Emerging Concept of Intracranial Arterial Diseases: The Role of High Resolution Vessel Wall MRI

Anatomy

The anatomical differences between the extracranial and intracranial arteries are as follows. The intracranial cerebral arteries are relatively small in diameter and travel in the subarachnoid space, floating in the cerebrospinal fluid (CSF). Intracranial arteries have many perforating branches that run perpendicularly, are more tortuous than extracranial arteries, and have abundant collateral circulation (Figure 1A). In contrast, extracranial arteries are relatively large in diameter, surrounded by connective tissue, lack perforating arteries, and are less tortuous.

Figure 1.

Overview of cerebrovascular anatomy and embryology. (A) Schematic drawing of the anatomy of the intracranial and extracranial arteries. The middle cerebral artery traverses through the subarachnoid space and extends perforators into the cerebral parenchyma. (B) In the early embryo, the three aortic arches (AA) link the ventral (VA) and dorsal aorta (DA). (C) As the first (I AA) and second aortic arches (II AA) regress, the DA and the third aortic arch (III AA) remain as the primitive internal carotid artery (ICA). The remnant I AA corresponds to the mandibular artery (MA), and the remnant II AA corresponds to the hyoid artery (HA). Other embryonic arteries arise, such as the primitive maxillary artery (PMA), trigeminal artery (TA), and primitive dorsal ophthalmic artery (PDOA). (D) The embryonic arteries that regress divide the primitive ICA into seven segments. The first segment corresponds to the remnant of the III AA. The second segment is the remnant of the DA between the II AA and III AA. The third to seventh segments are divided into the MA, PMA, PDOA, primitive ventral OA (PVOA), and ICA bifurcation. CrD, cranial division; CaD, caudal division; HypA, hypoglossal artery; PA, proatlantal artery; DAr, dorsal aorta right; DAl, dorsal aorta left; BA, basilar artery; ACA, anterior cerebral artery; PComA, posterior communicating artery; ECA, external carotid artery.

Embryology

As shown in Figure 1B-D, primitive intracranial arteries arise, regress, and mature during development. Lasjaunias [8] proposed dividing the internal carotid artery (ICA) into seven segments, each with distinct vulnerabilities to pathological changes based on segmental identity and evolutionary/embryological characteristics. There are examples of vascular diseases, such as serpentine arterial aneurysms, arterial dysplasia, dolichoarterial segments, and MMD, that involve only certain ICA segments [8].

In the posterior circulation, the proximal vertebral artery (VA) originates from seventh dorsal cervical segmental arteries, while the rostral portion originates from the bilateral longitudinal neural arteries [9]. The boundary between these two regions is near the site where the extracranial VA transitions to the intracranial vertebrobasilar system [10].

In addition, the endothelial cells of the intracranial arteries are of neural crest origin, whereas those of the extracranial ICA and the vertebrobasilar system, up to the superior cerebellar artery, are of mesodermal origin [11]. In intracranial arterial dolichoectasia, only the ICA, posterior communicating artery, and posterior cerebral artery, which originate from the neural crest, are affected, whereas the rest of the posterior circulation is spared, indicating that the disease is a neurocristopathy [11,12]. Thus, the intracranial arteries in each segment differ on an embryological basis, beyond mere regional or anatomic differences, which provides insight into the types of vascular pathologies that commonly occur in each segment.

Histology

Blood vessels consist of three layers: the inner intimal layer lined with endothelial cells, the middle tunica media containing smooth muscle cells, and the outer tunica adventitia (Figure 2A). The extracranial arteries, including the aorta and carotid arteries, are elastic with abundant elastin filaments in their thick tunica media and adventitia. The adventitial layer is continuous with the surrounding connective tissue. The vasa vasorum (VV) is abundant and supplies oxygen and nutrients to the blood vessels. In contrast, the intracranial artery is a muscular artery with relatively fewer elastic fibers [13]. The intracranial artery “floats” in the subarachnoid space, so the tunica adventitia cannot be directly extended to the connective tissue [14]. Therefore, the media and adventitial layers are relatively thin, and the VV is not extensively developed [14].

Figure 2.

