Asia-Pacific Transplant Oncology Symposium 2024: Expert Statements on the Current State and Future Direction of Hepatobiliary Surgical and Transplant Oncology

Abstract

Background

Hepatobiliary oncology is at an inflection point. Parallel advances in surgical technique, locoregional therapy, systemic treatment, and understanding of tumour biology have expanded the boundaries of curative-intent care yet have also exposed a critical reality that technical resectability is not synonymous with oncologic resectability. Technically successful resection or liver transplantation may be futile when tumour biology is aggressive, leading to early recurrence and limited survival benefit. Modern hepatobiliary surgical and transplant oncology must therefore move beyond anatomy-driven decision-making towards biologically informed strategies that incorporate tumour behaviour, treatment response, and temporal disease stability.

Summary

The Asia-Pacific Transplant Oncology Symposium (APTOS) 2024 examined how Eastern and Western paradigms intersect across hepatocellular carcinoma, cholangiocarcinoma, and colorectal liver metastases, identifying pathways towards integration. Eastern and Western programmes have historically approached hepatobiliary oncology through distinct and longstanding philosophies. Eastern centres, frequently operating within living donor liver transplantation-dominant systems characterized by organ scarcity and the ethical principle of double equipoise, have traditionally leveraged high operative volume and technical expertise in complex resections. These environments have favoured surgery-first approaches when resection appears technically feasible, alongside response-driven conversion strategies aimed at expanding operability. Western centres, shaped by deceased-donor allocation frameworks and trial-oriented oncology models, have more consistently prioritized biological selection and longitudinal validation of disease control before proceeding to curative-intent intervention, particularly in borderline-resectable disease. These differences are driven by infrastructural and ethical factors, including donor availability, transplant systems, and biomarker integration, rather than surgical capability. This article is an expert statement reflecting the opinion of an international multidisciplinary faculty of hepatobiliary surgeons, transplant surgeons, hepatologists, and medical and clinical oncologists from Eastern and Western centres. It is a narrative expert perspective and not a formal consensus or guideline.

Key Messages

This expert statement proposes that the next phase of hepatobiliary oncology requires harmonized, biology-driven definitions of resectability, transplantability, and futility. Comparative evaluation of regional strategies is essential to clarify how treatment sequencing affects recurrence and survival, an effort that is time-sensitive as advances in systemic therapy may soon obscure strategy-specific effects. Rather than competing models, Eastern procedural expertise and Western biology-centred frameworks should be viewed as complementary, supporting context-dependent or sequential approaches that integrate surgery with biology-guided systemic or locoregional therapy. Aligning these paradigms within ethical stewardship of donor organs can establish globally coherent, biology-driven standards that maximize durable survival while minimizing futile surgery and transplantation.

Introduction

Significant progress in hepatobiliary surgical and transplant oncology has been propelled by parallel advancements in surgical techniques, locoregional therapies (LRTs), systemic treatments, and an evolving understanding of tumour biology [1–8]. Historically, achieving an R0 resection was regarded as the cornerstone of cure, irrespective of operative extent or morbidity. However, it is now increasingly recognized that technical resectability does not necessarily equate to “oncologic resectability.” Poor tumour biology, characterized by aggressive molecular profiles or early recurrence patterns, can negate the benefits of radical surgery, leading to futile resections and diminished post-operative survival [1, 3, 9, 10]. In such scenarios, peri-operative or neoadjuvant therapies, and in selected cases, liver transplantation (LT), may represent more effective routes to long-term disease control.

Crucially, differences between Eastern and Western practice philosophies have shaped divergent paradigms in the management of hepatobiliary malignancies [11]. Eastern centres, particularly in Japan, South Korea, China, and Taiwan, have traditionally emphasized surgery-first strategies grounded in technical precision, high operative volume, and extensive expertise in living donor liver transplantation (LDLT). These systems often pursue extended or anatomical resections when negative margins appear technically achievable. By contrast, Western centres have more consistently adopted biology-driven or “conversion-first” strategies, prioritizing neoadjuvant therapy, tumour biology assessment, and multidisciplinary consensus to define resectability and sequence treatment. These differing philosophies reflect more than variations in culture or training. They stem from underlying structural factors, including reliance on LDLT in the East, broader deceased-donor access in the West, and variable integration of biomarker-guided oncology platforms across regions.

Such East-West contrasts underscore the pressing need for collaborative harmonization and evidence-based standardization as transplant oncology continues to evolve. The concept of “transplant oncology” was first introduced by Hibi et al. [12] to integrate transplantation medicine with oncologic principles. This framework was subsequently consolidated through the articulation of the four foundational pillars of transplant oncology, evolution, elucidation, exploration, and extension, by Hibi and Sapisochin [13]. These seminal contributions, together with related discourse within the International Liver Transplantation Society (ILTS), laid the groundwork for the first international consensus conference on transplant oncology held in Rotterdam [14–18]. Following the conference, the ILTS transplant oncology special interest group published a comprehensive position paper formally defining and contextualizing the “4Es” framework (Fig. 1) [19]. Together, these principles delineate the field’s translational trajectory: evolution integrates LT into multidisciplinary cancer care. Elucidation links tumour biology and transplant immunology to refine concepts of self and non-self-recognition. Exploration leverages access to explanted organs to interrogate carcinogenesis and develop transplant-specific endpoints. Extension applies transplantation to expand the boundaries of curative surgery beyond what resection alone can achieve. This shared conceptual scaffold provides a common language for comparing regional strategies and aligning diverse practices towards globally applicable, biology-driven standards in transplant oncology.

Fig. 1.

Differences in the 4Es of transplant oncology, comparing between the East and West (Created in BioRender. Kow, A. (2026) https://BioRender.com/rx0034d). MDT, multidisciplinary tumour board; LRT, locoregional therapy; HCC, hepatocellular carcinoma; BTC, biliary tract cancer; CRLM, colorectal liver metastasis; LT, liver transplantation; PVTT, portal venous tumour thrombosis.

