Hemorrhage and Recurrence of Obliterated Brain Arteriovenous Malformations Treated With Stereotactic Radiosurgery

Abstract

Background: Although complete nidal obliteration of brain arteriovenous malformations (AVM) is generally presumed to represent durable cure, postobliteration hemorrhage, and AVM recurrence have become increasingly recognized phenomena. The goal of the study was to define hemorrhage and nidal recurrence risks of obliterated AVMs treated with stereotactic radiosurgery (SRS). Methods: This is a retrospective cohort study from the International Radiosurgery Research Foundation comprising AVM patients treated between 1987 and 2020. Patients with AVM obliteration on digital subtraction angiography (DSA) were included. Outcomes were (1) hemorrhage and (2) AVM recurrence. Follow-up duration began at the time of AVM obliteration and was censored at subsequent hemorrhage, AVM recurrence, additional AVM treatment, or loss to follow-up. Annualized risk and survival analyses were performed. A sensitivity analysis comprising patients with AVM obliteration on magnetic resonance imaging or DSA was also performed for postobliteration hemorrhage. Results: The study cohort comprised 1632 SRS-treated patients with AVM obliteration on DSA. Pediatric patients comprised 15% of the cohort, and 42% of AVMs were previously ruptured. The mean imaging follow-up after AVM obliteration was 22 months. Among 1607 patients with DSA-confirmed AVM obliteration, 16 hemorrhages (1.0%) occurred over 2223 patient-years of follow-up (0.72%/y). Of the 1543 patients with DSA-confirmed AVM obliteration, 5 AVM recurrences (0.32%) occurred over 2071 patient-years of follow-up (0.24%/y). Of the 16 patients with postobliteration hemorrhage, AVM recurrence was identified in 2 (12.5%). In the sensitivity analysis comprising 1939 patients with post-SRS AVM obliteration on magnetic resonance imaging or DSA, 16 hemorrhages (0.83%) occurred over 2560 patient-years of follow-up (0.63%/y). Conclusions: Intracranial hemorrhage and recurrent arteriovenous shunting after complete nidal obliteration are rare in AVM patients treated with SRS, and each phenomenon harbors an annual risk of <1%. Although routine postobliteration DSA cannot be recommended to SRS-treated AVM patients, long-term neuroimaging may be advisable in these patients.