ABSTRACT
BACKGROUND: In the present study, the first aim was to compare the accuracy of three guidelines in the diagnosis of thyroid nodule malignancy. The second purpose was to find sonographic features potentially associated with the risk of malignancy. METHODS: In this cross-sectional study, we prospectively recruited patients referred with a diagnosis of thyroid nodule (≥ 1 cm) for fine-needle aspiration (FNA). Sonographic features were recorded and scored according to the American Thyroid Association (ATA-2015), the American College of Radiology-Thyroid Imaging Reporting and Data System (ACR-TIRADS), and the Korean TIRADS (K-TIRADS). FNA was conducted and cytological findings were reported. RESULTS: A total of 984 thyroid nodules were ultimately included, of which 144 (14.6%) were malignant and 840 (85.4%) were benign. The accuracy of ACR-TIRADS categories TR5 and TR4/5 was 88.3% and 69.3%, respectively. This rate for ATA-2015 classes High suspicion and Intermediate suspicion/High suspicion was 87.9% and 80.4%, respectively. For K-TIRADS classes 5 and 4/5, the diagnostic accuracy was 88.0% and 80.6%, respectively. The rate of unnecessary FNA was highest with ATA-2015 and K-TIRADS guidelines (53.9% and 53.7%, respectively), followed by ACR-TIRADS (32.0%). Significant direct associations were observed between malignancy and hypoechogenicity (odds ratio [OR] 5.78), fine calcification (OR = 6.7), rim calcification (OR = 2.56), ill-defined margin (OR = 3.31), and irregular margin (OR = 6.95). CONCLUSIONS: There are different strengths of ACR-TIRADS, K-TIRADS, and ATA-2015 guidelines in the prediction of malignant thyroid nodules, and clinicians and radiologists should consider these differences in the management of thyroid nodules.
