Efficacy of Dose-Escalated Hypofractionated Radiosurgery for Arteriovenous Malformations

There is considerable controversy about the management of arteriovenous malformations (AVMs) that are high risk for surgical resection. Stereotactic radiosurgery (SRS) has a reported success rate of less than 50% with unacceptably high rates of radiation necrosis with larger AVM volumes. Neither volume staging nor hypo-fractionated SRS have conclusively been demonstrated to improve results. We hypothesized that the failure of previous hypo-fractionation SRS trials was due to an insufficient biologically effective dose (BED) of radiation. We initiated a pilot study of treating AVM patients with a total dose divided into three or five fractions designed to deliver the equivalent BED of 20 Gy in a single fraction (α/β =3). We performed a retrospective analysis of 37 AVM patients who had a minimum of two years of follow-up or underwent obliteration. Patients were treated with 30 Gy/3 fractions, 33 Gy/3 fractions, or 40 Gy/5 fractions using a CyberKnife device (Accuracy Incorporated, Madison, Wisconsin, United States). The primary endpoint was complete AVM obliteration, determined by MRA imaging. Most obliterations were confirmed with diagnostic cerebral angiography. Secondary endpoints were post-radiosurgery hemorrhage and radiation-related necrosis. Kaplan-Meier analysis was used to determine obliteration rates. From 2013 to 2021, 37 patients fitting inclusion criteria were identified (62% male, average age at treatment = 48.88 years). Fifteen (41%) patients had prior treatment (surgery, radiosurgery, embolization) for their AVM, 32 (86%) had AVMs in eloquent locations, 17 (46%) had high-risk features, and 14 (38%) experienced AVM rupture prior to treatment. The average modified radiosurgery-based AVM score (mRBAS) was 1.81 (standard deviation (SD)= 0.52), and the mean AVM volume was 6.77 ccs (SD = 6.09). Complete AVM obliteration was achieved in 100% of patients after an average of 26.13 (SD = 14.62) months. The Kaplan-Meier analysis showed AVM obliteration rates at one, two, and three years to be 16.2%, 46.9%, and 81.1%, respectively. Post-operative AVM rupture or hemorrhage occurred in one (2.7%) patient, after nine months. Radiation necrosis occurred in four (11%) patients after an average period of 17.3 (SD =14.7) months. The SRS dose used in this study is the highest BED of any AVM hypofractionation trial in the published literature. This study suggests that dose-escalated hypofractionated radiosurgery can be a successful strategy for AVMs with acceptable long-term complication rates. Further investigation of this treatment regimen should be performed to assess its efficacy.