Long-Term Follow-Up of Adaptive Multistage Stereotactic Radiosurgery for Treatment of High-Grade Arteriovenous Malformations

Abstract

The optimal management of high-grade arteriovenous malformations (AVM) remains unclear. There is a paucity of literature addressing the efficacy and safety of radiosurgery for high risk patients. Our institution has adopted a multi-staged stereotactic radiosurgery (SRS) approach with treatment of the entire nidus using a lower dose per stage to achieve AVM obliteration with decreased complications. Retrospective review of patients with high grade Spetzler-Martin (S-M) III-V AVMs with at least two multi-stage SRS treatments between 1989 and 2013. Clinical outcomes of obliteration rate, adverse events and treatment data were collected. Patients with less than 12 months of follow-up after second multi-stage SRS treatment were excluded. Forty-two patients met inclusion criteria (n = 26, S-M III; n = 13, S-M IV; n = 3, S-M V). Median follow-up was 113.6 months after first SRS. Median number of staged SRS treatments was 2 and median interval between SRS treatment stages was 41.9 months. Median dose was 1400 cGy per stage and cumulative median dose at last follow-up or obliteration was 3350 cGy. 22 patients had prior embolization and 2 patients had subtotal resection before SRS. Complete AVM obliteration rate was 38.1%. AVMs that underwent pre-SRS embolization had a significantly lower rate of obliteration (23% vs 55%, p = 0.03). Median volume reduction for non-obliterated AVMs was 69.4%. A lower complete AVM obliteration rate was significantly associated with a higher S-M grade (p = 0.03). Additionally, AVM volume less than 10 cm3 (p = 0.22), first-stage SRS dose greater than 1400 cGy (p = 0.43) and cumulative SRS dose greater than 3200 cGy (p = 0.57) did not significantly influence the rate of obliteration. In total, 21 post-SRS adverse events occurred in 18 patients. Adverse events were defined as persistent headache (n = 3), new/persistent seizure activity (n = 8), symptomatic radiation necrosis (n = 1), post-SRS intracranial hemorrhage +/- neurologic deficits (n = 8, 4 with residual deficits), and neurologic deficits without antecedent intracranial bleed (n = 1). The post-SRS adverse event rate was 5.1% per year (21 events over 409.9 cumulative patient yrs) after initial SRS. A retrospective review of our experience treating high grade AVMs with adaptive multi-stage SRS to the entire nidus demonstrates that this may be an appropriate option as it results in AVM obliteration in a meaningful proportion of patients with an acceptable adverse event rate. Lower AVM obliteration rates were significantly associated with higher S-M grade and pre-SRS embolization. We continue to investigate whether this strategy may be improved with treatment at shorter intervals or additional treatment sessions over time. In spite of an acceptable toxicity profile, treatment of high grade AVMs should be considered with caution since partial obliteration does not protect from bleeding.