Vascular wall imaging in reversible cerebral vasoconstriction syndrome – a 3-T contrast-enhanced MRI study

Reversible cerebral vasoconstriction syndrome (RCVS) is a unifying term for a variety of clinical-radiological syndromes characterized by recurrent thunderclap headaches and reversible multifocal cerebral vasoconstrictions [1‐4]. RCVS is not uncommon and potentially devastating because it is associated with a high risk of complications, such as posterior reversible encephalopathy syndrome, ischemic stroke, intracerebral hemorrhage and cortical subarachnoid hemorrhage (SAH) [3, 5‐10]. The diagnosis is based primarily on angiography demonstrating cerebral vasoconstrictions and their reversibility, but its differentiation from central nervous system (CNS) vasculitis can be challenging [11, 12].

Conventional arterial imaging, such as computed tomography or magnetic resonance angiography (MRA), can be used to evaluate vascular stenosis in RCVS. However, the specificity of such imaging is limited by similar luminal defects being the result of other pathologies [13]. The small caliber and tortuosity of intracranial vessels hamper visualization of vascular walls by conventional imaging techniques [13]. After being used initially to characterize the luminal stenosis in carotid atherosclerotic disease [14], black-blood imaging techniques have been applied to intracranial vascular wall visualization and characterization of vascular wall pathologies, including intracranial atherosclerosis [15], vasculitis [16], arterial dissection [17], aneurysm [18] and RCVS [19, 20].

It is not known whether there are pathological vascular wall changes underlying RCVS vasoconstrictions. Generally, the limited histopathological data available do not support the presence of arterial wall inflammation in patients with RCVS [12, 21, 22]. However, in one case report, marked vascular wall enhancement was noted in a patient with cocaine vasculitis [23], and cocaine use has been considered to be an important etiology of RCVS [9, 12, 24]. In a recent case series, 3 patients with RCVS showed no apparent vascular wall enhancement on contrasted T1 fluid-attenuated inversion recovery (FLAIR) imaging, whereas marked vascular wall enhancement was found in 3 patients with CNS vasculitis and 1 patient with cocaine vasculopathy [19]. In another case series, 4 of 13 patients with RCVS had mild enhancement on T1-weighted sequences with fat suppression and a saturation band [20]; the remaining 9 patients had no enhancement. No congruous conclusions can be drawn from these studies. Therefore, we aimed to determine whether absence of arterial wall pathology on imaging is a universal finding in patients with RCVS or could be characteristic of a subgroup of RCVS patients, as well as to further refine these clinical-pathological syndromes into more specific disease entities.

In the present study, almost half of the RCVS patients showed some degree of enhancement in contrasted T1-FLAIR vascular imaging. In three fourths of the cases, the vascular wall enhancement was mild, and in the remaining fourth the enhancement was marked. The intensity of enhancement was not associated with MCA or ACA flow velocity. The enhancement of the vascular walls persisted at follow-up in a third of these patients, with a median follow-up duration of 7 months. The proportion of RCVS patients with vascular wall enhancement observed in this cohort was higher than that reported in previous studies [19, 20] and was persistent in some cases, suggesting that vascular wall enhancement may not be a reliable imaging sign as previously thought for clinical differentiation of RCVS from vasculopathy with an inflammatory component.

Differentiation from CNS vasculitis precludes unnecessary invasive brain biopsy, cerebral angiography, and lifelong immunosuppression in RCVS patients [12]. Previous studies with small numbers of patients found that arterial wall enhancement was mild (if present) in a minority of RCVS patients [20], as opposed to the strong wall enhancement frequently observed in vasculitis patients. In our present study, most of the vascular wall enhancement was also mild in the RCVS patients. However, the proportion of patients in which vascular wall enhancement was found was much higher than previously reported (47% vs. 31%) [20], and a fourth of the patients had strong vascular wall enhancement. Notwithstanding, the clinical hallmarks of recurrent thunderclap headaches and the reversibility of vasoconstriction without immunosuppressants in our patients support their being diagnosed with RCVS over CNS vasculitis.

Arterial wall enhancement in contrasted vascular imaging may reveal an inflammatory component of RCVS pathology. Although inflammation is not considered to play a key role in the pathogenesis of RCVS, prolonged vasoconstriction per se has been proposed to be associated with an inflammatory process [23, 27]. Of note, an inflammatory cascade has been reported in cerebral vasospasm in SAH [28]. Although the pathologies of RCVS and SAH would be expected to differ from each other, they might share some pathomechanisms. For example, oxidative stress and endothelial dysfunction, which contribute to the vascular wall inflammation, have also been noted in patients with RCVS [29, 30]. Additionally, a postulated mechanism of cocaine-induced vasculitis includes cerebrovascular smooth muscle cells apoptosis and promotion of leukocyte migration across cerebral vascular walls [23, 31, 32], producing vessel wall inflammation. Because cocaine-induced vasculitis is considered a spectral disorder of RCVS [9, 12, 24], it is reasonable to deduce that vascular wall inflammation exists in at least some patients with secondary RCVS. Prolonged vasoconstriction has been hypothesized to contribute to the development of secondary angiitis [27]; similar mechanisms might also contribute to prolonged vascular wall enhancement. However, these were purely speculative; the nature of the persistent/residual vascular wall enhancement remains to be elucidated.

