Identifying Cancer Patients at risk for Heart Failure using deep learning models

Abstract

The cardiotoxicity that may occur as a side effect of cancer treatments has emerged as a significant problem. Cancer patient’s quality of life may be improved if those at risk of cardiotoxicity are identified early and given prophylactic treatments before receiving cardio toxic drugs. The advancement of deep learning will help to support medical practitioners in their ability to make accurate. This study will focus on predicting the enhancement of heart failure in cancer patients. The purpose of this research is to determine whether historical data from electronic health records can accurately predict the occurrence of heart failure in cancer patients. We investigated deep learning algorithms by applying them to 300 cancer patient’s dataset drawn from the Seer database. We determined that there were a total of 300 eligible cases and matched them with controls according to gender age and the primary cancer type etc. Results from the tests suggest that techniques         based on deep learning may effectively capture clinical characteristics linked with heart failure in cancer patients.