Stereotactic radiosurgery for cerebellar arteriovenous malformations: an international multicenter study.

OBJECTIVE Cerebellar arteriovenous malformations (AVMs) represent the majority of infratentorial AVMs and frequently have a hemorrhagic presentation. In this multicenter study, the authors review outcomes of cerebellar AVMs after stereotactic radiosurgery (SRS). METHODS Eight medical centers contributed data from 162 patients with cerebellar AVMs managed with SRS. Of these patients, 65% presented with hemorrhage. The median maximal nidus diameter was 2 cm. Favorable outcome was defined as AVM obliteration and no posttreatment hemorrhage or permanent radiation-induced complications (RICs). Patients were followed clinically and radiographically, with a median follow-up of 60 months (range 7-325 months). RESULTS The overall actuarial rates of obliteration at 3, 5, 7, and 10 years were 38.3%, 74.2%, 81.4%, and 86.1%, respectively, after single-session SRS. Obliteration and a favorable outcome were more likely to be achieved in patients treated with a margin dose greater than 18 Gy (p < 0.001 for both), demonstrating significantly better rates (83.3% and 79%, respectively). The rate of latency preobliteration hemorrhage was 0.85%/year. Symptomatic post-SRS RICs developed in 4.5% of patients (n = 7). Predictors of a favorable outcome were a smaller nidus (p = 0.0001), no pre-SRS embolization (p = 0.003), no prior hemorrhage (p = 0.0001), a higher margin dose (p = 0.0001), and a higher maximal dose (p = 0.009). The Spetzler-Martin grade was not found to be predictive of outcome. The Virginia Radiosurgery AVM Scale score (p = 0.0001) and the Radiosurgery-Based AVM Scale score (p = 0.0001) were predictive of a favorable outcome. CONCLUSIONS SRS results in successful obliteration and a favorable outcome in the majority of patients with cerebellar AVMs. Most patients will require a nidus dose of higher than 18 Gy to achieve these goals. Radiosurgical and not microsurgical scales were predictive of clinical outcome after SRS.