Abstract
The purpose of drug repurposing is to identify alternative uses of FDA approved drugs, which significantly accelerates the drug development process. Meanwhile, clinical data illustrate the patterns and clinical outcomes of drug use, so they have been increasingly applied to support drug development, particularly for drug repurposing. The NIH Biomedical Translational Research Information System (BTRIS) is a resource which compiles deidentified patient data from clinical research done across NIH Institutes and Centers. In this study, we analyzed clinical data available from BTRIS to identify drug repurposing candidates, i.e., identifying drugs that were correlated with an increased survival rate for glioblastoma (GBM) patients. Specifically, we extracted all the administered drugs on GBM patients and fitted them to elastic-net penalized Cox proportional hazards (CPH) models, a regression model for investigating the association between the survival rate of patients and covariates (administered drugs in this study). We were able to identify several potential drug candidates for GBM to be further evaluated with other data types and by performing biological experiments.