Identifying Cancer Patients at Risk for Heart Failure Using Machine Learning Methods

Cardiotoxicity related to cancer therapies has become a serious issue, cancer treatment outcomes and quality of life. Early detection of patients at risk for cardiotoxicity before cardiotoxic treatments and preventive measures are potential solutions to improve cancer's quality of life. This study focuses on predicting the development of failure in cancer patients after cancer diagnoses using historical health record (EHR) data. We examined four machine learning using 143,199 cancer patients from the University of Florida Health(UF Health) Integrated Data Repository (IDR). We identified a total number of1,958 qualified cases and matched them to 15,488 controls by gender, age, race, major cancer type. Two feature encoding strategies were compared to encode as machine learning features. The gradient boosting (GB) based model the best AUC score of 0.9077 (with a sensitivity of 0.8520 and a of 0.8138), outperforming other machine learning methods. We also into the subgroup of cancer patients with exposure to chemotherapy drugs observed a lower specificity score (0.7089). The experimental results show machine learning methods are able to capture clinical factors that are to be associated with heart failure and that it is feasible to use learning methods to identify cancer patients at risk for cancer-related heart failure.