ACTR-17. EFFECT OF AGE ON OUTCOME IN THE GLARIUS TRIAL

Abstract

BACKGROUND: The phase II GLARIUS trial assigned patients with newly diagnosed, O-6-methylguanine-DNA methyltransferase (MGMT) non-methylated glioblastoma in a 2:1 ratio to experimental bevacizumab/ irinotecan (BEV/IRI) or standard temozolomide (TMZ). BEV/IRI prolonged progression-free survival but overall survival (OS) was similar in both treatment arms. Whether elderly patients derive benefit from BEV/ IRI treatment is a controversial issue. To address this, the association of age and BEV/IRI treatment was assessed in a subgroup analysis. METHODS: Patients included in the modified ITT population were grouped according to their age at tumor diagnosis into a <65 years or ≥65 years group. Characteristics (sex, Karnofsky Performance Score (KPS), extent of resection, and mini-mental status examination (MMSE) score) were then evaluated between groups. In a Kaplan-Meier analysis, we compared overall survival (OS) between age groups in each treatment arm. A Cox regression analysis served to detect whether age was an independent prognostic factor. RESULTS: In the BEV/IRI arm, 95 patients were younger than 65 and 21 were at least 65 years old. In the TMZ arm, 41 patients were <65 and 13 ≥65 years old. In each treatment arm, canonical prognostic factors (KPS, extent of resection) were balanced between age groups. Among patients treated with BEV/IRI, those <65 years survived significantly longer than those ≥65 (median OS, 17.5 vs. 13.4 months, p<0.001). Among TMZ treated patients, no significant difference was found between age groups (median OS, 20.0 vs. 17.3 months, p=0.567). In the Cox model, age emerged as an independent prognostic factor in BEV/IRI treated patients only (Hazard Ratio, 2.72, p<0.001). CONCLUSIONS: In the GLARIUS trial, patients <65 years derived significantly more benefit from BEV/IRI treatment as compared with elderly patients ≥65. However, the small sample size limits our analysis.