Operationalizing the 4Es across Regions: Shared Principles and Divergent Pathways

Evolution

Multidisciplinary cancer care has advanced globally, but its operational maturity differs. Western programmes have long embedded multidisciplinary tumour boards, structured governance pathways, and coordinated sequencing of systemic, locoregional, and surgical therapies. These frameworks integrate oncologists, hepatologists, transplant physicians, radiologists, and surgeons from the outset, enabling biologically informed treatment navigation. In the East, multidisciplinary practice is expanding but remains variably implemented. Decision-making for liver malignancies is still frequently surgeon-led outside major academic centres, although regional adoption is accelerating alongside broader access to systemic and locoregional treatment platforms.

Elucidation

The interplay between tumour and transplant immunology is foundational to modern transplant oncology. Western centres have driven mechanistic understanding of immune tolerance, rejection biology, and immunomodulation, which now inform peri-transplant immunotherapy (IO) strategies and graft management. In contrast, Eastern centres contribute complementary expertise derived from decades of experience with LDLT and ABO-incompatible grafting, providing a practical counterpoint that refines immunologic risk management in donor-constrained environments. Together, these perspectives illustrate how immunologic elucidation underpins divergent, but increasingly convergent, clinical strategies.

Exploration

Western programmes have generally driven the exploratory integration of advanced biologic and imaging modalities into transplant oncology, including genomics, radiomics, circulating tumour DNA (ctDNA) surveillance, immune profiling, and PET-based metabolic assessment. In colorectal liver metastases (CRLMs), molecular profiling is already embedded within guideline-based management and increasingly informs transplant candidacy through its influence on systemic therapy selection, risk stratification, and biological responsiveness. In contrast, for hepatocellular carcinoma (HCC) and cholangiocarcinoma (CCA), molecular factors remain largely investigational and have not yet been incorporated into formal transplant selection criteria, although they are gaining traction in prognostication and research-led decision support. Eastern centres are progressively adopting similar biologic tools as access and analytic capacity expand, complementing established operative strengths. Across regions, the gradual convergence of biologic insight with technical innovation is beginning to reshape how risk is conceptualized and how future selection thresholds may evolve.

Extension

The extension of oncologic boundaries through transplantation illustrates the most visible East-West divergence. Western programmes have broadened transplant indications for malignancy beyond HCC through highly selective frameworks that prioritize tumour biology, sustained disease stability, and stringent survival expectations in DDLT-driven systems. Conversely, Eastern centres, shaped by limited deceased-donor access, have historically prioritized anatomical resectability and technical innovation, advancing radical and extended resections while applying transplantation more selectively outside HCC. As biologic selection criteria diffuse globally and LDLT capacity matures, extension is increasingly defined not only by technical reach, but by the judicious application of LT to malignancies where resection cannot deliver durable control.

As LT indications expand across HCC, CCA, and CRLM, cross-regional alignment is increasingly necessary to prevent fragmentation in standards and access. Differences in donor availability, integration of systemic therapy, and uptake of biomarkers highlight the need for harmonized frameworks that retain regional adaptability while converging on shared definitions of resectability, transplantability, futility, and meaningful oncologic outcomes. Although recent international consensus statements, most notably the 2024 ILTS-ILCA Joint Consensus on LT for HCC and intrahepatic CCA (iCCA) [20] and the Consensus4pCCA recommendations [21], have advanced the field, regional perspectives on transplant oncology and their variable adoption between the East and West were only partially addressed.

The inaugural Asia-Pacific Transplant Oncology Symposium (APTOS 2024, Chengdu) was convened to specifically examine these East-West contrasts and their implications for the evolution of transplant oncology. This expert statement synthesizes areas of convergence and divergence, articulates shared principles for HCC, CCA, and CRLM, and outlines pathways towards integration. By aligning biologically informed surgical and transplant strategies with ethical stewardship of scarce donor organs, the field can move towards unified standards that maximize survival, minimize futile intervention, and expand curative potential worldwide. Since the conclusion of the conference, several pivotal studies have become available, including CARES-009 [22], ZSAB-TransGOLP [23] and TALENTop [24], all of which have been published in full in peer-reviewed journals, as well as ZSAB-neoGOLP [25], which have been presented as formal trial read-outs. Given their relevance to peri-operative systemic therapy and risk-adapted treatment selection in HPB malignancies, these studies have been subsequently incorporated into this statement with the agreement of all contributing authors, ensuring that the recommendations reflect the most current evidence base. These statements represent the considered expert opinion of the participating multidisciplinary faculty, drawn from high-volume Eastern and Western hepatobiliary and transplant centres, synthesized from the symposium discussions and informed by current literature. They do not constitute a formal consensus, Delphi, or guideline development exercise: no structured voting rounds, predefined agreement thresholds, or formal consensus methodology were undertaken, and none is claimed. This article should therefore be read as a narrative expert perspective rather than a formal consensus statement, and the term “consensus” is used throughout only in reference to published consensus documents or to the state of the literature.

Management of HCC: East-West Perspectives and Integration

Management of Resectable but “High-Risk” HCC (Statements 1-8)

Shared Principles

Across both Eastern and Western centres, the management of resectable but biologically high-risk HCC requires balancing technical feasibility with oncologic appropriateness (Fig. 2; Table 1). While resection offers a potential pathway to cure, it does not guarantee durable survival, particularly in tumours characterized by aggressive pathological features that predispose to early recurrence. In patients with technically resectable disease, factors such as tumour differentiation, vascular invasion, biomarker levels, and radiologic behaviour provide important prognostic information and help contextualize recurrence risk following upfront surgery.

Fig. 2.

