PDF(Pubmed)
Abstract：
BACKGROUND: Coronary status at one month after Kawasaki disease (KD) onset had a great significance. The present study aimed to establish a prediction model for coronary artery aneurysms (CAA) at one month in children with KD.
METHODS: Patients with a diagnosis of KD between May 2017 and Dec 2018 were enrolled as the development cohort to build a prediction model. The model was validated by internal and external validation. Patients between Jan 2019 and Dec 2019 were enrolled as the validation cohort. The adaptive least absolute shrinkage and selection operator (LASSO) was used to select the possible predictors. Receiving operating characteristic curve (ROC), calibration plots, and decision curve analysis (DCA) were used to evaluate the performance of the model. The performance of the Son score was also assessed.
RESULTS: LASSO regression demonstrated that age, sex, and CALs in the acute stage were predictors for CAA at one month. The area under the ROC (AUC) was 0.946 (95% confidence interval: 0.911-0.980) with a sensitivity of 92.5% and a specificity of 90.5%. The calibration curve and the DCA showed a favorable diagnostic performance. The internal and external validation proved the reliability of the prediction model. The AUC of our model and the Son score were 0.941 and 0.860, respectively (P < 0.001).
CONCLUSIONS: Our prediction model for CAA at one month after disease onset in KD had an excellent predictive utility.
METHODS: Patients with a diagnosis of KD between May 2017 and Dec 2018 were enrolled as the development cohort to build a prediction model. The model was validated by internal and external validation. Patients between Jan 2019 and Dec 2019 were enrolled as the validation cohort. The adaptive least absolute shrinkage and selection operator (LASSO) was used to select the possible predictors. Receiving operating characteristic curve (ROC), calibration plots, and decision curve analysis (DCA) were used to evaluate the performance of the model. The performance of the Son score was also assessed.
RESULTS: LASSO regression demonstrated that age, sex, and CALs in the acute stage were predictors for CAA at one month. The area under the ROC (AUC) was 0.946 (95% confidence interval: 0.911-0.980) with a sensitivity of 92.5% and a specificity of 90.5%. The calibration curve and the DCA showed a favorable diagnostic performance. The internal and external validation proved the reliability of the prediction model. The AUC of our model and the Son score were 0.941 and 0.860, respectively (P < 0.001).
CONCLUSIONS: Our prediction model for CAA at one month after disease onset in KD had an excellent predictive utility.
摘要：
背景：川崎病（KD）发病后1个月的冠状动脉状况具有重要意义。本研究旨在建立KD患儿1个月时冠状动脉瘤(CAA)的预测模型。
方法：纳入2017年5月至2018年12月诊断为KD的患者作为发展队列，以建立预测模型。通过内部和外部验证对模型进行了验证。2019年1月至2019年12月的患者被纳入验证队列。自适应最小绝对收缩和选择算子（LASSO）用于选择可能的预测因子。接收工作特性曲线(ROC),校准图,和决策曲线分析(DCA)用于评估模型的性能。还评估了儿子得分的表现。
结果：LASSO回归表明年龄，性别,急性期CAL是1个月时CAA的预测因子。ROC下面积（AUC）为0.946（95％置信区间：0.911-0.980），敏感性为92.5％，特异性为90.5％。校准曲线和DCA显示出良好的诊断性能。内部和外部验证证明了预测模型的可靠性。模型的AUC和Son评分分别为0.941和0.860(P<0.001)。
结论：我们的KD患者发病后1个月的CAA预测模型具有良好的预测效用。
方法：纳入2017年5月至2018年12月诊断为KD的患者作为发展队列，以建立预测模型。通过内部和外部验证对模型进行了验证。2019年1月至2019年12月的患者被纳入验证队列。自适应最小绝对收缩和选择算子（LASSO）用于选择可能的预测因子。接收工作特性曲线(ROC),校准图,和决策曲线分析(DCA)用于评估模型的性能。还评估了儿子得分的表现。
结果：LASSO回归表明年龄，性别,急性期CAL是1个月时CAA的预测因子。ROC下面积（AUC）为0.946（95％置信区间：0.911-0.980），敏感性为92.5％，特异性为90.5％。校准曲线和DCA显示出良好的诊断性能。内部和外部验证证明了预测模型的可靠性。模型的AUC和Son评分分别为0.941和0.860(P<0.001)。
结论：我们的KD患者发病后1个月的CAA预测模型具有良好的预测效用。
