2040 - Radiosurgery Outcomes for Patients with Significant Intra-Cranial Disease Burden with =50 Brain Metastases

Purpose/Objective(s): For patients with extremely high intracranial burden, WBRT is often favored but risks neurocognitive decline. We hypothesize that GKRS for patients with =50 brain metastases could be feasible, provide meaningful survival outcomes with low re-treatment rates, and reduce whole-brain radiation exposure compared to traditional WBRT approaches. We further identify factors associated with overall survival (OS) in this patient population.

Materials/Methods: A retrospective cohort study of 22 patients (2016–2025) with =50 cumulative lesions treated with GKRS (1,734 lesions) was performed at a single institution. Primary endpoint was OS from first GKRS treatment. Secondary analyses assessed prognostic factors (age, KPS, prior WBRT, initial lesion count and volume, histology, targetable mutations, etc) using Kaplan-Meier with univariate and multivariate Cox regressions. Whole-brain and organ-at-risk dosimetry were compared with a reference hippocampal-avoidance WBRT plan generated with IMRT. Dose de-escalation was defined as median per-lesion dose <15 Gy for predominantly sub-centimeter metastases when treating = 50 lesions. Number of re-treated lesions was recorded.

Results: Median OS was 17.56 months (95% CI, 13.66–24.91). On multivariable Cox modeling, =50 lesions at the initial GKRS independently predicted worse survival (HR 29.01, 95% CI 2.77–303.43, p=0.005). Prior WBRT and greater initial treated volume were associated with worse OS on univariate testing (p=0.036 and p=0.032, respectively), but significance was not retained in multivariable analysis; any WBRT trended toward inferior OS (p=0.074). Despite dose de-escalation, only 34 out of 1,734 lesions were re-treated (1.96%), with no radiation necrosis requiring intervention. Whole-brain and critical-structure doses were significantly lower with GKRS than with the HA-WBRT plan. Using median cut points, neither WBmax (=40.2 Gy) nor WBV12 (=16.08 cc) was associated with OS (p=0.97 and p=0.96). Median per-lesion prescription dose (=15 Gy vs <15 Gy) was not associated with OS.

Conclusion: In selected patients with =50 brain metastases, GKRS is feasible, maintains low lesion re-treatment rates with low toxicity, and reduces whole-brain exposure comparted to WBRT approaches traditionally used for this patient population. Survival is driven primarily by intracranial burden at initial treatment, with initial lesion count, rather than cumulative disease, emerging as the dominant predictor of OS. Dose de-escalation (median ~15 Gy) did not compromise OS, suggesting lower prescription doses to limit cumulative normal-brain exposure when treating high lesion counts. These findings support consideration of earlier GKRS intervention with dose de-escalation before extensive lesion burden develops.