目的 探讨腹腔镜膀胱根治性切除术(LRC)后输尿管肠吻合口狭窄(UES)发生的影响因素,并构建和验证列线图的预测模型.方法 回顾性分析2019年10月至2021年10月在本院接受LRC的315例膀胱癌患者的临床资料并设为建模组,另收集2021年10月至2022年10月在本院接受LRC的135例膀胱癌患者的临床资料并设为验证组,对列线图模型进行外部验证.根据患者是否发生UES将建模组患者分为UES组(35例)和非UES组(280例).采用logistic回归分析患者行LRC术后发生UES的影响因素,并构建列线图预测模型,分别通过受试者工作特征(ROC)曲线和校准曲线验证模型的区分度和校准度.结果 单因素分析结果显示,体质量指数(BMI)、年龄、术后留置D-J管时间、合并糖尿病、术后3 d C反应蛋白(CRP)水平与LRC术后发生UES具有相关性(均P＜0.05).多因素logistic回归分析结果显示,BMI、术后留置D-J管时间、合并糖尿病及术后3 d CRP水平为影响LRC术后发生UES的危险因素(均P＜0.05).构建列线图预测模型,结果显示,BMI每增加2 kg/m2将增加4.91分的权重、术后留置D-J管时间每增加2 d将增加7.5分的权重、术后3 d CRP水平每增加10 mg/L将增加12.5分的权重,合并糖尿病可得40分.内部验证的校准曲线显示列线图预测术后UES发生风险与实际发生风险之间具有良好的一致性(x2=7.357,P=0.214),ROC的曲线下面积为0.857(95％CI:0.813～0.894),灵敏度为88.57％,特异度为71.43％.外部验证校准曲线的预测概率与实际概率接近(x2=6.231,P=0.142),ROC 的曲线下面积为 0.840(95％CI:0.795～0.879),灵敏度为88.60％,特异度为69.29％.结论 BMI、术后留置D-J管时间、合并糖尿病及术后3 d CRP水平为影响LRC术后发生UES的危险因素.通过整合此危险因素构建的列线图预测模型表现出良好的预测性能,能为临床医师甄别高风险人群,并提供准确的风险评估.
Study on the influencing factors of ureteral enterostomy stenosis after laparoscopic radical cys-tectomy
Objective To analyze the influencing factors of ureteral enterostomy stenosis(UES)after laparoscopic radical cystectomy(LRC)and construct and verify the nomogram prediction model.Methods Clinical data of 315 patients with bladder cancer who received LRC in our hospital from October 2019 to October 2021 were retrospectively analyzed as the modeling group,and clinical data of 135 patients who received LRC in our hospital from October 2021 to October 2022 were collect-ed as the validation group for external verification of the nomogram model.Patients in the modeling group were divided into UES group(35 cases)and non-UES group(280 cases)according to whether they had UES.Lgoistic regression was used to analyze the factors affecting the incidence of UES after LRC.A nomogram prediction model was built with R software,the discrimination and calibration of the model were verified by receiver operating characteristic(ROC)curve and calibration curve,respec-tively.Results Univariate analysis showed that body mass index(BMI),age,postoperative retention time of D-J tube,complicated diabetes mellitus,and C-reactive protein(CRP)level 3 days after sur-gery were correlated with the incidence of UES after LRC(all P＜0.05).Multivariate logistic regres-sion analysis showed that BMI,time of D-J tube implantation,diabetes mellitus and CRP level 3 days after surgery were the risk factors for the development of UES after LRC(all P＜0.05).The results showed that every 2 kg/m2 had increased in BMI would increase the weight of 4.91 points,every 2 d had increased in postoperative retention time of D-J tube would increase the weight of 7.5 points,every 10 mg/L had increased in CRP level at 3 d after surgery would increase the weight of 12.5 points,and 40 points could be scored when diabetes mellitus was combined.There was a good agreement between the predicted risk of postoperative UES and the actual risk(x2=7.357,P=0.214),the area under ROC curve was 0.857(95％CI:0.813-0.894),the sensitivity was 88.57％,and the specificity was 71.43％.The predicted probability of external verification calibration curve was close to the actual probability(x2=6.231,P=0.142),the area under ROC curve was 0.840(95％CI:0.795-0.879),the sensitivity was 88.60％,and the specificity was 69.29％.Conclusions BMI,postoper-ative D-J tube indwelling time,diabetes mellitus and CRP level 3 days after surgery are risk factors for UES after LRC.By integrating this risk factor,a nomogram model for predicting UES after LRC is con-structed and shows good predictive performance,which can provide accurate risk assessment for clini-cians to identify high-risk groups.
NETL
NSTL
万方数据
