Contrast-enhanced ultrasound thyroid imaging reporting and data system outperforms K-TIRADS in diagnostic accuracy and reduces unnecessary fine-needle aspirations for thyroid nodules.

Background: Thyroid nodules are frequently detected on ultrasonography, yet only a small proportion are malignant. The Korean Thyroid Imaging Reporting and Data System (K-TIRADS) is widely used for risk stratification; however, its reliance solely on grayscale ultrasonographic features may result in suboptimal diagnostic performance and unnecessary fine-needle aspirations (FNAs). Contrast-enhanced ultrasound (CEUS) provides real-time visualization of tumor microvascular perfusion patterns, which may complement conventional ultrasound findings. Objective: This study aimed to evaluate a contrast-enhanced ultrasound-based Thyroid Imaging Reporting and Data System (CEUS TI-RADS) compared with K-TIRADS for differentiating benign from malignant thyroid nodules and to assess its ability to reduce unnecessary FNAs. Methods: A total of 175 thyroid nodules in patients who underwent FNA and/or surgical resection between January 2022 and December 2023 were retrospectively enrolled. Each nodule was classified according to both K-TIRADS and CEUS TI-RADS criteria, with pathological results serving as the reference standard. Receiver operating characteristic (ROC) curve analysis was performed to compare the diagnostic efficacy of the two systems. Sensitivity, specificity, accuracy, positive predictive value (PPV), negative predictive value (NPV), and unnecessary biopsy rate (UBR) were calculated. Multivariate logistic regression was used to identify independent predictors of malignancy. Results: Among 175 nodules, 84 were malignant and 91 were benign. CEUS TI-RADS demonstrated significantly higher diagnostic performance than K-TIRADS, with an area under the ROC curve (AUC) of 0.893 versus 0.804 (P < 0.05). CEUS TI-RADS achieved a substantially lower UBR at 18.6% compared with K-TIRADS 2016 at 44.8% and with K-TIRADS2021 at 44.8% (P < 0.01). Multivariate logistic regression identified hypoechoic echogenicity (P = 0.036), irregular or lobulated margins (P = 0.003), punctate echogenic foci (P < 0.01), and hypoenhancement (P = 0.007) as independent predictors of malignancy. Conclusion: CEUS TI-RADS offers superior diagnostic accuracy for differentiating benign from malignant thyroid nodules compared with K-TIRADS and effectively reduces the rate of unnecessary fine-needle aspirations. These findings support the clinical utility of incorporating CEUS-based features into thyroid nodule risk stratification.