GENO-18A 35 GENES SIGNATURE FOR PREDICTING OUTCOME IN GLIOMA PATIENTS WITH 1p/19q CODEL

Meta-analysis of glioma patients revealed that LOH of 1p/19q (IDHmutant-codel) is associated with favorable progression-free survival (PFS) and overall survival (OS). Heterogeneity in clinical outcomes exists within this subset of glioma. To determine biomarkers of treatment response, we assembled the expression profiles from 333 IDHmutant-codel patients. The dataset was split into a training dataset (n = 170) and a validation dataset (n = 163) through random sampling. An elastic net penalized Coxu0027s proportional hazards model was applied to build a classifier model through cross validation within the training dataset. This resulted in a signature of 35 genes significantly associated with outcome in IDH mutant-codels. The performance of the gene signature was determined in the validation dataset, which resulted in a high-risk (median survival of 68.3 months) and low-risk group (median survival of 118.2 months) that showed significant difference in survival (p= 0.0079, Log-rank test). The predictive power of the gene signature using the concordance index was 0.65. For time to death events, the risk score predicted by the above genes signature yielded HR = 2.296 (95% confidence interval [CI] = 1.279 to 4.121) after adjustment for age and stratification for grade. Gene Set Variation Analysis revealed that the genes in the signature were associated with increased acetylation activity. In summary, our study identified a 35-gene signature that distinguished between high risk and low risk groups of glioma patients with 1p/19q CODEL, which is able to discriminate low risk patients within 1p/19q CODEL cohort for whom aggressive therapeutic regimens might be omitted.