Abstract
OBJECTIVE: Kawasaki disease is an acute immune vasculitis that often has a poor prognosis when complicated by coronary artery lesions. Our study aims to construct a risk model for Kawasaki disease complicated by coronary artery lesions and validate it in different clinical characteristic subgroups, optimizing personalized and precise management of Kawasaki disease to improve patient outcomes. METHODS: First, we compared each factor between the groups with and without coronary artery damage. We then used LASSO analysis to further filter for factors that were more significant in predicting outcomes. The selected factors were used to construct the risk model. The model was evaluated using ROC curves, calibration curves, and DCA, and was internally validated using 5-fold cross-validation. Finally, we also conducted subgroup analyses based on factors such as age stages and sex. RESULTS: Through univariate analysis, LASSO analysis, and correlation analysis, we identified WBC, PLT, CRP, ALB, Na, Time to IVIG treatment, and symptoms of limb as the key factors for constructing the risk model. The model achieved an area under the curve of 0.815(95%CI: 0.779-0.851). Additionally, calibration curves, DCA, and 10-fold cross-validation demonstrated that the model has good predictive performance. The predictive efficacy of the model was also satisfactory across various subgroups. CONCLUSIONS: Our study has constructed a risk model for Kawasaki disease complicated by coronary artery lesions in the Chinese population that demonstrates good predictive performance, and it has been validated successfully across multiple subgroups.