Establishment And Application Of A Noninvasive Diagnostic Model For Liver Inflammation Activity In Chronic HBV Infected Patients

Background:Chronic hepatitis B infection can be character as chronic hepatitis B infectors,chronic hepatitis B(CHB)and cirrhosis.In clinical practice,HBV carriers can be defined as hepatitis B infection whose alanine aminotransferase(ALT)is less than two times the upper limit of normal(ALT <2 ULN).In chronic hepatitis B infection with ALT <2ULN,some patients only have normal or slightly inflammation or/and fibrosis in liver tissue,these person are truly HBV carriers,so they do not need antiviral treatment.But in some chronic hepatitis B infection,although ALT are normal or slightly elevated,there is significant inflammation or/and fibrosis in liver tissue,so they need accept antiviral therapy at once.ALT is the most widely used serum maker to assess liver inflammation in clinical practice.According to the authoritative guidelines,it is necessary to assess whether there is significant inflammation or/and fibrosis in chronic HBV infection and ALT <2 ULN.Nowadays,liver biopsy is still the gold standard for assessing liver inflammation and fibrosis;however,liver biopsy is an invasive test,complicated operation and high cost,so it does not meet the needs of chronic HBV infection screening.Therefore,more and more studies are now focused on the non-invasive assessment of liver tissue inflammation and fibrosis.In previous studies,some diagnostic models were established for the non-invasive diagnosis of fibrosis,but fibrosis is not accordance with liver inflammation,and the stage of fibrosis cannot represent liver the degree of inflammation.There are few models for the non-invasive diagnosis of inflammation,so a more valuable non-invasive diagnosis of inflammation model are urgently needed.The main reason of chronic HBV infection leads to liver inflammation is due to the damage by the body’s immune system.Studies have shown that inflammatory infiltration of liver tissue is accompanied by increased expression of chemokines.Morever,chemokines can change earlier than enzyme,so the chemokines may play more sensitivity.Therefor,chemokines may as a potential marker for liver inflammation.Objective:(1)Screening serum inflammatory markers with liver inflammation:detected and analyzed clinical virus index,biochemical index and chemokines with liver inflammation in order to screen serum predictors of significant liver inflammation;(2)Development of a non-invasive model: Selected predictive indexes were used to develop a non-invasive model;(3)Predictive value and clinical application: Selected chemokines were further verified in liver tissue.In addition,expanded the serum sample size and verified the application value of established model.Methods: Part one:298 patients with chronic HBV infection(ALT<2ULN)who were hospitalized in the department of Infectious Diseases,Shengjing Hospital affiliated to China Medical University from January 2012 to December 2016 were enrolled.All the patients underwent liver biopsy.Besides,all the patients had not received any antiviral therapy.Demographics data and serology data were recorded,including age,sex,family history,ALT,aspartate transaminase(AST),alkaline phosphatase(ALP),gamma glutamyl-transferase(GGT),cholinesterase(CHE),Albumin(ALB),total bilirubin(TBil),prothrombin time(PT),activated partial thromboplastin time(APTT),prothrombin time activity(PTA),white blood cell(WBC),hemoglobin(HGB),platelet(PLT),hepatitis B surface antigen(HBs Ag),hepatitis B e antigen(HBe Ag),hepatitis B e antibody(HBe Ab),hepatitis B core antibody(HBc Ab)and so on.Investigate the feature of liver pathology and find significant markers to predict liver inflammation through the clinical indicators analysis.Histologic assessment was based on the Scheuer’s classification,significant abnormality was defined as necroinflammation grade ≥ G2 and/or fibrosis stage ≥ S2.All the data analysis were used by SPSS 22.0 software.The data was tested by normality test at first.If the data was normal distribution,mean±standard deviation was used to express the variable,tow-sample t-test for independent samples was to compare between the two groups;If not,median(25%-75% percentile)was used to express the variable,rank sum test was used to compare.The area under the curve(AUC),negative predictive value(NPV),positive predictive value(PPV),sensitivity and specificity were used to assess the serum maker value.All the test are two-tailed test,P value< 0.05 has the statistical significance.Part two: 122 