ESTRO 2024 - Abstract Book

S3138

Physics - Autosegmentation

ESTRO 2024

Conclusion:

In this study, we first evaluated the accuracy of a DL-based method to automatically detect and label ten regions of interest in VFSS. Then, we proposed a novel approach to analyze and characterize dysphagia in HNC patients exploiting the information of the labeled VFSS, showing statistically significant differences with a control group of healthy individuals. This work smooths the way for further objective characterization of dysphagia from VFSS.

Keywords: DL, videofluoroscopy swallowing studies, dysphagia

2812

Digital Poster

Construction of clinically meaningful loss function for improved medical image auto-segmentation

Yizheng Chen, Michael Francis Gensheimer, Quynh-Thu Le, Bin Han, Lei Xing