Abstract
PURPOSE: Postoperative Gamma Knife Radiosurgery (GKRS) to resection cavities is standard for intracranial metastases, but the dynamics of these cavities after treatment and their relationship to local control remain incompletely defined. METHODS: We retrospectively reviewed 98 patients with 98 resection cavities treated with single- or hypofractionated Gamma Knife radiosurgery (GKRS) between 2011 and 2025. Cavity volumes were segmented from serial contrast-enhanced MRI and compared with GKRS planning volumes. Longitudinal changes in log-transformed volumes were modeled using linear mixed-effects regression. Local recurrence–free survival (LRFS) and overall survival (OS) were estimated using Kaplan–Meier analysis, and associations with outcomes were tested using Cox proportional hazards models. RESULTS: Median imaging follow-up was 9.1 months (IQR 4.9–16.5), and cavities showed significant volume reduction over time (β = −0.031 ln[cc]/month, p < 0.001). Modeling indicated a nonlinear pattern, with greater reduction early and relative stabilization thereafter. At 12 and 24 months, actuarial local control was 74.5% and 62.5%, respectively; median LRFS was 49.8 months. Volume regression occurred in both recurrent and non-recurrent groups, and the interaction between recurrence status and time was nonsignificant. Recurrence rates did not differ between single- and hypofractionated GKRS across preoperative volume strata. Higher Karnofsky performance status at surgery and at GKRS correlated with improved OS and smaller cavity volumes. CONCLUSIONS: Cavities treated with GKRS showed consistent involution over time, most pronounced early after treatment. Volumetric change alone was not associated with local control; interpretation of local failure requires dosimetric and target-definition context. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1007/s11060-026-05567-7.