AI in radiation therapy optimization: FOTELP-VOX program enhancement

Abstract

The intersection of Artificial Intelligence (AI) and medical physics heralds a new era for radiation therapy, promising enhancements in precision, safety, and outcomes for cancer patients. This introductory presentation explores the theoretical and practical facets of integrating AI technologies, particularly in the optimization of radiation treatment planning. As the medical community stands on the cusp of this technological revolution, understanding the potential applications, challenges, and benefits of AI in medical physics becomes paramount. Central to our discussion is the potential of AI to automate and refine the processes within the FOTELP-VOX program, a tool critical for simulating particle transport and interactions in radiation therapy. The traditional methodology, heavily reliant on manual optimization, is juxtaposed with AI-driven approaches, showcasing a future where treatment plans are not only devised more efficiently but with greater adherence to the dual objectives of maximizing tumor eradication and minimizing exposure to organs-at-risk (OARs). Our work aims to demystify AI' s role in medical physics, offering insights into Bayesian Optimization (BO) and Genetic Algorithms (GA) as pivotal technologies for enhancing the FOTELP-VOX framework. We address the technical and practical challenges associated with the adoption of AI in medical applications, such as computational costs and time consumption. Furthermore, we concentrate on the ethical dilemmas inherent in the utilization of AI in medicine, particularly concerning the preservation of personal data privacy. Finally, we emphasize the significance of interdisciplinary collaboration.