International journal of particle therapy · 2026
AI in Proton Therapy Treatment Planning: A Narrative Review.
СтатьяНазвание и аннотация: Перевод пока не подготовлен
Обработка статьи · этапы, ошибки и управление
Загрузка состояния…
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Ding Yuzhen, Feng Hongying, Bues Martin, Fatyga Mirek, Liu Tianming, Zhang Wei, Whitaker Thomas J, Lin Haibo, Lee Nancy Y, Simone Charles B, Patel Samir H, Ma Daniel J, Frank Steven J, Vora Sujay A, Ashman Jonathan B, Liu Wei
Аннотация
PURPOSE: Proton therapy provides superior dose conformity compared to photon therapy, but its treatment planning is challenged by sensitivity to anatomical changes, setup/range uncertainties, and computational complexity. This narrative review evaluates the role of artificial intelligence (AI) in improving proton therapy treatment planning, with emphasis on proton-specific requirements for dose-calculation accuracy, robustness, and adaptive decision-making.
MATERIALS AND METHODS: Recent studies on AI applications across the proton therapy planning workflow, including imaging, registration, segmentation, contour assessment, dose prediction/denoising, optimization, adaptive therapy, and large language model (LLM) and multimodal AI were reviewed and summarized by application domain and validation strategy. Both proton-therapy-specific studies and selected general radiotherapy AI studies were included when they introduced methods relevant to proton treatment planning, with attention to the additional validation needed for proton therapy.
RESULTS: AI has shown promise in enhancing imaging for dose calculation, automating contouring and contour assessment, accelerating dose calculation, predicting and denoising dose distributions, supporting robust optimization, and improving adaptive planning workflows. Compared with photon therapy, AI tools in proton therapy require more stringent evaluation because errors in AI-generated images, deformation fields, contours, dose estimates, or optimization outputs may propagate into water-equivalent path length, range uncertainty, scenario-specific dose distributions, and adaptive replanning decisions. Emerging applications, including LLMs and multimodal foundation models, may further support clinical-context integration, treatment-planning documentation, plan review, and workflow coordination. Key limitations include data scarcity, limited generalizability, physical interpretability, uncertainty management, lack of standardized proton-specific benchmarks, and limited prospective validation.
CONCLUSION: AI is emerging as a key enabler of efficient, consistent, and patient-specific proton therapy treatment planning. Clinical translation will require interpretable and physics-aware models, multi-institutional validation, workflow-level testing, and evaluation using proton-specific endpoints such as HU/RSP accuracy, range error, dose-calculation accuracy, robustness, deliverability, and adaptive decision consistency.
MATERIALS AND METHODS: Recent studies on AI applications across the proton therapy planning workflow, including imaging, registration, segmentation, contour assessment, dose prediction/denoising, optimization, adaptive therapy, and large language model (LLM) and multimodal AI were reviewed and summarized by application domain and validation strategy. Both proton-therapy-specific studies and selected general radiotherapy AI studies were included when they introduced methods relevant to proton treatment planning, with attention to the additional validation needed for proton therapy.
RESULTS: AI has shown promise in enhancing imaging for dose calculation, automating contouring and contour assessment, accelerating dose calculation, predicting and denoising dose distributions, supporting robust optimization, and improving adaptive planning workflows. Compared with photon therapy, AI tools in proton therapy require more stringent evaluation because errors in AI-generated images, deformation fields, contours, dose estimates, or optimization outputs may propagate into water-equivalent path length, range uncertainty, scenario-specific dose distributions, and adaptive replanning decisions. Emerging applications, including LLMs and multimodal foundation models, may further support clinical-context integration, treatment-planning documentation, plan review, and workflow coordination. Key limitations include data scarcity, limited generalizability, physical interpretability, uncertainty management, lack of standardized proton-specific benchmarks, and limited prospective validation.
CONCLUSION: AI is emerging as a key enabler of efficient, consistent, and patient-specific proton therapy treatment planning. Clinical translation will require interpretable and physics-aware models, multi-institutional validation, workflow-level testing, and evaluation using proton-specific endpoints such as HU/RSP accuracy, range error, dose-calculation accuracy, robustness, deliverability, and adaptive decision consistency.