Dosimetric comparison of photon and proton radiosurgery for large arteriovenous malformations

Radiosurgery for large arteriovenous malformations (AVMs) is limited by toxicity from dose spillage into normal brain, worsened by larger gradient indices in bigger targets. Proton radiosurgery theoretically reduces integral dose but has greater uncertainty due to robust planning constraints and fewer beam angles. This study evaluates whether pencil beam scanning radiosurgery (PBSR) improves normal brain dosimetry while maintaining comparable target coverage to photon radiosurgery.

The 10 largest AVMs treated with photon radiosurgery at a single institution were re-planned for PBSR in Varian Eclipse. Proton plans incorporated a 2% computed tomography (CT) density correction and 2 mm range uncertainty, with 2-4 beams per institutional practice. Plans prioritized isotoxicity to organs at risk (OARs) before matching target coverage. Dosimetric metrics were compared using a paired t-test.

Median AVM size was 36.6 cc (24.3-60.1 cc). All plans prescribed 1750 cGy in one fraction. PBSR had higher gradient indices and reduced target coverage near OARs compared to photon radiosurgery. While PBSR lowered normal brain mean dose and V _5 Gy, it increased V _12 Gy, a known predictor of toxicity.

PBSR, as planned with current institutional constraints, leads to greater high-dose spill into normal brain and inferior target coverage near OARs. Future improvements in PBSR, such as dual-energy CT, additional beam angles, and LET-based planning, may help mitigate these limitations.