Vasculitis affecting the central nervous system (CNS) occurs most commonly secondary to systemic vasculitis or other systemic inflammatory disorders. When confined to the CNS, it is referred to as primary angiitis of the CNS (PACNS). PACNS was initially described as a progressive, fatal disorder, characterised histologically by granulomatous necrotising vasculitis.1 Over time, it became apparent to workers in the field that PACNS was a heterogeneous entity, with different subsets based on clinical, laboratory, angiographic and pathological findings.2 Granulomatous angiitis of the CNS (GACNS) is considered the most severe subset of PACNS; it accounts for approximately 20% of all patients with PACNS.3 4 GACNS has also been described in association with cerebral amyloid angiopathy (CAA),5^–7 referred to as amyloid-β related angiitis (ABRA) or CAA-related inflammation. A less severe subset of PACNS was also delineated and termed benign angiopathy of the central nervous system (BACNS).8 Because BACNS closely resembles other cerebral vasospastic disorders, these disorders have recently been grouped under the term reversible cerebral vasoconstriction syndromes (RCVS), and are no longer considered as a form of PACNS.9 An under-recognised subset of PACNS is comprised of those patients that present with a solitary tumour-like mass lesion (ML). The aim of this work was to document the occurrence of PACNS presenting with ML and assess the utility of diagnostic testing and efficacy of treatment.

PATIENTS AND METHODS

Cases were identified from a database maintained at the Cleveland Clinic (CC), from a systematic medical record review of Massachusetts General Hospital (MGH) patients admitted with a diagnosis of CNS vasculitis between 1993 and 2003, and from the English language medical literature (up to April 2007) using the following keywords: cerebral vasculitis/central nervous system vasculitis/cerebral angiitis/central nervous system angiitis. The reference lists of articles obtained from the primary search strategy were checked for additional reports. Patients were included only if they presented with a solitary cerebral ML with histological proof of vasculitis. Relevant exclusions included evidence, upon initial assessment or follow-up, of local or systemic infection, malignancy, systemic vasculitis (eg, Wegener’s granulomatosis, Churg–Strauss syndrome), other multisystem autoimmune disease (eg, sarcoidosis, systemic lupus erythematosus) or other vascular disorders. Cases identified from the literature were excluded if they had been previously reported, or if there was insufficient detail to substantiate the diagnosis of PACNS. Variables collated included demographic variables, clinical features, results of relevant investigations, histological findings, therapeutic interventions, and clinical outcomes. The data were analysed with JMP statistical software.

RESULTS

Seven of 182 (3.8%) patients with CC and one of 20 (5%) patients with MGH with definite PACNS and 30 of 535 (5.6%) patients with PACNS identified from the medical literature had a ML presentation of PACNS (ML-PACNS).6 7 10^–32 One of the CC cases was previously reported.33 A salutary case is described in fig 1.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1 A 43-year-old previously well Caucasian female presented with a 3-day history of reduced co-ordination and visuospatial disorientation, with right inferior quadrantanopia on clinical examination. Magnetic resonance imaging (A) revealed a large contrast-enhancing left parietal mass with oedema and local mass effect. Because of the assumption that this represented a primary central nervous system malignancy, craniotomy and total excision of the mass was undertaken. Pathological examination of the mass revealed evidence of necrotising lymphocytic vasculitis, with a mixture of B and T lymphocytes and histiocytes (D); there was no evidence of lymphocytic atypia. No granulomatous changes were seen (haematoxylin and eosin, ×20, courtesy of Dr Susan Staugaitis, Cleveland Clinic). Microbiological studies performed on this tissue were negative. She received high-dose glucocorticoid therapy and mycophenolate mofetil. Glucocorticoids were tapered and discontinued over an 18-month period. She remains on mycophenolate mofetil 2.5 g/day. No recurrence has occurred over the 21 months since her initial presentation. Panels B and C depict magnetic resonance imaging findings after intervals of 5 months and 10 months respectively, showing only postoperative changes without evidence of recurrence.

Median age was 45.5 years (range 6–74). Nineteen patients were men, 19 women. Of 45 reported cases of amyloid-β related angiitis (ABRA), 13 (29%) presented with ML;6 7 13 14 20^–23 28 30 of these eight were men and five were women. The patients with ABRA were older with a median age of 66 years (range 49–74), as compared with a median age of 38 years (range 6–69) among the non-amyloid group.

The most common presenting features of patients with ML-PACNS were headache (74%), focal neurological deficit (64%), diffuse neurological deficit (50%), seizures (47%), nausea and vomiting (21%) and constitutional symptoms (12%). In contrast to patients with classic GACNS (ie, generalised cerebral involvement), who typically present on average 6 months after the onset of symptoms, 28 of 38 (74%) of ML patients presented within 1 month of onset of symptoms; only one had a symptom duration of 6 months or more. There were no major differences in frequency of presenting symptoms between those with and without ABRA.

