Role of early on-treatment serum HBV RNA declines in predicting hepatocellular carcinoma risk in patients with chronic hepatitis B.

ABSTRACT

BACKGROUND AND AIMS: Hepatocellular carcinoma (HCC) risk prediction models established in patients with chronic hepatitis B (CHB) receiving nucleoside analogue (NA) rarely included viral factors because of mediocre predictability of traditional viral markers. Here, we investigate the role of serum hepatitis B virus (HBV) RNA, a novel biomarker, in predicting HCC risk in NA-treated patients. METHODS: A total of 1374 NA-treated patients were enrolled from two prospective CHB cohorts. Serum HBV RNA was detected at baseline, year 1, 2 and 3 of treatment. Cox proportional-hazard model was used to investigate the association of HBV RNA kinetics with HCC risk. RESULTS: After a median follow-up of 5.4 years, 76 patients developed HCC. HBV RNA declines at year 1 (adjusted hazard ratio (aHR) = 0.70, P = .009) and 2 (aHR = 0.71, P = .016) were independently associated with HCC risk. Patients with less HBV RNA decline at year 1 (=< 0.4 log10 copies/mL) or 2 (=<0.6 log10 copies/mL) had 2.22- and 2.09-folds higher HCC risk, respectively, than those with more declines. When incorporating these early on-treatment HBV RNA declines into existing HCC risk scores, including PAGE B, mPAGE B and aMAP score, they could enhance their predictive performance [i.e. C-index, 0.814 vs. 0.788 (Model (PAGE B + year-1 HBV RNA decline)vs. PAGE B score based on baseline parameters)]. CONCLUSIONS: Serum HBV RNA declines at year 1 and 2 were significantly associated with on-treatment HCC risk. Incorporating early on-treatment HBV RNA declines into HCC risk prediction models can be useful tools to guide appropriate surveillance strategies in NA-treated patients.