Dear Editor,
We read with interest the
Clinical and Molecular Hepatology article by Yu et al. [
1] reporting a random-effects network meta-analysis (NMA) comparing post-liver transplantation (LT) outcomes in hepatocellular carcinoma (HCC) patients within the Milan criteria (MC) versus those beyond MC but meeting expanded criteria (EC), with validation in a large institutional cohort. The authors suggest that several morphology-based EC (e.g., University of California, San Francisco criteria [UCSF], Up-to-Seven, Hangzhou) were associated with inferior overall survival versus MC, while biologically integrated models (Metroticket 2.0 and an alpha-fetoprotein [AFP]-based model) appeared more favorable in ranking and cohort validation [
1]. We commend the authors for tackling a clinically important question and for attempting external validation. We nevertheless wish to highlight several points that may help readers translate these findings into listing and allocation decisions.
First, EC are applied at listing and during waitlist management, whereas the primary comparisons in the NMA are conditional on receiving LT (post-transplant overall survival [OS]/recurrence-free survival [RFS]). In deceased-donor LT (DDLT) settings, the clinically relevant question is often transplant benefit, incorporating waitlist dropout/progression and competing mortality, rather than post-LT outcomes among transplanted recipients alone [
2,
3]. We therefore encourage more explicit framing that the NMA estimates reflect “post-transplant treated” prognosis and may not directly map onto intention-to-treat outcomes from listing, particularly when waiting time and bridging strategies differ across systems.
Second, the validation cohort underscores the importance of dynamic tumor biology and treatment response. With predominance of hepatitis B virus (84.5%), living-donor LT (LDLT, 95.5%), and frequent pre-LT anti-HCC treatment (84.8%) [
1], eligibility is unlikely to be a static baseline state. In contemporary practice, bridging/downstaging and AFP kinetics often determine which MC-out candidates proceed to LT, and biomarker response may meaningfully stratify recurrence risk after LDLT [
4,
5]. We suggest that readers interpret comparisons across EC in light of this “response-based selection,” and that future work, where data permit, incorporate treatment response and AFP trajectories as key effect modifiers when judging the clinical utility of expanded criteria.
Third, while
p-scores provide a useful summary, ranking can be overinterpreted in star-shaped networks centered on MC, especially when some EC nodes are supported by limited direct evidence. To enhance clinical interpretability, it would be helpful to accompany rankings with concise absolute metrics (e.g., 3- and 5-year OS/RFS) and the amount of direct evidence informing each EC. In addition, because post-LT recurrence hazards may vary over time, reliance on a single hazard ratio can be difficult to translate at the bedside; milestone estimates and/or restricted mean survival time may provide more clinically intuitive summaries when non-proportional hazards are plausible [
6].
Finally, we agree that biologically integrated approaches are promising; however, “more favorable among EC” should not be conflated with clinical acceptability versus MC. In the validation, HCC-specific mortality remained higher than MC-in for most EC groups (except the radiology-based AFP model) [
1], emphasizing that relative ranking among EC does not automatically imply non-inferiority to MC. Explicitly separating comparative performance within EC from clinical acceptability versus MC would further strengthen the translational message.
In summary, Yu et al. [
1] provide valuable evidence supporting biological integration beyond morphology. We hope the above considerations, particularly the distinction between post-LT prognosis and listing-time benefit, the centrality of treatment response, and clinically interpretable reporting, will help guide safe and equitable expansion beyond MC.
FOOTNOTES
-
Authors’ contribution
Bo Ni: conceptualization; methodology; investigation; data curation; visualization; writing - original draft.
Hairui Wang: conceptualization; methodology; supervision; project administration; resources; validation; writing - review & editing; correspondence with the journal.
-
Conflicts of Interest
We declare there is no any conflict of interest.
Abbreviations
REFERENCES
- 1. Yu D, Hwang Y, Ju JS, Heo S, Kim SO, Choi SH, et al. Network meta-analysis and validation study of expanded liver transplantation criteria for hepatocellular carcinoma: significant role of alpha-fetoprotein. Clin Mol Hepatol 2026;32:751-771.
- 2. Kodali S, Kulik L, D'Allessio A, De Martin E, Hakeem AR, Lewinska M, et al. The 2024 ILTS-ILCA consensus recommendations for liver transplantation for HCC and intrahepatic cholangiocarcinoma. Liver Transpl 2025;31:815-831.
- 3. Mehta S, Trotter JF. Policy corner: HCC exception update. Liver Transpl 2023;29:1330-1331.
- 4. Norman JS, Li PJ, Kotwani P, Shui AM, Yao F, Mehta N. AFPL3 and DCP strongly predict early hepatocellular carcinoma recurrence after liver transplantation. J Hepatol 2023;79:1469-1477.
- 5. Chen IH, Hsu CC, Yong CC, Cheng YF, Wang CC, Lin CC, et al. AFP response to locoregional therapy can stratify the risk of tumor recurrence in HCC patients after living donor liver transplantation. Cancers (Basel) 2023;15:1551.
- 6. Corro Ramos I, Qendri V, Al M. Beyond hazard ratios: appropriate statistical methods for quantifying the clinical effectiveness of immune-oncology therapies - the example of the Netherlands. BMC Med Res Methodol 2024;24:260.
Citations
Citations to this article as recorded by
