Artificial intelligence-assisted risk stratification of thyroid nodules with atypia of undetermined significance
Authors' conclusion
AI-assisted ultrasound analysis offers supplementary diagnostic information for thyroid nodules with AUS cytology. Its high sensitivity and negative predictive value may assist clinicians in decision-making processes, particularly for small, low-risk thyroid nodules.
Study context
- Population
- Thyroid nodules with atypia of undetermined significance (AUS) cytology on fine-needle aspiration (FNA) (Jan 2019–Dec 2020), later confirmed benign or malignant, five institutions in Korea; median longest diameter 1.30 cm; malignancy rate 39%nodules analysed
- Participants
- 165 nodules
- Comparison
- AI-Thyroid deep learning model vs K-TIRADS
- Outcome
- Diagnostic performance for malignancy in AUS nodules
Methodological detailsmulticentre (five institutions in Korea)
- Multicentre (five institutions in Korea)
Performance
As reported in the abstract
Model sensitivity (binary classification)
0.91
Model NPV (binary classification)
0.87
AUC based on estimated malignancy risk
0.75
95% CI 0.68–0.83
AUC derived from K-TIRADS categories 2-5
0.76
95% CI 0.69–0.83
Further findings
Sensitivity in nodules <1.5 cm (subgroup)
- Model sensitivity in nodules smaller than 1.5 cm
- 98%
Original publication
Artificial intelligence-assisted risk stratification of thyroid nodules with atypia of undetermined significance. Eur Thyroid J. Published 15 January 2026. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →