ESTRO 2025 - Abstract Book

S3295

Physics - Intra-fraction motion management and real-time adaptive radiotherapy

ESTRO 2025

Conclusion: This analysis shows consistency between the 2D MRI tracking-based method and the spatially more detailed 3D-MRI registration approach with rotational components, with a mean median difference of less than 0.7 mm. However, some treatments required rotational registrations or had discrepancies of up to 3 mm in one direction. While these findings support the reliability of the current approach, further investigation is needed to determine if the individual discrepancies are due to the simplified linear motion extrapolation, which is needed to bridge the time differences between the 2D cine and 3D MRIs, or due to patient relaxation after treatment.

Keywords: MRgRT, prostate, accuracy

3058

Digital Poster An evaluation of deformable image registration impact on contour deformation and 4D dose calculations in clinical hemithorax proton therapy Martijn S Feenstra, Pietro Pisciotta, Daniele Cannavò, Erik W Korevaar, Bojan Štrbac, Sabine Visser, Stefan Both Department of Radiation Oncology, University of Groningen, Groningen, Netherlands Purpose/Objective: Deformable image registration (DIR) is extensively used in radiotherapy, but no standard methods exist for evaluating its accuracy on clinical patient datasets. This research investigates the accuracy of DIRs among 4DCT (breathing) phases within the RayStation treatment planning system and its impact on 4D dose calculations for proton therapy in thoracic cancer patients. Material/Methods: Forty thoracic cancer patients (20 lung, 20 esophageal) treated with protons were analyzed. DIRs were generated and evaluated on 4DCT datasets containing ten phases, where all 4DCT phases were matched to the 50% phase