Reply to: ‘A risk prediction model for hepatocellular carcinoma after hepatitis B surface antigen seroclearance: Has the correct patient cohort been targeted?’

A risk prediction model for hepatocellular carcinoma after hepatitis B surface antigen seroclearance: Has the correct patient cohort been targeted?Journal of HepatologyVol. 79Issue 4PreviewThe current goal of chronic hepatitis B (CHB) treatment is to prevent complications of cirrhosis and hepatocellular carcinoma (HCC), with an aspirational target of functional cure to reduce these sequelae. However, in those that do clear surface antigen, the literature surrounding the need for ongoing HCC surveillance is barren and decision-making in the clinic remains uncertain and challenging. Full-Text PDF A risk prediction model for hepatocellular carcinoma after hepatitis B surface antigen seroclearanceJournal of HepatologyVol. 77Issue 3PreviewAfter hepatitis B surface antigen (HBsAg) seroclearance, the risk of hepatocellular carcinoma (HCC) remains, and the optimal surveillance strategy has yet to be determined. Herein, we aimed to evaluate incidence and risk factors for HCC and establish a novel prediction model for HCC development after HBsAg seroclearance. Full-Text PDF We appreciate Carlson and colleagues’ keen interest in our work in which we developed and validated a risk prediction model for HCC after HBsAg seroclearance.[1]Yang H. Bae S.H. Nam H. Lee H.L. Lee S.W. Yoo S.H. et al.A risk prediction model for hepatocellular carcinoma after hepatitis B surface antigen seroclearance.J Hepatol. 2022; 77: 632-641Abstract Full Text Full Text PDF PubMed Scopus (15) Google Scholar Carlson et al. have raised several considerations regarding this model.[2]Carlson S. Vaz K. Peterson A. A risk prediction model for hepatocellular carcinoma after hepatitis B surface antigen seroclearance: has the correct patient cohort been targeted?.J Hepatol. 2023; (S0168-8278(23)00074-0)Abstract Full Text Full Text PDF PubMed Scopus (1) Google Scholar First, the goal of our study was not to identify patients who can safely avoid HCC surveillance but rather to identify a group of people who are at a high risk of developing HCC. Even for patients classified as having low risk in this model, surveillance is not considered unnecessary. Age and drinking habits are subject to changes over time; therefore, although it may not be necessary to monitor the low-risk group temporarily, monitoring may need to be reintroduced at some point in the future.[1]Yang H. Bae S.H. Nam H. Lee H.L. Lee S.W. Yoo S.H. et al.A risk prediction model for hepatocellular carcinoma after hepatitis B surface antigen seroclearance.J Hepatol. 2022; 77: 632-641Abstract Full Text Full Text PDF PubMed Scopus (15) Google Scholar Further validation is required to establish whether it is genuinely possible to exclude surveillance for the low-risk group identified by our model. Eventually, accurate HCC prediction may become more complex over time due to the evolving variables. Second, we acknowledge Carlson et al.'s concerns regarding patients without cirrhosis. The estimated annual incidence rate of HCC in the non-cirrhotic group was 0.43% (3,654 person-years, 95% CI 0.25-0.71). Within the non-cirrhotic group, the estimated annual incidence rates of HCC were as follows, according to the three risk groups: 0% (95% CI 0-0.31%) for patients with a low-risk score, 0.53% (95% CI 0.27-0.93%) for patients with an intermediate-risk score, and 1.73% (95% CI 0.47-4.45%) for patients with a high-risk score (p <0.001). Although the small number of HCC cases made it difficult to propose a new model for patients without cirrhosis, the time-dependent AUROCs of the original model for predicting HCC in patients without cirrhosis were 0.787 (95% CI 0.676-0.898), 0.766 (95% CI 0.649-0.883), and 0.791 (95% CI 0.636-0.946) at 5, 10, and 15 years, respectively. In addition, Harrell’s C-index was 0.768 (95% CI 0.658-0.883). Despite the original model demonstrating acceptable performance in the non-cirrhotic group, its generalizability is limited by the low incidence of HCC in this group. Therefore, larger validation studies are needed to determine whether our model still performs well for the prediction of HCC in patients without cirrhosis. Finally, we conducted a retrospective review of metabolic-associated fatty liver disease (MAFLD) in our cohort. A total of 139 patients (16.7%) were confirmed to have MAFLD, 578 patients (69.6%) were confirmed to have no MAFLD, and the MAFLD status could not be confirmed in the remaining 114 patients (13.7%) owing to missing data. Height and weight were not recorded in some cases, which was the primary cause of the missing data. Patients with MAFLD did not have a significantly different risk of HCC development compared to those without MAFLD (hazard ratio 0.644, p = 0.364). In the subgroup analysis of patients without cirrhosis, there was no significant difference in HCC risk between patients with or without MAFLD (hazard ratio 1.104, p = 0.867), which is consistent with the results of a previous study.