Recent international guidelines and Italian consensus documents have moved beyond the dichotomous Milan-in/Milan-out logic and now frame LT candidacy in hepatocellular carcinoma (HCC) as a multiparametric judgment integrating fitness, liver function, tumor biology, feasibility, and patient values^[7-10]. Within the European Association for the Study of the Liver (EASL) 2025^[8] and the Italian consensus^[9], candidacy is the outcome of this integrated assessment rather than the application of any single morphological cut-off. The Multiparametric Therapeutic Hierarchy (MTH)^[11] and the recent reappraisal of the Barcelona Clinic Liver Cancer (BCLC)^[12] framework extend the same logic across the broader HCC therapeutic hierarchy. In this multiparametric framework, the individual benefit of transplantation has long been formalized in the transplant-benefit literature as the difference between expected post-transplant survival and survival on the waiting list. Our group and others have shown that this construct better balances urgency and utility than morphology-based criteria alone^[13]. Building on this foundation, we previously proposed an additional step in which the individual transplant benefit is weighted against the cumulative harm that the same allocation creates for other candidates on the waiting list, a net-benefit reading of candidacy that makes explicit the equity dimension intrinsic to any deceased-donor allocation^[14].

In this commentary, we adopt the operational reading of the Grading of Recommendations Assessment, Development and Evaluation (GRADE) Evidence-to-Decision (EtD) framework recently proposed by our group for individual-level HCC decision-making^[15,16]. At the point of care, the four EtD-derived domains - technical feasibility, resources, equity, and values and acceptability - converge into a unified individual-level construct of feasibility, which describes whether a given treatment can be realized for a specific patient^[16]. Throughout Figure 1 and the analysis below, we use Feasibility in this integrative sense, alongside the three biology-related axes of fitness, liver function, and tumor biology, and alongside patient values, which we keep visible as a separate gate when it carries decisional weight in the case.

Zhang et al. describe a 71-year-old man with HBV-related cirrhosis, a 15 cm × 11 cm × 10 cm right-lobe HCC, Alpha-fetoprotein (AFP) > 100,000 ng/mL with no response to TACE plus TKI, an Indocyanine green retention rate at 15 min (ICG-R15) of 16.5% and a future liver remnant of 41.79%^[6]. Conventional and expanded morphological criteria (Milan, UCSF, Hangzhou)^[17-19] and the AFP-model^[20] were exceeded; the China Organ Transplant Response System (COTRS) allocation system declared him ineligible for allogeneic LT; living-donor liver transplant (LDLT) was unavailable in his social network; safe right hepatectomy and Associating Liver Partition and Portal Vein Ligation for Staged Hepatectomy (ALPPS) were precluded by liver function and rupture risk; systemic and locoregional therapies had failed. Under compassionate use, the right-lobe tumor was resected and a 10-gene-edited porcine liver was implanted in the right hepatic fossa as an auxiliary graft, intended to bridge towards regeneration of the native left lobe.

Read through the multiparametric framework, the candidacy pattern is mixed but, with respect to the standard transplant settings, decisive. Fitness, liver function, the integrated Feasibility domain, and patient values are favorable. tumor biology, however, is unequivocally aggressive and unresponsive to conversion attempts: an HCC of this size with very high AFP and no radiological response after combined locoregional and systemic therapy is, in transplant-oncology terms, the prototypical pattern of expected early recurrence and shortened post-transplant survival. In a deceased-graft setting, this combination yields a negative net-benefit: the small individual benefit cannot offset the substantial harm that allocation would impose on other waiting candidates with better post-transplant prospects. In a living-graft setting, the harm to the waiting list disappears, but the harm vector does not vanish, it is replaced by a residual donor-side risk (pooled mortality ≈ 0.06% across more than 60,000 donors)^[21] and a Feasibility constraint that, in this specific patient, was decisive: no suitable family donor was identified. Under the same framework that already governs candidacy decisions, both pathways correctly close [Figure 1A and B].

The xenograft does not extract organs from the deceased-donor pool, and it does not expose a healthy human to surgical harm. Therefore, within this framework, human-graft allocation harm and living-donor surgical harm are zero. Within the unchanged net-benefit framework, this represents a quantitative shift in the balance between benefit and harm: the expected individual benefit must still be weighed against the remaining risks and burdens of xenotransplantation, including operative risk, rejection and graft-related complications such as thrombotic microangiopathy (xTMA), tumor progression under immunosuppression, and resource implications [Figure 1C]. This conceptual structure allows an individualized assessment of whether a procedure with uncertain but plausible individual benefit may become ethically admissible after considering the full spectrum of residual risks and feasibility constraints.

That configuration is not generic. It requires that fitness, liver function, and the integrated feasibility domain (technical realisability, resources, equity, values and acceptability) all be favorable, and that the only realistic alternative for the specific patient be best supportive care, neither effective allotransplantation, nor curative resection, nor effective locoregional or systemic therapy. The patient described by Zhang et al. meets these conditions: imminent risk of rupture, exhausted alternatives, preserved fitness, established experimental gene-editing platform, and informed consent under compassionate use^[6]. Under such conditions, an experimental procedure with unknown individual benefit but zero deceased-donor allocation/living-donor harm is admissible, not because the framework has changed, but because, when applied honestly, the framework returns a different verdict when the deceased-donor allocation and living-donor surgical components of the harm term are zero.

The above is the methodological reading of the case. Three biological and structural frontiers determine whether xenotransplantation can become more than a single compassionate-use precedent.

Zhang et al. have produced a landmark biological proof of concept^[6]. The conceptual lesson the case carries, however, is narrower and more useful than the broad claims sometimes attached to it. The multiparametric net-benefit framework that governs LT candidacy in HCC does not change because of xenotransplantation. In that specific configuration, and only in that configuration, a procedure with uncertain individual benefit can be ethically admissible. The Zhang case is the first empirical illustration of this configuration, not a license to relax candidacy criteria more broadly. Preparing transplant oncology to use this new term in the equation responsibly, and to do so without confusing it with a structural change in the framework, is the methodological work that begins now.