Chen Kangjie, Wang Linpei, Wang Xianfeng, Yang Liang, Zhang Xiaodong, Lin Yonghua, Cao Linping
Department of Hepatobiliary and Pancreatic Surgery, The First Affiliated Hospital of Medical School, Zhejiang University, 79 Qingchun Rd., Shangcheng District, Hangzhou, 310003, China.
Department of General Surgery, The Second Affiliated Hospital of Fujian Medical University, Quanzhou, 362000, China.
Surg Endosc. 2025 Sep 17. doi: 10.1007/s00464-025-12169-3.
Common bile duct stones (CBDS) are the primary indication for endoscopic retrograde cholangiopancreatography (ERCP), yet post-ERCP pancreatitis (PEP) remains a significant complication due to its multifactorial etiology. This study aimed to identify core predictors and develop an optimized predictive model for PEP.
We retrospectively enrolled patients who underwent ERCP in three centers between March 2019 and March 2024. Potential predictors and their importance were evaluated with four machine learning (ML) algorithms. Predictive models were developed using logistic regression and assessed for discrimination, calibration, and clinical utility.
A total of 1758 patients were included in the training (n = 917), testing (n = 392), validation 1 (n = 366), and validation 2 (n = 83) cohorts. The incidences of PEP were 6.7%, 6.6%, 10.1%, and 12.0%, respectively, with no significant difference among them (p = 0.063). Using ML, eight critical predictors were identified: age, direct bilirubin, serum calcium, γGT, cannulation attempts, transpancreatic precut, pancreatic guidewire passage, and endoscopic papillary balloon dilation (EPBD) duration. Model 3, incorporating serum calcium (OR: 2.50, p = 0.002), transpancreatic precut (OR: 4.61, p < 0.001), pancreatic guidewire passage (OR: 3.62, p < 0.001), and EPBD duration (OR: 2.25, p = 0.009), exhibited the highest AUC (0.845) and superior sensitivity (83.2%). Internal and external validations confirmed robustness and generalizability of the model, demonstrating excellent predictive performance and clinical utility.
We established and validated an optimized predictive model for PEP using four key predictors, enhancing early identification and intervention after ERCP for patients with CBDS.
胆总管结石（CBDS）是内镜逆行胰胆管造影术（ERCP）的主要适应证，但由于其病因多因素，ERCP术后胰腺炎（PEP）仍然是一种严重的并发症。本研究旨在确定PEP的核心预测因素并开发优化的预测模型。
我们回顾性纳入了2019年3月至2024年3月期间在三个中心接受ERCP的患者。使用四种机器学习（ML）算法评估潜在预测因素及其重要性。使用逻辑回归开发预测模型，并评估其区分度、校准度和临床实用性。
训练队列（n = 917）、测试队列（n = 392）、验证队列1（n = 366）和验证队列2（n = 83）共纳入1758例患者。PEP的发生率分别为6.7%、6.6%、10.1%和12.0%，差异无统计学意义（p = 0.063）。使用ML确定了八个关键预测因素：年龄、直接胆红素、血清钙、γ-谷氨酰转移酶（γGT）、插管尝试次数、经胰预切开术、胰管导丝通过情况和内镜乳头球囊扩张（EPBD）持续时间。模型3纳入血清钙（OR：2.50，p = 0.002）、经胰预切开术（OR：4.61，p < 0.001）、胰管导丝通过情况（OR：3.62，p < 0.001）和EPBD持续时间（OR：2.25，p = 0.009），表现出最高的曲线下面积（AUC）（0.845）和较高的灵敏度（83.2%）。内部和外部验证证实了该模型的稳健性和可推广性，显示出优异的预测性能和临床实用性。
我们使用四个关键预测因素建立并验证了PEP的优化预测模型，增强了对CBDS患者ERCP术后的早期识别和干预。
