Genomic Copy Number Complexity Correlates With Overall Survival in Glioblastoma Patients

Abstract

Whole genome copy number analysis is increasingly available and now routinely performed for high-grade glial neoplasms at our institution. The purpose of this study is to determine if overall genomic complexity or specific gene copy number variations correlate with survival outcomes in patients with glioblastoma (GBM). Twenty-seven patients, diagnosed with GBM prior to February 2015 and analyzed using the Affymetrix OncoScan FFPE Assay for gene copy number variations, were identified and their charts reviewed. All patients received standard therapy consisting of focal radiation therapy and temozolomide. Actuarial overall survival (OS) and progression-free survival (PFS) were assessed by the Kaplan-Meier method and subgroups compared by the log-rank test. Univariate and multivariate analyses were performed using Cox proportional hazards regression. Standard demographic and tumor-related variables were evaluated. In addition, based on OncoScan data, tumors were classified as “complex” when whole genome copy number patterns exhibited multiple segmental gains and losses requiring multiple double strand breaks and “simple, or non-complex” when changes involved whole arm or chromosome losses or gains. Finally, exact copy numbers were determined for thirteen specific gene loci (CCNE1, CDK4, CDK6, EGFR, FGFR3, INSR, IRS1, IRS2, MDM2, MDM4, PDGFRA, PTEN, and SOX2) and each evaluated as individual variables. Statistical analyses were performed on the SPSS Statistics software, v23. Our cohort consisted of 11 women and 16 men. Median age was 63.2 years (range: 39.0-86.2 years). Median KPS was 80 (range: 60-90). Tumors were hypermethylated at MGMT in 13/27 (48.1%) and harbored mutant IDH1 in 2/27 (7.4%). When classifying by genomic complexity, distribution of simple:complex GBMs was 11:16. Median follow-up times for all and living patients were 14.7 and 15.8 months, respectively. Median PFS and OS were 9.4 and 20.2 months, respectively. OS was significantly better in the complex (median = 20.2 mos) versus simple (median = 14.7 mos) subgroup by log-rank test (P value = 0.034). Younger age, complex genomic organization and lower IRS2 (insulin receptor substrate 2) copy number were significant predictors of better OS by univariate analyses with P values of 0.036, 0.049, and 0.049, respectively. Complex genomic organization and lower IRS2 copy number retained prognostic significance for OS in multivariate analysis with P values of 0.026 and 0.011, respectively. Increased genomic complexity and lower IRS2 copy number were significantly correlated with longer OS. Both factors may serve as potential markers for increased sensitivity to therapy. Our findings are hypothesis generating and will require validation in a larger cohort.