This study was conducted using a cohort of 94 CHB patients receiving NAs monotherapy, who were prospectively recruited from Beijing YouAn Hospital, Capital Medical University (Beijing, China) between June 2007 and July 2008. CHB patients who were diagnosed by the American Association for the Study of Liver Diseases guideline (Terrault et al., 2016) were enrolled if they aged ≥ 16 years and were treat-naïve. The exclusion criteria were as follows: (a) co−infection with hepatitis C, hepatitis D virus, or human immunodeficiency viruses; (b) existence of the alcoholic liver disease or autoimmune liver disease; (c) with decompensated cirrhosis or HCC; (d) with a history of immunosuppressive therapy or organ transplantation; (e) pregnant or breastfeeding women.

Once recruitment, patients were given entecavir (ETV) or adefovir (ADV) and followed up. At each follow-up, serum samples were collected for HBV DNA quantification and liver function tests. The remaining serum specimens were stored at -80°C for subsequent research. At enrollment and Month 60, percutaneous liver biopsies were performed to evaluate the histology. With cryopreserved serum samples, HBsAg and HBcrAg levels at baseline, months 6 and 60, as well as HBV RNA levels at baseline, months 6, 12, 24, 36, 48, and 60 were retrospectively quantified.

This study followed the Declaration of Helsinki and was approved by the Institutional Review Board of Beijing YouAn Hospital, Capital Medical University (Beijing, China). All subjects provided written informed consent.

Serum HBsAg was quantified by Elecsys HBsAg II Quant reagent kits (Roche Diagnostics) with a lower limit of detection (LLD) of 0.05 IU/mL. Quantitative levels of HBcrAg were determined by chemiluminescent enzyme immunoassay in an automated analyzer system (Fujirebio Inc., Tokyo, Japan) with the LLD of 1,000 IU/mL and a linear range of 3-7 log IU/mL. Serum HBV RNA was isolated with the nucleic acid extraction or purification kit (Sansure Biotech, Changsha, China) and treated with DNase I (Thermo Fisher Scientific, Waltham, MA, USA). The LLD of the assay was 200 copies/mL. Details on HBV RNA assay can be found in Supplementary Materials . The serum HBV DNA level was determined using the Cobas HBV Amplicor Monitor assay (Roche Diagnostics, Pleasanton, CA, USA) with an LLD of 50 IU/mL.

At baseline and after 60 months of NAs treatment, a percutaneous liver biopsy was performed. A minimal 18mm length of liver tissue containing at least three complete portal tracts was obtained for pathological evaluation (Colloredo et al., 2003). All liver biopsies were reviewed continuously by an experienced pathologist blinded to treatment assignment and time of biopsy. Inflammation was assessed according to the Scheuer scoring system, which is entirely based on histology results. And histologic findings of portal inflammation, interface hepatitis, and lobular inflammation are assigned a score ranging from 0 to 4 (Scheuer, 1991). G≥3 was defined as having significant inflammation. Inflammation regression was defined as a ≥1-grade decrease according to the Scheuer scoring system.

For DNA extraction, about 30 μm formalin fixation and paraffin embedding (FFPE) liver biopsy tissue in sections of 6 μm each were used. QIAamp FFPE DNA Mini Kit (QIAGEN, GmbH, Hilden, Germany) was used to extract DNA according to the instructions of the manufacturer. HBV rcDNA, replicative dsDNA, and ssDNA were digested using T5 Exonuclease (New England Biolabs, USA). The reaction mixture contained 100 ng extracted DNA, 0.5 µL (10 units) T5 Exonuclease, 1 µL NEBuffer 4 (10×) with Nuclease-free H2O to a final volume of 10 μL. The digestion was conducted at 37°C for 1 h, and stop the reaction with EDTA to at least 11 mM. We combined 6.42 μL of digestion product obtained in the previous step, with 7.50 μL QuantStudio™ 3D Digital PCR Master Mix, 0.06 μL of TaqMan Probe-RC-MGB (50 μM), 0.06 μL TaqMan Probe-RNAseP-VIC (50 μM), 0.24 μL primer of rc-F, 0.24 μL primer of rc-R, 0.24 μL primer of RNaseP-F and 0.24 μL primer of RNaseP-R. 15μL of this sample mix was added to each chip and loaded on ProFlex™ 2x Flat PCR System. Absolute quantification was determined with QuantStudio™ 3D Digital PCR System (Thermo Fisher Scientific Inc., Waltham, Massachusetts, USA) and analyzed using QuantStudio 3D AnalysisSuite Cloud Software. (https://china.apps.thermofisher.com/quantstudio3d/). Intrahepatic HBV cccDNA values were normalized to cell number assessed by RNase P copy number assay.

Continuous variables were expressed as medians and ranges or means and standard deviations and categorical variables as frequencies. Differences between groups were analyzed using the student t or Mann-Whitney tests for continuous parameters and chi-square or Fisher exact tests for categorical parameters, as appropriate. The receiver operating characteristic (ROC) analysis assessed the markers’ diagnostic capacity for inflammation with the cut-off values determined using the Youden index. The 95% confidence interval of the area under the ROC curve (AUROC) was determined using a bootstrap method. The regression and Spearman correlation coefficients (r) were used to depict the correlation between the two variables. All analyses were performed with SPSS version 26.0, with a two-tailed p-value <0.05 considered statistically significant.

