Concordance Between TI-RADS and Bethesda System for Reporting Thyroid Cytopathology and Their Correlation with Histopathology

Abstract

Background: Thyroid nodules are increasingly detected with high-resolution ultrasonography (HRUS). Although most are benign, accurate malignancy risk stratification is essential to avoid unnecessary procedures. Assessing alignment between ultrasound and cytology-based classification systems may improve overall decision-making.

Objective: To assess concordance between ACR TI-RADS and revised 2017 Bethesda System (TBSRTC), compare their diagnostic performance against histopathology, and evaluate the value of combined interpretation.

Methods: This prospective observational study included 200 patients with thyroid nodules at Nizam's Institute of Medical Sciences, India. Nodules were assigned paired ACR TI-RADS categories (TR2–TR5) on HRUS and FNAC Bethesda categories II–VI. Bethesda I and incomplete paired data were excluded. Histopathology, supplemented by immunohistochemistry when indicated, served as the reference standard. Concordance and diagnostic indices (sensitivity, specificity, PPV, NPV, and accuracy) were calculated.

Results: Mean age was 45.4±8 years and 79% were female. TR4 (42.5%) and Bethesda II (55.5%) were most common. Malignancy rates increased with category (Bethesda II 18.5%, III 35.2%, IV 50.0%, V 71.4%, VI 93.3%; TI-RADS 2 12.5% to TI-RADS 5 87.5%). Concordant high-risk TR5 with Bethesda VI was 100% malignant (10/10). Papillary thyroid carcinoma and variants predominated. Bethesda outperformed TI-RADS in specificity (87.2% vs 53.9%) and accuracy (77.3% vs 65.3%), while TI-RADS showed higher sensitivity (77.8% vs 66.7%); both had acceptable NPV (~73–74%) and 0.71 (TI-RADS).

Conclusion: TI-RADS and Bethesda show limited concordance but demonstrate increasing malignancy risk with higher categories. Bethesda provides higher specificity as well as overall accuracy, while both systems maintain acceptable negative predictive value in the operated subset.