[3]Yu M.W. Lin C.L. Liu C.J. Wu W.J. Hu J.T. Huang Y.W. Metabolic-associated fatty liver disease, hepatitis B surface antigen seroclearance, and long-term risk of hepatocellular carcinoma in chronic hepatitis B.Cancers (Basel). 2022; 14Crossref Scopus (2) Google Scholar Since the term MAFLD has not yet been universally accepted,[4]De A. Ahmad N. Mehta M. Singh P. Duseja A. NAFLD vs. MAFLD - it is not the name but the disease that decides the outcome in fatty liver.J Hepatol. 2022; 76: 475-477Abstract Full Text Full Text PDF PubMed Scopus (16) Google Scholar it is difficult to determine the factors within this set of metabolic dysfunctions that influence HCC development, even if MAFLD has a significant impact. Additionally, while steatosis may inhibit hepatocarcinogenesis by inducing viral suppression of HBV,[5]Hu D. Wang H. Wang H. Wang Y. Wan X. Yan W. et al.Non-alcoholic hepatic steatosis attenuates hepatitis B virus replication in an HBV-immunocompetent mouse model.Hepatol Int. 2018; 12: 438-446Crossref PubMed Scopus (45) Google Scholar,[6]Liu Q. Mu M. Chen H. Zhang G. Yang Y. Chu J. et al.Hepatocyte steatosis inhibits hepatitis B virus secretion via induction of endoplasmic reticulum stress.Mol Cel Biochem. 2022; 477: 2481-2491Crossref PubMed Scopus (8) Google Scholar the progression of MAFLD may contribute to hepatocarcinogenesis by inducing chronic inflammation. Therefore, it is difficult to explain the effect of MAFLD on the development of HCC in patients with HBsAg seroclearance solely based on the contribution of metabolic dysfunction to HCC development.[7]Huang S.C. Liu C.J. Chronic hepatitis B with concurrent metabolic dysfunction-associated fatty liver disease: challenges and perspectives.Clin Mol Hepatol. 2023; 29: 320-331Crossref PubMed Scopus (7) Google Scholar Moreover, the oncogenic potential caused by HBV integration should not be dismissed, even after HBsAg seroclearance.[8]Jang J.W. Kim J.S. Kim H.S. Tak K.Y. Nam H. Sung P.S. et al.Persistence of intrahepatic hepatitis B virus DNA integration in patients developing hepatocellular carcinoma after hepatitis B surface antigen seroclearance.Clin Mol Hepatol. 2021; 27: 207-218Crossref PubMed Scopus (24) Google Scholar To put it differently, the pathophysiology of hepatocarcinogenesis in patients with MAFLD who have not been infected with HBV and those with MAFLD who have achieved HBsAg seroclearance after chronic HBV infection may be fundamentally distinct. The incidence of MAFLD-associated HCC is also significantly increasing.[9]Myers S. Neyroud-Caspar I. Spahr L. Gkouvatsos K. Fournier E. Giostra E. et al.NAFLD and MAFLD as emerging causes of HCC: a populational study.JHEP Rep. 2021; 3100231PubMed Google Scholar Carlson et al.'s inquiry is very intriguing as it emphasizes the need for further investigation on the impact of MAFLD upon HCC development after HBsAg seroclearance, which could not be demonstrated in our cohort. In conclusion, we are grateful to Carlson et al. for providing us with an opportunity to clarify the aforementioned important issues. We expect that our prediction model will provide reliable risk assessment for HCC and serve as a useful reference for decision-making regarding HCC surveillance and management of HBsAg-cleared patients. We eagerly await external validation of our prediction model. The authors received no financial support to produce this manuscript. All authors disclosure no conflict of interests. Please refer to the accompanying ICMJE disclosure forms for further details. Hyun Yang (Conceptualization: Equal; Data curation: Lead; Formal analysis: Lead; Investigation: Lead; Methodology: Lead; Projection administration: Equal; Resources: Lead; Software: Lead; Writing – original draft: Lead; Writing – review & editing: Lead); Ji Hoon Kim (Data curation: Supporting; Resources: Supporting); Ji Won Han (Data curation: Supporting; Resources: Supporting); Soon Kyu Lee (Data curation: Supporting; Resources: Supporting); Jeong Won Jang (Conceptualization: Lead; Data curation: Lead; Formal analysis: Lead; Funding acquisition: Lead; Investigation: Lead; Methodology: Lead; Projection administration: Lead; Resources: Lead; Software: Equal; Supervision: Lead; Writing – original draft: Lead; Writing – review & editing: Lead). The data used to conduct the research are available from the corresponding author upon request. This study was approved by the Institutional Review Board of The Catholic University of Korea (XC21RIDI0026). The following are the supplementary data to this article. Download .pdf (.31 MB) Help with pdf files Multimedia component 1