ESTRO 2025 - Abstract Book

S2716

Physics - Dose prediction, optimisation and applications of photon and electron planning

ESTRO 2025

36

Digital Poster Personalized radiotherapy planning tool with AI-guided optimization for patients with head and neck cancer Christian Velten 1,2 , Michael Bowers 3 , Patrik Brodin 1,2 , William Martin 1 , Rafi Kabarriti 1 , Madhur K Garg 1 , Julie Shade 3 , Todd McNutt 3,4 , Wolfgang A. Tomé 1,2 1 Radiation Oncology, Montefiore Medical Center, Bronx, USA. 2 Institute for Onco-Physics, Albert Einstein College of Medicine, Bronx, USA. 3 Oncospace, Oncospace, Inc., Baltimore, USA. 4 Radiation Oncology, Johns Hopkins University, Baltimore, USA

Purpose/Objective: Development of an automated planning tool utilizing AI generated patient-specific dose-volume histogram predictions for rapid H&N plan generation.

Material/Methods: Planning best-practices were developed, generalized, and subsequently implemented in C# using the Varian Eclipse API to automatically generate plans using the TPS’s dose calculation algorithms. To personalize treatment planning, organs-at-risk (OAR) dose-volume optimization objectives were automatically requested from a cloud-based AI DVH prediction algorithm (Oncospace, Baltimore, MD) to which CTs & RT Structures were previously exported followed by selection of a H&N treatment protocol. The tool was applied to the 2023 AAMD planning challenge’s H&N case (4-level SIB: 63/60/57/54Gy/33fx) available from Elekta-ProKnow, generating a 15-field IMRT plan on a Varian Halcyon. It was additionally benchmarked against 16 institutional plans generated by senior dosimetrists where the prescriptions were 69.96/59.4/54.12Gy/33fx. Plans were generated in batch mode for a Varian TrueBeam (TDS) and Halcyon (RDS), each with three techniques (11-/15-field IMRT, VMAT). Results: Times for plan generation (excluding protocol selection and setup) were as low as 2min for IMRT and up to 15min with VMAT using GPU-accelerated optimization and dose calculation. The automatically generated plan scored 132.4/150 total points on the 2023 AAMD plan challenge, above the median score of all submitted plans (130). On 18/23 evaluation metrics the plan’s scores were in the good or ideal category. Comparisons with all submitted plans are shown in Figure 1 for three metrics and the total score.