ML were identified by magnetic resonance imaging in 24 of 24 patients and by computed tomography in 29 of 30 cases in which these investigations were performed. Oedema, contrast enhancement and haemorrhage were associated with the ML in 17, 15 and five patients, respectively. Cerebrospinal fluid was abnormal in 14 of the 21 patients in which a lumbar puncture was performed (67%), generally elevation of protein ± cell count. Cerebral angiography was abnormal in eight of 14 patients; seven of eight showed a mass effect but no suggestion of vasculitis, in the remaining patient, angiographic evidence of “small vessel vasculitis” was reported. Acute phase reactants were abnormal in nine of 38 (24%) of patients. The diagnosis of PACNS was confirmed by histology in all cases, with granulomatous angiitis in 20 of 38 (53%) and lymphocytic vasculitis in 18 of 38 (47%). Positive staining for amyloid was found in 13 of 38 (34%); granulomatous angiitis was present in 10 of 13 (77%) of patients with ABRA, as compared with 10 of 25 (40%) of the non-amyloid-related group.

The outcomes of treatment are listed in table 1. The figures provided relate to the outcome of the treatment of the initial presentation only, and does not relate to delayed treatment with cytotoxic therapy or surgical excision or to the treatment of disease recurrence. Overall, 26 of 38 (68%) of patients entered remission, as defined by regression of the lesion and resolution/stabilisation of neurological symptoms. Six patients (16%) died during the period of follow-up; of these, there were five patients with ABRA. Of the patients that survived, 22 of 29 (76%) had residual deficits, most commonly visual field defects, seizures or cognitive deficits. Moderate or severe residual deficits were more common in those treated with corticosteroids (CS) alone (seven of 11; 64%) as compared with those treated with CS in combination with cyclophosphamide (one of 10; 10%). Included in the non-amyloid group were three patients age 18 years or younger; all three of these entered remission with high-dose CS alone, two had no residual deficits or recurrence at follow-up, while one had a seizure disorder but no disease recurrence.

DISCUSSION

Although rare, PACNS should be considered in the differential diagnosis of cerebral ML. The clinical features of ML-PACNS are non-specific; only a minority of patients have constitutional symptoms and/or laboratory evidence of an increased acute phase response. Furthermore, there are no features on neuroimaging, cerebral angiography or spinal fluid examination that can reliably distinguish ML-PACNS from other (more common) causes of ML. Therefore, biopsy is essential for ante-mortem diagnosis. Comprehensive testing of biopsy material is required to exclude alternative diagnoses even if features of vasculitis are demonstrated on routine histological examination, including staining and culture to rule out mycobacterial, fungal or other infections, and immunohistochemistry ± gene rearrangement studies to exclude lymphoproliferative disease.

No clinical trials have been performed in patients with CNS vasculitis, nor is it possible to draw firm conclusions from the current study because of the non-standardised nature of the treatment protocols used in these cases. Therefore, it is not possible to make clear recommendations regarding the treatment of patients with ML-PACNS. However, for the patients with non-amyloid-related ML, in particular, aggressive immunosuppressive therapy appears to be associated with more favourable outcomes than CS monotherapy, and may obviate the need for excision of the ML. Our approach has been to treat patients with ML in a similar fashion to those with GACNS, in accordance with the treatment paradigm established for systemic vasculitides such as Wegener’s granulomatosis.34 It is uncertain whether immunosuppressive therapy other than CS is required in patients that undergo resection of the ML, as the majority do not develop recurrence if the lesion has been entirely resected. This decision should be made on an individual basis depending on the size and location of the lesion and the general health status of the patient. As for all patients with PACNS, failure to respond to aggressive immunosuppressive therapy mandates re-evaluation of the diagnosis of PACNS, potentially including repeat biopsy of the ML, to exclude alternative diagnoses such as malignancy or infection.

Deposition of amyloid protein in cerebral blood vessels (cerebral amyloid angiopathy, CAA) has long been recognised as a cause of cerebral haemorrhage; this may or may not be associated with progression to Alzheimer’s disease. It has increasingly been recognised in recent years that CAA may be associated with a spectrum of severity of angiocentric inflammatory reaction, from little or no inflammation, to perivascular infiltrates, through to severe granulomatous angiitis. An interesting and unanticipated finding of this study was the considerably higher incidence of ML among patients with ABRA, as compared with those patients with PACNS without amyloid deposition. This apparent increased frequency of ML in the amyloid-related angiitis subset of PACNS may relate in part to publication bias, and to the greater likelihood of a brain biopsy being performed in an elderly patient with an ML as compared with those presenting with other manifestations of CAA. Nevertheless, this finding emphasises the importance of staining for amyloid in patients presenting with ML-PACNS. Patients with ML-PACNS in the setting of ABRA were older, and more likely to have other co-morbid illnesses; both of these factors are likely contributors to the poorer outcomes observed in patients with ABRA. While the features of angiitis in these patients may resolve with immunosuppressive treatment,21 28 the CAA generally remains unchanged; this also helps explain the poorer outcomes in this group. Immunosuppressive therapy should be judiciously employed in patients with ABRA, based on the estimated contribution of angiitis (as opposed to CAA alone) to the clinical features, any co-morbid conditions, and the potential for neurological recovery.

We describe ML as a rare subset of PACNS, itself a rare condition. The diagnosis of any cerebral mass as PACNS requires careful exclusion of other, more common conditions, including infectious, neoplastic and other inflammatory disorders. Nevertheless, greater awareness of this potential manifestation of PACNS may facilitate more prompt diagnosis and treatment, and thereby reduce the associated morbidity and mortality.