Key areas of convergence and divergence of transplant oncology in the East and West for hepatocellular carcinoma (Created in BioRender. Kow, A. (2026) https://BioRender.com/ft2v35s). FLR, future liver remnant; LRT, locoregional therapy; HCC, hepatocellular carcinoma; DFS/OS, disease-free survival/overall survival; LT, liver transplantation; PVTT, portal venous tumour thrombosis; TACE, transarterial chemoembolization; SBRT, stereotactic body radiation therapy; HAIC, hepatic artery infusion chemotherapy; RFA/MWA, radiofrequency ablation/microwave ablation.

Table 1.

Expert statements on HCC: Eastern and Western perspectives

HCC
(A) Management of resectable but “high-risk” HCC
Statement 1	Technical resectability does not necessarily equate to oncological resectability. Resection in resectable but “high-risk” HCC may be futile, as such tumours are likely to recur early, negating the potential benefits of surgical intervention
Statement 2	There is agreement between Eastern and Western centres on the paradigm shift towards emphasizing neoadjuvant therapy and assessing treatment response. This approach aims to better select patients who will benefit from resection, aligning oncological outcomes with surgical feasibility
Statement 3	A clear distinction must be made between neoadjuvant therapy for upfront resectable tumours and downstaging protocols intended for conversion to surgery in initially unresectable or borderline-resectable tumours. The intent of treatment should be explicitly documented at the time of initial diagnosis within an MTB discussion
Statement 4	Concerns regarding dropout rates from preoperative therapy in various studies may be attributed to the inclusion of borderline-resectable tumours rather than tumours intended for true neoadjuvant treatment. Patients whose borderline-resectable tumours fail to downstage and become unresectable after therapy likely represent a subgroup with biologically aggressive disease. For these patients, surgery would have been futile, as they are prone to early recurrence and poor OS
Statement 5	With regard to survival benchmarks for liver resection, a 5-year RFS rate of 40% is widely accepted. However, a 3-year RFS benchmark may serve as a more practical measure, as 5-year RFS could reflect de novo disease rather than true recurrence
Statement 6	For LT, a 5-year OS rate of 50% is agreed upon as the benchmark for acceptable outcomes
Statement 7	Patients failing to meet the above survival benchmarks for either liver resection or transplantation can be classified as being at “high risk of recurrence”
Statement 8	Genotyping and immunophenotyping of recurrent disease are crucial elements for research, as patterns may differ between the original disease, early recurrences, and late recurrences. These insights can guide tailored treatment strategies for biologically aggressive vs. indolent recurrences
(B) Definition of “high-risk” HCC in the context of neoadjuvant therapy
Statement 9	The definition of “high recurrence risk” HCC based solely on size and number is outdated. In addition to traditional “high-risk” classifications such as Milan and up-to-7 criteria, and the criteria established by IMbrave 050, future trials should incorporate novel methods of risk stratification. Future investigations should focus on identifying and validating novel biomarkers
Statement 10	Until reliable prognostic biomarkers are available, traditional clinical and histological parameters, along with treatment response assessments like mRECIST/RECIST 1.1 and prognostic models (e.g., Shanghai score), should guide the use of neoadjuvant therapies in “high-risk” patients
Statement 11	Histological features of microvascular invasion and PET scan findings are not routinely obtained preoperatively and are unlikely to be useful in selecting patients for neoadjuvant therapy
(C) Definition of “borderline-resectable” HCC and downstaging therapy
Statement 12	Upfront curative liver resection for HCC is contraindicated in the presence of extrahepatic disease or tumour thrombus involving the main portal vein (Vp4), inferior vena cava (Vv3), or common bile duct (B4). These features preclude upfront resection. Highly selected patients may, however, be reconsidered for conversion (downstaging) surgery or, in rare cases, transplantation after demonstrable biological response to therapy (see statement 13)
Statement 13	For tumours considered borderline-resectable, downstaging therapy may be considered in scenarios involving tumour thrombus in the portal vein (Vp3, and in highly selected cases Vp4), hepatic vein (Vv2, and in highly selected cases Vv3), and bile duct (B3, and in highly selected cases B4) or limited local invasion of surrounding structures, with the intent of conversion to curative resection or, in selected cases, transplantation following demonstrable biological response
Statement 14	Some centres support upfront resection despite contested criteria. There is no clear consensus on the cut-off for high tumour burden (defined by tumour size, number, and tumour burden score), anatomical extent of portal vein tumour thrombus (VP1 to VP4), AFP levels, PIVKA levels, or distance of tumour from major vascular structures. These factors reflect variability in surgical decision-making practices
(D) Role of peri-operative IO in HCC
Statement 15	Patient selection for ICIs should take into account the likelihood of therapeutic response. While evaluating tumour-immune status or subtype may help predict response to IO, no validated predictive biomarkers currently exist for guiding ICI selection. This highlights a critical need for further research, particularly in identifying T-cell-inflamed tumours and understanding mechanisms of immune resistance, to optimize patient outcomes
Statement 16 (neoadjuvant/downstaging)	Neoadjuvant systemic therapy is commonly delivered as approximately two cycles followed by a rest period before surgery. However, treatment duration and the preoperative interval are not standardized and depend on the agents used. A washout of 4–6 weeks is typical, and a longer interval is required when anti-VEGF therapy (e.g., bevacizumab) is included, in order to mitigate bleeding and wound-healing risk
Statement 17 (downstaging)	A combination of LRT and systemic therapy is a rational downstaging strategy for both resection and transplantation, often beginning with LRT followed by early initiation of systemic therapy. No specific combination or sequence has, however, been proven superior in the absence of direct comparative evidence
Statement 18 (adjuvant)	Adjuvant therapy for HCC remains uncertain in its ability to improve OS or RFS. The IMbrave 050 trial likely did not select for sufficiently high-risk patients, as ~90% of the cohort had a solitary tumour and only ~50% had a lesion ≥5 cm.
Statement 19 (adjuvant)	The optimal hazard ratio target for adjuvant therapy trials remains undefined, as does the ideal duration of adjuvant therapy. These are critical aspects for future studies
Statement 20 (peri-op)	A “sandwich” protocol combining peri-operative systemic therapy, rather than relying solely on neoadjuvant or adjuvant therapy, may be more effective
Statement 21	Therapeutic target endpoints for adjunct therapies should include a response rate >40% as assessed by mRECIST and pCR rates of 30–40%
Statement 22	Even though some studies have correlated pCR to OS, the significance of pCR in HCC remains uncertain, as it may not directly correlate with survival outcomes. Further research is required to clarify its role
(E) Role of peri-LT IO in HCC
Statement 23	Pre-transplant use of ICIs may be feasible in carefully selected patients within protocolized settings. While the optimal washout period before LT is debated, a washout of approximately 3 months has been associated with a lower risk of acute allograft rejection. However, a clinically relevant risk of severe, occasionally fatal, rejection remains, and pre-transplant ICI use should currently be regarded as investigational
Statement 24	The influence of ICI washout periods and post-operative immunosuppressive regimens on tumour recurrence remains unclear and warrants further investigation
Statement 25	A combination of ICIs and LRT is a promising strategy for pre-transplant bridging or downstaging therapy, although comparative evidence establishing its superiority over alternative approaches is currently lacking
Statement 26	Post-transplant use of ICIs for recurrence carries a high risk of graft rejection. Patients experiencing rejection often had shorter intervals between LT and ICI initiation (2–3 years). Conversely, responders to IO typically had longer intervals (4–6 years). Careful timing and monitoring are essential to minimize rejection risks
Statement 27	High PD-1/PD-L1 expression in the graft appears to predict the risk of rejection following post-LT ICI therapy. This remains an investigational rather than a clinically validated tool
Statement 28	T-cell immunomonitoring is a promising but currently investigational strategy that may help reduce graft rejection risk with ICI therapy
Statement 29	Patient selection for peri-transplant IO remains uncertain. Criteria may include high-risk pre-LT imaging features, adverse post-LT explant histology, or adaptive strategies based on response to bridging or downstaging therapies, assessed via mRECIST or MRD status post-LRT or LT
Statement 30	Key parameters and endpoints for assessing locoregional and systemic therapies in downstaging and bridging include progression-free survival, treatment response rates, disease control rates, duration of response, toxicity profiles, waitlist time, and dropout rates
Statement 31	Defining upper limits for tumour burden amenable to downstaging and establishing extended criteria for LT (LDLT and DDLT) are critical. Standardized protocols based on tumour size, number, and biology are necessary for consistent application across centres

