S1829
Physics - Dose prediction/calculation, optimisation and applications for photon and electron planning
ESTRO 2026
Purpose/Objective: Current treatment guidelines recommend defining CTV and PTV margins to account for microscopic infiltration and setup errors, respectively [1]. Both margins are defined in a binary way and do not explicitly account for the probabilistic nature of the uncertainties. Therefore, they might be overly conservative. Additionally, the sequential combination of these margins is not statistically well-founded, making it hard to determine total patient coverage. This project implements a treatment optimization approach that explicitly accounts for probabilistic tumor presence and geometric errors. A Clinical Target Map (CTM), based on an infiltration model as described by Buti et al. [2], was combined with expected value robust optimization for setup errors. For evaluation, the microscopic infiltration was modelled by different isotropic expansions of the GTV with their respective probability of being tumorous. Material/Methods: The treatment plans for five non-small cell lung cancer patients were designed for VMAT treatment on the Halcyon™ by Varian. All plans were evaluated and compared based on a probabilistic evaluation of over 300 uncertainty scenarios (10 infiltration scenarios combined with 33 setup error scenarios). The infiltration scenarios were inferred from a truncated Gaussian with mean 3.4mm and standard deviation 2.8mm, while the setup error was assumed to be distributed according to a Gaussian with mean 0mm and standard deviation 3mm. The evaluation returned mean and median values for each of the prescribed clinical goals, together with a passing rate representing the total probability mass of scenarios passing the clinical goal. All treatment optimizations and evaluations were executed in the RayStation 2023b [3] treatment planning system (research license) using the scripting environment. Results: For all patients, the D98%>95% prescribed dose was achieved with more than 90% probability. The OARs experienced a reduction in dose, especially the heart and lungs, as both lowered their mean dose by on average 2Gy (see Figure 1 for target and lung evaluation results). The probabilistic evaluation provided a more comprehensible, statistics-based (e.g., passing rate) assessment of patient population coverage, compared to conventional plan evaluation metrics (e.g., CTV dose-volume metrics, etc.).
Conclusion: The probabilistic planning strategy proposed in this project encourages further exploration of a probabilistic framework for radiotherapy treatment planning. Probabilistic evaluation enables probabilistic objectives and constraints, which opens the path to statistically better-founded and more realistic treatment optimization strategies. Considering clinically well-accepted target coverage objectives, significant OAR sparing can be achieved. References: [1] “The International Commission on Radiation Units and Measurements,” J. Int. Comm. Radiat. Units Meas., vol. 10, no. 1, p. NP, Apr. 2010, doi: 10.1093/jicru/ndq001.[2] G. Buti, K. Souris, A. Maria Barragán Montero, J. Aldo Lee, and E. Sterpin, “Introducing a probabilistic definition of the target in a robust treatment planning framework,” Phys. Med. Biol., vol. 66, no. 15, p. 155008, Aug. 2021, doi: 10.1088/1361-6560/ac1265.[3] “RaySearch releases RayStation 2023B,” RaySearch Laboratories. Accessed: Oct. 21, 2025. [Online]. Available: https://www.raysearchlabs.com/media/press- releases/2023/raysearch-releases-raystation-2023b/ Keywords: probabilistic planning, clinical target map Digital Poster Highlight 1338 Evaluation of automated dosimetry performance in RayStation for Head and Neck treatment Sofia Jebbari 1 , Soukaina Moujahid 1 , Toussaint Tarwobgo 2 , Dounia KAMAL 1,3 , Abdessalam Bouk 1 , Adam Shulman 4 1 Radiotherapy, Ryad Oncologia Clinic, Casablanca, Morocco. 2 Radiotherapy, Hospital Center of Nganda, Kinshasa, Congo, the Democratic Republic of the. 3 Laboratory of Immunogenetics and Human Pathologies, Faculty of Medicine and Pharmacy, Casablanca, Morocco. 4 Radiotherapy, Rayos Contra Cancer, Nashville, USA Purpose/Objective: The purpose of this study is to evaluate the performance of the Shulman optimization method in combination with an automated radiotherapy planning workflow in RayStation Planning System, by comparing it with conventional manual planning.The automation involves the development of scripts for
Made with FlippingBook - Share PDF online