Vasa vasorum of proximal cerebral arteries after dural crossing — potential imaging confounder in diagnosing intracranial vasculitis in elderly subjects on black-blood MRI

Vasa vasorum is a term referring to typically microvascular structures or rather vessel networks, consisting of arteries, capillaries, and veins, physiologically occurring in the walls of larger extracranial arteries and veins, supplying them with oxygenated blood and nutrients and removing metabolic substances [1]. Both intra- and extracranial arteries are composed of three layers: the intima, media, and adventitia. Nutrition of the inner portions of the vessel walls is mainly ensured by diffusion from the luminal blood flow. Outer parts of extracranial vessel walls, however, do not benefit from the diffusing nutrients in the bloodstream, as they are out of reach. Therefore, these outer portions require additional supply with nutrients by vasa vasorum [2, 3]. Beyond, vasa vasorum also play a role in vascular pathology by favoring hypertrophy of the intima-media complex, atherosclerotic vessel wall changes, plaque hemorrhage, and dissection via rupture [4].

Vasa vasorum of intracranial arteries, in contrast, are rare, mainly due to the special anatomic construction and surrounding of intracranial arteries [2, 5, 6]. Nourishment of the outer parts of intracranial arteries is probably partly ensured by the nutrient-rich cerebrospinal fluid typically surrounding their adventitial surface [5]. Diffusion of nutrients from the luminal blood flow and the surrounding cerebrospinal fluid through the vessel wall is facilitated in intracranial arteries compared to extracranial arteries of comparable size due to their thinner media and adventitia [6]. Intracranial vasa vasorum usually develop with age, and proliferation is favorably influenced by different vasculopathies, particularly atherosclerosis [2], but also by inflammatory changes such as vasculitis. Intracranial vasa vasorum occur predominantly along the proximal intracranial vessel segments [7], shortly after the vessels’ dural crossing and usually fading in the further distal vessel course. In vessel wall imaging (VWI) studies using black-blood MRI, vasa vasorum typically cause a concentric mural contrast-enhancement and slight vessel wall thickening [8]. This type of vessel wall enhancement (VWE) and slight wall thickening may easily be mistaken for otherwise characteristic findings of inflammation related to cerebral vasculitis [9]. Primary angiitis of the central nervous system typically affects younger patients with no prior history of cardiovascular diseases with a mean age of the initial manifestation at the end of the fourth decade of life [10]. In this patient population, no relevant VWE caused by vasa vasorum is to be expected. Secondary central nervous system vasculitis, however, usually occurs in the framework of any other underlying pathology, often in the presence of other autoimmune diseases such as systemic forms of vasculitis, for example giant cell arteritis. The typical patient population with suspected giant cell arteritis and potential involvement of the intracranial arteries usually includes elderly patients, often with pre-existing vascular conditions. Therefore, concentric VWE of the proximal intracranial arteries related to the presence of vasa vasorum may be considered as a possible and frequent confounder of large artery vasculitis in the VWI assessment of intracranial arteries, potentially causing false positive results [9].

We hypothesized that concentric VWE may be commonly found on MRI VWI studies in elderly subjects. Thus, our aim was to determine the frequency and longitudinal extension of concentric VWE within the proximal cerebral arteries in a non-vasculitic cohort of elderly subjects, which may most likely relate to vasa vasorum. We used high-resolution 3D black-blood VWI with modern compressed-sensing (CS) technology for analysis of VWE in the proximal intradural portions of the internal carotid (ICA) and vertebral arteries (VA) after their dural crossings in order to build a reference base for future comparison with patients suspected of intracranial vasculitis.

The main results are summarized in Table 1. Sole VWE of the ICA was never observed. Patients with VWE of the ICA always showed coexisting VWE of both VAs. In 13 of 39 cases (33%), the V5 segment of the vertebral artery was affected of at least one side. In 26 of 39 cases (66%), the VWE of the VA was limited to the V4 segment. The length of VWE of the proximal intradural artery segments was significantly longer in both VAs compared to the ICAs (see below Fig. 6). In 14 of 39 patients (36%) with proximal intradural VWE, the character of contrast-enhancement was smooth, delicate, and concentric likely related to the presence of vasa vasorum. In the remaining 25 patients (64 %), irregular thickening and mural enhancement of the vessel walls with eccentric plaques was disclosed, so that additional atherosclerotic lesions were assumed. Four of 43 (9 %) patients did not show any VWE at all (Figs. 2 and 3).

In 7 out of 43 patients (16 %), the results of assessment by readers 1 and 2 were significantly discordant due to a long-segmental VWE and a tortuous vessel course in the respective proximal intradural vessel segment. Reliable measurability of the curvy enhancing vessel segment is only possible to a limited extent with conventional methods. Therefore, these cases were re-analyzed after post-processing using a dedicated software algorithm allowing straightening of the respective vessel structures (Fig. 4).

Concentric VWE within the proximal 2 mm of the ICA and within the proximal 13 mm of the VA after dural entry was detected in a substantial number (30 % in the ICA, and 91 % in the VA) of non-vasculitic elderly patients (aged ≥ 50 years) with extracranial malignant disease. This type of VWE is likely related due to vasa vasorum enhancement of the proximal intradural portions of the large cerebral arteries and should not be confused with VWE related to inflammation in intracranial vasculitis [2, 7, 9].

