A STATE-OF-THE-ART REVIEW OF ADAPTIVE SCHEDULING AND OPTIMIZATION IN RADIOTHERAPY: BRIDGING OPERATIONAL RESEARCH AND CLINICAL PRACTICE

Abstract

Background: Radiotherapy (RT) scheduling is an important element in managing the schedule for treatment and reimbursement, as well as managing scarce resources to benefit as many patients as possible. Standard scheduling algorithms fail to consider dynamic factors such as time, working machine availability, each patient's unique situation, and interruptions that require powerful algorithms. Objective: The study aims to discuss the existing techniques that help optimize RT scheduling, emphasizing adaptation and Artificial Intelligence (AI) to reduce operational constraints, enhance the flexibility of the scheduling approach, and incorporate real-time decision-making. Methodology: Many deterministic, stochastic, and AI-based scheduling methods have been reviewed for suitability in the RT workflow. The paper discusses hybrid optimization methods, simulation-based scheduling, federated learning for distributed optimization, and policies for advocating scheduling solutions. Results: There is evidence that enhancing the integration of AI and Machine Learning and using real-time optimization in the work schedule improves scheduling efficiency. Enhancing the deterministic models with AI increases efficiency, whereas real-time scheduling, such as rolling-horizon and multi-agent systems, contributes to dynamic decision-making. However, challenges like computational difficulty, issues relating to compatibility, and clinician unwillingness act as barriers. Conclusion: Bringing research into clinical practice calls for collaboration and interaction between different professionals, common practice protocols, and exponentially growing AI-based schedule solutions. Future work should also aim to develop real-time adjustments and address the ethical practice of integrating AI into clinical work to enhance RT scheduling.