Deep Reinforcement Learning for Multi-Drug Therapy Optimization in Rare and Refractory Cancers

DOI

10.2991/978-94-6463-948-3_13How to use a DOI?

Keywords

Precision Oncology; Tumor Heterogeneity; Rare and Complex Cancers; Biomarker Discovery; Adaptive Radiotherapy

Abstract

Deep reinforcement learning (DRL) is found to be an interesting model to optimize the complex multi-drug treatment regimens especially in the rare and treatment-resistant cancers. These tumors are usually characterized by shortage in standard treatment approaches, high level of patient-patient het-erogeneity, and small clinical data which makes traditional treatment planning challenging. The question that this paper seeks to answer is how DRL can tackle these complexities responding to the patient specifically, modelling and enabling advanced decision-making in the therapeutic process. The existing situation of multi-drug therapy optimization is discussed, in particular: the problem of predicting the interactions of drugs, the problem of over-coming the adverse effect of treatment with drugs, and the problem of searching the huge dimensionality of combinations of possible drugs. The key feature of DRL is its capability to develop the adaptive strategies of patient-specific treatment by combining various sources of data genomics, transcriptomics, proteomics), as well as its prediction of therapeutic out-comes by analysis through simulation. Despite its potential, some limitations still exist: the lack of quality clinical data, substantial computational requirements of training, and necessity of strict clinical validation. The discussion of the strategies of getting rid of these problems and expanding the impact of DRL is provided, with one of their solutions being the integration of DRL with digital twins to model patients, federated learning to facilitate training across facilities, and the possibility of utilizing quantum computers to address both the size and difficulty of simulations of treatment. Its results indicate that DRL could revolutionize precision oncology and that it could enable more effective, personal and safe treatment regimens to support clinical practice in patients with rare and hard to treat cancers.

Copyright

© 2025 The Author(s)

Open Access

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

TY  - CONF
AU  - Bhupender Singh
AU  - Arvind Kakulte
AU  - Sampathi Sunitha
AU  - Jagadish V. Tawade
AU  - Nitiraj V. Kulkarni
PY  - 2026
DA  - 2026/01/06
TI  - Deep Reinforcement Learning for Multi-Drug Therapy Optimization in Rare and Refractory Cancers
BT  - Proceedings of the International Conference on Sustainable Innovation with Artificial Intelligence and Machine Learning 2025 (ICSIAIML 2025)
PB  - Atlantis Press
SP  - 195
EP  - 205
SN  - 1951-6851
UR  - https://doi.org/10.2991/978-94-6463-948-3_13
DO  - 10.2991/978-94-6463-948-3_13
ID  - Singh2026
ER  -