A new thyroid imaging reporting and data system for nodules: Based on grayscale and color Doppler ultrasonography

Objective: To construct and validate a new thyroid imaging reporting and data system (TI-RADS) based on radiating blood flow and grayscale US features. Materials and methods: This study enrolled patients from 4 hospitals from January 2018 to November 2023 retrospectively and prospectively. All US features associated with malignant thyroid nodules were assessed by multivariable logistic regression to construct baseline US TI-RADS (BUS TI-RADS), which was tested with internal validation set, external validation set and prospective validation set. Additionally, its potential of reducing biopsy was assessed. Results: There were 2932 patients (2153 female, age: 42.83 +/- 11.71; 779 male, age: 42.36 +/- 11.78) with 3940 nodules (2831 malignant nodules and 1109 benign nodules). Independent predictive factors included composition, echogenicity, shape, margin, suspicious extrathyroidal extension, punctate echogenic foci, and radiating blood flow. Compared with American College of Radiology (ACR) TI-RADS and Chinese TI-RADS (C-TIRADS), the BUS TI-RADS had higher AUCs of 0.96 (95 % CI: [0.95, 0.97]; P < 0.001), 0.93 ([0.90, 0.97]; P < 0.001), 0.91 ([0.86, 0.96]; P < 0.003) and 0.95 ([0.93, 0.97]; P < 0.001) for the training set, internal validation set, external validation set and prospective validation set respectively. Decision curve analysis demonstrated higher net benefit for the BUS TI-RADS. And the BUS TI-RADS (4.1 %; 18.1 %) had lower percentage of biopsy and false negative rate compared with the ACR TI-RADS (31.2 %; 20.9 %) and C-TIRADS(33.1 %; 58.5 %). Conclusion: BUS TI-RADS was created according to the simplified regression coefficients of radiating blood flow and grayscale ultrasonography features with excellent diagnostic performance and could reduce unnecessary biopsy with lower missed malignancy.