Introduction Arteriovenous malformations (AVMs) are a heterogeneous group of lesions that can rupture and cause severe morbidity or even death. However, there is no consensus with respect to their natural history or proper treatment. When safely feasible, microsurgical resection is the primary treatment, but certain features like size and location can render lesions unresectable. Embolization can be performed to obliterate high-risk features like aneurysms or reduce the lesion nidus to make it more amenable to stereotactic radiosurgery. This study describes the experience at two centers with embolization performed to selectively target aneurysms associated with AVMs.

Materials and methods Prospectively maintained records from two medical centers were retrospectively analyzed to identify patients undergoing targeted embolization for AVM aneurysm between January 2002 and February 2015. Patients were selected for analysis if embolization was only targeted at the aneurysm. Patients undergoing repeat embolization or surgery after targeted embolization were excluded from analysis. For included patients, age, gender, and presenting symptom were recorded. Spetzler-Martin and supplemental AVM scores were recorded, as were the components of each of these scales. Aneurysm location (feeding artery versus nidus), size, treatment type, technical outcome, and any complications were recorded. Follow-up data examined were hemorrhage, seizure, and modified Rankin score at 30 days, 90 days, 1 year, and point of last contact. Chi-square tests were performed to identify any correlation between demographic, treatment, and clinical outcome variables.

Results 31 patients (14 men, 7 women; mean age at treatment 43.8 years, IQR 23.7–43.8) were identified who met inclusion criteria. Mean aneurysm size was 5.6 mm. 16 (51.6%) aneurysms were present in arterial feeders, while 15 (49.4%) were located in the AVM nidus. 18 (81.8%) presented with intracranial hemorrhage_16 (51.6%) with parenchymal hemorrhage and 19 (61.3%) with subarachnoid or intraventricular hemorrhage. The remaining 4 (18.2%) presented with seizures. 24 (77.4%) AVMs had deep drainage, 23 (74.2%) involved eloquent territory, and 11 (35.5%) demonstrated diffuse nidus. 13 (41.9%) were treated with Onyx, 13 (41.9%) with coils, 3 (9.7%) with n-BCA, and 2 with ethanol (6.5%). Technical success occurred in all cases. 1 procedure was complicated by thrombus formation and was successfully treated with abciximab, 1 aneurysm ruptured during coiling but was further coiled without incident, and another procedure was complicated by dissection that was self-limited and caused no untoward effects. 23 (74.2%) lesions were subsequently treated with stereotactic radiosurgery. No lesions hemorrhaged following treatment. 1 patient (4.5%) developed seizures after treatment after presenting with hemorrhage. Mean follow-up time was 759 days (IQR 125–1167). All patients had improved mRS at each point of follow up. Lesions with diffuse nidus were more likely to have complication during embolization (OR 6.8, p = 0.037). No other statistically significant relationships were found.

Conclusion AVMs are complex lesions whose natural history and optimal treatment are incompletely understood. For unresectable lesions with aneurysms, targeted aneurysm embolization appears to be safe and effective. More formal analysis of this approach and direct comparison to other forms of treatment are warranted.

Disclosures M. Alexander: None. D. Cooke: None. D. Hallam: None. S. Hetts: None. L. Kim: None. H. Kim: None. B. Ghodke: None.