ESTRO 2025 - Abstract Book

S3065

Physics - Image acquisition and processing

ESTRO 2025

Cohen’s-d value of 1.716 indicates a large effect size. Chi-Square analysis showed a significant association (statistic: 48.13,p≤0.0001,Cramér’s-V=0.366) between the reconstruction method and anatomical correctness. The segmented OARs showed a Dice score >0.95 for both reconstructions (greatest deviations for heart and right middle lung lobe). OAR-volume differences were 0.4±0.4% (range: spinal cord, 0.2% to right middle lung lobe, 1.1%). Differences in CT-values were <5HU. Conclusion: ZeeFree demonstrated significant improvements in image quality and anatomical correctness. The small to negligible differences observed in OARs confirm that anatomical integrity is maintained. In future studies, the impact of the algorithm on target volume definition and dose calculation needs to be evaluated. References: 1. Wasserthal, J., et al., TotalSegmentator: Robust Segmentation of 104 Anatomic Structures in CT Images. Radiol Artif Intell, 2023. 5 (5): p. e230024. 2. Szkitsak, J., et al., First clinical evaluation of breathing controlled four-dimensional computed tomography imaging. Phys Imaging Radiat Oncol, 2021. 20 : p. 56-61. 3. Werner, R., et al., Intelligent 4D CT sequence scanning (i4DCT): Concept and performance evaluation. Med Phys, 2019. 46 (8): p. 3462-3474. Keywords: artifact-correction, 4DCT, stack-artifacts

3844

Digital Poster Accelerated Diffusion-weighted MRI enabled by deep learning on a 1.5T Unity MR-Linac Prashant P Nair 1 , Yu Xiao 2,1 , Bastien Lecoeur 1 , Magali Nuixe 1 , Joan Chick 3 , Alison Tree 3 , Uwe Oelfke 1 , Mathew D Blackledge 1 , Andreas Wetscherek 1 1 Joint Department of Physics, The Institute of Cancer Rese and and The Royal Marsden NHS Foundation Trustarch, London, United Kingdom. 2 St. John's College, University of Oxford, Oxford, United Kingdom. 3 The Royal Marsden Hospital, The Institute of Cancer Research, London, United Kingdom Purpose/Objective: The apparent diffusion coefficient (ADC) is a potential biomarker for radiotherapy response assessment and could aid in daily treatment planning, particularly on hybrid MR-Linacs 1 but suffers from poor repeatability 2 and long