ESTRO 2024 - Abstract Book

S3029

Physics - Autosegmentation

ESTRO 2024

Conclusion:

Even though HyperSight CBCT shows better image quality, it did not lead to an improved contour correction time and autosegmentation performance in the pelvic region compared to conventional Ethos CBCT. Only for the seminal vesicles, a significantly improved interobserver agreement was observed on HyperSight CBCT. A possible explanation for the small differences might be that the current AI-based clinical Ethos model was not trained on HyperSight CBCT images. Nevertheless, HyperSight CBCT clearly enhanced user confidence in organ delineation for prostate, seminal vesicles and rectum.

Keywords: Online adaptive radiotherapy, Hypersight, Ethos

905

Digital Poster

Adaptive auto-segmentation for MRI-guided online adaptive radiotherapy of cervical carcinoma

Shouliang Ding

Sun Yat-sen University Cancer Center, Radiation Oncology, Guangzhou, China