MRCP predicts the distribution of benign and malignant biliary lesions in pancreaticobiliary maljunction and high confluence of pancreaticobiliary ducts, as well as the risk factors for biliary tract cancer
GAO Feng,GONG Yan,WENG Ting-ting
Department of Radiology, Tianjin Nankai Hospital, Tianjin300100, China
Abstract:
Objective To explore the distribution characteristics of pancreaticobiliary maljunction (PBM) and high confluence of pancreaticobiliary ducts (HCPBD) in benign and malignant biliary lesions and analyze the predictive factors related to the incidence of biliary tract cancer. Methods A retrospective analysis was conducted on 4321 patients after biliary tract surgery admitted to Nankai Hospital in Tianjin from January 2021 to May 2024. Among them, there were 4065 cases of benign biliary tract lesions, 98 cases of gallbladder cancer, and 158 cases of cholangiocarcinoma. Analysis of magnetic resonance imaging pancreatic and all patients magnetic resonance cholangiopancreatography (MRCP) image data and clinical data, analysis of PBM and HCPBD in biliary distribution characteristics of benign and malignant lesions. Multivariate logistic regression was used to analyze the independent influencing factors of biliary tract cancer, and a nomogram prediction model was constructed. The predictive value of each index was analyzed through the receiver operating characteristic curve (ROC) and correction curve. Results The detection rates of PBM and HCPBD in patients with biliary tract cancer were significantly higher than those in the benign lesion group (P <0.001). PBM has the highest detection rate (16.3%，16/98) in gallbladder cancer. The age, common channel length and the proportion of common bile duct dilation in PBM and HCPBD patients with biliary tract cancer were higher than those in PBM and HCPBD patients without biliary tract cancer, and the differences were statistically significant (P <0.001), mainly affecting females. Multivariate Logistic regression analysis showed that age (OR=1.070, P =0.004), biliary dilation (OR=5.472, P =0.004), and common duct length (OR=1.161, P =0.012) were independent predictors of biliary tract cancer. The three risks were combined to construct a nomogram prediction model. ROC curve analysis showed that the AUC of the nomogram model for predicting biliary tract cancer was 0.92(95%CI:0.85-0.98). The correction curve indicates that the model's prediction rate has a high consistency with the actual probability. Conclusion PBM and HCPBD are closely related to the occurrence of biliary tract cancer, especially being more common in gallbladder cancer. Age, whether the biliary tract is dilated and the length of the common channel are independent risk factors for predicting biliary tract cancer in patients with PBM and HCPBD. The nomogram model constructed based on the above factors can help identify high-risk patients and improve diagnostic accuracy.