ALT<2ULN chronic HBV infection patients were enrolled form the whole cohort.The cohort was divided into three groups: model group(52 cases)and validation group(70 cases).In the model group,a multi-factor detection kit was used to evaluate expression of chemokines in peripheral blood.13 chemokines include Eotaxin/CCL11,thymus activation regulated chemokine(TARC/CCL17),Monocyte chemotactic protein-1(MCP-1/CCL2),regulated upon activation normal T cell expressed secreted(RANTES/CCL4),Epithelial neutrophil activating peptide 78(ENA-78/CXCL5),chemokine growth-regulated oncogene-α(GRO-α),interferon-γ-inducible T cell α chemoattractant(I-TAC/CXCL11),monokine induced by gamma-interferon(MIG/CXCL9),macrophage inflammatory protein 3α(MIP-3α/CCL20),interleukin-8(IL-8/CXCL8),interferon-γ-inducible protein-10(IP-10/CXCL10),macrophage Inflammatory Protein-1α(MIP-1α/CCL3)and macrophage inflammatory Protein-1β(MIP-1β/CCL4).Relativity analysis and difference analysis to confirm significant independent predictors for liver necroinflammation.Then we develop a non-invasive model to predict liver inflammation in chronic HBV infection patients by logistic analysis.AUC,NPV,PPV,sensitivity and specificity were used to assess the model value.Part three: Immunohistochemical was used to detect the expression of CCL2 and CCL17 in liver tissue.Furthermore,expend sample and verified the model value for the assessment of liver inflammation both in whole group and validation group.AUC,NPV,PPV,sensitivity and specificity were used to assess the diagnostic value in whole group and validation group.Results:1.298 patients whose ALT<2ULN,including Hbe Ag positive 177 cases and Hbe Ag negative 121 cases;129 cases(43.3%)had inflammation grade ≥2(G≥2),74 cases(24.8%)had fibrosis grade ≥2(S≥2).The occurrence rate of significant necroinflammation is higher than the rate of significant fibrosis.2.Age,sex,ALB,TBIL and HGB had no relationship with liver inflammation.In Hbe Ag positive group,ALT,AST,ALP,GGT,APTT and Hbc Ab have a positive correlation with the degree of liver inflammation,PLT,HBs Ag and HBV-DNA have a negative correlation with the degree of liver inflammation.In Hbe Ag negative group ALT,AST,ALP,PT,HBc Ab and HBV-DNA have a positive correlation with the degree of liver inflammation,CHE,PTA,PLT have a negative correlation with the degree of liver inflammation.3.ALT,AST,ALP and PLT have significant difference in the degree of liver inflammation.AST has the highest valuable for the assessment of liver inflammation.The AUC were 0.696.4.CCL11,CCL17,CCL2,CXCL5,CXCL1 and CXCL11 in significant necroinflammation group were significant higher than the patients with normal or slightly liver inflammation.5.CCL17,CCL2,CXCL5,CXCL1,ALT and AST were significant independent predictors for liver necroinflammation.We develop a diagnostic model for predict liver inflammation by logistic analysis:G = 0.009×CCL2(pg/ml)+ 0.015×CCL17(pg/ml)-5.272.6.In model group,when use the lower cut-off point-3.699,the sensitivity was 96.15%,the specificity was 46.15%,NPV was 92.3% and the PPV was 64.1%;When use the higher cut-off point-0.546,the sensitivity was 46.15%,the specificity was 100%,the NPV was 100% and the PPV was 65%.In the validation group,when use the lower cut-off point-3.699,the sensitivity was 100%,the specificity was 9.1%,the NPV was 100%% and the PPV was 41.18%;When use the higher cut-off point-0.546,the sensitivity was 53.75%,the specificity was 85.71%,the NPV was 73.47% and the PPV was 71.43%.In HBV infection group,when use the lower cut-off point-3.699,the sensitivity was 98.15%,the specificity was 22.06%,the NPV was 93.76%,the PPV was 50%;When use the higher cut-off point-0.546,the sensitivity was 50%,the specificity was 91.18%,the NPV was 69.67%,the PPV was 81.82%.Conclusion:1.In chronic HBV infection patients whose ALT<2ULN,about half patients with significant necroinflammation and/or fibrosis.The occurrence rate of significant necroinflammation is higher than the rate of significant fibrosis.2.CCL2 and CCL17 was positive correlated with liver necro-inflammation,and can regard as a sensitive marker to predict the liver necroinflammation.3.The expression of CCL2 and CCL17 were significantly elevated in both serum and liver tissue in CHB patients with significant liver necro-inflammation.CCL2 and CCL17 can be used to develop a non-invasive model to predict liver inflammation in chronic HBV infection patients.4.The model is valuable for the assessment of liver necroinflammation in clinical diagnosis.It can let part of patients avoid liver biopsy.