ESTRO 2023 - Abstract Book

S184

Saturday 13 May

ESTRO 2023

relation between T and known BODY deformations averaged over the seven proton patients. A known deformation of 8 mm and larger within the patient group showed a significant difference compared to no deformation between the T values.

Conclusion Breast surface anatomy deformation is highly corelated with target dosimetrical indicators and therefore is potentially a reliable predictor for the need of repeat CT acquisition and plan adaptation in proton breast cancer patients. Future work is warranted to determine accurate thresholds in daily clinical workflow. PD-0249 Automatic deep learning treatment planning of gradient-optimized match fields for PBS proton therapy G.A. Helgason 1 , M. Arvola 2 , D. Maes 3 , F. Löfman 2 , L. Glimelius 1 1 RaySearch Laboratories, Physics department, Stockholm, Sweden; 2 RaySearch Laboratories, Machine learning department, Stockholm, Sweden; 3 University of Washington, School of Medicine, Department of Radiation Oncology, Seattle, USA Purpose or Objective Gradient optimization is a technique used for beam matching in proton PBS and is often used for large targets that extend beyond the maximum field or to drive OAR sparing. In this technique, dosimetric gradients between match fields are designed to be sufficiently gradual to ensure plan robustness with respect to setup uncertainties. This work investigates the use of a deep learning model to infer beam-specific dose distributions to automate robust PBS treatment planning of the prostate and pelvic nodes using gradient-optimized field matching. Materials and Methods A deep learning model has been trained to predict individual beam dose distributions based on binary volumes of patient geometries, such as target and OAR. The architecture used was a neural network based on the 3D U-Net. The model was trained on a data set consisting of 30 manually planned prostate patients with pelvic nodes, where 25 were used for training and five for validation. An additional set of five patients was used to test the model by performing automatic planning in a research version of RayStation 11B. It consisted of deep learning prediction of individual beam doses and robust mimic optimization where optimization function penalties are based on quadratic differences to the predicted beam doses. The full optimization problem combines dose level-based optimization functions and mimic optimization functions. All plans have been evaluated with respect to OAR sparing and robust target coverage (3 mm setup uncertainty along the inferior-superior and anterior-posterior axes and 3% range uncertainty in 16 scenarios).