2044 Background: Glioblastoma (GBM) is the most aggressive and common primary brain tumor. Nomograms are prediction models that help form individualized risk scores for cancer patients, which are valuable for treatment decision-making. The aim of this study is to create a refined nomogram by including novel molecular variables beyond MGMT promoter methylation. Methods: Clinical data and miRNA expression data were obtained from 226 newly diagnosed GBM patients. Clinical data included age at diagnosis, sex, Karnofsky performance status (KPS), extent of resection, O6-methylguanine-DNA methyltransferase ( MGMT) promoter methylation status, IDH mutation status and overall survival. Due to low representation of less than 13 cases each, IDH mutant glioblastomas and patients submitted to biopsy-only were excluded. Total RNA was isolated from formalin-fixed paraffin-embedded (FFPE) tissues; miRNA expression was subsequently measured using the NanoString human miRNA v3a assay. A Cox regression model was developed using glmnet R package with the elastic net penalty while adjusting for known prognostic factors. A dichotomized genomic score was created by finding the optimal cutpoint (maximum association with survival) of the linear combination of the selected. A nomogram was generated using known clinical prognostic factors, specifically age, sex, KPS, and MGMT status along with the dichotomized genomic score. Results: Four novel miRNAs were found to significantly correlate with overall survival and were used to create the dichotomized miRNA genomic score (GS). This score split the cohort into a poor performing group (GS_high) and a better performing group (GS_low) (p = 0.0031). A final nomogram was created using the Cox proportional hazards model (Figure 1). Factors that correlated with improved survival included younger age, KPS > 70, MGMT methylation and a low genomic score. Conclusions: This study is a proof of concept demonstrating that integration of molecular variables beyond MGMT methylation improve existing nomograms to provide individualized information about patient prognosis. Future directions include a more comprehensive analysis, including proteomic and methylation data, and subsequent validation in an external cohort. Finally, network analysis integrating molecular signatures of poor performers will help identify therapeutic targets.