Assessing agreement with a single-center expert consensus: artificial intelligence-assisted teleultrasound for thyroid nodules categorized as C-TIRADS 4A or higher
Objective:
To evaluate the diagnostic agreement between AI-assisted teleultrasound and expert consensus in thyroid nodules classified as C-TIRADS 4A or higher during the study period from January 2024 to June 2024.
Key Findings:
- Poor diagnostic consistency between community medical institutions and teleultrasound (kappa = 0.20 [95% CI: −0.04 to 0.44]).
- Good diagnostic consistency between teleultrasound and the AI system (kappa = 0.80 [95% CI: 0.67 to 0.93]).
- AI system achieved a sensitivity of 97.1% [95% CI: 90.2 to 99.2] and specificity of 100.0% [95% CI: 72.2 to 100.0] at the ≥C-TIRADS 4A threshold.
Interpretation:
The AI system significantly improves diagnostic agreement for thyroid nodules classified as C-TIRADS 4A or higher, aligning closely with expert assessments, which may enhance clinical decision-making.
Limitations:
- Study limited to a single center, which may affect generalizability.
- Retrospective design may introduce selection bias, potentially impacting the reliability of the findings.
Conclusion:
AI-assisted teleultrasound enhances diagnostic agreement in evaluating thyroid nodules, achieving consistency comparable to expert assessments, which is crucial for improving patient outcomes.
