Related Experiment Video
Updated: Jul 10, 2025
A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images
Published on: April 21, 2023
US Risk Stratification System for Follicular Thyroid Neoplasms.
Jianming Li1, Chao Li1, XiaoHui Zhou1
1From the Department of Interventional Ultrasound, Fifth Medical Center of Chinese PLA General Hospital, 28 Fuxing Road, Beijing 100853, China (J.L., J.Y., P.L.); Department of Ultrasound, The First Affiliated Hospital of Henan University of CM, Henan, China (C.L.); Department of Ultrasound Diagnostics, The Affiliated Changsha Central Hospital, Hengyang Medical School, University of South China, Hunan, China (X.Z.); Department of Ultrasound, Peking University Third Hospital, Beijing, China (J.H.); Department of Ultrasound, Beijing Friendship Hospital, Capital Medical University, Beijing, China (P.Y.); Department of Otolaryngology Head and Neck Surgery, Chinese PLA General Hospital, Beijing, China (Y.C.); Department of Ultrasound, Tianjin Medical University General Hospital, Tianjin, China (H.Z.); Department of Otolaryngology-Head & Neck Surgery, The Second Affiliated Hospital of Guilin Medical University, Guangxi, China (R.H.); Department of Ultrasound, Traditional Chinese Medical Hospital of Xinjiang, Xinjiang, China (Y.M.); Department of Pathology, First Medical Center of Chinese PLA General Hospital, Beijing, China (X.G.); and Department of Pathology, Affiliated Hospital of Hebei Engineering University, Hebei, China (Y.Z.).
A new ultrasound risk stratification system (F-TI-RADS) effectively differentiates follicular thyroid adenoma from follicular thyroid carcinoma. This novel system shows improved performance over existing methods for preoperative assessment of follicular thyroid neoplasms.
Area of Science:
- Radiology
- Oncology
- Endocrinology
Background:
- Preoperative assessment of follicular thyroid neoplasms presents challenges with current ultrasound risk stratification systems (RSSs) designed for papillary thyroid neoplasms.
- Distinguishing between follicular thyroid adenoma (FTA) and follicular thyroid carcinoma (FTC) preoperatively is crucial for patient management.
Purpose of the Study:
- To develop a novel ultrasound (US) feature-based RSS for differentiating FTA from FTC in biopsy-proven follicular neoplasms.
- To compare the performance of the new RSS against existing systems.
Main Methods:
- Retrospective multicenter study involving adult patients with follicular thyroid neoplasms.
- Development of a prediction model and RSS (F-TI-RADS) based on qualitative US features using logistic regression.
- Comparison of F-TI-RADS with ACR TI-RADS, K-TI-RADS, and C-TI-RADS using area under the receiver operating characteristic curve (AUC) in a validation dataset.
Main Results:
- The F-TI-RADS demonstrated superior performance in differentiating FTA from FTC, with an AUC of 0.81.
- F-TI-RADS outperformed ACR TI-RADS (AUC, 0.74), K-TI-RADS (AUC, 0.69), and C-TI-RADS (AUC, 0.68) in the validation dataset.
- Statistical significance was observed in the performance comparison (P < .05 for all comparisons).
Conclusions:
- The developed F-TI-RADS is a more effective tool for differentiating follicular thyroid carcinoma from follicular thyroid adenoma compared to existing RSSs.
- This new system has the potential to improve preoperative risk stratification for follicular thyroid neoplasms.
