Retrospective study from a single center to comparison of diagnostic value of three thyroid imaging reporting and data systems in Bethesda III/IV thyroid nodules.

To compare the diagnostic value of ACR Thyroid Imaging Reporting and Data System (TI-RADS), K-TIRADS, and C-TIRADS in Bethesda III/IV thyroid nodules. This single-center retrospective study classified 80 Bethesda stage III/IV thyroid nodules from 80 patients between January 2020 and July 2023 according to three different systems. Diagnostic performance was evaluated using receiver operating characteristic curves, with histopathological diagnosis serving as the reference standard. Overall, 41/80 (51.2%) nodules were malignant and 39/80 (48.8%) were benign. The malignancy rates for Bethesda type III and IV nodules were 50.7% and 55.6%, respectively. The malignancy risk in thyroid nodules increased with higher TI-RADS categories (P<0.001). Optimal cutoff values for ACR-, K-, and C-TIRADS were categories 5, 5, and 4C, respectively. Area under the curve (AUC) for ACR-, K-, and C-TIRADS was 0.782, 0.767, and 0.842, respectively, with C-TIRADS showing a significantly higher AUC than ACR-TIRADS and K-TIRADS (all P<0.05). C-TIRADS demonstrated the highest sensitivity, accuracy, and positive predictive value, whereas ACR TI-RADS showed the highest specificity and negative predictive value. Furthermore, the AUC, sensitivity, specificity, and accuracy of ACR TI-RADS, K-TIRADS, and C-TIRADS were higher in nodules >1 cm than in those ≤ 1 cm." All three TI-RADS systems have diagnostic value in differentiating benign from malignant Bethesda III/IV nodules, With C-TIRADS showing the highest area under the curve(AUC), suggesting its potential utility in clinical evaluation and management of such nodules, particularly in Chinese populations.