ACR TI-RADS Modified by CEUS Features and Nodule Location for Diagnosis and Biopsy Decision in Thyroid Nodules of TR4 or Higher.

Abstract

Objectives: To develop a modified American College of Radiology (ACR) thyroid imaging, reporting and data system (TI-RADS) incorporating contrast-enhanced ultrasound (CEUS) features and nodule location to improve diagnostic performance and fine-needle aspiration (FNA) recommendations for ACR TR4 or higher thyroid nodules.

Methods: This prospective study enrolled patients who underwent ultrasound and CEUS at 3 hospitals between June 2024 and January 2026. CEUS and location features were analyzed using multivariable logistic regression to construct the modified system. Diagnostic performance and biopsy recommendation efficiency were compared with ACR TI-RADS.

Results: Four hundred and eight patients with 442 nodules were included. Independent predictors included enhancement intensity, enhancement direction, ring enhancement, enhancement area and nodule location. In the training set (n = 334), the modified system achieved higher AUC and sensitivity than ACR TI-RADS (AUC: 0.90 [95% CI: 0.86-0.93] versus 0.75 [95% CI: 0.71-0.79]; sensitivity: 88.5% versus 60.3%; p < .01), higher accuracy of malignant nodule biopsies (AMB: 67.1% versus 49.4%), lower unnecessary biopsy rate (UBR: 32.9% versus 50.6%), and lower biopsy rate of benign nodules (BRB: 20.8% versus 31.2%), with no unsafe downgrading. Similar results were observed in the temporal validation set (n = 108): AUC 0.95 (95% CI: 0.90-0.99) versus 0.73 (95% CI: 0.66-0.79); sensitivity 87.0% versus 49.3%; AMB 83.3% versus 53.3%; UBR 16.7% versus 46.7%, and BRB 5.1% versus 17.9%.

Conclusions: A modified ACR TI-RADS integrating CEUS features and nodule location improved diagnostic performance and FNA recommendations for ACR TR4 or higher nodules. Pending external validation, this system may provide a practical approach for refining risk stratification and reducing unnecessary biopsies.