血清学指标在慢性乙型肝炎肝纤维化诊断中的应用评价及神经网络纤维化诊断模型的建立

目的对17项肝纤维化血清学指标诊断肝纤维化的价值进行评价，并筛选出有意义的指标用于构建肝纤维化神经网络诊断模型。方法留取254例本院经病理诊断明确为肝纤维化的慢性乙型肝炎患者的血清，并对17项与肝纤维化相关的血清学指标进行检测及统计学分析。利用多因素Logistic回归分析筛选出6项诊断价值较高的血清学指标作为诊断模型的输入指标，建立神经网络诊断模型。结果共13项指标与纤维化程度具有相关性（P〈0．05），但单项指标的诊断价值均不理想[受试者工作特征曲线下面积（alea under receiver characteristic curve，AUROC）〈0．75]。将年龄、γ-谷氨酰转移酶、PLT、PT金、属蛋白酶组织抑制因子-1和DBIL这6项指标组合为诊断模型，其诊断价值优于任何单项指标（AUROC--0．834），以此建立的神经网络模型诊断正确率、灵敏度和特异度分别为77．9％、82．7％和75．4％。结论通过本实验筛选的6项指标组合具有较高的诊断效率，同时以此为基础建立的神经网络模型也具有很好的临床诊断灵敏性和特异性，具有较好的临床应用前景。

Evaluation of serum markers in the diagnosis of liver fibrosis in patients with chronic hepatitis B and the development of an artificial nerual network diagnosis model for liver fibrosis

Objective To evaluate the diagnostic value of 17 fibrosis serum markers,and to develop an artificial nerual network(ANN) model for the diagnosis of liver fibrosis with screened meaningful markers.Methods Totally 254 serum samples from the patients with chronic hepatitis B(CHB),who were confirmed to have developed liver fibrosis with biopsy in our hospital,were collected,and 17 serum markers correlated with liver fibrosis were detected and analyzed statistically.Multivariable regression analysis was conducte...