CT评估胰腺癌不同神经通路侵犯的预后价值及模型构建

Using CT to Evaluate the Invasion of Different Neural Pathways in Pancreatic Cancer and Model Construction

目的 本研究旨在探讨CT评估胰腺癌侵犯周围不同组神经通路的预后价值,并结合其他CT影像学特征,建立预测胰腺癌预后的列线图模型。方法 回顾性分析96例胰腺癌患者的临床与影像学资料。根据肿瘤侵犯不同神经通路分为A组(侵犯前路神经通路)和B组(侵犯胰丛1/胰丛2/肠系膜根部神经通路),比较两组患者的临床及CT影像特征,采用Kaplan-Meier生存曲线和Log-rank检验比较两组患者的生存差异。采用单因素和多因素COX比例风险回归模型分析胰腺癌患者预后的影响因素,并基于此构建列线图模型,通过受试者工作特征曲线下面积来评估列线图的性能。结果 B组中位总生存期长于A组(P=0.002)。COX模型分析结果显示,静脉畸形是影响胰腺癌预后的独立危险因素(HR=2.585, 95%CI:1.397~4.785,P=0.003),B组相较于A组是影响胰腺癌预后的独立保护因素(HR=0.537, 95%CI:0.323~0.893,P=0.017)。训练集和验证集的ROC曲线体现了列线图模型具有良好的预测能力。结论 侵犯前路神经通路的胰腺癌患者总生存期较短。基于CT评估的静脉畸形和周围不同组神经通路侵犯的列线图模型可以协助预测胰腺癌患者的预后。

Objective To investigate the prognostic value of CT in evaluating different groups of neural pathways around pancreatic cancer invasion,and to establish a nomogram model for predicting prognosis combined with other CT imaging features.Methods A retrospective analysis was conducted on clinical and imaging data of 96 patients with pathologically confirmed pancreatic ductal adenocarcinoma.Based on tumor invasion of different neural pathways,patients were categorized into Group A(invasion of the anterior neural pathway)and Group B(invasion of pancreatic plexus 1/pancreatic plexus 2/mesenteric root neural pathways).Clinical and CT imaging features were compared between the two groups,and Kaplan-Meier survival curves and Log-rank tests were employed to assess survival differences.Univariate and multivariate COX proportional hazard regression models were used to analyze prognostic factors,and a nomogram model was further constructed.The performance of the nomogram was evaluated using the receiver operating characteristic(ROC)curve area under the curve.Results Group B exhibited a longer median overall survival compared to Group A(P=0.0024).COX model analysis revealed venous malformation as an independent adverse prognostic factor for pancreatic cancer(HR=2.585,95%CI:1.397-4.785,P=0.003);Group B was an independent protective factor for pancreatic cancer prognosis(HR=0.537,95%CI:0.323-0.893,P=0.017)compared to Group A.The ROC curves for the training and validation sets demonstrated the nomogram model′s good discriminative capabilities.Conclusion Pancreatic cancer patients with invasion of the anterior neural pathway have a shorter overall survival.The nomogram model,based on CT assessment of venous malformation and perineural invasion via different pathways,can assist in predicting the prognosis of pancreatic cancer patients.