【摘要】 目的构建食管癌术后辅助化疗期间肺部感染的logistic回归分析风险预测模型，评估其应用价值。方法 分析2010年1月-2020年12月在海口市人民医院行手术治疗的135例食管癌患者，根据患者术后1个月内是否发生肺部感染，分为感染组（n=50）和未感染组（n=85）。采集患者的呼吸道分泌物并对病原菌进行分析，记录患者性别、年龄、手术时间、吻合口瘘、术后误吸、喉返神经损伤、合并疾病等临床资料，分别进行单因素和多因素logistic回归分析，构建logistic回归分析风险预测模型，绘制受试者工作特征（ROC）曲线分析该模型对食管癌术后辅助化疗期间肺部感染的预测价值。结果 共有50例患者术后发生了肺部感染，术后肺部感染的发生率为37.04%。共分离出病原菌41株，主要为铜绿假单胞菌（24.39%）、金黄色葡萄球菌（19.51%）、肺炎克雷伯菌（17.07%）和肺炎链球菌（17.07%）。单因素和多因素logistic回归分析显示，吻合口瘘、有吸烟史、低血浆白蛋白水平、低第1秒呼气量/用力肺活量（FEV1/FEC）水平是影响食管癌术后辅助化疗期间肺部感染的独立危险因素（P<0.05），得到logistic回归分析风险预测模型，P=1/[1+e⁽（16.663+1.568×（吻合口瘘）+3.114×（吸烟史）-0.113×（血浆白蛋白）-0.218×（FEV1/FVC））]。ROC曲线分析显示，logistic回归分析风险预测模型预测食管癌术后辅助化疗期间肺部感染的曲线下面积（AUC）为0.821（95%CI:0.753～0.890,P<0.05），敏感性为84.0%，特异性为69.4%。结论 Logistic回归分析风险预测模型可以有效预测食管癌术后辅助化疗期间肺部感染的发生。更多还原

【Abstract】 Objective To construct the risk prediction model based on logistic regression analysis for lung infection during adjuvant chemotherapy after esophageal cancer surgery,and evaluate its application value.Methods A retrospective analysis was performed on the 135 patients with esophageal cancer undergoing surgical treatment in Haikou People’s Hospital between January 2010 and December 2020. According to the presence or absence of lung infection within 1 month after surgery,they were divided into infection group（n=50）and non-infection group（n=85）. The respiratory secretions were collected for the pathogen analysis. The clinical data such as gender,age,operation time,anastomotic fistula,postoperative aspiration,recurrent laryngeal nerve injury and complications were recorded and analyzed by univariate and multivariate logistic regression analysis. The risk prediction model based on logistic regression analysis was constructed. The predictive value of the model for lung infection after esophageal cancer surgery was analyzed by receiver operating characteristic（ROC）curve.Results Of the 135 patients,there were 50 cases（37.04%）with postoperative lung infection. And there were 41strains of pathogens,mainly Pseudomonas aeruginosa（24.39%）,Staphylococcus aureus（19.51%）,Klebsiella pneumoniae（17.07%）and Streptococcus pneumoniae（17.07%）. Univariate and multivariate logistic regression analysis showed that anastomotic fistula,smoking history,low levels of plasma albumin and forced expiratory volume in the first second/forced vital capacity（FEV1/FEC）were independent risk factors of lung infection after esophageal cancer surgery（P<0.05）. Logistic regression analysis risk prediction model was as follow:P=1/[1+e⁽（16.663+1.568×（anastomotic fistula）+3.114×（smoking history）-0.113×（plasma albumin）-0.218×（FEV1/FVC））].ROC analysis showed that area under curve（AUC）,sensitivity and specificity of the risk prediction model based on logistic regression analysis for predicting lung infection after esophageal cancer surgery were 0.821（95%CI:0.753-0.890,P<0.05）,84.0% and 69.4%,respectively.Conclusion The risk prediction model based on logistic regression analysis could effectively predict the occurrence of lung infection after esophageal cancer surgery.更多还原