A Machine Learning Approach to Chemotherapy-induced Remodeling of the Tumor

Project summary

All patients battling ovarian cancer will undergo chemotherapy and surgery. Unfortunately, most patients will experience disease recurrence leading to secondary treatment interventions. Thus, to improve patient outcomes there is a significant need to advance the predictive power of when a tumor will recur and provide a tumor-tailored secondary treatment. The Ovarian Cancer Research Group at The University of Colorado, recently reported that the predictive signatures and secondary treatment information can be inferred from how the tumor responds to upfront chemotherapy. Using cutting-edge molecular profiling of patient-matched tumors before and after chemotherapy, the investigative team honed in on several key biological processes including transcriptional regulation (CEBP/b) and cell signaling (Wnt) that are critical for predicting disease recurrence. Moreover, using advanced computational analysis these inferences can be used to provide meaningful clinical information. The proposed research aims to expand the Ovarian Cancer Research Group’s findings and to mechanistically determine critical attributes that predict disease recurrence. The Ovarian Cancer Research Group consists of a diverse and synergistic team of clinicians, basic science researchers, computer scientists, and biostatisticians. The research group is highly capable of completing the proposed work and translating the findings into a clinical trial. The long-term vision is to be able to perform advanced molecular analysis and computer modeling on patient tumors before and after chemotherapy in a clinically meaningful timeframe so that the gynecologic oncologist would have both a reliable estimate of recurrence and a secondary treatment option tailored to the tumors.