ESTRO 2025 - Abstract Book

S4379

RTT - Treatment planning, OAR and target definitions

ESTRO 2025

Conclusion: This study aims to assess the dosimetric accuracy of the MRCAT Brain product and demonstrate its clinical feasibility for MRI-only radiotherapy treatment planning for the brain. MRCAT. Overall, it demonstrates dosimetric differences within a clinically acceptable level of ±2%. In some cases, the MRCAT failed to generate acceptable sCT, where the inclusion criteria for MRCAT generation were not met, which will be individually assessed for exclusion during the commissioning stage.

Keywords: MRI-only pathway, Brain, Synthetic-CT

3434

Digital Poster Utilization of three AI-driven anatomical organ delineation software and evaluation of their efficacy in radiotherapy treatment planning

Gergely Bungyi, Tibor Major, Zoltan Takacsi-Nagy, Tamas Pocza Radiotherapy, National Oncology Institute, Budapest, Hungary

Purpose/Objective: To evaluate the quality of three contouring software that employs artificial intelligence algorithms, and to define the organs that are delineated with the highest consistency, and those with significant variations in segmentation made by the three software. Material/Methods: OAR segmentations were performed by three contouring software (MVision, ART-Plan, Limbus) in four anatomical regions on CT series of ninety-three patients. MVision contours were selected as references, and contours created by the other two systems were compared to them based on spatial similarity (dice similarity coefficient, DSC) and conformity (conformal index, COIN) indices. We ranked the indices, defining the organs contoured by the highest and lowest consistency by the three software. Results: The order of DSC values is presented in Figure 1. For twenty-five organs out of thirty-eight, the DSC reached 0,8 on average. The order of the organs with the best agreement: lungs, brain, liver, spleen, stomach, heart, eyes, mandible, kidneys, spinal cord, breasts, and bladder. The largest deviations were found at small volume organs: thyroid, chiasma, left anterior descending artery, and pituitary. Large variations were found in the prostate due to its poor visualization on CT images. The dissimilarities in trachea delineations were due to different interpretations of the anatomical boundaries by the three software (Figure 2.).