ESTRO 2025 - Abstract Book

S2968

Physics - Image acquisition and processing

ESTRO 2025

Fig. 2 Comparison of the pCT and sCT (green) to the intensity matched CBCT (purple).

Conclusion: In this study, we have developed a robust sCT generation method via DIR for patients with HNC. Our method met the AAPM Task Group 132 guidelines, with the TRE below the maximum voxel dimension (3 mm). 2

Keywords: deformable image registration, synthetic CT, CBCT

References: 1. Thirion, J. P. (1998). Image matching as a diffusion process: an analogy with Maxwell's demons. Medical image analysis , 2 (3), 243-260. 2. Brock, K.K., et al. (2017). Use of image registration and fusion algorithms and techniques in radiotherapy: Report of the AAPM Radiation Therapy Committee Task Group No. 132. Medical physics , 44 (7), e43-e76.

1255

Digital Poster Evaluation of synthetic CT generated from MRI by using Vision transformer based cGAN model for proton therapy planning in nasopharyngeal carcinoma Hongdong Liu, Shouliang Ding, Fanghua Li, Honghu Song, Guangyu Wang, Wenzhao Sun, Li Chen, Xiaoyan Huang Department of Radiation Oncology, Sun Yat-sen University Cancer Center, Guangzhou, China Purpose/Objective: To investigate the dosimetric accuracy of synthetic CT (sCT) from MRI generated using Vision transformer (VIT) versus convolution neural network (CNN) based conditional GAN (cGAN) model in proton therapy treatment planning of nasopharyngeal carcinoma (NPC). Material/Methods: Paired pretreatment T2w MRI and planning CT were retrospectively collected from 131 NPC patients in this study. For each patient, both the MRI and CT images were scanned on the same day. Preprocessing such as rigid image registration, N4 bias field correction, image cropping, image resampling, excluding the couch and positioning