Schematic diagram displaying the heterogeneity of intracranial arterial diseases. (A) The structure of a normal intracranial vessel wall consists of endothelial cells, intima, internal elastic lamina (IEL), media, and adventitia. There is no external elastic lamina in intracranial arteries. Intracranial arteries have fewer elastin fibers in the media, thinner adventitia, and thicker IEL compared to extracranial arteries. (B) Atherosclerosis is characterized by a lipid core covered by fibrous caps. Infiltrating macrophages are found within the lipid core. More advanced forms of atherosclerosis are infrequently seen in intracranial arteries. (C) Intracranial arteries in moyamoya disease undergo negative remodeling due to fibrosis and thickening of the intima and proliferation of smooth muscle cells. There is also infolding and chronic contraction of the IEL. (D) Intracranial arterial dissection. The intima is torn inward to form an intimal flap and a double lumen. Intramural hematoma is sometimes seen. (E) Mononuclear and granulomatous adventitial inflammation and focal fibrin thrombus formation in vasculitis. Concentric thickening of the vessel wall is seen. (F) In reversible cerebral vasoconstriction syndrome, the diameter of the entire vessel wall is reduced due to the contraction of smooth muscle in the media layer.

Histological differences between the intracranial and extracranial arteries likely stem from the different embryological origins of each arterial segment [13]. It has been proposed that the thick adventitia surrounding the extracranial cerebral arteries undergo significant transition as they travel through the skull and pierce the dura. Their histological characteristics gradually transition to those of the intracranial arteries as they travel through the intracranial subarachnoid space [10].

ICAD: atherosclerosis

Atherosclerotic plaques and the resulting arterial stenosis of the intracranial arteries show histological differences from those of the extracranial cerebral arteries. Intracranial atherosclerosis more often forms a relatively fibrous plaque, while the formation of a lipid-rich necrotic core, often observed in extracranial atherosclerotic plaques, is less frequently observed or restricted to the most proximal intracranial segments [13,15]. In addition, intracranial cerebral arteries after ICA bifurcation do not frequently have calcifications except at the carotid siphons or proximal intracranial VAs (Figure 2B). Interestingly, intracranial atherosclerosis develops and progresses at a relatively later age than that in the coronary and extracranial carotid arteries [15]. In the systemic arterial vasculature, a fatty streak is detected in the aorta during the second decade, and atherosclerosis begins in the carotid arteries during the third decade.

The reasons for these differences in the atherogenic processes in the intracranial and extracranial cerebral arteries are not clearly understood. Several biological and histological differences are believed to play a role in this process. First, the intracranial artery “floats” in the subarachnoid space. The formation of the VV is weak, and it is possible that oxygen and nutrients are supplied via direct diffusion from CSF [14]. Consequently, neovascularization of the VV, which may have a role in the maturation of atherosclerotic plaques, is limited. Intracranial VV does not exist at birth but may develop later in adulthood; however, extracranial vascular beds, including proximal ICA and VA, contain VV-ish structures from birth [16]. Intracranial VV further develops in related to vessel wall thickening, decreased blood flow, and decreased oxygen utilization, which occur in older individuals and those with hypertension [17]. In one study, intracranial atherosclerosis was associated with age, hypertension, and diabetes, whereas hyperlipidemia and smoking were not [18]. It could be interpreted that hypertension and age could promote neovascularization of the VV in the adventitia, whereas hyperlipidemia and smoking do not directly impact the VV because they damage the intima at the initial stage [16]. Second, intracranial arteries have an embryological origin that is distinct from that of extracranial arteries and appear to be relatively insensitive to sympathomimetic stimulation [10]. Third, the shear stress on the intracranial arteries may be partially ameliorated by autoregulation in the distal cerebral vascular beds. Lastly, molecular differences, such as augmented expression of matrix metalloproteinase-9 and fewer caveolae, may explain the differences in the atherogenic process [19].

ICAD: RNF213 spectrum disorder or MMD spectrum

MMD is an uncommon inherited condition that causes progressive steno-occlusion of the distal ICA during early adulthood. It has a relatively high prevalence in East Asia (1.61–18.1/100,000 individuals) (Figure 2C) [20,21].

According to the diagnostic criteria proposed in 2021, MMD can be diagnosed by MR or cerebral angiography when significant steno-occlusion of the terminal ICA with a rich abnormal vascular network in the basal ganglia or perivascular white matter is documented, and other potential causes are excluded, such as autoimmune disease, meningitis, brain tumor, neurofibromatosis, or radiation vasculopathy [22].

