年龄≤30岁转氨酶正常的慢性乙型肝炎肝纤维化患者无创诊断模型的建立

Establishment of a noninvasive diagnostic model for chronic hepatitis B liver fibrosis patients with normal aminotransferases aged≤30 y

目的从慢性乙型肝炎(CHB)患者的临床、生化及影像学等常用无创伤性指标中筛选出特定指标,构建年龄≤30岁、ALT水平正常CHB患者肝纤维程度无创性诊断模型,在一定程度上避免该类患者行有创的肝穿刺活检,从而指导其抗病毒治疗时机。方法收集2019年1月—2022年1月在深圳市第三人民医院及长沙市第一医院行肝穿刺活检的年龄≤30岁ALT正常的CHB患者251例,其中模型组175例,验证组76例,根据临床经验及相关文献筛选出临床常用指标,计量资料组间比较采用成组t检验或两独立样本Mann-Whitney U秩和检验;计数资料采用χ^(2)检验。经Spearman等级相关逐一分析各指标与肝纤维化之间的关联性,筛选出有关联的指标(P<0.01,r>0.200),采用Logistic回归创建无创诊断模型,使用受试者工作特征曲线(ROC曲线)评价其能力并进行验证,将该模型与目前广泛使用的模型APRI及FIB-4进行比较。病理结果一致性采用诊断试验的Kappa一致性检验。结果筛选17项临床常用指标,其中9项指标[ALT、AST、GGT、铁蛋白(FERR)、血小板计数(PLT)、Ⅲ型前胶原氨基端肽(PⅢP)、Ⅳ胶原(CⅣ)、HBV DNA以及脾厚]与肝纤维化有关联(P<0.01,r>0.232),根据以上指标建立预测模型为:P=1/(1+e^(-γ)),γ=-1.902+0.106×AST-0.011×PLT-0.265×HBV DNA+0.059×PⅢP(P为预测患者肝纤维化≥S2的概率值,γ为预测指数),单项指标与模型比较提示该模型曲线下面积最大为0.852,其敏感度为92.7%,特异度为76.9%;以76例患者进行验证,准确度为77.6%。该模型与模型APRI及FIB-4进行比较,显示该模型有较好的准确性。结论本研究创建的无创诊断模型相较于国际上常用模型APRI及FIB-4能更准确地判断该类患者肝纤维化程度,在一定程度上可替代肝组织活检,并指导其抗病毒治疗时机。

Objective To establish a noninvasive diagnostic model for liver fibrosis in chronic hepatitis B(CHB)patients with normal alanine aminotransferase(ALT)and an age of≤30 years by selecting specific indicators from the commonly used noninvasive indicators such as clinical,biochemical,and imaging indicators,to avoid invasive liver biopsy in such patients to some extent,and to guide the timing of antiviral therapy.Methods A total of 251 CHB patients with normal ALT and an age of≤30 years who underwent liver biopsy in Shenzhen Third People’s Hospital and The First Hospital of Changsha from January 2019 to January 2022 were enrolled,with 175 patients in the model group and 76 patients in the validation group,and commonly used clinical indicators were obtained based on clinical experience and related articles.The two-independent-samples t test or the twoindependent-samples Mann-Whitney U rank sum test was used for comparison of continuous data between groups,and the chisquare test was used for comparison of categorical data between groups.A Spearman rank correlation analysis was used to investigate the correlation between each indicator and liver fibrosis and identify the indicators with correlation(P<0.01,r>0.200);a Logistic regression analysis was used to establish a noninvasive diagnostic model,and the receiver operating characteristic(ROC)curve was used to evaluate its performance and perform validation of the model;this model was then compared with the widely used models of aspartate aminotransferase-to-platelet ratio index(APRI)and fibrosis-4(FIB-4).The Kappa consistency test was used to investigate the consistency of pathological results.Results A total of 17 commonly used clinical indicators were obtained,among which 9 indicators(ALT,aspartate aminotransferase[AST],gamma-glutamyl transpeptidase[GGT],ferritin[FERR],platelet count[PLT],procollagen typeⅢamino-terminal peptide[PⅢP],collagenⅣ[CⅣ],HBV DNA,and spleen thickness)were correlated with liver fibrosis(P<0.01,r>0.232).Based on the above indicators,the predictive model was established as P=1(/1+e^(-γ)),γ=-1.902+0.106×AST-0.011×PLT-0.265×HBV DNA+0.059×PⅢP,in which P was the probability for predicting≥S2 liver fibrosis andγwas the predictive index.The comparison between each indicator and the model showed that the model had the largest area under the ROC curve of 0.852,with a sensitivity of 92.7%and a specificity of 76.9%.The model was validated in 76 patients and showed an accuracy of 77.600%.The model was compared with APRI and FIB-4,and the results showed that the model has good accuracy.Conclusion Compared with the models of APRI and FIB-4 commonly used in the world,this model can more accurately judge the degree of liver fibrosis in such patients,thereby replacing liver biopsy to some extent and guiding the timing of antiviral therapy.