Predictors of Survival in Patients Aged ≥70 with Glioblastoma: A Time-Dependent Multivariable Analysis

Background: Glioblastoma (GB) carries a dismal prognosis, with survival outcomes particularly poor in older patients. With the fastest-growing global demographic being those aged over 65, the incidence of GB is expected to rise. Objective: To evaluate predictors of survival in patients aged ≥70 years with histologically confirmed GB, focusing on surgical resection, adjuvant therapy, and comorbidities. Methods: A retrospective review was performed of all patients aged ≥70 undergoing index surgery for GB between January 2021 and March 2025 at a single tertiary neurosurgical centre. Demographics, pre-operative fitness scores (Karnofsky Performance Status [KPS]., Charlson Comorbidity Index [CCI].), tumour characteristics, extent of resection, adjuvant treatment, and survival were analysed. Tumour volume was estimated using the ABC/2 method. Survival outcomes were assessed using Kaplan-Meier curves and multivariable Cox proportional hazards regression. Results: A total of 124 patients aged ≥70 years (median 74 years, range 70-86) were included. Median overall survival was 8 months (IQR 4-15). On multivariable analysis, adjuvant chemoradiotherapy (HR = 0.30, 95% CI 0.17-0.52; p < 0.001) and gross total resection (GTR) (HR = 0.41, 95% CI 0.20-0.86; p = 0.019) were independently associated with improved survival. Smoking history was associated with increased hazard of death (HR = 2.02, 95% CI 1.07-3.81; p = 0.029), an effect robust to multiple sensitivity analyses. No significant associations were found for age, pre-operative KPS, comorbidity index, tumour volume, or methylation status (all p > 0.10). Tests for non-proportional hazards indicated that the survival benefit of adjuvant therapy diminished over time (interaction p = 0.0002), with early post-operative benefit (HR ≈ 0.35 at 1 month) that attenuated towards unity by 6-12 months. The effects of GTR and smoking were time-invariant. RMST analysis suggested a modest, non-significant absolute survival advantage of GTR over STR (mean difference = 2.0 months at 18 months; p = 0.11). After exclusion of early post-operative deaths (<6 weeks), adjuvant therapy (HR = 0.34; p < 0.001) and GTR (HR = 0.33; p = 0.005) remained independent predictors of improved survival. Conclusions: Among patients aged ≥70 years with glioblastoma, adjuvant therapy and extent of resection remain key independent predictors of survival, while smoking is associated with poorer outcomes. The survival benefit of adjuvant chemoradiotherapy is strongest in the early post-operative period and diminishes over time, underscoring the importance of early multidisciplinary intervention. These findings highlight that aggressive multimodal treatment may confer survival advantage even in older patients.