Variants in several susceptibility genes are involved in MMD development. Among them, RNF213 14429G>A (rs112735431, resulting in Arg4810Lys) is a major founder mutation in East Asian countries and is found in approximately 80%–90% of families with familial traits of MMD [23]. However, the RNF213 14429G>A variant has a prevalence of approximately 1%–1.5% in the general population in Korea and Japan [24]. Patients with this RNF213 variant do not always present with the typical MMD-like distal ICA occlusions or abundant collaterals. Instead, patients with this variant often present normal intracranial arteries or asymptomatic narrowing of the long segment of the intracranial artery without collateral channels [25]. Although RNF213-related intracranial atherosclerosis is more likely to be associated with negative remodeling and vulnerable to hemodynamic compromise [25], it is difficult to distinguish this condition from atherosclerotic stenosis, which is typically found in middle-aged individuals.

Bang et al. [26] proposed the concept of the RNF213 spectrum disorder. In this disorder, aging, susceptibility genes such as RNF213, and environmental factors interact to develop and progress to ICAD. Their hypothesis suggested that ICAD is likely a result of a mixed pathology; consequently, the progression of ICAD may vary between individuals. The unifying suggestion of RNF213 spectrum disorder is a practical way to evaluate pre-symptomatic ICAD cases and comprehend the pathogenesis of ICAD, particularly for patients of East Asian origin, given the relatively high percentage of carriers of RNF213 variants in the population [26]. However, no pathological studies have been reported, and further research is needed to test the validity of this concept. Additionally, a minority of patients with ICAD have typical MMD-like vascular pathology without an RNF213 mutation [27], and RNF213 polymorphism is reported to be found in intracranial dissection cases and multiple intracranial stenosis [28,29]. More research correlating genetic, pathological, and serial imaging data is needed to better understand this phenotype.

ICAD: intracranial dissection

Arterial dissection involves the disruption of each layer of the arterial wall, which can lead to an intimal flap, intramural hematoma, pseudoaneurysm, and even artery rupture. Similar to intracranial atherosclerosis and MMD, intracranial dissection is reported more frequently in Asian patients [30]. Cervicocephalic arterial dissection is the leading cause of ischemic stroke in children and young adults. Most pathological studies on intracranial arterial dissection have revealed disruption of the internal elastic lamina and media, followed by subarachnoid hemorrhage. Hemorrhagic stroke can occur in up to 50%–60% of cases after intracranial dissection, but this could be an overestimate considering the selection bias in these studies, which were mainly conducted by neurosurgeons who are consulted only when there is evidence of subarachnoid hemorrhage [31,32]. Intracranial dissection without arterial rupture can cause ischemic stroke via several mechanisms. This includes luminal narrowing due to intramural hematoma, distal embolization of the thrombus formed within a false lumen, or occlusion of the perforating arteries by the dissected segment. A high clinical suspicion by a physician is warranted for a correct diagnosis. Due to the small caliber of the intracranial arteries, it is difficult to detect imaging clues for diagnosing arterial dissection using standard imaging tools [33]. Typically, several weeks after arterial dissection, the dissected segment is replaced by granulation tissue and subsequent neovascularization occurs in the thinned intima, potentially leading to a fusiform aneurysm or a healed vessel wall [32]. This necessitates caution when interpreting HR-VWI to evaluate ICAD, as discussed in the following section (Figure 2D).

ICAD: vasculitis

Inflammation of the cerebral arterial system is described as central nervous system (CNS) vasculitis [34]. A very large number of conditions are associated with CNS vasculitis, and the pathogenesis and mechanism of this disease entity is expansive. Although CNS vasculitis can manifest as large- and medium-vessel inflammation, which is easily detected on conventional lumen-based imaging techniques and HR-VWI, small arteries and leptomeningeal vessels can also be affected, which are beyond the imaging spatial resolution of conventional techniques. Thus, the negative predictive value of imaging for CNS vasculitis is low, and a negative imaging examination does not exclude the possibility of CNS vasculitis [35]. Several studies using serial imaging have reported that contrast enhancement of the vessels reflects vasculitic activity and is correlated with clinical outcomes in terms of recurrence and treatment effects (Figure 2E) [36-39].

On HR-VWI, CNS vasculitis may show periadventitial enhancement extending beyond the borders of the vessel wall itself in addition to concentric vessel wall thickening and enhancement [40]. This imaging feature may help distinguish between infectious CNS vasculitis and autoimmune/inflammatory CNS vasculitis and atherosclerosis. However, owing to the rarity of these infectious vasculitides, the discriminatory power of this imaging finding has not been well studied, and studies with larger sample populations are needed.

