Development of Model Predictive Control by Reinforcement Learning to treat Cancer with Mixed of Chemotherapy and Anti-Angiogenic

Abstract

Sarina Haghighatdoust, Mohsen Hadian, Mobina Ghajar, Mandana Sadat Ghafourian, Sara Tarkiani and Amin Ramezani

Cancer is a major global health concern and one of the leading causes of death worldwide. As such, the study of cancer treat- ment and control has become an essential area of research in biomedical engineering. One critical aspect of cancer treatment is the prevention of cancer cell proliferation. This article proposes a novel approach in a cancer model with zero initial conditions once and another with random initial conditions by a Model Predictive Controller (MPC) developed by Reinforcement Learning. Models are treated by 3 different treatment methods including chemotherapy, anti-angiogenic and the combination of both treat- ment methods is used to evaluate the effectiveness and reduction of the number of cancer cells and the improvement of disease outcomes. Our results show that although chemotherapy is necessary to weaken cancer cells, the combination of both treatment methods reduces the number of cancer cells by approximately 65%, and this shows the effectiveness of the combination of two treatment methods with the help of a Model Predictive Controller (MPC) developed by Reinforcement Learning by reducing the number of tumor cells in the targeted location.

PDF