Contemporary Prognostic Signatures and Refined Risk Stratification of Gliomas: An Analysis of 4,400 Tumors.

Background: With the significant shift in the classification, risk stratification, and standards of care for gliomas, we sought to understand how the overall survival of patients with these tumors is impacted by molecular features, clinical metrics, and treatment received.

Methods: We assembled a cohort of patients with a histopathologically diagnosed glioma from The Cancer Genome Atlas, Project Genomics Evidence Neoplasia Information Exchange, and Dana-Farber Cancer Institute/Brigham and Women's Hospital. This incorporated retrospective clinical, histological, and molecular data alongside prospective assessment of patient survival.

Results: 4,400 gliomas were identified: 2,195 glioblastoma, 1,198 IDH1/2-mutant astrocytoma, 531 oligodendroglioma, 271 other IDH1/2-wildtype glioma, and 205 pediatric-type glioma. Molecular classification updated 27.2% of gliomas from their original histopathologic diagnosis. Examining the distribution of molecular alterations across glioma subtypes revealed mutually exclusive alterations within tumorigenic pathways. Non-TCGA patients had significantly improved overall survival compared to TCGA patients, with 26.7%, 55.6%, and 127.8% longer survival for glioblastoma, IDH1/2-mutant astrocytoma, and oligodendroglioma respectively (all p<0.01). Several prognostic features were characterized, including NF1 alteration and 21q loss in glioblastoma, and EGFR amplification and 22q loss in IDH1/2-mutant astrocytoma. Leveraging the size of this cohort, nomograms were generated to assess the probability of overall survival based on patient age, the molecular features of a tumor, and the treatment received.

Conclusions: By applying modern molecular criteria, we characterize the genomic diversity across glioma subtypes, identify clinically applicable prognostic features, and provide a contemporary update on patient survival to serve as a reference for ongoing investigations.