恶性胆道梗阻ERCP术后早期胆道感染预测模型的建立与评价  

Establishment and evaluation of early biliary infection prediction model after ERCP in malignant biliary obstruction

作　　者：马勇新 张旭升 柳科军 刘伊敏 周红才 魏鹏 陈本栋[1] MA Yongxin;ZHANG Xusheng;LIU Kejun;LIU Yimin;ZHOU Hongcai;WEI Peng;CHEN Bendong(Department of Hepatobiliary Surgery,General Hospital of Ningxia Medical University,Yinchuan 750004,China;School of Clinical Medicine,Ningxia Medical University,Yinchuan 750004,China)

机构地区：[1]宁夏医科大学总医院肝胆外科,宁夏银川750004 [2]宁夏医科大学临床医学院,宁夏银川750004

出　　处：《中国普通外科杂志》2023年第8期1208-1217,共10页China Journal of General Surgery

摘　　要：背景与目的:早期胆道感染(EBI)作为恶性胆道梗阻(MBO)患者行内镜逆行胰胆管造影(ERCP)联合胆道支架植入术后的常见并发症,其对患者的生存时间和生活质量产生重要影响,目前的研究主要关注在胆道感染的危险因素方面,而有关EBI发生风险预测模型的研究少见。因此,本研究基于术前临床资料构建MBO患者行ERCP联合胆道支架植入术后EBI的风险预测模型,以期通过术前临床资料早期精准干预,降低患者EBI的发生率。方法:回顾性分析2018年1月—2021年9月在宁夏医科大学总医院肝胆外科行ERCP联合胆道支架植入术的285例患者临床资料(纳入的所有患者经影像资料或病理证据诊断为MBO)。研究终点为ERCP术后30 d内发生胆道感染。按照7∶3随机分为建模组及验证组。建模组资料经过单变量分析及多变量Logistic回归分析构建预测模型,人工神经网络(ANN)评价预测变量重要性。对模型进行内外部验证,绘制受试者工作曲线(ROC)及校正曲线评估检验模型。结果:共纳入285例患者临床资料,随机分组后建模组200例,验证组85例。单变量及多变量分析结果显示,梗阻位置(OR=5.942,95%CI=2.507~14.081,P<0.001),胆结石(OR=4.821,95%CI=2.087~11.138,P<0.001),糖尿病(OR=5.407,95%CI=2.067~14.148,P=0.001),梗阻长度(OR=1.058,95%CI=1.028~1.089,P<0.001)为MBO患者ERCP术后EBI的独立危险因素,通过独立危险因素构建Logistic回归模型并以列线图形式将模型可视化。利用ANN评估预测变量所占权重由高到低依次为:梗阻长度(46.8%)、梗阻位置(18.6%)、糖尿病(18.1%)、胆结石(16.5%)。Logistic模型经内外部验证,曲线下面积(AUC)分别为0.807和0.831,C指数分别为0.807和0.831,Hosmer-Lemeshow拟合优度评估模型预测值与实值之间无明显偏差(建模组:P=0.845,验证组:P=0.197)。结论:所构建的Logistic模型可以较好地预测ERCP术后EBI的发生风险,经ANN评估梗阻长度是最重要的预测�Background and Aims:Early biliary infection(EBI),as a common complication after endoscopic retrograde cholangiopancreatography(ERCP)combined with biliary stenting in patients with malignant biliary obstruction(MBO),has a significant impact on the survival time and quality of life of patients,and the current research mainly focuses on the risk factors for biliary tract infections,whereas studies on the risk prediction model for the occurrence of EBI are rare.Therefore,based on preoperative clinical data,this study was conducted to construct a risk prediction model of EBI after ERCP combined with biliary stent placement in MBO patients to reduce the incidence of EBI in patients through early and precise preoperative intervention.Methods:The clinical data of 285 patients who underwent ERCP combined with biliary stent placement in the Department of Hepatobiliary Surgery of the General Hospital of Ningxia Medical University from January 2018 to September 2021 were retrospectively analyzed(all patients included were diagnosed with MBO by imaging data or pathological evidence).The study endpoint was biliary infection within 30 d after ERCP.The patients were randomized into modeling and validation groups in a 7∶3 ratio.The data of the modeling group were analyzed by univariate analysis and multivariate Logistic regression analysis to construct a predictive model,and artificial neural network(ANN)was used to evaluate the importance of predictor variables.The model was internally and externally validated,and receiver operating characteristic(ROC)and calibration curves were generated to evaluate and test the model's performance.Results:The clinical data of 285 patients were included in this study,with 200 cases assigned to the modeling group and 85 cases to the validation group after randomization.Results from univariate and multivariate analyses indicated that location of obstruction(OR=5.942,95%CI=2.507-14.081,P<0.001),gallstones(OR=4.821,95%CI=2.087-11.138,P<0.001),diabetes mellitus(OR=5.407,95%CI=2.067-14.148,P=0.001)

关 键 词：胆汁淤积 胰胆管造影术 内窥镜逆行 胆道感染 危险因素 列线图 神经网络 计算机 

分 类 号：R657.4[医药卫生—外科学]