Abstract
Radiotherapy (RT) planning is complex, subjective, and time-intensive.Advances with artificial intelligence (AI) promise to improve its precision andefficiency, but progress is often limited by the scarcity of large,standardized datasets. To address this, we introduce the Automated Iterative RTPlanning (AIRTP) system, a scalable solution for generating high-qualitytreatment plans. This scalable solution is designed to generate substantialvolumes of consistently high-quality treatment plans, overcoming a key obstaclein the advancement of AI-driven RT planning. Our AIRTP pipeline adheres toclinical guidelines and automates essential steps, including organ-at-risk(OAR) contouring, helper structure creation, beam setup, optimization, and planquality improvement, using AI integrated with RT planning software like VarianEclipse. Furthermore, a novel approach for determining optimization parametersto reproduce 3D dose distributions, i.e. a method to convert dose predictionsto deliverable treatment plans constrained by machine limitations is proposed.A comparative analysis of plan quality reveals that our automated pipelineproduces treatment plans of quality comparable to those generated manually,which traditionally require several hours of labor per plan. Committed topublic research, the first data release of our AIRTP pipeline includes ninecohorts covering head-and-neck and lung cancer sites to support an AAPM 2025challenge. To our best knowledge, this dataset features more than 10 timesnumber of plans compared to the largest existing well-curated public dataset.Repo: https://github.com/RiqiangGao/GDP-HMM_AAPMChallenge.