ESTRO 2025 - Abstract Book

S3136

Physics - Inter-fraction motion management and offline adaptive radiotherapy

ESTRO 2025

Conclusion: This study explored decision support for daily adaptive MR-guided prostate radiotherapy to assist in determining the need for additional replan before beam-on. The decision support balances target coverage and the need for motion correction, while being straightforward to implement in a clinical setting.

Keywords: MR-Linac, intra-fraction motion, margin, prostate

References: [1] S. Klein, M. Staring, K. Murphy, M. A. Viergever, and J. P. Pluim, ‘elastix: a toolbox for intensity-based medical image registration’, IEEE Trans Med Imaging , 2010.

2393

Digital Poster Visibility, streak artefacts and impact on auto-segmentation of rectum of fiducial markers for image-guided radiotherapy of prostate cancer Barbora Hostová 1,2 , Vladimír Vojtek 1 , Pavol Dubinský 1,3 , Miroslav Olejár 1 , Michal Boldižár 1 , Martin Jasenčak 1,2 1 Department of Radiation Oncology, East Slovakia Institute of Oncology, Košice, Slovakia. 2 Faculty of Science, Pavol Jozef Šafárik University, Košice, Slovakia. 3 Faculty of Health, Catholic University, Ružomberok, Slovakia Purpose/Objective: This study aims to evaluate the radiographic visibility, streak artefacts and the impact on rectal auto-segmentation of three types of fiducial markers used in image-guided stereotactic body radiotherapy for prostate cancer. Material/Methods: Clinically used fiducial markers were evaluated in a cohort of 35 patients divided into 3 groups: 13 patients with BiomarC Gold markers (gold cylinder), 12 patients with BiomarC (barbell-shaped carbon) and 10 patients with BiomarC PRO (carbon rod) markers. Four fiducials of a single type were implanted into each patient’s prostate. The visibility of fiducials on kV X-ray images obtained using the ExacTrac Dynamic (ETD) system was analysed. The severity of streak artefacts on CT images was quantified using the streak index (SI). To assess the impact of streak artefact on auto-segmentation, rectum contours delineated on CT images by a radiation oncology expert (serving as the ground truth) were compared with contours generated by two auto-segmentation systems, Contour+ (MVision AI) and syngo.via (Siemens Helthineers). For this comparison, the average Hausdorff distance (aHD) and Dice coefficient were used. Results: On kV X-ray ETD images, barbell-shaped carbon fiducials were not visible and did not allow image registration, while both carbon rods and gold cylinders were detectable and both suitable for image registration and prostate tracking. The mean SI index (with standard deviation) for gold, barbell-shaped carbon and carbon rod fiducials were 171 (46.4), 29.3 (7.6) and 60.9 (17.6), respectively with significant difference (p<0.001) for each comparison. These results reflected visual assessment of CT images, where streak artefacts caused by gold markers were more prominent than carbon markers. The mean aHD and Dice coefficients for syngo.via and Contour+ are listed in Table 1 and Table 2. There was a significant difference in mean aHD and Dice coefficients between barbell-shaped carbon and carbon rod fiducials for syngo.via (p = 0.028 and 0.011, respectively).