MTB, multidisciplinary tumour board.

In contrast, for patients with borderline-resectable disease, integration of systemic therapy or LRT serves not only as cytoreduction but also as a means of biological assessment, helping identify tumours with favourable response dynamics that are more likely to derive durable benefit from resection or transplantation. A clear conceptual distinction should therefore be made between neoadjuvant therapy and downstaging protocols. Neoadjuvant therapy is directed at patients with upfront resectable disease, aiming to eradicate micrometastases and reduce recurrence risk. In contrast, downstaging therapy targets borderline or initially unresectable tumours to achieve sufficient tumour regression for curative resection or transplantation. Reported dropout rates in preoperative therapy trials often reflect inclusion of these borderline cases, which are intrinsically associated with more aggressive disease biology and poorer prognosis, rather than failure of neoadjuvant therapy itself. Within the context of resectable but high-risk HCC, three peri-operative systemic treatment strategies are currently under active investigation, and there remains no clear consensus regarding the optimal approach:

• 1.Neoadjuvant therapy alone, administered prior to surgery with the aim of tumour priming, early biologic risk stratification, and reduction of occult metastatic burden.
• 2.Adjuvant therapy alone, delivered after curative resection to mitigate recurrence risk in patients with adverse pathological or clinical features.
• 3.Peri-operative therapy, combining neoadjuvant and adjuvant treatment to maximize both early tumour control and sustained post-operative suppression of residual disease.

Each of these strategies has demonstrated feasibility and biological activity in early studies, yet their relative contributions to long-term outcomes remain uncertain. Accordingly, these approaches will be discussed individually in the subsequent sections.

At APTOS 2024, experts from both regions were in broad agreement that tumour biology should guide surgical decision-making, highlighting the growing role of neoadjuvant and peri-operative therapy in refining patient selection and improving oncological outcomes, and emphasizing the need for continued development of prospective clinical trials and structured treatment protocols. Insights from other malignancies, including gastroesophageal, rectal, and breast cancers, demonstrate the survival benefits of pre-operative and peri-operative approaches and provide a translational framework for their adaptation to HCC [26–28].

Regional Perspectives (Divergence)

Approaches to the management of HCC vary across regions, reflecting differences in clinical philosophy, disease epidemiology, trial culture, and health system structure rather than differences in technical capability. In East Asia, particularly in Japan and China, high-volume centres have traditionally pursued aggressive, surgery-oriented strategies when an R0 resection appears technically feasible. This approach is supported by population-based surveillance, extensive experience with anatomical liver resection, and broad utilization of multimodal therapies to facilitate curative-intent treatment.

The concept of conversion therapy, using systemic and/or LRT to render initially unresectable tumours amenable to curative intervention, originated in East Asia and has been most clearly articulated through the atezolizumab-bevacizumab-conversion approach described by Kudo [29]. In this paradigm, treatment response serves to expand surgical options, with early reassessment for resection or transplantation once technical feasibility is achieved. Beyond resection, some Eastern centres also apply downstaging strategies to facilitate LDLT, leveraging donor availability and institutional expertise.

