P12.10.A RETHINKING TREATMENT DE-ESCALATION IN ELDERLY GLIOBLASTOMA: LESSONS FROM A MULTI-CENTER REAL WORLD COHORT

Abstract BACKGROUND Glioblastoma (GBM) in elderly patients (≥65 years) is associated with increased frailty, treatment toxicity, and poor prognosis. Consequently, many elderly patients receive de-escalated treatment regimens, often guided by MGMT promoter methylation status. However, real-world data comparing outcomes across different treatment modalities remain limited. MATERIAL AND METHODS We conducted a retrospective cohort study including 573 elderly GBM patients treated between 2009 and 2023 at two tertiary medical centers in Israel. Patients were categorized according to postoperative treatment: (1) chemoradiotherapy (CRT; 60 Gy/30 fractions or 40.05 Gy/15 fractions), (2) upfront temozolomide (TMZ) alone, (3) radiotherapy (RT) alone, or (4) best supportive care. MGMT methylation status and Karnofsky Performance Status (KPS) were analyzed when available. Survival outcomes were assessed using Kaplan-Meier and Mantel-Cox tests. RESULTS Median overall survival (mOS) was significantly longer for patients treated with chemoradiotherapy (14 months) compared to TMZ alone (8 months) or RT alone (8 months). Among MGMT-methylated patients, combined therapy resulted in an mOS of 23 months, versus 8 months for TMZ alone. In TMZ-only patients, salvage treatment upon disease progression was associated with improved survival (10 vs. 5 months). TMZ-related toxicity was manageable, with grade 3/4 hematologic and non-hematologic toxicities occurring in 6% and 12% of patients, respectively. Lower KPS at diagnosis correlated with reduced treatment benefit. CONCLUSION Fit elderly patients benefit significantly from standard-of-care therapy, particularly in the presence of MGMT methylation. Age alone should not determine treatment intensity. Our findings underscore the importance of integrating clinical, molecular, and functional assessments into treatment planning, emphasizing the need for real-world validation of geriatric oncology algorithms.