Related Experiment Video
Updated: Jun 12, 2026

A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images
Published on: April 21, 2023
Predicting malignancy in solid thyroid nodules with no or punctate-linear enhancement on contrast-enhanced
Ruijun Kang1, Jiajia Tang2, Yichan Zhang1
1Department of Ultrasound, China-Japan Friendship Hospital, Beijing, China.
Background:
The 2020 Chinese Thyroid Imaging Reporting and Data System (C-TIRADS) regards no or scant punctate-linear enhancement on contrast-enhanced ultrasound (CEUS) as a benign feature in thyroid nodules. However, emerging evidence suggests that some malignant nodules may also exhibit this enhancement pattern. Few systematic studies focusing on the differential diagnosis of such nodules have been conducted. Therefore, this study aimed to identify the ultrasound (US) and CEUS characteristics of malignant solid thyroid nodules with no or punctate-linear enhancement and to evaluate their diagnostic value in distinguishing between benign and malignant lesions.
Methods:
A retrospective analysis was conducted on clinical and imaging data from 107 patients with 108 pathologically confirmed solid thyroid nodules exhibiting no or punctate-linear enhancement on CEUS, with 73 benign and 35 malignant nodules ultimately being analyzed. Punctate-linear enhancement was further categorized as internal or peripheral based on the location of microbubbles. The basic characteristics, as well as US and CEUS characteristics, were compared between benign and malignant nodules. Multivariate logistic regression analysis was performed to identify independent predictors for malignancy, with odds ratio (OR) and 95% confidence interval (CI) calculated. Receiver operating characteristic (ROC) curve analysis was used to evaluate diagnostic performance, and the area under the curve was calculated with 95% CI, as well as the sensitivity and specificity.
Results:
Significant differences were observed between the two groups in terms of location, margin, orientation, echogenic foci, enhancement pattern, and enhancement margin (all P<0.05). Multivariate analysis identified that the independent predictors of malignancy were isthmic location (OR =28.288; 95% CI: 4.515-177.250, P<0.001), ill-defined/irregular margin (OR =10.221; 95% CI: 1.929-54.156, P=0.006), microcalcifications (OR =6.516; 95% CI: 1.392-30.497, P=0.017), and peripheral punctate-linear enhancement (OR =15.587, 95% CI: 2.932-82.855, P=0.001), which yielded areas under the curve of 0.702 (95% CI: 0.587-0.816), 0.729 (95% CI: 0.628-0.830), 0.687 (95% CI: 0.572-0.803), and 0.765 (95% CI: 0.674-0.856), respectively; meanwhile the combined model achieved an area under the curve of 0.944 (95% CI: 0.905-0.984), with a sensitivity of 0.943 and a specificity of 0.836.
Conclusions:
Solid thyroid nodules with no or punctate-linear enhancement on CEUS are not exclusively benign. The combination of isthmic location, ill-defined/irregular margin, microcalcifications, and peripheral punctate-linear enhancement may accurately indicate malignancy. These findings may assist in identifying malignant lesions and reducing unnecessary invasive procedures for benign nodules.
