Validation of ACR TI-RADS for Predicting Cytological Outcomes in Thyroid Nodules: A Large FNA Cohort Study.

Purpose

To evaluate the diagnostic performance of ACR TI-RADS for predicting malignant cytology in thyroid nodules undergoing fine-needle aspiration (FNA) in a large tertiary centre.

Methods

A retrospective analysis was performed of 3368 thyroid nodules undergoing ultrasound-guided FNA. Nodules were classified according to ACR TI-RADS (TR1-TR5). Cytological outcomes were recorded using the Bethesda System. Malignant cytology was defined as Bethesda V-VI and Bethesda IV-VI was evaluated as a secondary composite cytological outcome. The distribution of cytology across TI-RADS categories was assessed, and malignancy rates were calculated. Diagnostic performance was evaluated using TR4-TR5 as a positive threshold.

Results

Nodules were classified as TR1 (1.4%), TR2 (5.3%), TR3 (23.7%), TR4 (51.0%) and TR5 (18.7%). Malignant cytology increased across categories from 0.0% in TR1 to 6.5% in TR5, while Bethesda IV-VI cytology increased from 0.0% to 12.6%. Among TR3-TR5 nodules, 81.0% met the corresponding ACR size threshold for FNA; malignant cytology occurred in 2.6% meeting the threshold and 1.7% below threshold (p = 0.24). TI-RADS category was significantly associated with cytological outcome (p < 0.001). Using TR4-TR5 as a threshold, sensitivity was 88.2% and negative predictive value 99.1%.

Conclusion

ACR TI-RADS demonstrates a clear association with cytology outcomes, with increasing malignancy risk across categories. In this large FNA-based cohort, TI-RADS showed excellent negative predictive value, supporting its role in ultrasound-based risk stratification and optimisation of FNA decision-making.