Whole liver transplantation has been considered the gold standard operative technique globally. Deceased-donor transplantation has been the principal organ source in North America and Western Europe. Living-donor transplantation has become popularised in Asian countries^[3]. In an attempt to overcome organ shortage, the criteria for organ suitability have been relaxed, to permit inclusion of organs previously considered unsuitable (e.g., presence of steatosis, older age, viraemic donors, donation after circulatory death)^[3,8,9]. Complications of whole organ transplantation include allograft dysfunction, biliary complications, and the morbidity due to lifelong immunosuppression^[10-12].

Split-liver transplantation (SLT) involves dividing a single deceased-donor liver into two partial grafts, allowing for greater organ availability^[13]. This approach has proven particularly suited to pediatric recipients and selected adult patients^[14,15]. SLT is considered technically more challenging than whole liver transplantation^[8,14]. Higher rates of biliary and vascular complications have been observed with SLT^[14,16].

Ex vivo liver perfusion (EVLP) is an emerging preservation approach in liver transplantation, in which the graft is perfused via an extracorporeal circuit through the hepatic vasculature. This technique has been shown to reduce ischemia-reperfusion injury. It also allows viability testing of the organ prior to transplantation. The two approaches used for EVLP are normothermic machine perfusion (NMP) and hypothermic oxygenated perfusion (HOPE)^[17,18]. Both have been shown to improve metabolic function of the graft and extend organ preservation time before surgical transplantation^[19-21].

Xenotransplantation involves the use of organs from other species, predominantly genetically engineered pigs. Advances in gene editing targeting glycan xenoantigens (alpha-Gal and other non-Gal xenoantigens) have reduced the risk of hyperacute rejection, while immunosuppressive strategies such as costimulation blockade have further improved short-term xenograft compatibility and survival^[22,23]. However, the immunogenicity of these organs remains the limitation to this approach currently^[24,25]. Immunologically mediated coagulation-related complications, such as thrombotic microangiopathy, can result in graft failure^[26,27]. Other risks of xenotransplantation include the theoretical risk of zoonotic transmission, and underproduction of bile and porcine albumin for sustained functional support^[25,26]. Although potentially promising as a future adjunct, xenotransplantation remains experimental at present.

These inherent limitations of organ transplantation have promoted research into liver scaffolds as suitable alternatives.

A number of different source tissues have been used in liver decellularization over the last decade. These have included rodent tissue, porcine liver and discarded human liver tissue from ineligible donors^[29,30]. Porcine liver has been used largely because of its size and anatomical compatibility with human liver. As alluded to previously, a pitfall of xenogeneic scaffolds is their immunogenicity due to xeno-epitopes not found in humans. Certain epitopes have proved easier to remove than others (such as the alpha-gal epitope)^[7].

In contrast, discarded human liver-derived ECM may offer lower immunogenicity^[31]. It is derived from a more heterogenous population, where age and disease-related ECM remodeling (e.g., fibrosis), can affect its resistance to enzymatic processing of the ECM^[32].

Decellularization commonly involves a combination of physical, chemical, and enzymatic approaches. The aim is to disrupt cell membranes, induce cell death and facilitate removal of cellular material^[33]. It is a delicate balance between the use of aggressive agents that achieve a more complete decellularization process, and the resultant damage to key ECM components and vascular architecture^[7].

In liver bioengineering, recellularization may be achieved by one of two methods, either direct injection into the parenchyma or by infusion into the vascular system^[7]. With its unique portal circulation, perfusion-based delivery is well-suited to achieve widespread cell distribution. Ex vivo machine perfusion is the process by which the liver scaffold is perfused through an extracorporeal circulation. Similarly, perfusion bioreactors support oxygen and nutrient delivery through preserved vascular conduits under controlled perfusion pressure, while providing an ex vivo environment for graft maturation and pre-implantation functional assessment^[53].

Table 2 summarises the most recent studies that employ perfusion recellularization in rodent and porcine liver models. In some instances, the experimental liver constructs underwent ex vivo vascular patency testing and transplantation. This approach allows comparisons to be made as to how perfusion recellularization protocols relate to in vivo liver function.

