ESTRO 2022 - Abstract Book

S471

Abstract book

ESTRO 2022

dataset consisting of 15 CT and 7 CBCT volumes. Evaluation was performed using the Dice similarity coefficient (DSC) and the 95 th percentile of the Hausdorff distance (HD95). Results Mean DSC and HD95 values for the trained models are summarized in table 1 . Our improved, domain adapting model, resulted in improved mean DSC and HD95 values for all evaluated CBCT structures, when compared to the baseline model (segmentation examples shown in figure 1 ). For target and rectum e.g., mean DSC values could be increased from (mean ± sd) 0.81 ± 0.05 and 0.80 ± 0.04, to 0.86 ± 0.03 and 0.87 ± 0.03, respectively. Both models resulted in similar, acceptable, CT segmentation performance.

Conclusion A model trained with a combination of standard, additional, and domain specific augmentations demonstrated improved CBCT segmentation performance, while maintaining CT segmentation quality. The proposed domain adaptive training strategy indicates that multipurpose segmentation models can be improved, and potentially benefit adaptive prostate radiation therapy treatments. The impact of each individual augmentation technique will be explored further in future projects.

MO-0547 A federated learning IT-infrastructure to support the Dutch model-based approach for proton therapy

P. Kalendralis 1 , M. Sloep 1 , J. Snel 2 , N.M. George 1 , M. Veening 3 , A. Van der Schaaf 3 , J.A. Langendijk 4 , A. Dekker 1 , J. Van Soest 5 , R. Fijten 1 1 Department of Radiation Oncology (Maastro), GROW School for Oncology, Maastricht University Medical Centre+, Maastricht, The Netherlands, Department of Radiation Oncology (Maastro), Maastricht, The Netherlands; 2 Brightlands Institute of Smart Society (BISS), Faculty of Science and Engineering, Maastricht University, Heerlen, The Netherlands, Brightlands Institute of Smart Society (BISS), , Heerlen, The Netherlands; 3 Department of Radiation Oncology, University of Groningen, University Medical Centre Groningen. Groningen, The Netherlands, Department of Radiation Oncology, Groningen, The Netherlands; 4 Department of Radiation Oncology, University of Groningen, University Medical Centre Groningen. Groningen, The Netherlands, Department of Radiation Oncology, University , Groningen, The Netherlands;