基于白蛋白-胆红素指数的预测肝内胆管细胞癌根治性切除术后生存率的列线图模型构建  

Construction and evaluation of a nomogram prediction model for survival after radical surgical resection of intrahepatic cholangiocarcinoma based on the albumin-bilirubin index

作　　者：张号枫 李青山 黄冠 杨镇玮 任志元 余海波 Zhang Haofeng;Li Qingshan;Huang Guan;Yang Zhenwei;Ren Zhiyuan;Yu Haibo(Department of Hepatobiliary and Pancreatic Surgery,People’s Hospital of Zhengzhou University(Henan Province People’s Hospital),Zhengzhou 450003,China)

机构地区：[1]郑州大学人民医院(河南省人民医院)肝胆胰腺外科,郑州450003

出　　处：《中华肝胆外科杂志》2023年第6期428-433,共6页Chinese Journal of Hepatobiliary Surgery

基　　金：河南省重点研发与推广专项(222102310709)。

摘　　要：目的基于白蛋白-胆红素指数(ALBI),构建预测肝内胆管细胞癌(ICC)根治性手术切除术后生存率的列线图模型,并对其预测效能进行评估。方法回顾性收集2016年1月至2020年1月在郑州大学人民医院行根治性手术切除的170例ICC患者的临床资料,其中男性90例,女性80例,年龄(58.5±10.6)岁。采用随机数字表法,按照7∶3的比例,将170例ICC患者分为训练集(n=117)与验证集(n=53)。训练集用于列线图模型构建,验证集用于模型验证与评估。采用门诊复查及电话联系的方式进行随访。采用Kaplan-Meier法进行生存分析,基于多因素Cox回归分析中P<0.05的变量绘制列线图,通过训练集和验证集的一致性指数(C-index)、校准曲线以及临床决策曲线分析对预测模型的预测强度进行评估。结果多因素Cox回归分析结果显示,肿瘤糖类抗原19-9(CA19-9)≥37 U/ml(HR=1.99,95%CI:1.10~3.60,P=0.024)、ALBI≥-2.80(HR=2.43,95%CI:1.40~4.22,P=0.002)、脉管癌栓(HR=2.34,95%CI:1.40~3.92,P=0.001)、第8版美国癌症联合委员会(AJCC)N1分期(HR=2.18,95%CI:1.21~3.95,P=0.010)是影响ICC患者根治性手术切除术后生存的独立危险因素。对基于上述变量构建的预测模型进行评估发现模型C-index为0.76,同时校准曲线显示该列线图模型预测ICC患者术后3年生存率的曲线与代表实际生存的45°对角线拟合良好。临床决策曲线分析结果显示,该模型在训练集及验证集中具有明显的正向净收益。结论本研究建立的基于ALBI、CA19-9、脉管癌栓和第8版AJCC N分期共4个变量的ICC根治性手术切除术后生存率的列线图模型,能够为ICC更准确的预后评估和治疗方案的选择提供参考。Objective To construct a nomogram prediction model for survival after radical surgical resection of intrahepatic cholangiocarcinoma(ICC)based on the albumin-bilirubin index(ALBI),and to evaluate its predictive efficacy.Methods From January 2016 to January 2020,170 patients with ICC who underwent radical surgical resection at the People's Hospital of Zhengzhou University were retrospectively analyzed.There were 90 males and 80 females,aged(58.5±10.6)years old.Based on a ratio of 7∶3 by the random number table,the patients were divided into the training set(n=117)and the internal validation set(n=53).The training set was used for nomogram model construction,and the validation set was used for model validation and evaluation.Follow up was conducted through outpatient reexamination and telephone contact.The Kaplan-Meier method was used for survival analysis,and a nomogram was drawn based on variables with a P<0.05 in multivariate Cox regression analysis.The predictive strength of the predictive model was evaluated by analyzing the consistency index(C-index),calibration curve,and clinical decision curve of the training and validation sets.Results Multivariate Cox regression analysis showed that carbohydrate antigen 19-9(CA19-9)≥37 U/ml(HR=1.99,95%CI:1.10-3.60,P=0.024),ALBI≥-2.80(HR=2.43,95%CI:1.40-4.22,P=0.002),vascular tumor thrombus(HR=2.34,95%CI:1.40-3.92,P=0.001),and the 8th edition AJCC N1 staging(HR=2.18,95%CI:1.21-3.95,P=0.010)were independent risk factors affecting postoperative survival of ICC patients after curative resection.The predictive model constructed based on the above variables was then evaluated,and the C-index of the model was 0.76.Calibration curve showed the predicted survival curve of ICC patients at 3 years after surgery based on the model was well-fitted to the 45°diagonal line which represented actual survival.Clinical decision curve analysis showed that the model had a significant positive net benefit in both the training and validation sets.Conclusion The nomograph model for surviv

关 键 词：胆管上皮癌 白蛋白-胆红素指数 列线图 总生存 

分 类 号：R735.8[医药卫生—肿瘤]