Mitigation of the risk of organ thrombosis remains a key requirement for the translational feasibility of decellularized liver scaffolds.

Re-endothelialization represents an important strategy for improving vascular patency and reducing scaffold thrombogenicity. It allows restoration of the vascular lining and limits direct coagulation contact with the exposed ECM^[7]. This approach has been explored in large-animal models with encouraging translational progress. Shaheen et al. seeded decellularized liver scaffolds with HUVECs and achieved sustained perfusion in vivo after heterotopic transplantation for up to 15 days^[5]. Similarly, Higashi et al. repopulated the vascular compartment of a porcine liver scaffold with endothelial cells, and reported extended survival of the transplanted liver graft up to 28 days^[50]. Liver sinusoidal endothelial cells (LSECs) may also be ideal for reconstructing the sinusoidal vascular network, however large-scale applications have not yet been established^[50]. These findings suggest endothelialization as a promising method to improve revascularization, however further optimisation of endothelial cell source and long-term vascular stability is still required.

Heparin has been used as an anti-thrombotic agent. Yadav et al. found it to reduce thrombogenicity and platelet aggregation during ex vivo blood perfusion, while also improving endothelial adhesion^[65]. Fibronectin coating has been explored as another potential anti-thrombotic agent. Afrin et al. found that fibronectin coating on the liver scaffolds resulted in improved endothelial cell adhesion within the vascular network, reduced thrombogenicity, and increased HepG2 proliferation^[54]. This model has not been trialled in a xenotransplant model, so its in vivo performance remains uncertain. In a different approach, Horie et al. applied the biomimetic polymer 2-methacryloyloxyethyl phosphorylcholine (MPC) to blood-contacting ECM surfaces and reported reduced platelet deposition and suppressed thrombus formation during ex vivo blood perfusion, with similar antithrombotic effects maintained in a heterotopic transplantation model^[51].

Collectively, these studies suggest potential approaches to targeting scaffold thrombogenicity. However, longer term hemocompatibility testing in large-animal transplantation models remains necessary before clinical translation can be considered.

Successful liver transplantation fundamentally depends on a well-functioning graft. However, current bioengineered liver constructs remain unable to consistently produce fully functional grafts. It is primarily hindered by the number of cells required and the profound difficulty in maintaining their functional maturity. To achieve a clinically relevant liver mass for an adult human liver, a cell amount of approximately 150 to 350 billion cells is required^[30]. For a 70-kg patient, it requires at least 84 billion functional hepatocytes to achieve the 20%-30% hepatic mass necessary to sustain baseline physiological functions^[7]. Although primary hepatocytes are the gold standard in most studies, their limited donor availability constrains clinical application. Moreover, primary hepatocytes still face donor-recipient immunological mismatch. In addition to these limitations, primary hepatocytes undergo rapid dedifferentiation ex vivo, with downregulation of metabolic enzyme synthesis reported within 24 h of culture^[62].

To overcome the shortage of primary hepatocytes, patient specific iPSCs have been trialled. However, iPSC-derived hepatocytes frequently arrest development at a fetal phenotype and demonstrate incomplete functional maturation in critical parameters like cytochrome P450 activity^[7,57]. Furthermore, the lack of an intact and functional biliary tree in most models leads to the accumulation of cytotoxic bile, hindering long-term survival^[7,57].

As a result, current bioengineered liver grafts suffer from short in vivo survival times. Early preclinical models failed to maintain continuous perfusion for more than 3 days^[5]. Although recent advancements utilizing dynamic bioreactors and immunosuppressed large animal models have extended the graft survival to 15 days and even 28 days, the overall hepatic function remains partial, and delayed graft failure due to cellular senescence and vascular stenosis remains unresolved^[6,50].

Although many small animal studies of bioengineered liver transplantation have demonstrated short-term graft survival and partial liver function, the clinical translation to human models requires validation in large animal models, particularly porcine models^[6,67]. Several surgical technical problems remain in the transplantation of bioengineered livers. Firstly, the vessels of the decellularized liver scaffold are usually more fragile than those of native liver tissue. They are easily torn during vascular clamping, traction, or suturing, which makes vascular anastomosis more difficult.

