Related Experiment Video
Updated: Jul 26, 2025
A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images
Published on: April 21, 2023
Generalizability and Diagnostic Performance of AI Models for Thyroid US
WenWen Xu1, XiaoHong Jia1, ZiHan Mei1
1From the Department of Ultrasound, Ruijin Hospital, Shanghai Jiaotong University School of Medicine, 197 Ruijin Er Road, 200025, Shanghai, China (W.W.X., X.H.J., Z.H.M., W.W.Z., T.L., H.T.Z., Y.J.D., J.Q.Z.); Department of Scientific Research, Shanghai Aitrox Technology Corporation Limited, Shanghai, China (X.L.G., Y.L., C.C.F., K.Y.Z., Q.F., C.H.); Department of Ultrasound, The First Affiliated Hospital of Zhengzhou University, Zhengzhou, China (R.F.Z.); Department of Medical Ultrasound, Affiliated Hospital of Guizhou Medical University, Guiyang, China (Y.G., X.C.); Department of Medical Ultrasound, Yunnan Cancer Hospital & The Third Affiliated Hospital of Kunming Medical University, Kunming, China (X.M.L.); Department of Ultrasound, Yunnan Kungang Hospital, The Seventh Affiliated Hospital of Dali University, Anning, China (N.L.); Department of Ultrasound, Affiliated Hospital of Yan'an University, Yan'an, China (B.Y.B.); Department of Ultrasound, Tangdu Hospital, Fourth Military Medical University, Xi'an, China (Q.Y.L.); Department of Ultrasound, Shanxi Provincial People's Hospital, Taiyuan, China (J.P.Y.); Department of Ultrasound, Traditional Chinese Medical Hospital of Xinjiang Uygur Autonomous Region, Urumqi, Xinjiang Uygur Autonomous Region, China (H.Z.); Department of Ultrasound, Gansu Provincial Cancer Hospital, Lanzhou, China (L.G.); Department of Ultrasound, Jilin Central General Hospital, Jilin, China (B.G.); and College of Health Science and Technology, Shanghai Jiaotong University School of Medicine, Shanghai, China (J.Q.Z.).
This study developed artificial intelligence (AI) models for thyroid nodule ultrasound (US) assessment, demonstrating high diagnostic performance and improved radiologist accuracy in thyroid cancer detection.
Area of Science:
- Radiology
- Artificial Intelligence
- Medical Imaging
Background:
- Artificial intelligence (AI) models show promise in improving ultrasound (US) assessment of thyroid nodules.
- Limited generalizability of current AI models hinders their widespread clinical application.
Purpose of the Study:
- To develop robust AI models for thyroid nodule segmentation and classification using diverse nationwide datasets.
- To evaluate the impact of these AI models on diagnostic performance in thyroid cancer detection.
Main Methods:
- Retrospective analysis of 10,023 patients with pathologically confirmed thyroid nodules from 208 hospitals across China.
- Development of AI models for detection, segmentation, and classification using data from 12 vendors.
- Performance evaluation using precision, recall, Dice coefficient, and area under the receiver operating characteristic curve (AUC).
- Comparison of radiologist performance with and without AI assistance.
Main Results:
- AI models achieved high performance: detection (precision 0.98), segmentation (Dice 0.86), and classification (AUC 0.90).
- Nationwide and mixed-vendor trained models showed superior performance (Dice 0.91, AUC 0.98).
- AI models outperformed senior and junior radiologists; rule-based AI assistance significantly improved radiologist diagnostic accuracy (P < .05).
Conclusions:
- AI models developed from diverse datasets demonstrate high diagnostic performance for thyroid US in the Chinese population.
- Rule-based AI assistance enhances radiologist capabilities in diagnosing thyroid cancer.
- These findings support the integration of AI into clinical practice for improved thyroid nodule assessment.
