Development and clinical validation of an artificial intelligence based model for thyroid nodule malignancy risk assessment using C-TIRADS guidelines - Summary - MDSpire
Advertisement
Development and clinical validation of an artificial intelligence based model for thyroid nodule malignancy risk assessment using C-TIRADS guidelines
To develop and clinically evaluate an AI-driven model for assessing the malignancy risk of thyroid nodules using C-TIRADS criteria.
Approach:
Model Development: A C-TIRADS-guided computer-aided diagnostic framework was created, consisting of a nodule detection module, a feature classification module, and a risk scoring module.
Clinical Validation: The model was trained on task-specific datasets and validated on an independent cohort of 303 thyroid nodules/images from 284 patients.
Key Findings:
The AI model achieved an overall accuracy of 0.862 [95% CI, 0.822–0.901] in the independent clinical validation set.
Physician accuracy improved from 0.705 [95% CI, 0.657–0.756] without AI assistance to 0.845 [95% CI, 0.802–0.884] with AI assistance.
Interpretation:
The findings suggest the potential value of the proposed system as a C-TIRADS-guided decision-support tool.
Limitations:
Larger blinded multicenter studies are required before routine clinical application.
Conclusion:
The C-TIRADS-guided framework can assist in localizing thyroid nodules, classifying ultrasound features, and providing standardized malignancy risk stratification.