Further, the liver scaffold must be able to immediately withstand physiological portal and arterial pressure after implantation. Many current large animal models have relied on auxiliary rather than total orthotopic transplantation, largely because the current bioengineered livers do not provide sufficient liver function to support complete liver replacement.

Once the graft is re-perfused with blood, exposed vascular ECM may rapidly activate the coagulation cascade and platelet aggregation, leading to acute thrombosis within minutes if the graft is not adequately re-endothelialized^[6,49]. Even with several re-endothelialization strategies, achieving complete and uniform coverage of the complex sinusoidal microvascular network remains difficult, causing the graft to face a hostile environment of increased vascular resistance, incomplete microcirculation, and eventual ischemic failure^[7,49].

Even if surgical anastomosis is successful and acute thrombosis is avoided, the bioengineered graft faces severe immune responses derived from a “dual-source immunogenicity”: the decellularized scaffold and the repopulating cells.

From the scaffold perspective, xenogeneic matrices (e.g., porcine livers) may retain highly immunogenic carbohydrate epitopes, notably galactose-alpha-1,3-galactose (α-Gal), N-glycolylneuraminic acid (Neu5Gc), and Sda antigens^[36]. Furthermore, structural damage to the ECM during decellularization releases damage-associated molecular patterns (DAMPs), such as low molecular weight hyaluronic acid and fibronectin fragments, which strongly activate pattern recognition receptors (PRRs) on host macrophages, driving them toward a pro-inflammatory (M1) phenotype^[36].

From the cellular perspective, if the cells utilized for recellularization were not recipient-specific cells, the adaptive immune rejection may be provoked. Preclinical studies have shown that repopulating porcine scaffolds with human cells, such as HUVECs, triggers intense complement-mediated cytotoxicity and robust antibody responses. Without immunosuppression, the humanized grafts are rapidly destroyed within three to seven days post-transplantation^[5]. Theoretically, recellularization with recipient-specific cells may reduce or even eliminate the need for immunosuppressive therapy^[6]. In reality, however, even an autologous-derived cell source should not be considered entirely immune-privileged. MSCs are generally considered to have relatively low immunogenicity, yet they are not completely immunologically inert^[68,69]. Similarly, autologous iPSC-derived cells may still elicit immune responses^[70,71].

Regulations are established to protect public health and patient safety by strictly controlling the quality, safety, and efficacy of medical products. However, when emerging products or clinical scenarios fall outside the existing regulations, the researchers and the industry developers may lack clear guidance on study design, manufacturing processes, and defining appropriate quality-control and translational strategies^[72]. The whole-organ bioengineered liver produced via decellularization and recellularization techniques is a complex product containing decellularized scaffold, implanted cells, and vascular structures sourced from either humans or animals. It also requires ex vivo processing to produce a functional graft. These features make whole-organ bioengineered liver grafts fundamentally more complex than conventional cell therapies or scaffold-based constructs^[73].

In the United States, there is no specific regulatory framework established for whole organ engineering. For the bioengineered liver composed of human-source scaffolds and cells, the regulation may fall within the Food and Drug Administration (FDA) framework of human cells, tissues, and cellular and tissue-based products (HCT/P)^[7,72,74]. Such products would exceed section 361 of the Public Health Service Act (PHS Act)^[75]. In contrast, for animal based decellularized matrices seeded with human cells, the regulatory pathways remain unclear. Decellularized and recellularized bioengineered organs would more likely fall under section 351 of the PHS Act and/or the Federal Food, Drug, and Cosmetic Act (FD&C Act) as drug, device, and/or biological products^[75]. They would be subject to stricter regulatory pathways, including current good tissue practice (cGTP), current good manufacturing practice (cGMP), and would generally require an Investigational New Drug (IND) application before clinical investigation in humans^[72,75].