Clinical considerations on the image acquisition of HR-VWI

Due to the high submillimeter spatial resolution achieved by current HR-VWI techniques, the detailed vessel wall anatomy and disease processes of the intracranial arteries can be resolved. The inner luminal diameter of middle cerebral artery (MCA) is typically 2 to 4 mm, and the wall thickness of a normal, healthy MCA ranges from 0.2 to 0.3 mm, which is smaller than the clinically achievable HR-VWI imaging voxel size (Figure 3) [41]. High magnetic field strength, preferably at 3 T, is recommended to achieve a high signal-to-noise ratio (SNR) with the recommended goal of 0.5-mm isotropic resolution [41]. To accurately track an intracranial artery, which is often tortuous, three-dimensional (3D) volumetric image acquisition is recommended to perform multiplanar reformats and visualize the artery in the orthogonal plane. However, it is worth noting that there are many different protocols and pulse sequence combinations using 3D versus 2D imaging for imaging the intracranial arteries. Suppression of blood and CSF signals is necessary and achievable on T1-weighted sequences. Multiple tissue weightings, with and without contrast agents, are recommended for characterizing intracranial vessel walls and their diseases. A proton density (PD) sequence is superior to a T1-weighted image as it achieves a better SNR to visualize the anatomy of the intracranial vessels; however, contrast enhancement may be less obvious, and signal intensities of the CSF and vessel wall are similar [41]. Postcontrast T1-weighted images can be used to characterize various wall pathologies by visualizing contrast enhancement, which may be interpreted as inflammation from atherosclerotic plaque, fibrosis from chronic pathology, or VV (Table 1).

Figure 3.

A representative image of high-resolution vessel wall imaging of normal intracranial arteries. (A) Time-of-flight image. (B-D) Proton density (PD), T1-weighted, and T1-weighted with gadolinium (T1GD) coronal images showing the left middle cerebral artery (MCA). (E-G) PD, T1-weighted, and T1GD coronal image showing a cross-sectional view of the left MCA.

Table 1.

Imaging characteristics of vascular wall pathologies on different pulse sequences

The PD sequence has a high SNR and is used to measure the outer wall diameter [42,43]. Fibrotic tissue and lipid cores may appear as hyperintense signals on PD-weighted images. In T1-weighted images, the inner lumen and vessel wall are well differentiated; therefore, the luminal area can be identified [41,44]. T1-weighted imaging is also useful in identifying atherosclerosis because fibrotic tissue and the lipid core (without fat suppression) appear distinct with a hyperintense signal, and intraplaque hemorrhage appears hyperintense, especially in the subacute period [13,45]. Contrast-enhanced T1-weighted imaging also shows a well-demarcated inner lumen and vessel wall, with enhancement of the vessel wall due to inflammation, neovascularization, and VV. 41 T2-weighted imaging may reveal an intimal flap in dissection and perivascular edema in vasculitis with a hyperintense signal [41].

HR-VWI of intracranial arteries has several limitations. The first is the “slow-flow artifact,” which can cause vessel walls to appear falsely thick owing to inadequate blood signal suppression along the periphery due to slow laminar flow [46,47]. It is particularly ambiguous in certain vascular segments such as the Sylvian cistern near the MCA, dilated arteries, and petrous and cavernous ICA [46,47]. This artifact can be misinterpreted as arterial wall enhancement, and cross-checking with time-of-flight MR angiography (TOF MRA) and the pattern of the hyperintense signal is warranted [46,47]. Another interpretive pitfall is enhancement of the VV, which may mimic arterial wall thickening or vasculitis, especially in the presence of aging or atherosclerotic risk factors. Furthermore, the extracranial VV extends along the proximal intracranial segments of the ICA and VA immediately after dural penetration; this mild enhancement of the VV can be misinterpreted as vessel wall enhancement [41,46,47]. Finally, the venous plexus adjacent to the arteries may also be falsely interpreted as arterial wall enhancement [41]. For example, the petrous ICA is surrounded by the venous plexus of Rektorzik (e.g., the petrooccipital venous plexus) [48]. Ideally, these problems can be addressed by improving the spatial resolution and optimally suppressing blood and CSF flow, but often at the cost of increased acquisition time [46]. Efforts are underway to improve the resolution while reducing the acquisition time through the development of new sequences (e.g., motion-sensitized-driven equilibrium [MSDE] and delay alternating with nutation for tailored excitation [DANTE]), efficient imaging techniques (efficient k-space sampling trajectories), and hardware improvements [46,49-52].

