ESTRO 2025 - Abstract Book

S3276

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

ESTRO 2025

2651

Poster Discussion Prostate CTV tracking and intra-fraction motion evaluation on 2D-cine-MRI in MRgRT

Nicole Wernlein 1 , Ivan Ćorić 1 , Alexander Köhler 1 , Simon Böke 2 , Cihan Gani 2 , Maximilian Niyazi 2 , Daniela Thorwarth 1,3 1 Section for Biomedical Physics, Department of Radiation Oncology, University Hospital Tübingen, Tübingen, Germany. 2 Department of Radiation Oncology, University Hospital Tübingen, Tübingen, Germany. 3 Cluster of Excellence “Machine Learning”, Eberhard Karls University Tübingen, Tübingen, Germany Purpose/Objective: Intra-fraction tracking of prostate movement during radiation therapy is essential for accurately administering ultra hypo-fractionated (UHFX) treatments while preserving surrounding healthy tissue. MR-guided radiotherapy (MRgRT) offers a promising solution to these challenges by enabling real-time MR imaging and motion control. This work investigates residual prostate CTV margins to compensate intra-fraction motion during UHFX-MRgRT by considering the option of a baseline-shift correction. Material/Methods: The study included 15 prostate cancer patients treated between 02/2021 and 03/2024 in 5-7 fractions at the 1.5T MR-Linac, resulting in a total of 207.764 2D-cine-MRI frames. 2D-cine-MRIs were acquired in the sagittal and transversal plane during UHFX-MRgRT delivery. The segmentation of the prostate contours was performed by an in house trained neural network (nnUNetV2) with an optimized inference pipeline for real-time segmentation. The model was pre-trained and evaluated on 260 independent cine-MRIs and label maps. Intra-fraction motion was assessed by determining the center of area (COA) shift of the prostate contour on sagittal and transversal cine-MRIs for anterior-posterior (AP), superior-inferior (SI) and right-left (RL) directions. First, the intra-fraction motion was calculated and assessed cumulatively, starting from the pre-treatment reference image without applying a baseline shift correction. In a next step, a baseline-shift correction was simulated for displacement vectors d = √x²+y²+z² > 3 mm by repositioning the prostate COA to zero for all axes. 5 th - and 95 th -Percentiles (P5%/P95%) were used to determine the necessary margins to compensate for intra-fraction motion with baseline-shift correction. Results: Intra-fraction motion analysis without applying a baseline-shift correction showed for AP-movement P5%/P95% of - 3.75 mm/2.66 mm over all patients and fractions. SI- and RL-motion P5%/P95% were found to be -3.65 mm/1.81 mm and -1.44 mm/1.21 mm. Using baseline-shift correction the displacement was reduced to P5%/P95% -1.51 mm/1.64 mm for AP-, -1.53 mm/1.56 mm for SI- and -0.68 mm/0.71 mm for RL-directions.