Ovarian cancer is a rare but extremely deadly disease affecting approximately 22,000 American women per year. There is a critical need for tools that would assist with chemotherapy prediction and selection. The standard of care is platinum based chemotherapy, to which approximately 15% of patients are resistant. If resistance could be identified earlier, before giving chemotherapy, alternative treatments could be given immediately, improving their survival. The goal of this project is to utilize an artificial intelligence based approach, combined with a unique cohort of patients, to develop an innovative, practical, tool to predict which women will experience resistance to platinum based chemotherapy. The primary competitive advantage to this project is the access to a unique cohort of ovarian cancer patients. This cohort, PREDICTOVAR, consists of 527 women, is fully annotated with 96 clinical variables, and includes all pathology and radiology imaging. Dr. Carlson is the pathology representative on the steering committee, guaranteeing access to this material. Dr. Liu has extensive experience with machine learning in health care. There is significant potential for commercialization from this project, and the PIs have begun discussion with the USC Stevens Center for Innovation regarding licensing. Although numerous companies are exploring machine learning for therapy prediction, few systems have arrived to market, and none within ovarian cancer. This project combines a ready-to-go clinical cohort with the deep machine learning expertise of Dr. Liu to propose an algorithm that could significantly impact women with ovarian cancer.