However, the criteria used for downstaging to LDLT across Eastern programmes are heterogeneous and, in some settings, relatively permissive. While this flexibility enables access to curative therapy for patients beyond conventional thresholds, it also raises concerns regarding oncologic futility, post-transplant recurrence, and equitable donor risk. Increasingly, there is recognition within the region that downstaging strategies, particularly those leading to LDLT, should be accompanied by more rigorous biological selection, incorporating treatment response, biomarker dynamics, and temporal stability, rather than relying on technical downstaging alone.

In contrast, Western centres have generally emphasized biologic risk stratification and longitudinal validation as prerequisites for curative-intent therapy, especially LT. Response to systemic or LRT is interpreted as a surrogate of tumour aggressiveness rather than an automatic trigger for conversion. This philosophy is shaped in part by reliance on DDLT, where organ scarcity necessitates stringent selection and prolonged observation, and where non-surgical therapies often function as bridging or selection tools rather than direct enablers of conversion.

As systemic therapies achieve higher response rates and durability, these regional approaches are increasingly converging. Contemporary management of HCC is moving towards a hybrid framework in which conversion, downstaging, resection, and transplantation are guided by both technical feasibility and biological behaviour, with growing consensus that stricter, biology-informed criteria are essential, particularly when extending curative strategies to transplant candidates.

Pathways towards Integration (Convergence)

Contemporary studies evaluating systemic and LRT in the pre- and peri-operative setting have provided important proof of concept that tumour biology can be modulated and interrogated prior to definitive intervention, without compromising surgical safety [30–35]. However, these data also underscore a key limitation in current paradigms: early signals of radiologic or pathological response do not reliably translate into durable oncologic benefit. Across available studies, there remains no consistent evidence that incorporating additional therapies around surgery uniformly reduces post-operative recurrence or improves overall survival (OS) compared with surgery alone. This highlights the central challenge in managing high-risk resectable HCC, namely, distinguishing patients with aggressive disease biology who may benefit from treatment intensification from those with favourable biology who risk overtreatment [36].

Accordingly, future collaborative efforts should focus on biologically informed patient selection, integrating into decision-making for resection, downstaging, or transplantation. Integration of dynamic treatment response, biomarker kinetics (such as AFP or PIVKA-II trajectories), radiologic behaviour, and temporal disease stability into treatment algorithms will be critical to refining decisions around resection, treatment sequencing, downstaging, or transplantation. Multiregional trials and collaborative registries incorporating systemic therapy, LRT, and surgery, designed with harmonized endpoints and adequate long-term follow-up, will be essential to define global standards. In parallel, international platforms such as APTOS can facilitate real-world validation, enabling convergence between East Asian conversion-driven paradigms and Western biology-informed sequencing frameworks, and ultimately improving outcomes for patients with resectable but high-risk HCC.

Definition of “High-Risk” HCC in the Context of Neoadjuvant Therapy (Statements 9-11)

Shared Principles

Emerging evidence from early-phase prospective studies demonstrates that neoadjuvant therapy in resectable HCC is feasible, safe, and associated with meaningful radiologic and pathological responses [32–35]. In light of recent data, including positive peri-operative trial results such as CARES-009 [22], which demonstrated the efficacy of systemic therapy delivered around surgery, neoadjuvant therapy should no longer be viewed as merely exploratory, but rather as a biologically active strategy with potential clinical benefit, albeit one that remains to be refined and validated. Both Eastern and Western centres recognize that careful definition of “high-risk” disease is essential to identify patients most likely to benefit from neoadjuvant approaches while avoiding overtreatment of biologically favourable tumours. Historically, morphologic criteria such as the Milan, up-to-7, and tumour burden-based definitions have been used to stratify risk in resectable HCC. However, these criteria alone incompletely capture tumour aggressiveness and propensity for early recurrence [37–39]. There is growing consensus that neoadjuvant strategies should be considered primarily for patients with high recurrence risk despite technical resectability.

Regional Perspectives (Divergence)

Eastern centres have traditionally defined “high-risk” resectable HCC predominantly based on tumour burden, morphology, and vascular involvement, reflecting pragmatic surgical experience, high volumes of anatomical resection, and established integration of LRTs. This approach has yielded strong outcomes in selected patients and remains central to practice across much of Asia.

Western centres, while sharing the same therapeutic goals, have increasingly emphasized histologic and biochemical markers, such as microvascular invasion, poor tumour differentiation, tumour size or multiplicity, and elevated serum markers including AFP and PIVKA-II, to refine patient selection and recurrence risk assessment. Molecular and multi-omics factors are increasingly recognized as potentially important, but they are not yet routinely or uniformly incorporated into clinical decision-making in either region, owing to limited standardization, validation, and accessibility.

Accordingly, contemporary neoadjuvant trial designs in both East and West largely rely on morphologic, histopathologic, and biochemical criteria to define high-risk populations, rather than molecular selection. Neoadjuvant-specific studies [30, 33] reflect this pragmatic, and stage- and risk-based approach to patient selection.

Pathways towards Integration (Convergence)

A growing body of ongoing and recently completed clinical trials is actively investigating neoadjuvant strategies in resectable but high-risk HCC, with the aims of refining patient selection, interrogating tumour biology prior to surgery, and ultimately reducing recurrence risk [40–47]. Collectively, these studies have established that preoperative IO-based regimens can be delivered without compromising surgical safety and can induce biologically meaningful tumour responses. Considering recent peri-operative data, the question has shifted from whether neoadjuvant therapy is biologically active to which patients derive durable benefit, and how neoadjuvant strategies should be optimally integrated into curative-intent pathways.

At present, no validated biomarker reliably predicts neoadjuvant benefit or recurrence risk in resectable HCC. Until such tools are established, clinicians must rely on multifactorial risk assessment, integrating clinical presentation, tumour morphology, histopathology, serum biomarkers (including AFP and PIVKA-II) [48], and radiologic behaviour to identify candidates most likely to benefit from neoadjuvant therapy. Radiologic assessment using RECIST 1.1 or mRECIST remains central for evaluating tumour dynamics and treatment response in this setting [49].

