Divya V. – BASIS Scottsdale

Project Abstract

Cancer is a leading cause of death around the world. However, chemotherapy, one of the most common cancer treatments, has significant efficacy issues arising from the uncertainty of what the most effective chemotherapy regimens, or combinations of chemotherapeutic drugs, are for individual patients. Physically testing regimen efficacy through clinical trials is lengthy and impractical; however, computational techniques, the use of statistical models that can quickly identify trends in large inputs of data, overcome the limitations of physical testing. Thus, the objective of this study is to create a computational model to determine the most effective regimens for breast cancer patients by analyzing the molecular and genomic mechanisms responsible for regimen efficacy. This study implemented an outcomes analysis, molecular analysis, and genomic analysis. The outcomes analysis used meta-analysis statistics to rank a set of tested breast cancer regimens from most to least effective based on outcomes data from clinical trials. From the effective regimens in the outcomes analysis, the molecular analysis identified untested regimens with similar molecular features. Finally, the genomic analysis predicted whether the untested regimens identified in the molecular analysis would actually be effective against various breast cancer mutations. The model will allow the medical community to identify regimens that will likely be most effective for individual breast cancer patients by understanding molecular and genomic mechanisms responsible for regimen efficacy. Thus, this research transforms chemotherapy from a general to targeted treatment model, improving likelihood of treatment success and the lives of cancer patients around the world.

Automated Prediction Model for Chemotherapy Treatment Regimens through Molecular and Genomic Feature Selection