The enhancement of diseased vessels in RCVS was suggested to be reversible in a study completed in the USA [20]. In that study, 8 out of 9 RCVS patients showed complete resolution of their initial vascular wall imaging findings, with only 1 having minimal residual wall thickening after a median follow-up period of 3.5 months [20]. In contrast, one third of our patients with follow-up vascular imaging had persistent mild enhancement after a median period of 3 months (longest period, 21 months). Even among the 10 patients with some level of reduced enhancement on follow-up imaging (follow-up range, 55–95 months), four had residual enhancement. Slower resolution or greater persistence of the enhancement has been observed in atherosclerosis of the intracranial vessels [33], particularly if the enhancement is eccentric and heterogeneous with mild to moderate intensity [13]. Although atherosclerosis risk factors were not commonly present in our patients with persistent or residual enhancement (one patient had hypertension, and one patient had hypertension and diabetes), we could not completely exclude the possibility of subclinical atherosclerosis in our patients. Particularly, it has been found that subclinical atherosclerosis can be present in as high as 50% of patients with low cardiovascular risk [34] and about 60% of asymptomatic patients [34, 35], and that intracranial atherosclerosis is more prevalent in Asians [36]. Compared with the study by Mossa-Basha et al. [37], we focused more on the reversibility of vascular wall enhancement in RCVS, finding that the vascular wall enhancement was not always reversible in patients with RCVS, probably due to etiological heterogeneity. Hence, although vascular wall imaging is a powerful and reliable tool for evaluating diseases involving intracranial vessels, the use of it as an ancillary diagnostic tool for RCVS required deliberation. The persistence of vascular wall enhancement in RCVS did not depreciate the value of vessel wall imaging for differentiation of nonocclusive intracranial vasculopathies, but instead reminded the clinicians not making the diagnosis solely based on reversibility of the enhancement.

The headache characteristics of RCVS [38] are distinct from the primary headaches such as migraine [39] or cluster headache [40]. Although 20% of the patients with RCVS have pre-existing migraine [38], the cardiovascular or neurological comorbidities known to be associated with migraine [41‐43] have not been well explored in patients with RCVS. Because both migraine [44] and RCVS [45] are associated increased risks of white matter hyperintensities, there could be some shared mechanisms between these two disorders. A higher headache frequency and long-term migraine may worsen the cardio-metabolic profile in migraineurs [44], which might partially be mediated by circulating microRNAs associated with vascular function [46, 47]. Whether similar mechanisms could contribute to RCVS pathogenesis or the imaging findings disclosed in this study deserve further investigation.

The present study had several limitations. First, we did not include a control group because doing so would involve unnecessary exposure of subjects to the potential risks of gadolinium deposition in the brain [48]. Second, because we did not recruit patients with secondary causes of RCVS, one should be cautious to extrapolate the findings to the general pathogenesis of RCVS. Given that secondary causes of RCVS are far less common than idiopathic ones in Asian patients [38, 49], elucidating the pathogenesis of the latter was our major concern. Third, although follow-up MRA evaluations for confirming vasoconstriction reversibility were obtained for all of our patients, the retention rate for vascular imaging was 70.8%, mainly due to the undesirable requirement of contrast injection. Fourth, the magnitudes of vascular wall enhancement observed in the present study do not correspond precisely with severity levels defined in the previous reports. However, vascular wall enhancement level differences across studies may reflect the particular machines, settings, and protocols used. In our study, we focused more on the presence of vascular wall enhancement, and the temporal change of the enhancement, both may have little to do with the degrees of the initial vascular wall enhancement. Fifth, these patients received the same treatment (nimodipine) but the resolution of enhancement was heterogeneous, so we cannot be sure if that enhancement of vascular wall imaging is altered by medical treatment. Sixth, the reluctance of patients to undergo spinal tap in Taiwanese society precluded CSF studies in many patients; therefore, the differential diagnosis of RCVS in our practice heavily relied on the presence of the clinical hallmark of RCVS (i.e. recurrent thunderclap headaches) and imaging findings (to demonstrate the reversibility of vasoconstrictions and to exclude SAH or other secondary causes of thunderclap headaches by susceptibility weighted imaging or other MR sequences [38, 45]. Nevertheless, the characteristics of the patients who received spinal tap were not different from those who did not receive spinal tap. Finally, our study could not confirm how long the persistent or residual enhancement could last. Studies with a longer follow-up period are needed.

Demographics, clinical profiles, and cerebral artery flow velocities were similar across patients with versus without vascular wall enhancement. Half of the RCVS patients had enhancement of diseased vessels and it was persistent for one third of them, so vascular wall enhancement may not be a reliable imaging marker for differentiating RCVS from central nervous system vasculitis or subclinical atherosclerosis. The clinical implication of our findings is that the differentiation of RCVS from other intracranial vasculopathy should not be made solely based on vascular wall imaging.