3141 - Size-Stratified Dosimetric Impact of Gamma Knife Lightning Inverse Planning for Brain Metastases in Real-World Practice

Purpose/Objective(s): Gamma Knife radiosurgery planning has undergone a paradigm shift with the adoption of Lightning inverse planning. While this technology automates shot placement and optimization, real-world plan quality reflects interactions between algorithmic output and provider-driven selection. We compared dosimetric parameters of Gamma Knife plans for brain metastases before and after Lightning integration, stratified by lesion size.

Materials/Methods: Retrospective single-institution study of consecutive intact brain metastases from lung or breast primaries treated with Gamma Knife from 2020–2024. Postoperative cavities were excluded. Lesions were categorized as pre-Lightning (before January 1, 2022) or Lightning-era (on or after). Dosimetric parameters—coverage, selectivity, gradient index (GI), prescription isodose line, and beam-on time (BOT)—were compared overall and stratified by volume: small (<1cc), medium (1–4cc), and large (>4cc).

Results: 640 brain metastases were analyzed (pre-Lightning: 362; Lightning: 278) with comparable size distributions (small: 281 vs 209; medium: 53 vs 50; large: 28 vs 22). Overall, Lightning plans demonstrated higher selectivity (0.570 vs 0.510; p<0.001), GI (4.38 vs 3.55; p<0.001), and prescription isodose (70% vs 60%; p<0.001). Coverage remained equivalent (0.995 vs 0.990; p=0.057). By size: small lesions showed higher GI (4.9 vs 3.7; p<0.001), selectivity (0.503 vs 0.455; p<0.001), isodose levels (74% vs 62%; p<0.001), and longer BOT (16.1 vs 12.1 min; p=0.018). Medium lesions showed improved selectivity (0.615 vs 0.562; p=0.012) and modestly higher GI (4.21 vs 3.78; p=0.041) without significant BOT change. Large lesions showed improved selectivity (0.712 vs 0.658; p=0.010) without significant differences in GI (p=0.27), isodose level (p=0.12), or BOT (p=0.61). Coverage remained uniformly high across all strata.

Conclusion: Conclusion(s): Lightning adoption produced size-dependent dosimetric shifts. Medium and large lesions achieved improved selectivity without worsening GI or prolonging treatment, supporting seamless integration. Small lesions showed improved selectivity and higher prescription isodose lines—indicating reduced high-dose normal brain exposure—but at the cost of significantly worse GI and longer BOT, reflecting provider optimization priorities favoring conformity over dose falloff. These findings highlight the importance of plan-selection awareness, particularly regarding gradient index for small targets. The clinical significance of this trade-off, particularly regarding radionecrosis risk, requires prospective evaluation and longer-term follow-up.