A simplified version of TI-RADS exhibits comparable diagnostic performance in malignancy risk stratification of thyroid nodules
Abstract Evaluation of thyroid focal lesions using thyroid imaging reporting and data system (TIRADS) has been proceeding for a decade, but there is no consensus on any version of TIRADS. The purpose of this study was to validate a recently launched simplified Chinese version of TIRADS, with a compare to the American College of Radiology version of TIRADS. A total of 1306 patients with 1389 thyroid nodules were reviewed and assessed according the two TIRADSs, and the histopathological results were taken as golden standard. The results showed there were 973 benign nodules and 416 malignant nodules. The highest accuracies(AUCs)of C-TIRADS 4C and ACR-TIRADS 5 were taken as the optimized cut-off values for diagnosis. The sensitivity, specificity, PPV, NPV and AUC by C-TIRADS 4C and ACR-TIRADS 5 for thyroid nodule evaluation were 87.39%, 89.92%, 75.00%, 95.38% and 0.89, and 85.58%, 91.88%, 81.84%, 93.71% and 0.89, respectively, (P > 0.05 for all). We concluded that C-TIRADS and ACT-TIRADS have very good diagnostic performance in differentiating malignant from benign thyroid nodules by each optimized cut-off value, and the diagnostic performance of the fewer parameters based C-TIRADS 4C is comparable to the multiple parameters based ACR-TIRADS 5.