-
1
Thyroid nodules are common lesions, with a prevalence of up to 68%, and 7 to 15% of them being malignant.
-
2
Ultrasound is the primary imaging modality for evaluating thyroid nodules, providing high sensitivity for identifying characteristics and malignancy risk.
-
3
An AI-driven C-TIRADS-guided diagnostic framework was developed, including modules for nodule detection, feature classification, and risk scoring.
-
4
In clinical validation, the AI model achieved an accuracy of 0.862, significantly improving physician accuracy from 0.705 to 0.845 with AI assistance.
-
5
The proposed AI framework can assist in thyroid ultrasound interpretation, but larger multicenter studies are needed before routine clinical application.
