Hepatitis B core‐related antigen levels predict recurrence‐free survival in patients with HBV‐associated early‐stage hepatocellular carcinoma: results from a Dutch long‐term follow‐up study

Abstract Prognosis of hepatitis B (HBV)‐associated hepatocellular carcinoma (HCC) is poor due to high rates of HCC recurrence and progression of underlying liver disease. We studied whether serum hepatitis B core‐related antigen (HBcrAg) levels could predict HCC recurrence and outcome in HBV associated. Higher HBcrAg levels at HCC diagnosis were independently associated with reduced overall and recurrence‐free survival in patients with early, but not advanced, stage HCC.


| INTRODUC TI ON
Chronic hepatitis B virus infection (HBV) affects more than 250 million people globally and is a leading cause of liver cirrhosis, liver failure and hepatocellular carcinoma (HCC). HCC is the fifth most common cause of cancer, and its long-term outcome is poor even in patients undergoing potentially curative treatment, which is due to HCC recurrence and/or progression of the underlying liver disease. 1 Currently, antiviral therapy with nucleo(s)tide analogues (NA) improves prognosis but does not cure HBV nor prevent HCC recurrence. The poor prognosis in HBV-associated HCC may be partially attributed to persistence of intrahepatic replication and ongoing oncogenic processes that are not influenced by NA therapy mediated HBV DNA suppression. Hepatitis B core-related antigen (HBcrAg) is an emerging serological biomarker that correlates with replicational activity of the intrahepatic cccDNA pool 2 and has shown to predict HCC development in both NA treated and untreated patients. [3][4][5][6] Whether serum HBcrAg levels predict the recurrence-free survival in patients with HBV-associated HCC is unknown.

| Data acquisition
Data were collected retrospectively from the electronic medical records. The diagnosis of HCC was either based on histopathology or imaging criteria in accordance with the international guidelines.
We calculated follow-up time from the date of HCC diagnosis until the date of death (ascertained through the national mortality database), or (for analyses on recurrence-free survival) the date of disease recurrence based on noninvasive imaging-based criteria or histology or the last date of the evaluation which was 20 November 2018.
Serum levels of HBcrAg (log U/mL) were measured using the Lumipulse G HBcrAg assay (Fujirebio Europe, Ghent, Belgium) on a LUMIPULSE G1200 analyser (Fujirebio Inc, Tokyo, Japan). The lower limit of detection was 2 log U/mL, with a validated lower limit of quantification of 3 log U/mL. 7 All statistical tests were two-sided and evaluated at the .05 level of significance.

| Patient characteristics
A total of 119 patients with HBV-associated HCC were included in the study, of whom 77% were males, mean age was 55 years, 68% were HBeAg-negative, and 48% were on NA therapy at time of HCC diagnosis. Patients were Caucasian (51%), Asians (30%), or Africans
In multivariate analysis, only BCLC stage (P < .0001) and AFP levels (P = .003) were independently associated with survival.

| D ISCUSS I ON
Prognosis of HBV-associated HCC remains poor due to limited options for curative treatment, high rates of post-therapy recurrence and progression of underlying liver disease. In our cohort, of whom the majority were Caucasian, 5-year survival was less than 37%, with BCLC stage being the main predictor for long-term survival in the overall study population.
Recent studies have consistently shown a persistently elevated risk of HCC despite successful viral suppression in patients with chronic hepatitis B. This may be explained by ongoing intrahepatic viral replication that is not suppressed with NA therapy. This persistent replicational activity may be reflected by higher serum HBcrAg levels. 2 Recent studies showing higher rates of HCC occurrence and reoccurrence in patients with higher HBcrAg levels during antiviral therapy provides further support for this hypothesis. 6,8 In the current study, HBcrAg levels at HCC diagnosis were not associated with HCC outcome in the overall population. Because the prognosis of patients with advanced-stage HCC (BCLC stage B-D) was dismal, with a median survival of less than 5 months, we also performed separate analyses in the subset of patients with early-stage HCC. Among patients with early-stage HCC, higher HBcrAg levels were independently associated with higher rates of HCC recurrence and higher risk of mortality. These findings in the subset of patients with early-stage HCC are important, since they provide further support for the concept that complete viral eradication is necessary to improve outcomes, and it confirms the association between HBcrAg and HCC risk. Also, HBcrAg levels might potentially help identify patients with high rates of disease recurrence who may benefit more from liver transplant than local therapy because of their high recurrence risk after local therapy. Our findings also suggest that among those with advanced-stage HCC, virological factors are probably less important and prognosis is dependent on the (often already advanced) stage of liver disease.
Our study is quite unique since it enrolled mainly non-Asian patients with HBV-associated HCC with a very long follow-up duration. However, there are some limitations. The most important is the relatively limited number of patients, especially after stratification by BCLC stage. However, our findings are robust, which is most clearly reflected in the comparable outcomes regarding both HCC recurrence and overall survival. Nevertheless, this was a single-centre study and further external validation of our findings is therefore required. Given that we mainly enrolled Caucasian patients, confirmation is also required in patients with other ethnicities.
In conclusion, serum HBcrAg levels are independently associated with overall and recurrence-free survival in patients with early, but not advanced, stages of, HBV-associated HCC.

ACK N OWLED G EM ENTS
The authors would like to thank Laura Vernoux for the helpful and interesting discussions. This study was supported by the