原发性肝癌干预前并发慢性肝衰竭的风险预测模型建立与评价  

Establishment and Evaluation of a Risk Prediction Model for Chronic Liver Failure Complicated by Primary Hepatocellular Carcinoma Before Intervention

作　　者：王远珍 魏红艳 田仁海 陈永臻 许丹青 张映媛 常丽仙 刘春云 刘立 WANG Yuanzhen;WEI Hongyan;TIAN Renhai;Chen Yongzhen;XU Danqing;ZHANG Yingyuan;CHANG Lixian;LIU Chunyun;LIU Li(Department of Liver Diseases and Immunology,The Third People’s Hospital of Kunming,Yunnan Province Communicable Disease Clinical Medical Center,Kunming Yunnan 650041;School of Public Health,Dali University,Dali Yunnan 671000,China)

机构地区：[1]昆明市第三人民医院肝病免疫科,云南省传染病临床医学中心,云南昆明650041 [2]大理大学公共卫生学院,云南大理671000

出　　处：《昆明医科大学学报》2025年第3期139-147,共9页Journal of Kunming Medical University

基　　金：昆明市卫生健康委基金项目(2022-03-08-008)。

摘　　要：目的 分析原发性肝癌(PHC)干预前并发慢性肝衰竭的影响因素,建立列线图风险预测模型并进行评价。方法 收集PHC的一般资料和入院24 h内临床检测资料等进行回顾性分析,使用单因素分析和Lasso回归进行变量筛选后进行多因素Logistic回归分析,确定PHC干预前并发慢性肝衰竭的独立影响因素,据此建立PHC干预前并发慢性肝衰竭的列线图风险预测模型。使用Hosmer-Lemeshow检验、绘制受试者工作特征曲线(ROC)、校准曲线、临床决策曲线、临床影响曲线对模型进行评价。结果 共收集PHC患者353例,肝衰竭组(n=153)和非肝衰竭组(n=200),患病率43.3%。Lasso回归筛选出的变量有消化道出血、凝血酶原时间(PT)、白蛋白(ALB)、总胆红素(TBIL)、γ谷氨酰转移酶(GGT),进行多因素Logistic回归分析结果显示:消化道出血(OR=13.549, 95%CI2.899~63.322, P=0.001)、 PT(OR=1.599, 95%CI1.282~1.995, P<0.001)、TBIL(OR=1.016,95%CI 1.006~1.025,P=0.002)、GGT(OR=1.002,95%CI 1.000~1.003,P=0.028)是PHC干预前并发慢性肝衰竭的独立影响因素,据此建立列线图风险预测模型。Hosmer-Lemeshow检验显示该模型具有较好的拟合度(χ^(2)=6.152,P> 0.05);ROC下面积为0.902(0.869~0.934),灵敏度80.4%,特异度87.5%。校准曲线显示该模型预测PHC干预前并发慢性肝衰竭具有较好的一致性。临床决策曲线和临床影响曲线显示在一定阈值范围内该模型具有较好的临床实用性。结论 有消化道出血、PT≥16.05 s、TBIL≥37.80 mmol/L、GGT≥99.00 U/L是PHC干预前发生慢性肝衰竭的独立危险因素,建立的列线图风险预测模型在预测PHC干预前发生慢性肝衰竭的风险时具有一定的临床应用价值。Objective To analyze the influencing factors of chronic liver failure in patients with primary hepatic carcinoma(PHC)before intervention,and to establish and evaluate a nomogram risk prediction model.Methods A retrospective analysis was conducted by collecting general data and clinical test data within 24 hours of admission for PHC patients.Univariate analysis and Lasso regression were used for variable selection,followed by multivariate logistic regression analysis to identify independent influencing factors for CLF before PHC intervention,leading to the establishment of a nomogram risk prediction model.The model was evaluated using the Hosmer-Lemeshow test,receiver operating characteristic(ROC)curve,calibration curve,clinical decision curve,and clinical impact curve.Result A total of 353 cases of PHC patients were collected,including 153 cases in the liver failure group and 200 cases in the non-liver failure group,with a prevalence rate of 43.3%.Variables selected by Lasso regression included gastrointestinal bleeding,prothrombin time(PT),albumin(ALB),total bilirubin(TBIL),and gamma glutamyl transferase(GGT).Multivariate logistic regression analysis showed that gastrointestinal bleeding(OR=13.549,95%CI:2.899~63.322,P=0.001),PT(OR=1.599,95%CI:1.282~1.995,P<0.001),TBIL(OR=1.016,95%CI:1.006~1.025,P=0.002),and GGT(OR=1.002,95%CI:1.000~1.003,P=0.028)were independent risk factors for chronic liver failure prior to PHC intervention,leading to the establishment of a nomogram risk prediction model.The Hosmer Lemeshow test showed that the model had a good fit(χ^(2)=6.152,P>0.05);the area under ROC was 0.902(0.869-0.934),with a sensitivity of 80.4%and a specificity of 87.5%.The calibration curve indicated that the model predicts chronic liver failure prior to PHC intervention with good consistency.Clinical decision curve analysis and clinical impact curve analysis showed that the model has good clinical utility within a certain threshold range.Conclusion Gastrointestinal bleeding,PT≥16.05s,TBIL≥37.80 mmol/L,and GGT�

关 键 词：原发性肝癌 慢性肝衰竭 影响因素 预测模型 

分 类 号：R735.7[医药卫生—肿瘤]