Development and clinical validation of an artificial intelligence based model for thyroid nodule malignancy risk assessment using C-TIRADS guidelines
By
Rongzhou Ye
Yao Liu
Xiuming Wu
Kangjian Wang
July 15, 2026
Clinical Scorecard: Creation and clinical evaluation of an AI-driven model for assessing malignancy risk in thyroid nodules based on C-TIRADS criteria
At a Glance
Category Detail
Condition Thyroid nodules Key Mechanisms C-TIRADS ultrasound features classification and AI-driven risk scoring Target Population Patients with thyroid nodules Care Setting Clinical evaluation and diagnostic imaging
Key Highlights
AI model achieved an overall accuracy of 0.862 in malignancy risk assessment. Physician accuracy improved from 0.705 to 0.845 with AI assistance. The model classifies nodules based on C-TIRADS ultrasound features. Guideline-Based Recommendations
Diagnosis
Utilize ultrasound imaging to assess thyroid nodules. Apply C-TIRADS criteria for risk stratification. Management
Benign nodules managed with follow-up observation. Malignant nodules typically treated with surgery. Monitoring & Follow-up
Regular follow-up for benign nodules. Risks
7 to 15% of thyroid nodules may be malignant. Patient & Prescribing Data
Patients with detected thyroid nodules via ultrasound.
AI-assisted evaluation may enhance diagnostic accuracy.
Clinical Best Practices
Incorporate AI tools to assist in ultrasound interpretation. Focus on high-risk ultrasound features for malignancy assessment. Related Resources & Content
