TMIC-81. DECODING THE ROLE OF MICROENVIRONMENT ON DISTINCT FAILURE PATTERNS OF MALIGNANT GLIOMAS

Abstract

Abstract Most malignant gliomas experience recurrence with distinct failure patterns. The spatial relationship between the primary and recurrent tumors usually represents the genomic distanc; however, the molecular backgrounds of these failure patterns have not been evaluated yet. Our aim was to investigate the molecular profiles underlying the distinct failure patterns of malignant gliomas. This study involved 272 malignant glioma patients, with transcriptomic data available for 166 of these cases. The failure pattern was defined according to the spatial relationship between the primary and recurrent pairs including local recurrence, remote recurrence, and leptomeningeal seeding. When the recurrent tumor exhibits more than one failure pattern simultaneously, we define this as a combined failure pattern. Local recurrence was the most frequent pattern, followed by combined failure pattern (62.9% & 22.8%, respectively). Multi-variate cox analysis showed that combined failure pattern was a significant prognostic factor (odds ratio= 3.34, p-value < 0.001). We defined a genomic signature for malignant gliomas with a combined failure pattern (comb-GS) and validated its worse prognostic impact in other cohorts (GLASS, TCGA). Using public scRNA-seq data with regional annotation, we found that two distinct non-tumor components, immune and neuronal components, contributed to the development of the combined failure pattern. This finding was well correlated with the variability of comb-GS in different anatomic regions in IVY-gap dataset. In summary, this study demonstrated that both immune and neuronal components are involved in the progression of malignant gliomas. These two microenvironmental elements exhibit a shared negative clinical outcome, despite presenting distinct phenotype. To decipher the exact biological role of each component, experimental validation is necessary.