In the European Union, the bioengineered liver would most likely fit within the advanced therapy medicinal product (ATMP) framework, particularly as a tissue-engineered product (TEP), according to Regulation (EC) No 1394/2007^{[72,73,76,77]}. Under Regulation (EC) No 1394/2007, a Committee for Advanced Therapies (CAT) within the European Medicines Agency (EMA) was established to prepare a draft opinion on the quality, safety and efficacy of each ATMP. Although the ATMP framework serves as the closest regulatory pathway for bioengineered products, the organ-level standards remain undeveloped. Whole-organ bioengineering products would still require case-by-case evaluation^[77].

In Japan, two acts for human regenerative medicine were enacted in 2014: the Act on the Safety of Regenerative Medicine (ASRM) and the Pharmaceuticals, Medical Devices, and Other Therapeutic Products Act (PMD Act)^[72,78,79]. The ASRM mainly regulated the clinical use and medical research for regenerative medicine in clinical institutions, while the PMD Act is more related to the development and commercialization of cell-containing bioengineered products^[80]. This dual-track system in Japan appears relatively favourable in regenerative medicine research and product development compared with other regions. Nevertheless, both acts were originally designed for cell-based therapy, rather than for whole-organ-scale bioengineered grafts.

In general, current regulatory systems were primarily developed for simpler cell therapies, tissue-engineered constructs, or traditional drugs, rather than complex organ-scale bioengineered grafts^[7]. The lack of standardized biosafety criteria for decellularized matrices and recellularized qualities, together with rigorous good manufacturing practice (GMP) demands, continues to hinder commercialization and delay clinical translation^[7,73,81]. As a result, more practical, organ-specific regulatory frameworks are needed to better accommodate the unique biological complexities of whole-organ engineering and to accelerate its clinical application^[73].

Beyond scientific and regulatory hurdles, the high financial costs associated with whole-organ bioengineering may threaten its commercial viability. The high manufacturing costs will likely restrict research and development to a few well-funded centres, slowing dissemination of the technology^[63]. Further, the therapy would need to prove its worth and cost-effectiveness compared to traditional liver transplantation.

Decellularized liver scaffolds may help address donor organ shortage by enabling modular bioengineering strategies that do not require a whole-organ graft. In the context of split/segment liver perfusion, the liver is anatomically divided into lobes and segments with independent vascular inflow and outflow, and these units may retain nearly equal metabolic functionality. Supporting this concept, Kanani et al. showed that isolated healthy liver segments from resected cancerous livers could remain viable under ex vivo machine perfusion^[82]. If scaffold bioreactor perfusion can be considered an analogue of ex vivo machine perfusion, these findings point the way to a future potential. Resected liver segments from diseased or porcine liver could be decellularized and re-endothelialized, generating auxiliary grafts for partial liver support. Such an approach would increase graft availability, particularly for patients unsuitable for whole-organ transplantation.

Collectively, these lines of work suggest that coupling segmental liver scaffolds with robust vascular reconstruction may benefit patients with liver cancer or end-stage liver disease, and may also provide partial functional support for selected patients with liver cirrhosis thereby reducing metabolic demand.

The future scope of bioengineered liver scaffolds may also extend beyond transplantation towards bioartificial liver systems^[83] and disease-modelling platforms. Liu et al. developed a perfused ex vivo human liver model using liver-specific biomatrix scaffolds and reported that the recellularized liver constructs reproduced essential hepatic functions, including metabolic and enzymatic activity^[52]. Furthermore, the model was capable of simulating diclofenac-induced liver injury and generating precise hepatic responses. These findings highlight the capacity of such models in supporting predictive drug discovery, thereby broadening the scope of liver scaffold engineering beyond transplantation to preclinical drug testing.

In parallel, Chen et al. developed a tumor organoid-like tissue model using liver-derived decellularized ECM recellularized with HepG2 cells, reporting its value as a predictive platform for evaluating antitumor agents and investigating epithelial-mesenchymal transition (EMT) dynamics in tumor cells^[84]. In the context of hepatocellular carcinoma, such models may provide a more physiologically relevant system for studying tumor-matrix interactions, antitumor drug toxicity, and therapeutic response in tumor progression.