肝胆胰外科杂志2026，Vol.39Issue(4)：229-238,10.DOI:10.11952/j.issn.1007-1954.2026.04.001
基于Lasso回归分析构建乙型肝炎相关HCC术后早期复发的预测模型
Construction of a prediction model for early postoperative recurrence of hepatitis B-related hepatocellular carcinoma based on Lasso regression analysis
摘要
Abstract
Objective To analyze the risk factors for early recurrence of hepatitis B-related hepatocellular carcinoma(HCC)after radical resection and construct a nomogram prediction model.Methods The clinical and pathological data of patients with HBV-related HCC who underwent radical hepatectomy at the First Affiliated Hospital of Xinjiang Medical University from August 2017 to June 2023 were collected retrospectively.Lasso regression,univariate and multivariate Logistic regression analyses were used to screen the risk factors for early recurrence of hepatitis B-related HCC after radical resection,and a nomogram prediction model was established.The area under the receiver operating characteristic(AUC)curve,Hosmer-Lemeshow goodness-of-fit test,calibration curve,1 000 Bootstrap resampling test,decision curve analysis(DCA),and clinical impact curve(CIC)were used to evaluate and internally validate the model performance.The applicability of the model in different populations was verified by plotting ROC curves of different characteristic subgroups.Results A total of 245 patients were included in this study,among whom 125 cases(51%)had early recurrence.Through Lasso regression,univariate and multivariate Logistic regression analyses,it was finally determined that HBV-deoxyribonucleic acid(DNA),gamma-glutamyl transferase(GGT),alpha-fetoprotein(AFP),surgical margins,intact capsule,and tumor differentiation degree(Edmondson grade)were important predictors for early recurrence of hepatitis B-related HCC after radical resection.A nomogram prediction model for predicting early postoperative recurrence in hepatitis B-related HCC patients was constructed based on the above 6 indicators.The AUC of this model was 0.764(95%CI 0.704 to 0.823).The P value of the Hosmer-Lemeshow goodness-of-fit test was 0.839.The calibration curve was close to the ideal standard line.Internal validation using the 1 000 Bootstrap resampling test indicated that the model had good stability,and both the DCA curve and CIC curve showed that the model provided greater clinical benefits.ROC curves and AUC values of various subgroups showed that the model had good discriminatory ability for populations with different characteristics.Conclusion HBV-DNA≥1 000 IU/mL,GGT,AFP≥400 ng/mL,narrow surgical margins,intact capsule,and EdmondsonⅢ/Ⅳ grade are important predictors for early recurrence of hepatitis B-related HCC after radical resection.The nomogram prediction model constructed based on these 6 indicators has good predictive efficacy and certain guiding significance for individualized prevention and treatment strategies.关键词
肝细胞癌/早期复发/危险因素/列线图/预测模型/乙型肝炎Key words
hepatocellular carcinoma/early recurrence/risk factors/nomogram/prediction model/hepatitis B分类
医药卫生引用本文复制引用
杨宇航,鲁雪梅,丛赟,邵英梅..基于Lasso回归分析构建乙型肝炎相关HCC术后早期复发的预测模型[J].肝胆胰外科杂志,2026,39(4):229-238,10.基金项目
新疆维吾尔自治区自然科学青年科学基金(2023D01C216) （2023D01C216）
国家自然科学基金(82360111) （82360111）
省部共建中亚高发病成因与防治国家重点实验室开放课题(SKL-HIDCA-2023-2). （SKL-HIDCA-2023-2）
