Deep learning to diagnose Hashimoto’s thyroiditis from sonographic images

• Published in: Nature communications Vol. 13; no. 1; p. 3759
• Main Authors: Zhang, Qiang, Zhang, Sheng, Pan, Yi, Sun, Lin, Li, Jianxin, Qiao, Yu, Zhao, Jing, Wang, Xiaoqing, Feng, Yixing, Zhao, Yanhui, Zheng, Zhiming, Yang, Xiangming, Liu, Lixia, Qin, Chunxin, Zhao, Ke, Liu, Xiaonan, Li, Caixia, Zhang, Liuyang, Yang, Chunrui, Zhuo, Na, Zhang, Hong, Liu, Jie, Gao, Jinglei, Di, Xiaoling, Meng, Fanbo, Zhang, Linlei, Wang, Yuxuan, Duan, Yuansheng, Shen, Hongru, Li, Yang, Yang, Meng, Yang, Yichen, Xin, Xiaojie, Wei, Xi, Zhou, Xuan, Jin, Rui, Zhang, Lun, Wang, Xudong, Song, Fengju, Zheng, Xiangqian, Gao, Ming, Chen, Kexin, Li, Xiangchun
• Format: Journal Article
• Language: English
• Published: London Nature Publishing Group UK 29.06.2022; Nature Publishing Group; Nature Portfolio
• Subjects: 631/114/1305; 692/163/2743/1841; Datasets; Deep learning; Diagnosis; Humanities and Social Sciences; Hypothyroidism; Medical imaging; multidisciplinary; Science; Science (multidisciplinary); Thyroid; Thyroiditis; Video data
• ISSN: 2041-1723; 2041-1723
• DOI: 10.1038/s41467-022-31449-3

Hashimoto’s thyroiditis (HT) is the main cause of hypothyroidism. We develop a deep learning model called HTNet for diagnosis of HT by training on 106,513 thyroid ultrasound images from 17,934 patients and test its performance on 5051 patients from 2 datasets of static images and 1 dataset of video data. HTNet achieves an area under the receiver operating curve (AUC) of 0.905 (95% CI: 0.894 to 0.915), 0.888 (0.836–0.939) and 0.895 (0.862–0.927). HTNet exceeds radiologists’ performance on accuracy (83.2% versus 79.8%; binomial test, p  < 0.001) and sensitivity (82.6% versus 68.1%; p  < 0.001). By integrating serologic markers with imaging data, the performance of HTNet was significantly and marginally improved on the video (AUC, 0.949 versus 0.888; DeLong’s test, p  = 0.004) and static-image (AUC, 0.914 versus 0.901; p  = 0.08) testing sets, respectively. HTNet may be helpful as a tool for the management of HT. Hashimoto’s thyroiditis (HT) is the main cause of hypothyroidism. Here the authors develop a deep learning model for diagnosis of HT on a large multi-site dataset including image and video data.