Moving forward, East-West collaborative efforts should prioritize both external validation and calibration of existing clinicopathologic risk models, including tools such as the ERASL score [50], and the development of next-generation risk models that better capture contemporary treatment contexts and biological heterogeneity. As emerging biomarkers such as ctDNA [51–53], radiomic signatures [54, 55], and immune profiling [2, 5, 56–58] mature, they should be evaluated not merely as add-ons but as candidate components of integrated, prospectively validated prediction frameworks. The goal is to deliver risk stratification that is precise enough to guide treatment intensification (including neoadjuvant strategies) while remaining scalable and generalizable across diverse populations and health systems.

Definition of Borderline-Resectable HCC and Downstaging Therapy (Statements 12-14)

Shared Principles

Across both Eastern and Western centres, the definition of borderline-resectable HCC remains evolving and institution-dependent [59–61]. With advances in locoregional and systemic therapies, including IO [62], the role of downstaging therapy has expanded as an alternative pathway to curative surgery [29, 63–66]. Conceptually, this approach is distinct from neoadjuvant therapy applied to upfront resectable disease: downstaging is directed at tumours that are technically marginal at presentation, with the aim of reducing tumour burden, improving technical feasibility, and enabling oncologically sound resection.

Borderline-resectable HCC typically includes tumours that are technically operable but associated with high oncologic risk without prior therapy, or those constrained by insufficient future liver remnant volume or function, raising concern for post-operative liver failure. Such cases may involve tumour thrombus in the portal vein (Vp3–Vp4), hepatic vein and inferior vena cava (Vv2–Vv3), and bile duct (B3–B4) [59], limited local invasion, or other indicators of aggressive disease biology. Downstaging functions as a biological selection strategy, identifying patients with favourable treatment response while excluding those with progressive disease unlikely to benefit from major surgery. This risk-adapted approach supports parenchymal-sparing strategies and reduces the morbidity associated with futile resections. IO may further contribute through immunomodulatory effects on the tumour microenvironment, potentially altering tumour behaviour and improving post-treatment disease control [67–69]. Upfront curative resection remains contraindicated in the presence of extrahepatic disease or tumour thrombus in the main portal vein, inferior vena cava, or common bile duct (Vp4, Vv3, or B4) [59], where durable oncologic benefit is unlikely and downstaging or conversion strategies should instead be considered.

The recently published study by Lin et al. [70] evaluating preoperative nivolumab plus ipilimumab represents an important contribution to this evolving field. Although described by the investigators as a neoadjuvant IO study, the enrolled population largely comprised patients with borderline-resectable or locally advanced disease, many of whom would more accurately be classified under a downstaging paradigm. This distinction is not merely semantic: it has implications for trial design, interpretation of outcomes, and comparison across studies. The study reinforces the need for precise terminology to differentiate treatment applied to upfront resectable disease from response-adapted strategies aimed at enabling resection in marginal cases. Notably, this study is among the first in HCC to incorporate detailed immunologic endpoints, including longitudinal immune profiling and tumour microenvironment analyses, alongside conventional radiologic and pathologic assessments. These data provide rare mechanistic insight into how preoperative IO reshapes host-tumour interactions in the downstaging setting, complementing clinical outcome measures. However, as with other early-phase studies, survival data remain immature, and the generalizability of immunologic findings across broader populations requires further validation.

In parallel, the TALENTop trial represents a key development [24] as a true downstaging study to evaluate patients with locally advanced HCC. In this study, HCC with macrovascular invasion is treated with induction systemic therapy prior to consideration of resection. In the published trial, 489 patients entered induction and 201 were randomised; median time to treatment failure was 20.4 months with resection versus 11.8 months with maintenance therapy (HR 0.60, 95% CI 0.39–0.91; p = 0.015), supporting response-guided resection in selected patients, although overall survival data are awaited.

Taken together, these studies underscore both the promise and the complexity of downstaging strategies in HCC. They highlight the importance of clearly distinguishing neoadjuvant from downstaging intent, integrating biological response into surgical decision-making, and incorporating immunologic and translational endpoints to better understand mechanisms of benefit. As the field matures, harmonized definitions and endpoint frameworks will be essential to enable meaningful cross-trial comparison and guide clinical adoption.

Regional Perspectives (Divergence)

Definitions and treatment thresholds for borderline-resectable HCC differ substantially between regions. Eastern centres, especially high-volume programmes in Japan, Korea, India, and China, often adopt a surgery-first or aggressive downstaging approach, supported by extensive expertise in vascular resection, portal vein thrombectomy, and complex anatomical hepatectomy. In this context, integration of LRT with systemic treatment is increasingly employed to facilitate resection in patients with locally advanced disease. This approach is reflected in various Asian-led studies [71–73] reporting successful downstaging and subsequent resection following combinations of LRT and systemic therapies in selected patients with locally advanced HCC.

The STRATUM study in HCC (AHCC09) exemplifies this regional philosophy [74]. STRATUM is an actively recruiting, Asia-led trial evaluating the integration of selective internal radiation therapy (Y-90 resin microspheres) followed by atezolizumab plus bevacizumab for patients with locally advanced HCC. While no efficacy results have yet been reported, the trial highlights an ongoing commitment within the region to response-guided downstaging strategies aimed at expanding curative-intent options.

In contrast, Western centres tend to define borderline resectability less by anatomic feasibility alone and more by biologic behaviour and durability of disease control. Selection for downstaging and subsequent surgery is typically guided by radiologic or biomarker-based evidence of disease control, reflecting a philosophy that tumour biology, rather than technical resectability, ultimately determines the value of operative intervention.

These divergent approaches are shaped by broader structural factors. Eastern programmes function largely within LDLT-dominant systems and high-volume surgical environments, allowing earlier consideration of surgery following favourable response. Western centres, operating within deceased-donor allocation frameworks and trial-driven treatment paradigms, generally require more prolonged biological validation before proceeding to resection or transplantation, prioritizing long-term oncologic benefit over early technical conversion.

