Integrating Artificial Intelligence in Thyroid Nodule Management: Clinical Outcomes and Cost-effectiveness Analysis

Area of Science:

• Medical Imaging
• Artificial Intelligence in Medicine
• Endocrinology

Background:

• Increasing incidence of thyroid nodules (TN) leads to concerns about overdiagnosis and overtreatment.
• Accurate risk stratification of TNs is crucial for appropriate management.
• Artificial intelligence (AI) tools are emerging to aid in medical decision-making.

Purpose of the Study:

• To evaluate the clinical and economic impact of KOIOS, an FDA-approved AI tool, for the management of thyroid nodules.
• To compare the diagnostic performance of KOIOS AI with expert and novice operators using ACR-TIRADS.
• To assess the cost-effectiveness of KOIOS AI in managing TNs.

Main Methods:

• Retrospective analysis of 176 patients undergoing thyroid surgery.
• Independent evaluation of ultrasound images by expert and novice operators using ACR-TIRADS.
• KOIOS AI provided AI-adapted risk stratification for TNs.
• Diagnostic performance assessed using sensitivity, specificity, and ROC analysis.
• Cost-effectiveness evaluated using Incremental Cost-Effectiveness Ratio (ICER) and sensitivity analyses.

Main Results:

• KOIOS AI demonstrated diagnostic performance comparable to expert operators (AUC: 0.794 vs. 0.784) and significantly outperformed novice operators (AUC: 0.619).
• ICER analysis indicated KOIOS AI as a dominant and cost-saving strategy (-€8,085.56 per additional optimal care decision).
• Sensitivity analyses confirmed KOIOS AI as the optimal strategy, with surgical costs being a key driver of variability.

Conclusions:

• KOIOS AI presents a cost-effective solution for thyroid nodule management.
• The AI tool shows potential in reducing overdiagnosis and overtreatment of benign thyroid nodules.
• Further prospective studies are recommended to validate findings and assess long-term implications.