慢性丙型肝炎肝纤维化无创性诊断模型的建立

Establishment of the non-invasive diagnostic model of liver fibrosis in chronic hepatitis C

目的建立由血清学指标组成的慢性丙型肝炎(丙肝)肝纤维化无创性综合诊断模型。方法收集慢性丙肝肝纤维化患者91例,随机分成模型组(68例)和验证组(23例),对年龄、性别以及27项血清学参数行单因素方差分析和多因素Logistic回归分析,筛选出独立预测因子并构建指数模型(FMIC模型),采用受试者工作特征曲线(ROC曲线)评估模型的诊断能力。同时将FIMC模型与其他模型进行比较。结果建立了一个由血小板(PLT)、总胆红素(T-Bil)和透明质酸(HA)3项指标构成的判别丙肝肝纤维化程度的诊断模型FMIC。肝纤维化S2、S3、S4分期的cut off值分别为3.1、4.7、8.9,其ROC曲线下面积(AUC^(ROC))分别为0.900、0.943、0.860;FMIC模型联合指标的AUC^(ROC)明显高于单独使用PLT、T-Bil、HA指标(AUC^(FMIC)=0.905,AUC^(PLT)=0.611,AUC^(T-Bil)=0.805,AUC^(HA)=0.789);FIMC模型的诊断效率高于其他模型(AUC^(FMIC)=0.943,AUC^(Forns index)=0.629,AUC^(APRI)=0.446,AUC^(CDS)=0.682,AUC^(API)=0.582,AUC^(AAR)=0.405)。结论建立的丙肝肝纤维化程度的诊断模型FMIC,提高了诊断效率,避免了不必要的肝脏活检。

Objective To establish a comprehensive non-invasive diagnostic model of liver fibrosis in chronic hepatitis C based on hematology markers. Methods Ninety-one chronic hepatitis C patients with liver fibrosis were enrolled and randomly divided into a model group（ n = 68） and a verification group（ n = 23）. The age,sex and 27 hematology markers in the model group were performed the univariate ANOVA analysis and multivariate Logistic regression analysis,respectively,to screen the independent predictive factors and construct the diagnostic model of fibrosis in chronic hepatitis C（ FMIC model）. Then,the diagnostic efficiency of the FMIC model was assessed by receiver operating characteristic curve（ ROC curve）,and compared with other models. Results The established FMIC model included 3 markers,platelet count（ PLT）,total bilirubin（ T-Bil） and hyaluronidase（ HA）. The cut-off values and the area under the ROC curve（ AUC^（ROC）） in stages S2,S3 and S4 of liver fibrosis were 3. 1 and 0. 900,4. 7 and 0. 943,and 8. 9 and 0. 860,respectively. The AUC^（ROC）value in the FMIC model constructed from PLT,T-Bil and HA（ AUC^（FMIC）= 0. 905） was significantly higher than that from single marker（ AUC^（PLT）= 0. 611,AUCT-Bil = 0. 805,AUC^（HA）= 0. 789）. The diagnostic efficiency of the FIMC model（ AUC^（FMIC）= 0. 943） was higher than that of other models（ AUC^（Forns）index = 0. 629,AUC^（APRI）= 0. 446,AUC^（CDS）= 0. 682,AUC^（API）=0. 582,AUC^（AAR）= 0. 405）. Conclusion The established FMIC model for the diagnosis of liver fibrosis in chronic hepatitis C improves the diagnostic efficiency,and avoids the unnecessary liver biopsy.