Pathways towards Integration (Convergence)

The absence of universally accepted criteria for borderline-resectable HCC underscores the urgent need for standardized definitions and shared endpoints [59, 75]. Traditional metrics, tumour size, serum markers (AFP, PIVKA-II), composite scores (e.g., Shanghai score [76]), and classifications of portal or hepatic vein involvement [77, 78], remain inconsistently applied and lack validated cut-offs.

Future consensus efforts should therefore pursue a hybrid definition of borderline resectability that integrates both anatomic extent and biological risk. Incorporation of emerging precision tools, such as radiomics [54, 55, 79], molecular and immune-related profiling [56, 80, 81], and tumour microenvironment or TP53-associated signatures [82, 83], holds promise for refining patient selection and contextualizing treatment response, although these approaches remain investigational and are not yet ready for routine clinical implementation.

With respect to systemic therapy, no specific neoadjuvant or downstaging regimen has yet been proven superior in this setting. Multiple combination strategies, including immune checkpoint inhibitors (ICIs) paired with anti-angiogenic agents [84, 85], ICIs combined with TKIs [86, 87], and dual immune checkpoint blockade, have demonstrated biological activity and encouraging response rates in various studies. Cross-trial comparisons suggest that dual-ICI regimens, such as nivolumab plus ipilimumab [40, 88], may achieve higher objective response rates, but such comparisons are inherently limited by differences in study design, patient selection, and endpoints, and do not establish comparative efficacy. Accordingly, the relative merits of ICI-TKI combinations, ICI-anti-angiogenic regimens, and dual-ICI approaches in the neoadjuvant or downstaging setting remain unresolved and require prospective, head-to-head evaluation.

Ultimately, harmonization will depend on the development of multidisciplinary, risk-adapted algorithms that integrate imaging, clinicopathologic features, and emerging biological insights, rather than reliance on any single platform. Multinational East-West collaborative trials will be essential to define optimal sequencing strategies and ensure consistent, equitable management of borderline-resectable HCC across diverse practice environments.

Role of Peri-Operative IO in HCC (Statements 15-22)

Shared Principles

Peri-operative IO has emerged as an important strategy in HCC, with the aim of improving tumour control, modulating the tumour-immune microenvironment, and reducing micrometastatic relapse. In this context, the peri-operative “sandwich” strategy seeks to combine the potential advantages of both neoadjuvant and adjuvant therapy by targeting distinct phases of tumour-host interaction [30, 33, 35, 89]. Across both Eastern and Western centres, there is broad consensus that appropriate patient selection is critical, as indiscriminate peri-operative treatment risks overtreatment without durable oncologic benefit.

Although tumour-immune phenotypes and inflammatory states are biologically relevant, no validated predictive biomarkers currently guide ICI use [90–94]. Ongoing studies are evaluating candidate markers, including PD-L1 expression [95], tumour mutational burden [96], immune gene signatures [58, 97], ctDNA [51–53], and tumour-infiltrating lymphocytes [98, 99], primarily as prognostic and response-monitoring tools, rather than definitive treatment selectors.

Neoadjuvant IO aims to favourably modify tumour biology prior to resection, eradicate occult micrometastatic disease, and provide early biological risk stratification. Typical protocols involve short courses of systemic therapy (often two cycles), followed by a defined washout period (often 4–6 weeks) before surgery. Treatment duration and the preoperative interval are not standardized, however, and depend on the agents used, with a longer interval required when anti-VEGF therapy such as bevacizumab is included, to mitigate bleeding and impaired wound healing. Early-phase studies combining ICIs with LRTs, such as transarterial chemoembolization, Y-90 radioembolization, or stereotactic body radiotherapy, have demonstrated feasibility, acceptable peri-operative safety, and meaningful radiologic and pathological responses, while preserving surgical outcomes [100, 101]. Importantly, these approaches enable in vivo assessment of therapeutic and immune response, supporting adaptive peri-operative decision-making [30, 33, 35, 89].

Adjuvant and peri-operative strategies extend this concept beyond surgery. While the initial information from the IMbrave050 trial demonstrated an improvement in recurrence-free survival (RFS) with adjuvant atezolizumab-bevacizumab in pathologically high-risk HCC, establishing proof of principle that systemic therapy can modify post-resection outcomes, subsequent follow-up analyses did not show a clear OS benefit [102, 103]. This underscores the limitations of adjuvant therapy when applied to broadly defined risk populations and reinforces the importance of refined biological enrichment. On the other hand, recent peri-operative studies have shifted the paradigm. For example, the CARES-009 study demonstrated significant improvement with peri-operative IO in high-risk but resectable HCC [22]. Notably, the short event-free survival observed in the control arm suggests effective enrichment for biologically aggressive disease, supporting the concept that stringent high-risk selection may be key to demonstrating benefit. While these findings represent the strongest evidence to date that systemic therapy delivered around surgery can improve outcomes, they are derived predominantly from Asian populations and require validation in Western cohorts. Another study to note, the TALENTop trial [24], represents a true downstaging-to-resection strategy in locally advanced HCC with macrovascular invasion. TALENTop reinforces the role of response-adapted peri-operative decision-making, rather than reliance on upfront anatomic classification alone.

These studies emphasize that peri-operative (“sandwich”) IO can confer meaningful benefit in carefully selected, biologically high-risk patients, although the precise definition of this population continues to evolve. The optimal regimen, sequencing, and duration of therapy remain undefined, and broader validation is required.

Regional Perspectives (Divergence)

Eastern centres, building on extensive experience with LRT and LDLT, are increasingly adopting neoadjuvant and “sandwich” IO protocols integrating ICIs with transarterial chemoembolization or TARE to enhance tumour regression before surgery. These approaches are supported by large procedural volumes and flexibility in sequencing surgery with IO, enabling early real-world insights into feasibility and safety.

