LCBDE + LC和ERCP分期联合LC治疗胆囊结石合并胆总管结石(≤10 mm)的临床疗效和术后并发症的独立危险因素并构建相关预测模型

3Clinical Efficacy and Independent Risk Factors for Postoperative Complications of LCBDE + LC and ERCP Staging Combined with LC in the Treatment of Gallstones with Common Bile Duct Stones (≤10 mm), and Construction of Relevant Predictive Models

摘要: 目的：比较胆囊结石合并胆总管结石(≤10 mm)患者采用腹腔镜下胆囊切除(LC)分期联合内镜逆行胰胆管造影(ERCP)与腹腔镜下胆囊切除合经胆总管胆道探查术(LCBDE + LC)胆总管一期治疗的临床效果及术后并发症的发生预测并构建相关预测模型。方法：选取2023年6月~2025年6间就诊于安徽医科大学第四附属医院的202例胆囊结石合并胆总管结石(≤10 mm)患者，根据治疗方案分组和术后有无并发症分组，分析基线资料表，采用LASSO筛选出有意义的特征变量，进一步用多因素Logistic回归分析探讨患者并发症的影响因素，构建预测模型，采用受试者工作特征(ROC)曲线评估模型预测价值，并使用Bootstrap法进行500次重抽样进行十倍交叉验证。结果：在治疗胆囊结石合并胆总管结石(≤10 mm)上，LCBDE + LC和ERCP + LC两种手术方式应该按患者的具体条件而定，其治疗方案才会更具针对性，从而使患者从中获益更多。年龄 ≥ 70岁(OR, 1.059; 95% CI, 1.022~1.099)、手术方式(OR, 0.077; 95% CI, 0.017~0.359)、手术时长(OR, 1.013; 95% CI, 1.000~1.026)、住院时间(OR, 1.288; 95% CI, 1.041~1.593)是术后发生并发症的独立危险因素，Hosmer-Lemeshow检验显示，模型拟合度良好(P = 0.2394)。联合预测的曲线下面积(Area Underthe Curve, AUC)为0.862 (95% CI: 0.793~0.933)，模型的敏感度为84.38%，特异度为79.82%。并使用Bootstrap法进行500次重抽样进行十倍交叉验证，AUC值为0.796，95% (CI: 0.708~0.884)。上述因素构建的预测模型对胆囊结石合并胆总管结石(≤10 mm)患者术后发生并发症风险预测具有指导作用。结论：LC分期联合ERCP组住院时间、术中出血及术后恢复时间相对较短，LCBDE + LC胆总管一期缝合治疗总体花费低，总体并发症的发生更少。并且手术方式、年龄、住院时间、手术时间为术后并发症的独立危险因素，本模型根据独立危险因素建立模型能较好地预测并发症的发生。

Abstract: Objective: Compare the clinical efficacy and prediction of postoperative complications between laparoscopic cholecystectomy (LC) staging combined with endoscopic retrograde cholangiopancreatography (ERCP) and laparoscopic cholecystectomy combined with transcholedochal exploration (LCBDE + LC) for patients with gallstones and common bile duct stones (≤10 mm) in the first stage of common bile duct treatment, and construct relevant prediction models. Methods: 202 patients with gallstones and common bile duct stones (≤10 mm) who visited the Fourth Affiliated Hospital of Anhui Medical University from June 2023 to 2025 were selected. They were grouped according to the treatment plan and the presence of postoperative complications. The baseline data table was analyzed, and meaningful feature variables were selected using LASSO. Further, multiple logistic regression analysis was used to explore the influencing factors of patient complications, and a prediction model was constructed. The receiver operating characteristic (ROC) curve was used to evaluate the predictive value of the model. Bootstrap method was used for 500 resamples for ten-fold cross validation. Results: In the treatment of gallstones combined with common bile duct stones (≤10 mm), the two surgical methods of LCBDE + LC and ERCP + LC should be determined according to the specific conditions of the patient, so that the treatment plan can be more targeted and benefit the patient more. Age ≥ 70 years (OR, 1.059; 95% CI, 1.022~1.099), surgical method (OR, 0.077; 95% CI, 0.017~0.359), duration of surgery (OR, 1.013; 95% CI, 1.000~1.026), and length of hospital stay (OR, 1.288; 95% CI, 1.041~1.593) are independent risk factors for postoperative complications. Hosmer Lemeshow test showed good model fit (P = 0.2394). The Area Under the Curve (AUC) of the joint prediction is 0.862 (95% CI: 0.793~0.933), the sensitivity of the model is 84.38%, and the specificity is 79.82%. And Bootstrap method was used for 500 resamples for ten fold cross validation, with AUC values of 0.796 and 95% (CI: 0.708~0.884). The predictive model constructed based on the above factors has a guiding role in predicting the risk of postoperative complications in patients with gallstones and common bile duct stones (≤10 mm). Conclusion: The LC staging combined with ERCP group had relatively shorter hospitalization time, intraoperative bleeding, and postoperative recovery time. The overall cost of LCBDE + LC common bile duct primary suture treatment was lower, and the incidence of overall complications was less. Moreover, surgical method, age, length of hospital stay, and duration of surgery are independent risk factors for postoperative complications. This model, established based on these independent risk factors, can effectively predict the occurrence of complications.