An interpretable machine learning model for early risk stratification of medium-to-giant coronary artery aneurysm in Kawasaki disease: development and external validation - PubMed
5 hours ago
- #Kawasaki disease
- #coronary artery aneurysm
- #machine learning
- Development of an interpretable machine learning model for early risk stratification of medium-to-giant coronary artery aneurysm (MGCAA) in Kawasaki disease.
- Model uses six routinely collected clinical variables: hemoglobin, time to diagnosis, oral mucosal changes, rash, triglycerides, and neutrophil percentage.
- Achieved AUC of 0.70 in internal validation and 0.75 in external validation.
- SHAP analysis used for model interpretability, supporting clinical relevance of predictors.
- Decision curve analysis (DCA) indicated potential net benefit within low-to-moderate threshold probabilities.
- Model deployed as a web-based tool for individualized risk estimation.
- External validation and intercept-only recalibration improve generalizability across populations.
- Ethical approval obtained with waived informed consent due to retrospective design and de-identified data.