Objective To establish a Nomogram model for predicting the risk of Kawasaki disease (KD) complicated with coronary artery lesions (CAL).

Methods The clinical data and hematological indexes of children with KD were retrospectively analyzed.According to the results of coronary artery detection by echocardiography, the children were divided into coronary artery lesions group (CAL group) with 70 cases and non-coronary artery lesions group (NCAL group) with 95 cases.The risk factors of KD complicated with CAL were analyzed regressively by minimum absolute value convergence and selection operator as well as the least absolute shrinkage and selection operator (LASSO) algorithm, and these factors were incorporated intomultivariate Logistic regression for establishing the prediction model and constructing the Nomogram.The model was verified and the advantages and disadvantages were evaluated by receiver operating characteristic (ROC) curve, correction curve and decision curve analysis.

Results Five predictors were screened by LASSO regression, includingduration of fever ≥10 d, mycoplasma infection, white blood cell (WBC) count >20×10⁹/L, platelet (PLT) count, C-reactive protein (CRP) to serum albumin (ALB) ratio.The five predictors were included in Logistic regression to construct a Nomogram; after verification, the area under the curve (AUC) of Nomogram ROC curve was 0.841(95%CI, 0.782 to 0.900), the sensitivity was 85.7%, and the specificity was 71.6%.

Conclusion The established prediction model in the study for KD complicated with CAL has good discrimination and accuracy, which is helpful for clinical workers to screen out high-risk children with CAL.