Histological and pathological correlation of HR-VWI findings

Interpretation of HR-VWI findings should consider the clinical context of the patient’s laboratory and clinical examination findings. To date, there are few correlative radiology–pathology data, given the challenge of performing biopsies of diseased intracranial arteries. In lieu of pathology, serial HR-VWI can also add to our knowledge by revealing the presence of ICAD.

For example, a young ischemic stroke patient with a history of sudden severe headache may be clinically diagnosed with intracranial dissection. HR-VWI may reveal a hyperintense lesion on a precontrast T1-weighted image, which may be interpreted as an intramural hematoma. However, imaging findings are nonspecific, as vulnerable plaques with intraplaque hemorrhage or high lipid content may manifest similarly. Serial HR-VWI can help distinguish between intramural hematomas and atherosclerotic lipid components because subacute to chronic intramural hematomas show diminished T1 hyperintense signals as they become hypointense. In contrast, the high T1-weighted signal intensity of atherosclerotic plaques remain unchanged over time.

Postcontrast HR-VWI provides valuable diagnostic information. Strong vessel wall enhancement suggests vessel wall inflammation and may be a sign of an ongoing pathological process. However, because persistent enhancement can be caused by neovascularization in the fibrotic tissue, the enhancement itself lacks specificity, and the clinical history can add valuable context to these cases.

When interpreting HR-VWI results, it is necessary to check the intracranial arteries in the orthogonal plane to review the eccentricity of the lesion [41]. For example, atherosclerotic plaques are thought to predominantly show eccentric vessel wall thickening, whereas vasculitis tends to show concentric wall thickening [40]. Additionally, it is crucial to remember that normal or reactive findings can mimic pathological findings. For example, poor blood suppression due to slow intraluminal laminar flow can be misinterpreted as enhancement, particularly on postcontrast T1-weighted images because it appears slightly hyperintense on T1-weighted images. This pitfall can be especially problematic in branches distal to the M2 segment and arteries adjacent to the veins. The VV, which is sparse in the intracranial arteries, can develop with aging or disease processes and may appear to be slightly enhanced. This finding can mimic vasculitis or be mistaken for plaques, particularly in the proximal intracranial ICA or VA, where the VV is often present after dural penetration [53]. Finally, after thrombectomy during recanalization treatment using a stent retriever, the resulting endothelial injury may result in strong enhancement in the acute period [54].

Clinical interpretation of HR-VWI findings in ICAD

ICAD: atherosclerosis

When symptomatic ICAD is identified in a patient with atherosclerotic risk factors, such as advanced age, a long history of hypertension, diabetes, dyslipidemia, and smoking, vascular physicians typically first consider atherosclerosis as an etiology. It has been suggested that atherosclerotic ICAD does not share the same histopathological characteristics with extracranial arteries. Histological analysis has shown that intracranial atherosclerosis has more fibrous components and fewer necrotic lipid-rich cores than extracranial atherosclerosis [13,15]. Therefore, it is not easy to identify the typical imaging features of HR-VWI for atherosclerotic ICAD. Instead, strong enhancement on postcontrast T1-weighted images is a feature common to vasculitis and may be the sole feature visualized in ischemic stroke. In the correct clinical setting and history, HR-VWI can increase the diagnostic confidence of vascular physicians of atherosclerotic origin (Figure 4A). Intracranial atherosclerotic plaque characteristics can differ, leading to differences in the mechanisms of stroke. For example, intracranial atherosclerotic plaques can develop and progress by “creeping” along the superior/dorsal MCA vessel walls, leading to occlusion of the lenticulostriate arteries and ischemia. In contrast, intracranial atherosclerotic plaques may progress with vulnerable features, leading to plaque rupture, thrombus development, and embolism. Prior work has described this as branch-occlusive (B)-type versus coronary (C)-type atherosclerosis [55]. With HR-VWI, identifying B- versus C-type atherosclerosis can add precision to diagnosis, as these two different types of plaques may warrant different tailored therapies to reduce the risk of recurrent stroke.

Figure 4.

