摘　要：为了识别儿童早期社区获得性肺炎（CAP）合并心肌损伤的高危患儿，构建和评估早期CAP合并心肌损伤的列线图，回顾性分析了385例CAP患儿病历资料，并将其分为建模组和验证组，采用LASSO回归和Logistic回归筛选预测因子，并构建预测合并心肌损害的列线图，通过接受者操作特征（ROC）曲线、校准曲线、临床决策（DCA）曲线评估模型。统计学分析表明，年龄、早期胸片特征、肺外并发症、最高体温、咳嗽类型、淋巴细胞百分比、血氧饱和度（SaO2）、C反应蛋白（CRP）和支原体感染是儿童早期CAP合并心肌损伤的预测因子。列线图的两组ROC曲线下面积（AUC）值分别为0.827和0.815，校准曲线吻合度较高。DCA曲线显示，使用列线图可在较大诊断阈值范围内获益。本研究列线图模型的区分度和准确性良好,可识别早期CAP患儿心肌损伤的风险，有助于临床医生早期预警和及时干预。
关键词：儿童；社区获得性肺炎；心肌损伤；预测；列线图模型
中图分类号：R4                                     文献标志码：A                  DOI：10.3969/j.issn.1007-7146.2022.04.012

Abstract: In order to identify the high-risk children with early community-acquired pneumonia (CAP) complicated with myocardial injury, the nomogram of CAP complicated with myocardial injury was constructed and evaluated, the nomogram of early CAP complicated with myocardial injury was constructed and evaluated. The medical records of 385 children with CAP were retrospectively analyzed and divided into training set and validation set. The predictors were screened by LASSO regression and Logistic regression, and the nomogram for predicting myocardial injury was developed. The nomogram was evaluated by ROC curve, calibration curve, DCA curve. Age, early chest radiograph characteristics, type of cough, extrapulmonary complication, the highest temperature, lymphocyte percentage, SaO2, CRP and mycoplasma infection were the predictors of early CAP with myocardial injury in children. The AUC value of nomogram in two set were 0.827 and 0.815. The calibration curve was in good agreement with the actual curve. And the DCA curve showed that the use of nomogram can benefit in a wide diagnostic thresholds range. The nomogram with good discrimination and accuracy can identify the risk of myocardial injury in children with early cap and help clinicians to give early warning and timely intervention.
Key words: children; community-acquired pneumonia; myocardial injury; prediction; nomogram model
（Acta Laser Biology Sinica, 2022, 31(4): 379-384）