中国实用外科杂志

中国实用外科杂志 ›› 2025, Vol. 45 ›› Issue (03): 322-328.DOI: 10.19538/j.cjps.issn1005-2208.2025.03.16

• 论著 • 上一篇    下一篇

肝内胆管癌术后生存预后预测列线图模型预后评估效能研究

韩智强1，尹毅青1，王凯元1，宋天强1，黄纪伟2，陈    璐1，3   

  1. 1天津医科大学肿瘤医院 国家肿瘤临床医学研究中心 天津市恶性肿瘤临床医学研究中心 天津市消化系统肿瘤重点实验室 药物成药性评价与系统转化重点实验室，天津300060；2四川大学华西医院肝脏外科，四川成都610041；3日本国家国际医疗研究中心（NCGM），日本东京1628655
  • 出版日期:2025-03-01 发布日期:2025-03-27

  • Online:2025-03-01 Published:2025-03-27

摘要： 目的    基于多中心临床数据，探讨肝内胆管癌（ICC）根治性切除术后的独立危险因素，构建并验证预测术后生存的列线图模型。方法    回顾性分析2011年1月至2018年6月天津医科大学肿瘤医院及四川大学华西医院332例接受ICC根治性切除术的病人资料，将之分为建模队列（131例）和验证队列（201例）。通过单因素及多因素COX回归分析筛选预后独立危险因素，构建列线图模型，采用一致性指数（C-index）、受试者工作特征曲线（ROC曲线）及校准曲线评估模型的预测性能。结果    多因素COX回归分析显示，糖类抗原19-9（CA19-9）、肿瘤直径、微血管侵犯、分化程度及TNM分期是影响ICC病人术后生存的独立危险因素（均P＜0.05）。基于上述因素构建的列线图模型在训练队列中C-index为0.807，ROC曲线下面积（AUC）为0.786；验证队列中C-index为0.818，AUC为0.857。校准曲线显示模型预测生存率与实际生存率高度一致。结论    整合CA19-9、肿瘤直径、微血管侵犯、分化程度及TNM分期的列线图模型具有良好的预测效能，可为ICC根治性切除术后病人的个体化预后评估及临床决策提供参考。

关键词: 肝内胆管癌, 生存预后, 列线图, 预测模型

Abstract: To explore the independent risk factors after radical resection of intrahepatic cholangiocarcinoma (ICC) and construct a survival prognostic nomogram based on multicenter clinical data. Methods    Data of 332 patients who underwent radical resection for ICC at Tianjin Medical University Cancer Institute & Hospital and Sichuan University West China Hospital between January 2011 and June 2018 were collected. Patients were divided into modeling and validation cohorts based on the institution, with 131 in the modeling cohort and 201 in the validation cohort. COX regression analysis was used to identify independent risk factors affecting the prognosis of ICC patients and to construct the nomogram. The predictive performance of the nomogram was evaluated using the concordance index, ROC curve, and calibration curve. Results    Multivariate COX regression analysis indicated that CA19-9, tumor diameter, vascular invasion, differentiation, and TNM stage are independent risk factors influencing prognosis after radical resection for ICC (all P<0.05). A nomogram was constructed based on these five factors, showing a C-index of 0.807 and AUC of 0.786 in the training cohort, and a C-index of 0.818 and AUC of 0.857 in the validation cohort. Conclusion    The predictive model based on CA19-9, tumor diameter, vascular invasion, differentiation, and TNM stage demonstrates good prognostic prediction accuracy, providing a reference for individualized treatment in clinical practice.

Key words: intrahepatic cholangiocarcinoma, survival prediction, nomogram, prediction model