2882 - Spatially Fractionated Radiosurgery for Large AVMs: An Early Experience of Dosimetry and Patient Outcomes

Purpose/Objective(s): SRS achieves obliteration rates of 70-80% for arteriovenous malformations (AVMs), with the highest success being in patients with smaller AVMs (<4 cc), which is around 80-85%. The success rate is typically less in patients with larger or higher-grade AVMs. Grade III lesions have obliteration rates around 76%, while grade IV-V lesions have obliteration rates in the 62-68% range. The purpose of this study was to explore spatial fractionation as a technique to treat larger AVMs.

Materials/Methods: In this study, we looked at the outcomes of 11 patients who received spatially fractionated SRS for their AVMs between 2013-2026. Patients were chosen for spatially fractionated SRS if their AVM was >15 cc near an eloquent site, >20 cc in a non-eloquent site, or pediatric patients whose AVM occupied a large proportion of brain parenchyma, though not >15 cc. Initially, the entire AVM was contoured to determine overall volume. AVMs were then divided into equally sized sections. If the AVM was smaller than 30 cc, it was divided axially into two sections. AVMs that were larger than 30 cc were subdivided into three sections, focusing on the eloquent area first. This was done by contouring the eloquent structure abutting the AVM, expanding into the AVM to create a volume that was ~10 cc. The remainder of the AVM was equally divided. A 3 mm optimization structure was created from each subsection to avoid hotspots where the structures met. Each subsection was treated to 1750 cGy/ 1 fx with 4-6 weeks between each treatment. Patients were followed until obliteration while tracking serious adverse events.

Results: The age range of patients who were treated with spatially fractionated SRS was 7 to 68 years of age. There were 7 females and 4 males who were treated between 2013 to 2026. One patient had a Spetzler Martin Grade 2 AVM, 3 patients had a Spetzler Martin Grade 3 AVM, 6 patients had a Spetzler Martin Grade 4 AVM, and 1 patient had a large dural AVF, which is not graded on the Spetzler Martin system. The size range for the treated AVMs was 6.31-59 cc, with the median size being 23.62 cc. The 6.31 cc AVM was in a pediatric patient. Nine patients received 2 fractions of spatially fractionated SRS, while two patients received 3 fractions of spatially fractionated SRS. One patient still needed surgical embolization for AVM obliteration, one patient reached total obliteration that was confirmed by angiography, one patient experienced a serious adverse event in the form of hemorrhage, and eight patients have not yet reached the 3-year confirmatory angiography, though there is MRI evidence of improvement.

Conclusion: We have established a novel methodology for spatially fractionated SRS for large AVMs that avoids hot spots and results in similar outcomes to that of single fraction SRS for smaller AVMs. Patients are followed for 3-years prior to undergoing confirmatory angiography with neurosurgery. The majority of the patients in our cohort have not reached 3-year follow-up yet, but interval MRI scans suggest decrease in size of the AVMs.