Representative high-resolution vessel wall imaging (HR-VWI) images of various intracranial arterial diseases. (A) HR-VWI findings of atherosclerosis in a 45-year-old male with prehypertension and hypercholesterolemia. A stenotic portion (arrow) at the left middle cerebral artery (MCA) is identified on the time-of-flight (TOF) image. The proton density (PD) image shows no total vessel wall diameter stenosis. The T1 image shows eccentric vessel wall thickening with isointense signal suggesting a lipid core or fibrous cap. The area appears to be enhanced on the T1-weighted with gadolinium (T1GD) image. (B) HR-VWI in a 35-year-old female with moyamoya disease presenting with right MCA infarction. A heterozygous RNF213 c.14429G>A mutation was found. The TOF image shows stenosis at both distal internal carotid artery (ICA) to proximal A1 and M1. The arrow indicates the most stenotic part of the right MCA. Concentric wall thickening involving the right MCA with enhancement and negative remodeling was noted on PD, T1, and T1GD images. (C) HR-VWI of a left MCA dissection in a 48-year-old male who underwent endovascular treatment due to a left MCA infarction 4 months earlier. TOF-magnetic resonance angiography (MRA) shows a stenotic portion of the left MCA (arrow). The intimal flap is clearly visible on the PD axial image. Angiography at the time of endovascular treatment shows the intimal flap (arrow). PD, T1, and T1GD sagittal images show a double lumen and an intramural hematoma. A lesion with hyperintense PD and hyperintense T1 signal and no enhancement suggests a late subacute stage of the intramural hematoma. (D) HR-VWI in a 47-year-old female with vasculitis. TOF image shows a severely stenotic portion of the left MCA (arrow). PD, T1, and T1GD coronal images show wall thickening and prominent enhancement of the left M1. Sagittal images show concentric wall thickening with negative remodeling with strong enhancement. The stenosis later progressed despite steroid treatment. (E) HR-VWI in a 27-year-old female with reversible cerebral vasoconstriction syndrome presenting with severe headache. TOF-MRA axial image shows multifocal stenoses of both MCAs. Reconstructed TOF-MRA shows multifocal mild to moderate stenoses at the distal ICA, MCA, anterior cerebral artery, and basilar artery. Sagittal images of the right MCA show normal appearance of the vessel wall. Coronal images show stenotic segments without any abnormal vessel wall signal. Intracranial vessels were significantly improved on follow-up HR-VWI.

Clinical significance of follow-up HR-VWI

Serial HR-VWIs in a patient with ICAD can provide valuable information (Table 2). An increase in plaque burden and enhancement ratio on follow-up HR-VWI is associated with an increased risk of recurrent stroke [60]. Culprit lesions tend to maintain their enhancement in stroke patients and serial HR-VWI can provide information on whether the lesion is likely to be the culprit [61]. Statin is known to decrease wall enhancement in intracranial atherosclerosis [62]. Antiplatelets and statins can reduce plaque burden in both non-culprit and culprit lesions [63]. Recently, a study showed a decrease in plaque enhancement after balloon angioplasty or drug-coated balloon angioplasty on follow-up HR-VWI, indicating plaque stabilization [64]. Another study showed that reduced enhancement on early follow-up HR-VWI after balloon angioplasty and stenting could predict fewer recurrent ischemic events [65].

Table 2.

Reports of follow-up HR-VWI for various ICADs

In a study investigating changes in intracranial dissection, intramural hematomas began to disappear within 6 months, and most damaged walls healed within 12 months [66]. This finding provides insights into deciding whether to discontinue antiplatelet therapy, especially in young patients, who are more susceptible to intracranial dissections. In MMD, a study showed that the progression of negative remodeling on follow-up HR-VWI was slower when cilostazol was administered than when other antiplatelet agents were used [56]. Another study showed that rapid progression of arterial stenosis increased the risk of recurrent stroke in patients with MMD [67]. In vasculitis, the clinical course and degree of enhancement are usually well correlated, and changes in follow-up HR-VWI can guide management to continue or taper steroids or other immunotherapies [36-38]. A recent case series revealed temporal changes in antiphospholipid syndrome-associated proliferative vasculopathy using serial HR-VWIs [68]. An atypical case of RCVS without typical thunderclap headache was also diagnosed using serial HR-VWI [69].

Determining the appropriate timing for follow-up HR-VWI is critical and should be etiology-specific. Our unpublished data (Kang DW, 2024) revealed that intracranial dissections exhibit notable changes in vessel wall morphology over a shorter time frame than atherosclerosis. We suggest a follow-up of approximately 3–6 months for dissection and 6–12 months for atherosclerosis. For RCVS, a 12-week follow-up would be reasonable in line with the diagnostic criteria [70]. MMD can be monitored for approximately 24 months to assess progression unless otherwise indicated [56]. For vasculitis, imaging should be performed when critical treatment decisions are pending and when a clinical–radiological correlation is needed. Therefore, follow-up imaging may be important for the diagnosis and treatment of ICAD, assessment of treatment response, and clinical decision-making.