Abstract
BACKGROUND: The preoperative differential diagnosis of follicular thyroid neoplasms remains a major challenge in clinical practice. This study aimed to analyze the differences in ultrasonic characteristics between follicular thyroid carcinoma (FTC) and follicular thyroid adenoma (FTA), and establish a ultrasound-based risk stratification system (RSS) for follicular thyroid neoplasms. METHODS: This was a retrospective multi-center study. Data from patients who underwent surgical treatment and were definitively diagnosed with FTC or FTA based on postoperative pathology at each center between 2017 and 2024 were included. The subjects were randomly divided into a training set and a validation set at a ratio of 8:2. In the training cohort, an RSS tailored for follicular neoplasms (termed predictive score for follicular tumors, F-Score) was developed based on the multivariate logistic regression analysis. The diagnostic performance of the F-Score was subsequently assessed in the validation cohort, and compared with the commonly used Thyroid Imaging Reporting and Data Systems (TIRADSs). RESULTS: A total of 448 patients were included in this study. Multivariate logistic regression analysis showed that hypo/markedly hypoechoic, calcifications (any type), ill-defined/irregular margins, uneven/absent halo, and heterogeneous echotexture were independent risk factors for FTC (all P values <0.05). In the training set and the validation set, the area under the curve (AUC) of the F-Score for differentiating follicular neoplasms was 0.878 [95% confidence interval (CI): 0.842-0.914] and 0.871 (95% CI: 0.789-0.930), respectively. In both cohorts, the AUCs of F-Score were significantly higher than that of the TIRADS published by the American College of Radiology and the Chinese TIRADS (both P<0.01). Additionally, it performed significantly better than the follicular TIRADS (P=0.03) in the validation set. CONCLUSIONS: The ultrasound-based RSS can effectively differentiate between FTC and FTA, providing an excellent approach for the preoperative diagnosis of follicular thyroid neoplasms.