检验医学与临床2026，Vol.23Issue(1)：112-119,8.DOI:10.3969/j.issn.1672-9455.2026.01.019
基于腹水、IBIL、Ca、ALB构建慢性乙型肝炎肝硬化服用芪苓益肝颗粒的疗效预测模型
Based on ascites,IBIL,Ca and ALB,a predictive model for the efficacy of Qiling Yigan granules in hepatitis B cirrhosis was constructed
摘要
Abstract
Objective To construct and verify a predictive model for the efficacy of Qiling Yigan granule in the treatment of chronic hepatitis B cirrhosis based on ascites,indirect bilirubin,serum calcium and albumin,so as to provide decision support for clinical precision medication.Methods A total of 135 patients with chro-nic hepatitis B cirrhosis treated with Qiling Yigan granules for more than 3 months in Shanxi Academy of Traditional Chinese Medicine/Shanxi Provincial Hospital of Traditional Chinese Medicine from May 2020 to April 2023 were selected as the internal training set,and 30 patients from May 2023 to May 2024 were selected as the external validation set.The potential predictors were screened by univariate analysis,and the Gaussian naive Bayes prediction model was established by combining the Lasso regression feature selection and Akaike information criterion optimization.The receiver operating characteristic(ROC)curve was used to evaluate the discrimination,the Bootstrap method calibration curve was used to evaluate the calibration of the model,and the clinical practicability was verified by decision curve analysis.Results Based on the clinical significance and univariate analysis,18 factors including red blood cell count in the internal training set were included in the Lasso linear regression,and the results showed that ascites(β=0.842),indirect bilirubin(β=0.004),serum calcium(β=-0.237)and albumin(β=-0.066)were the key predictors.According to the above four predic-tors,nine prediction models were constructed,among which the Gaussian naive Bayes model was the best,with an AUC of 0.866(95%CI:0.699-1.033)in the internal training set and 0.869(95%CI:0.705-1.000)in the external validation set.The calibration curve showed that the prediction probability of the model was good(Brier Score=0.144).Based on this,an online prediction tool was developed,and four factors including the de-gree of ascites were input to obtain the prediction conclusion:if the disease probability was≤39.2%,the pre-diction was effective,and Qiling Yigan granules were recommended,and the converse is invalid.Conclusion The online prediction tool for efficacy based on ascites,indirect bilirubin,serum calcium and albu-min can reliably predict the efficacy of Qiling Yigan granules in patients with hepatitis B cirrhosis.关键词
慢性乙型肝炎肝硬化/芪苓益肝颗粒/疗效/预测模型/高斯朴素贝叶斯模型/腹水/间接胆红素/血钙/清蛋白Key words
chronic hepatitis B cirrhosis/Qiling Yigan granule/efficacy/prediction model/gaussi-an naive bayesian model/ascites/indirect bilirubin/blood calcium/albumin分类
医药卫生引用本文复制引用
杨萌,王晓玲,王莹,刘利,谢玺媛,周宇捷,蔺薛瑞,秦硕,李怀长,姜楠..基于腹水、IBIL、Ca、ALB构建慢性乙型肝炎肝硬化服用芪苓益肝颗粒的疗效预测模型[J].检验医学与临床,2026,23(1):112-119,8.基金项目
山西省中医药管理局科研课题计划(2023ZYYDA2001) （2023ZYYDA2001）
山西省中医药研究院院级课题(202301) （202301）
山西省中医药研究院项目(202404) （202404）
山西省基础研究计划项目(202303021222369). （202303021222369）