Diagnosing cerebral vasculitis using MRI VWI remains a challenge despite substantial technical progress: on the one hand due to the special anatomic conditions of intracranial arteries, being very thin vessel structures with a mostly tortuous course, rarely fitting standard reconstructions, and on the other hand due to a number of confounders and differential diagnoses with similar imaging findings [2, 9]. One important confounder is the presence of vasa vasorum along intracranial arteries as it can mimic concentric VWE otherwise characteristic for vasculitic affection [2, 9]. Vasa vasorum–related VWE typically occurs in the proximal vessel segments of intracranial arteries, right after their dural crossing [2]. There are no reference values in the literature regarding the length of intradural extension of VWE of the ICA and VA caused by vasa vasorum or recommendations defining the distance from these arteries’ dural entry to the point where a concentric VWE may be considered as rather suggestive of vasculitis. However, a benchmark for the typical length of VWE of the proximal intracranial artery segments caused by vasa vasorum might help in the differentiation from inflammatory changes. In most cases, the VWE due to the vasa vasorum is clearly differentiable from the normal appearance of the intracranial arterial vessel wall that is depicted isointense or mildly hyperintense on T1w black-blood VWI sequences [2, 9]. Sometimes mural enhancement caused by vasa vasorum tapers in the further vessel course resulting in a very smooth transition between enhancing and non-enhancing vessel wall segments so that the exact length of the VWE caused by vasa vasorum is visually hardly definable, which may explain a significant inter-reader variability.

Particularly in the elderly patient population that often suffers from increased vascular comorbidities, there are frequently additional atherosclerotic vessel wall changes that might cause difficulties in the interpretation [2, 9]. Moreover, it can be assumed that the presence of vasa vasorum is associated with the progression of atherosclerosis [7, 21]. While vasculitic affection and VWE caused by the presence of vasa vasorum usually manifest as smooth concentric lesions, atherosclerotic vessel wall lesions typically show an eccentric plaque-like configuration [9]. Therefore, it is essential to assess any thickened and enhancing vessel segments in all three spatial dimensions in order to differentiate the underlying pathology [9]. However, in practice and especially in elderly patients, VWE and wall irregularities are often due to a combination of both, the presence of vasa vasorum and additional atherosclerotic vessel wall changes. In our cohort, VWE in 25 of 39 patients (64 %) showed an irregular, asymmetric, or eccentric configuration, supposedly related to a combination of vasa vasorum and atherosclerotic lesions.

Using black-blood VWI MRI, assessment of intracranial vessel wall structures with a mostly tortuous vessel course requires special technical features. Measurements were implemented on a dedicated sagittal 3D T1-weighted SPACE-sequence, optimized for intracranial VWI with ultra-high resolution due to the compressed-sensing technique as described previously [11]. In some cases, multiplanar reconstruction of high-resolution 3D VWI data is not sufficient to allow reliable measurement of the respective enhancing vessel segment. We used a dedicated post-processing tool enabling reconstruction and visualization of the vessel in stretched and straightened form, thus facilitating a precise measurement of the vessel segment in question.

Our study was not without limitations: First, the retrospective nature and the relatively small sample size makes it prone to potential selection bias. Thus, further studies with larger non-vasculitic cohorts are needed to support the results of this study. However, the analyzed patient cohort with an age ≥ 50 years represents an age-matched control group to the patients typically affected by giant cell arteritis with potential involvement of the cerebral arteries and with comparable atherosclerotic disease burden and associated vessel wall changes. The analyzed patients all had a history of extracranial malignancy not known to be associated with pathologic vessel wall changes (Table 2). Second, the length of vessel wall enhancement is difficult to be precisely measured in some cases due to the sometimes very smooth transition between enhancing and non-enhancing vessel wall segments and the oftentimes tortuous vessel course. Application of the dedicated post-processing tool facilitated length measurement of the vessel segment in question. Furthermore, meaningfulness of the study is limited due to the lack of histological proof which may be hard to obtain in patients with extracranial malignant disease but of otherwise good neurological health status. However, all studied patients neither revealed clinical signs or suspicion of vasculitis in history nor had laboratory signs of vasculitis at the time of the MRI scan, so that the vessel wall changes can be plausibly attributed to the presence of vasa vasorum and atherosclerotic changes

Previous histopathologic studies, however, have revealed results that make the imaging results of our study plausible: One histopathologic study revealed that vasa vasorum occur significantly more often in the proximal intracranial vessel segments than in the distal vessel segments, particularly in the VA and the ICA [7]. Another histopathologic study examined 30 human autopsy materials and found vasa vasorum in the proximal 1–2 cm of all VA and ICA [22], comparable to our imaging results. Another histopathologic study demonstrated that vasa vasorum were found in nearly half of all examined arteries with a significant higher prevalence in the VA compared to the basilar artery or the middle cerebral artery as well as a correlation between the existence and amount of vasa vasorum and the atherosclerotic disease burden in the VA [21].

Concentric VWE within the proximal 2 mm of the intradural ICA and within the proximal 13 mm of the intradural VA portions was found in non-vasculitic elderly patients (aged ≥ 50 years). This finding presumably represents vasa vasorum–related VWE without pathological significance and should not be misinterpreted as intracranial vasculitis. Distal to these segments, VWE is more likely related to pathologic conditions such as vasculitis.