EXTENT OF SURGERY IMPROVES THE OVERALL SURVIVAL INDEPENDENTLY OF IMAGING CHARACTERISTICS IN NEWLY DIAGNOSED GLIOBLASTOMA

Introduction: In glioblastoma, the extent of surgery affects neurological condition, survival, and the efficacy of postsurgical systemic treatment. We analyzed the impact of surgical extent on overall survival in function of baseline tumor radiological characteristics (size, eloquence of adjacent brain, and location) and known prognostic factors. Material and Methods: Patients received standard radiotherapy and concomitant/adjuvant temozolomide (GLIOCAT Study). We used an MRI-based topographical glioblastoma staging system (stages I, II, and III) integrating tumor location (A, B, C), size (≤5 cm or >5cm), and eloquence of adjacent brain. Evaluations were accepted when kappa was near 1; only evaluations of T1Gd sequences are reported. Surgery was classified as complete resection (CR=disease-free postoperative MRI), partial resection (PR), or biopsy (B). Age, Karnofsky performance score (KPS), Mini-Mental State (MMS), isocitrate dehydrogenase mutations (mIDH), and O6-methylguanine-DNA methyltransferase promoter methylation (methMGMT) were recorded. Results: In 280 patients, Location: A = 133 (47.5%) // B = 110 (39.3%) // C = 37 (13.2%); Eloquence: NO=196 (70%)//YES= 84(30%); size: ≤5cm: 150(54.5%) // > 5cm: 125(45.5%); Stage: I: 100(43.9%)//II: (22.8%)//III: 76(33.3%). Clinical factors KPS (<70, ≥70), age (<50, ≥50), MMS (<27, ≥27), mIDH, and methMGMT were uniformly distributed across stages. Surgical type: CR 16%//PR 65.7%//B 14.6%. CR was more frequent in stage I (80%) and biopsy in stage III (46.3%) (p=0.000). Univariate analyses found type of surgery (p=0.000), KPS ≥70 (p=0.02), methMGMT (P=0.01), mIDH (P=0.03), stage I vs II vs III (P=0.000), location A vs B vs C (P=0.000), and size (P=0.005) but not brain eloquence (P=0.39) were associated with longer overall survival. In multivariate analyses, methMGMT, KPS≥70% and resection (CR or PR) were independent prognostic factors, and eloquence, size, and location lost their prognostic value. Conclusions: In glioblastoma, PR or CR improves overall survival independently of size, eloquence, or tumor location. Overall survival is best in patients with good KPS and methMGMT. The present study is supported by a Grant of the Fundació TV3 (Project: 665/C/2013). The radiological platform is supported by a citizens initiative: Associació Mongat en Acció.