Diagnostic Performance of ACR TI-RADS and ATA Guidelines in the Prediction of Thyroid Malignancy: A Prospective Single Tertiary Center Study and Literature Review.

AIM

This study sought to compare two common risk stratification systems in terms of their diagnostic performance for the evaluation of thyroid malignancy.

METHODS

The American College of Radiology (ACR) Thyroid Imaging, Reporting and Data System (TI-RADS) and the American Thyroid Association (ATA) guidelines were compared among 571 thyroid nodules with definitive fine needle aspiration (FNA) cytology or postoperative histopathology. Ultrasound characteristics such as composition, echogenicity, shape, margin, size, and vascularity were assessed for each thyroid nodule. Diagnostic performance measures were determined and compared through receiver operating characteristic (ROC) curves, and decision curve analysis (DCA).

RESULTS

Of 571 nodules, 65 (11.4%) were malignant. The AUC, sensitivity, specificity, positive predictive value, and negative predictive value were 0.691, 49.2%, 84.9%, 29.6%, and 92.8% for ATA guideline, and 0.776, 72.3%, 79.2%, 30.9%, and 95.7%, for ACR TI-RADS, respectively. ACR TI-RADS was more sensitive ( = 0.003), while the ATA guideline was more specific ( < 0.001). DCA demonstrated that the ACR TI-RADS provided a greater net benefit than the ATA guideline. In addition, the net reduction in unnecessary biopsies is higher for ACR TI-RADS than ATA guidelines. The total number of indicated biopsies and unnecessary FNA rates were lower in ACR TI-RADS compared to ATA guideline (293 vs. 527 and 80.2 vs. 87.8). ACR TI-RADS presented no biopsy indication in seven malignant nodules (all categorized as TR2), whereas ATA guideline missed one. Hypoechogenicity was the most significant predictor of malignancy (OR = 8.34, 95% CI: 3.75-19.45), followed by a taller-than-wide shape (OR = 6.73, 95% CI: 3.07-14.77).

CONCLUSIONS

Our findings suggest that each system has particular advantages in the evaluation of thyroid nodules. ACR TI-RADS reduces unnecessary FNA rates, however, malignant nodules categorized as TR2 might be missed using this system. Further evaluation of this group of nodules using Doppler and other ultrasound modalities is recommended.