To evaluate the integration of an AI-based ultrasound model with molecular testing for improving diagnostic metrics and patient outcomes in patients with indeterminate thyroid nodules.
Key Findings:
Sensitivity remained at 95% with both ThyroSeq and the integrated approach.
Specificity improved from 45% (ThyroSeq) to 60% (integrated approach).
Positive predictive value (PPV) increased from 66% to 72% with the combined method.
Area under the receiver operating characteristic curve (AUC) rose from 0.70 to 0.78.
Negative predictive value (NPV) improved to 99% with the combined approach.
Interpretation:
The integration of AI-driven imaging with molecular diagnostics may enhance precision medicine in thyroidology, potentially improving patient outcomes, including surgical complications and quality of life.
Limitations:
Small sample size limited statistical significance of findings, affecting the reliability of results.
Classification of nodules without surgical pathology as benign may have introduced bias, potentially inflating specificity and NPV.
Results may not generalize to other molecular testing platforms, necessitating further validation.
Conclusion:
Further validation in larger, diverse populations is necessary to confirm the benefits of the combined approach and its impact on surgical rates and patient outcomes.
