Analyzing Patient Symptoms and Treatment Effectiveness of Lung Cancer Drug Therapies with Machine Learning Algorithms

Abstract

Cancer patients often experience various severe side effects from cancer treatment. Specifically, lung cancer patients commonly experience Cisplatin-induced nephrotoxicity–deterioration in kidney function due to toxic effects of the chemotherapy drug Cisplatin. However, it is difficult to predict how a patient will respond to Cisplatin treatment, as patients respond differently depending on their unique clinical-demographic features. This project analyzes data for lung cancer patients undergoing Cisplatin treatment by using machine learning methods to determine the most important clinical features that affect nephrotoxicity, and to predict the probability of a patient experiencing toxic symptoms given their characteristics. Three different clinically relevant patient groups were used, and age, dose given, and pre-treatment GFR were found to be positively correlated and the most relevant features for determining treatment effectiveness and possibility of nephrotoxicity. This information can help doctors and patients predict treatment response and optimize treatment plans.