efficacious for both ICAD and moyamoya disease (Fig. 48.6). In a study of patients with long arterial arrival times, velocity-selective ASL had the best performance overall compared with gold standard xenon CT CBF, and in particular, the errors had the least dependence on arrival time.²¹