A Learning-Driven Automatic Planning Framework for Proton Pencil Beam Scanning Treatments of Head and Neck Cancers

Proton pencil beam scanning treatment planning for head and neck cancers involves numerous conflicting objectives, requiring iterative objective parameter adjustments to balance multiple clinical goals. We propose a learning-driven inverse optimizer and integrate it into a proximal policy optimization (PPO)-based planning framework to automatically generate high-quality plans for patients with diverse treatment requirements.

The inverse optimizer is a learning-to-optimize (L2O) method that predicts update steps by learning from task-specific data distributions. For the first time, long-context processing techniques developed for large language models are used to address the scalability limitations of existing L2O methods, enabling simultaneous optimization over a substantially large set of variables. The PPO framework functions as an outer-loop virtual planner, autonomously adjusting objective parameters through a policy network, and the inner-loop L2O inverse optimizer computes machine-deliverable spot monitor unit values based on the PPO-refined objectives. Moreover, a Swin UNetR dose predictor is trained with prescription- and beam-specific information to estimate the initial objective parameters. In our experiments, a total of 97 patients with bilateral or ipsilateral head and neck cancers were included for training and testing.

Compared with second-order gradient-based methods, our L2O optimizer improves the effectiveness and efficiency of the time-consuming inverse optimization by 22.97% and 36.41%, respectively. In conjunction with the PPO-based virtual planner, plans are generated within clinically acceptable times, that is, 2.55 hours on average, and show improved or comparable organs at risk sparing with superior target coverage compared with human-generated plans.

The proposed inverse optimizer is the first L2O model applied to radiation therapy treatment planning and achieves promising performance. The high-quality plans generated for patients with variable prescription dose levels, multiple target volumes, and patient-specific beam angles highlight the strong potential of the proposed automatic planning framework for practical clinical use.