We sincerely appreciate the insightful letter by Dr. Zhao and Liu and thank them for their thoughtful comments on our study evaluating a steatosis-related polygenic risk score (PRS) for hepatocellular carcinoma (HCC) risk stratification after hepatitis C virus eradication with direct-acting antiviral therapy 1. Dr. Zhao and Liu raise several important issues regarding the clinical translation of PRS-based risk assessment, including the identification of clinically actionable risk thresholds, the implementation of surveillance strategies for non-cirrhotic populations, and the incremental value of PRS-5 beyond established clinical risk models 2. We welcome the opportunity to clarify these points. First, the threshold of PRS-5 ≥ 0.66 was not selected solely based on a percentile approach. Rather, restricted cubic spline analysis demonstrated that HCC risk remained relatively stable at lower PRS-5 values but increased sharply beyond 0.66 after adjustment for potential confounders, suggesting a non-linear relationship and a potential inflection point in the association between PRS-5 and HCC risk. The purpose of this analysis was therefore to identify a risk transition point rather than to establish a clinically definitive threshold for risk stratification. However, due to the limited number of HCC events, we were unable to evaluate optimal thresholds balancing sensitivity and specificity or derive stable absolute risk estimates across risk strata. Larger multicenter cohorts will therefore be necessary to determine clinically actionable thresholds 3. Second, the observation that PRS-5 remained predictive of HCC risk among non-cirrhotic individuals represents a novel finding. However, this subgroup analysis should be interpreted as preliminary evidence rather than an immediately actionable basis for clinical surveillance. Current guideline recommendations regarding HCC surveillance in non-cirrhotic patients following SVR remain uncertain, despite their clear clinical relevance 4, 5. Further well-designed studies focusing on non-cirrhotic populations, with sufficient event numbers and external validation, are therefore needed. However, given the relatively low incidence of HCC in non-cirrhotic populations, substantially larger sample sizes will be required to achieve adequate statistical power 6, 7. In addition, the technological and economic demands associated with genetic testing may present further challenges. Nevertheless, our study provides preliminary evidence that may help inform and streamline future investigations. Third, genetic predisposition is not the only determinant of cancer risk and should be considered alongside established clinical factors within comprehensive risk prediction models. The incorporation of genetic risk scores into existing clinical prediction models has been demonstrated in several areas. For example, PRS have been incorporated into non-genetic prediction models in the fields of breast cancer and cardiovascular disease 8, 9. In HCC risk stratification, Nahon et al. incorporated a genetic risk score into established clinical prediction models in patients with cirrhosis 10. In our study, age, metabolic disorders and liver fibrosis also contributed to HCC development after SVR, as demonstrated in the multivariable analyses. We agree with Dr. Zhao and Liu that integrating PRS-5 into a validated clinical prediction model represents a logical and pragmatic next step, as this would allow formal evaluation of its incremental predictive value and help determine whether PRS-5 meaningfully improves risk stratification beyond existing clinical tools. Yu-Sheng Lin: writing – original draft, conceptualization. Ying-Cheng Lin: writing – review and editing, supervision. Teng-Yu Lee: conceptualization, writing – original draft, writing – review and editing, supervision. The authors' declarations of personal and financial interests are unchanged from those in the original article 1. The authors have nothing to report. Teng-Yu Lee: research support from Gilead, Merck and Roche diagnostics; consultant of BMS, Gilead and AstraZeneca; and speaker of Abbvie, BMS, Eisai, Gilead, Roche and AstraZeneca. This article is linked to Lin et al. papers. To view these articles, visit https://doi.org/10.1111/apt.70566 and https://doi.org/10.1111/apt.70618. Data sharing not applicable to this article as no datasets were generated or analysed during the current study.
Lin et al. (Fri,) studied this question.