BIOM-32. PRECISION PROGNOSTICATION IN DIFFUSE GLIOMAS: MULTI-COHORT VALIDATION OF CDKN2A/B, MET, AND PIK3R1 ALTERATIONS

Abstract BACKGROUND Despite advances in the molecular classification of diffuse gliomas, limited data exist on the prognostic implications of specific genetic alterations. We integrated molecular profiles of diffuse glioma subtypes in multi-center cohort and investigated prognostic significance especially in glioblastoma, IDH-wildtype. METHODS We analyzed 470 primary diffuse gliomas from a hospital-based cohort using next-generation sequencing, followed by multivariable Cox regression and Kaplan-Meier survival analysis. For independent validation of associations between molecular profiles and overall survival, 840 primary glioblastoma IDH-wildtype cases were obtained from four external cohorts: CPTAC (n = 90), GLASS (n = 106), MSKCC (n = 410), and TCGA (n = 234). RESULTS Distinct genetic landscapes were demonstrated in oligodendroglioma, IDH-mutant, 1p/19q co-deleted (n = 68), diffuse astrocytoma, IDH-mutant (n = 63), and glioblastoma, IDH-wildtype (n = 339). In glioblastoma, PIK3R1 mutation (HR [95%CI]=1.53 [1.06-2.20], P=0.022), CDKN2A/2B deletion (HR [95%CI]=1.41 [1.09-1.83], P<0.001), and MET amplification (HR [95%CI]=2.78 [1.40-5.51], P=0.003). were significant prognostic markers along with negative MGMT promoter methylation status (HR=1.51 [1.14-2.00], P<0.001). External validation confirmed prognostic impact of PIK3R1 in CPTAC (HR=3.72, P=0.007) and MET amplification in MSKCC (HR=2.74, P=0.027), though CDKN2A/2B showed cohort-specific effects. Tumor mutational burden lacked prognostic significance across all cohorts. MTAP deletion was observed in 45.4% of glioblastomas but was not significantly associated with overall survival. CONCLUSIONS This study highlights CDKN2A/2B deletion, MET amplification, and PIK3R1 mutation as independent markers of poor prognosis in glioblastoma, which enables molecular-based risk stratification.