RESUMO
OBJECTIVE: To compare 2 ultrasound-based risk stratification systems in malignancy risk assessment of thyroid nodules and the clinical applicability of these guidelines in Brazil. METHODS: We retrospectively reviewed the ultrasound findings of 314 patients (473 thyroid nodules) who underwent fine-needle aspiration (FNA) biopsy and/or surgery between February 2018 and March 2019. All nodules were classified using 2 systems: the Thyroid Imaging, Reporting, and Data System (TIRADS) of the American College of Radiology (ACR-TIRADS) and the TIRADS of the European Thyroid Association (EU-TIRADS). Both risk stratification systems were analyzed. We identified the diagnostic predictive values that yielded optimal sensitivity, specificity, positive predictive value, negative predictive value, and accuracy. RESULTS: Of the 473 nodules, all underwent FNA, and histopathology was performed for 332 nodules. The agreement between the ACR-TIRADS and EU-TIRADS results and that between cytology and histopathology findings was 92.6% (kappa = 0.84) and 86.7% (kappa = 0.73), respectively. The area under the curve for the ACR-TIRADS and EU-TIRADS was 0.871 and 0.828, respectively (P < .001). The EU-TIRADS had the best sensitivity and negative predictive value, whereas the ACR-TIRADS had the best specificity, positive predictive value, and accuracy. Of the 473 nodules studied, only 158 (33.4%) followed the FNA size criteria suggested by the ACR-TIRADS. CONCLUSION: ACR-TIRADS and EU-TIRADS had good diagnostic performances. However, most aspirated nodules did not follow the TIRADS indication; thus, the overuse of FNA as a diagnostic tool was observed.
