Comparison of ACR TI-RADS and ATA Ultrasound Risk Stratification for Thyroid Nodules – Data from a Tertiary Hospital in the United Arab Emirates
Abstract
Abstract Objective Accurate thyroid nodule risk stratification is critical to guide biopsy decisions and minimize unnecessary procedures. This study compared the diagnostic performance of the American Thyroid Association thyroid nodule risk stratification system (ATA-RSS) and American College of Radiology Thyroid Imaging Reporting and Data System (ACR TI-RADS) systems and assessed their independent predictive value for malignancy. Methods We retrospectively analyzed 557 patients with 693 thyroid nodules undergoing fine-needle aspiration biopsy (FNAB). Baseline characteristics, Bethesda cytology, and biopsy recommendations per ATA-RSS and TI-RADS were recorded. The primary outcome for statistical framework analysis was malignancy, defined using a composite reference standard: surgical histopathology when available ( n = 151, 21.8%) and Bethesda category VI in non-surgical cases. Indeterminate categories (Bethesda III, IV, and V) were retained in the dataset. Diagnostic performance metrics and multivariate logistic regression models adjusting for age, gender, and nodule size were evaluated. Results Mean age was 46.1 years; 78.2% were female. ATA-RSS recommended biopsy for 90.5% of nodules versus 53.7% by TI-RADS. Overall malignancy rate was 15.2%. ATA-RSS and TI-RADS showed excellent correlation ( ρ = 0.931), but ATA-RSS assigned higher risk categories more frequently. ATA-RSS demonstrated higher sensitivity (96.2 vs. 89.5%) but markedly lower specificity (10.5 vs. 52.7%), resulting in lower positive predictive value (PPV 16.1 vs. 25.3%). TI-RADS maintained superior specificity and PPV across subgroups. Regression analysis confirmed both systems as independent predictors of malignancy, with TI-RADS showing stronger discrimination (odds ratio [OR] = 23.2, 95%CI: 10.2–52.9, p <0.001) than ATA-RSS (OR = 6.0, 95%CI: 1.5–23.5, p = 0.01). Conclusion TI-RADS provides more balanced diagnostic performance and superior independent predictive value compared with ATA-RSS, reducing unnecessary biopsies while maintaining high sensitivity. This study contributes real-world data from the Middle East, where local comparison of ATA-RSS and ACR TI-RADS remains limited.