A total of 94 CHB patients who performed liver biopsies at baseline were analyzed in this study, including 74 HBeAg-positive and 20 HBeAg-negative patients. Table 1 summarizes the characteristics of this population, which was predominantly male (n=74, 78.7%) with a median age of 35.5 years and a median BMI of 23.8. Forty-two patients had available HBV genotype data, with 71.4% (30/42) genotype C. Of the 94 patients, 47.9% (45/94) were treated with ETV, and 52.1% (49/94) were treated with ADV after recruitment. The median serum ALT and AST were 69.8 U/L and 44.1 U/L, respectively. Median levels of baseline serum HBV RNA, HBV DNA, HBsAg and HBcrAg were 5.08 log10copies/mL, 6.27 log10IU/mL, 3.52 log10IU/mL and 6.72 log10IU/mL, respectively. Median levels of intrahepatic HBV DNA and cccDNA are 6.44 and 4.73 log10 copies/10⁵ cells.

Compared to HBeAg-negative CHB patients, HBeAg-positive patients have significantly higher serum HBV DNA, HBV RNA, HBsAg and HBcrAg, as well as intrahepatic HBV DNA and cccDNA.

Spearson’s correlation coefficients (r) were used to evaluate the correlation of serum and intrahepatic HBV markers, ALT and AST with inflammation grade according to the Scheuer scoring system. As shown in Figure 1 , serum HBsAg and HBcrAg levels weakly negatively correlated with inflammation grade (r=-0.39, P=0.008; r=-0.34, P=0.02), while ALT and AST levels weakly positively correlated with inflammation grade (r=0.39, P<0.001; r=0.38, P<0.001) in HBeAg-positive patients at baseline. Nonetheless, after 60 months of NAs treatment, these correlations disappeared, and intrahepatic HBV DNA and cccDNA got a positive correlation with inflammation grade (r=0.29, P=0.04; r=0.29, P=0.04) ( Figure 2 ).

In HBeAg-negative patients, all these serum and intrahepatic HBV markers, ALT and AST had no relevancy with inflammation grade neither at baseline nor after 60 months of NAs treatment ( Supplementary Figures 1, 2 ).

At baseline, HBeAg-positive patients with G1, G2, G3 and G4 were 29(39.2%), 25(33.8%), 17(23.0%) and 3(4.1%) respectively ( Figure 3 ). We defined G≥3 as significant inflammation, and the associated factors are depicted in Table 2 . The results showed that patients with significant inflammation had lower serum HBsAg and HBcrAg, higher serum ALT and AST, as well as older age. To evaluate the abilities of these variables to diagnose significant inflammation, ROC analyses were conducted. HBsAg performed best with the AUROC of 0.784 (95%CI 0.655-0.912; P=0.002). While the AUROCs of ALT, AST, HBcrAg and age were 0.733 (95%CI 0.601-0.865; P=0.002), 0.725 (95%CI 0.586-0.865; P=0.003), 0.728 (95%CI 0.574-0.883; P=0.02) and 0.677 (95%CI 0.541-0.812; P=0.02), respectively. Moreover, the combination of these markers would further improve the performance, with AST>101(U/L) plus HBsAg<4(log10IU/mL) having the best AUROC of 0.896 (95% CI 0.802-0.990; P<0.001), of which the sensitivity and specificity are 57.1% and 100%, respectively ( Table 3 ).

HBeAg-negative patients with G1, G2, G3 and G4 at baseline were 4(20.0%), 11(55.0%), 5(25.0%) and 0(0.0%), respectively ( Supplementary Figure 3 ). All these serum and intrahepatic HBV markers, ALT and AST had no difference between patients with or without significant inflammation ( Supplementary Table 1 ).

Figure 3 shows 59 out of 74 HBeAg-positive patients have undergone liver biopsies after 60 months of NAs therapy, with 94.9% (56/59) of G1 and 5.1% (3/59) of G2. Among the 59 patients, 32 (54.2%) achieved inflammation regression at month 60, including 17 (28.8%) had 1- grade regression, 12 (20.3%) had 2- grade regression, and 3 (5.1%) had 3-grade regression ( Table 4 ). No patients had inflammation progression.

17 out of 20 HBeAg-negative patients have undergone liver biopsies at Month 60, with 94.1% (16/17) of G1 and 5.9% (1/17) of G2 ( Supplementary Figure 3 ). Among the 17 patients, 14 (82.4%) achieved inflammation regression, including 10 (58.8%) had 1- grade regression and 4 (23.5%) had 2- grade regression ( Supplementary Table 2 ). No patients had inflammation progression.

The dynamic changes of serum and intrahepatic HBV markers, ALT and AST from baseline to Month 60 were investigated. All these markers, including serum HBV DNA, HBV RNA, HBsAg and HBcrAg, intrahepatic HBV DNA and cccDNA, as well as ALT and AST, declined in both patients with or without inflammation regression. Moreover, serum ALT, AST and HBV DNA declined most quickly in the first 6th months after initiating NAs therapy ( Figure 4 ).

Similar findings were observed in HBeAg-negative patients ( Supplementary Figure 4 ).