论文部分内容阅读
目的建立一个慢性乙型病毒性肝炎(CHB)肝组织炎症活动度无创诊断模型。方法选择CHB患者294例,随机分成建模组(224例)和验模组(70例),均接受肝组织病理检查,同时记录20项血清学指标和超声检查的结果。在建模组对各指标进行单因素分析和多因素Logistic回归分析,筛选出与研究终点相关的独立危险因素,在此基础上构建肝组织炎症活动度无创诊断模型,最后用验模组的相关指标验证模型的诊断效率。结果在单因素分析有统计学意义的14个指标中,筛选出4个指标——高密度脂蛋白(HDL)、碱性磷酸酶(AKP)、天门冬氨酸氨基转移酶(AST)和肝中静脉(MHV),并建立了由这4个指标组成的肝组织炎症活动度无创诊断模型(NDMGLTI=6.599-0.611×logeHDL+0.758×logeAST+0.963×logeAKP-1.098×logeMHV),NDMGLTI评分与肝组织炎症活动度分级呈正相关(rs=0.515,P<0.01),建模组不同肝组织炎症活动度患者的NDMGLTI评分间差异有统计学意义(F=23.577,P<0.01)。受检者工作特征曲线(ROC)分析显示,在建模组NDMGLTI评分诊断显著炎症的曲线下面积(AUC)为0.770,最佳诊断值为10.98,其诊断的敏感度为76.47%,特异度为64.81%,准确度为73.66%,阳性预测值为87.25%,阴性预测值为46.67%,Youden指数为41.28%。验模组NDMGLTI评分诊断显著炎症的AUC为0.849,其敏感度为83.67%,特异度为76.19%,准确度为81.43%,阳性预测值为89.13%,阴性预测值为66.67%,Youden指数为59.86%。结论应用NDMGLTI评分评估CHB患者的肝组织炎症程度具有较好的准确度和可重复性,可以替代肝组织病理检查来指导CHB患者的抗病毒治疗。
Objective To establish a noninvasive diagnostic model of chronic hepatitis B (CHB) liver inflammation. Methods 294 CHB patients were randomly divided into model group (n = 224) and model group (n = 70). All patients underwent histopathological examination, and 20 serological markers and ultrasound findings were recorded. Univariate analysis and multivariate logistic regression analysis of each index in the modeling group, screening out the independent risk factors related to the end of the study, on the basis of which noninvasive diagnosis model of liver tissue inflammatory activity was constructed, and finally with the correlation of the experimental group Indicators verify the diagnostic efficiency of the model. Results Among the 14 indicators that were statistically significant in univariate analysis, four indicators were screened - HDL, AKP, AST and (NDMGLTI = 6.599-0.611 × logeHDL + 0.758 × logeAST + 0.963 × logeAKP-1.098 × logeMHV). The NDMGLTI score and liver (Rs = 0.515, P <0.01). There was significant difference in NDMGLTI scores among the different groups of hepatic inflammation activity in the model group (F = 23.577, P <0.01). The receiver operating characteristic curve (ROC) analysis showed that the area under the curve (AUC) of the NDMGLTI score for diagnosis of significant inflammation in the modeling group was 0.770, the best diagnostic value was 10.98, the diagnostic sensitivity was 76.47% and the specificity 64.81%, accuracy was 73.66%, positive predictive value was 87.25%, negative predictive value was 46.67%, Youden index was 41.28%. The AUC of NDMGLTI in diagnosis of significant inflammation was 0.849, the sensitivity was 83.67%, the specificity was 76.19%, the accuracy was 81.43%, the positive predictive value was 89.13%, the negative predictive value was 66.67%, the Youden index was 59.86 %. Conclusion The NDMGLTI score can be used to evaluate the degree of liver inflammation in CHB patients with good accuracy and repeatability. It can be used as an alternative to liver biopsy to guide the antiviral treatment of CHB patients.