Hepatitis C virus infection is a major cause of hepatocellular carcinoma (HCC), with a significantly increased risk in patients with advanced liver fibrosis or cirrhosis. Although the introduction of direct antiviral drugs has dramatically reduced the incidence of HCC, some high-risk patients (e.g., those with advanced liver fibrosis or other oncogenic factors) still have a residual risk of developing HCC after achieving a sustained virologic response (SVR) [1]. In view of this, how to screen high-risk patients and optimize HCC follow-up strategies based on noninvasive indices has become an urgent clinical question. Lee et al. [2] evaluated the performance of transient elastography (TE) and fibrosis-4 index (FIB-4) in predicting HCC risk before and after SVR through a systematic review and meta-analysis that included a total of 169,911 hepatitis C virus-infected patients who achieved SVR. It was found that both TE and FIB-4 had more than moderate predictive ability at different stages (both area under the curve >0.7), and the optimal cut-off values were further determined (e.g., 12.6 kPa for TE before SVR, 11.2 kPa for TE after SVR, and 3.25 for FIB-4). In addition, the study emphasized the importance of dynamic evaluation of noninvasive markers and found that the predictive power of TE was superior to that of FIB-4. This study not only provides a theoretical basis for the development of personalized HCC surveillance strategies, but also highlights the potential of low-cost noninvasive tools in screening high-risk populations.

However, this study has the following unemphasized limitations that need to be further highlighted and corrected. First, although the authors emphasized in the article that there was no language restriction for the inclusion of the study, only three English databases including EMBASE, CENTRAL, and MEDLINE were searched. Other common English databases such as Google Scholar, Web of Science, Scopus, and KoreaMed, as well as common Chinese databases such as China Knowledge and Wanfang, should have been searched extensively to minimize the omission of target literature. Second, most of the included studies were from East Asia, so would this result in a significant geographical bias? In addition, large heterogeneity was found in the forest plot of risk ratios, but this was not analyzed and discussed in detail by the authors. Therefore, we suggest that subgroup analyses and meta-regression analyses can be performed for different demographic variables. Third, the quality assessment figures presented in the article show only low and unknown risks, which makes us wonder. Could it be that none of the projects in so many retrospective studies were high risk? Furthermore, what is the established boundary between unknown risk and high risk? Finally, the text lacks an assessment of the level of evidence for the pooled results, which is an integral and important part of a high-quality meta-analysis. In addition, an assessment of publication bias is not found in the text.