Western centres, on the other hand, have focused on trial-based validation and biomarker discovery, emphasizing rigorous patient selection, pathological response metrics, and prospective evaluation of surrogate endpoints such as pathologic complete response (pCR) and 3-year RFS. The West’s strong clinical trial infrastructure has facilitated detailed immunologic correlative studies, such as T-cell profiling and immune gene expression analyses, that are informing mechanistic understanding of tumour-immune interactions. Together, these complementary approaches reflect regional strengths: the East’s procedural pragmatism and rapid clinical translation, and the West’s data-driven standardization and biomarker sophistication.

Pathways towards Integration (Convergence)

Despite encouraging results, several practical and scientific challenges continue to limit the widespread adoption of peri-operative IO in HCC. Treatment-related toxicity or disease progression during therapy may delay potentially curative surgery, and a subset of patients may discontinue treatment prior to resection, affecting overall feasibility. In addition, institutional requirements for histologic confirmation before initiating IO in some settings may further complicate patient selection and trial implementation.

Assessment of treatment response also remains an evolving area. Radiologic changes following IO can be subtle or atypical, complicating response interpretation. Importantly, emerging data increasingly suggest that major pathologic response (MPR) and pCR are clinically meaningful endpoints in the neoadjuvant setting, with growing evidence of correlation with event-free outcomes, as highlighted in recent analyses and prospective studies, including CARES-009 and contemporary meta-analytic work [22, 104–106]. Nonetheless, the strength and consistency of these correlations across different regimens, disease stages, and populations remain incompletely defined, underscoring the need for further validation. A limited understanding of how neoadjuvant immune oncologic therapies reshape the TME and systemic immunity over time continues to constrain optimization of treatment sequencing and duration [35, 107, 108].

The next phase of peri-operative IO development will require greater harmonization of trial design, response definitions, and biomarker frameworks. Multinational studies should adopt standardized definitions of pCR, MPR, and immune-related radiologic response to enable robust cross-trial and cross-regional comparison. While early-phase studies have proposed radiologic response rates exceeding ∼40% by mRECIST and pCR rates of 30–40% as clinically meaningful benchmarks, emerging data suggest that higher depths of response may translate into greater survival benefit, as exemplified by the pathologic response rates observed in selected high-risk cohorts such as CARES-009 [22]. Ultimately, 3- and 5-year RFS and OS remain the definitive efficacy endpoints, and continued efforts to validate pCR and MPR as reliable surrogate markers are essential [105, 109].

Parallel efforts should prioritize biologically aggressive patient subsets where the risk-benefit balance of peri-operative IO is most favourable and the potential for demonstrating survival benefit is greatest [110, 111]. Progress will also depend on integrating multi-omics approaches to refine patient selection and elucidate mechanisms of resistance. Dynamic biomarkers such as ctDNA, tumour-infiltrating lymphocyte profiling, and radiomic signatures hold promise for real-time monitoring and response-adapted escalation or de-escalation of therapy [51–53]. Broader global trial participation across diverse healthcare systems will be essential to ensure generalizability. Key unresolved questions include optimal treatment duration, appropriate hazard ratio targets, and the role of biomarkers such as ctDNA in guiding therapy continuation or intensification [104]. Ultimately, by combining the technical expertise and pragmatic innovation of Eastern centres with the trial discipline and biologically informed frameworks of Western programmes, the international HCC community can establish evidence-based, biology-integrated standards for peri-operative IO, translating emerging biological insights into durable and globally applicable clinical benefit.

Role of Peri-Transplant IO in HCC (Statements 23-31)

Shared Principles

Peri-transplant IO represents a rapidly evolving strategy to optimize outcomes for patients with HCC undergoing LT. This paradigm includes pre-transplant and post-transplant applications, which serve fundamentally different purposes and carry distinct risk-benefit considerations. Importantly, current clinical practice and available evidence support a much stronger role for IO in the pre-transplant setting, whereas post-transplant IO use remains largely discouraged and investigational.

Pre-Transplant IO: Downstaging vs. Bridging. In the pre-transplant setting, the use of ICIs remains selective and largely confined to clinical trials or highly protocolized pathways, rather than routine practice across most centres. Where employed, ICIs have been explored more frequently as part of downstaging strategies rather than bridging. Downstaging aims to reduce tumour burden in patients who initially exceed transplant criteria to achieve eligibility, whereas bridging therapy seeks to maintain tumour control in patients already within criteria while awaiting transplantation. Although ICIs may theoretically contribute to both objectives, current evidence and prevailing clinical practice support their use primarily within carefully defined downstaging protocols in biologically aggressive disease, rather than indiscriminate bridging [112–114].

Pre-transplant IO may function as a biological selection tool, identifying patients with unfavourable tumour biology who progress despite therapy and are therefore unlikely to benefit from LT, even if morphologic criteria are met. In this way, ICIs may help avoid futile transplants and improve overall transplant utility. Importantly, this conceptual benefit underpins the rationale for evaluating ICIs within prospective studies rather than routine clinical use [115].

Available data suggest that pre-transplant ICI exposure may be feasible in highly selected patients, provided that strict eligibility criteria and an adequate washout period, typically in the order of 3 months, are observed to lower the risk of acute graft rejection, although a clinically relevant risk of severe, occasionally fatal, rejection remains [116–118]. Accordingly, most contemporary protocols, including large prospective efforts such as VITALITY [119], mandate stringent inclusion criteria, predefined washout intervals, and close immunologic monitoring, reinforcing that pre-transplant IO should currently be restricted to structured study settings.

Combination approaches integrating ICIs with LRT have shown encouraging early signals with respect to tumour control and downstaging success, particularly in selected patients with aggressive disease biology [34, 120–123]. However, evidence remains limited, heterogeneous, and largely non-randomized. While the VITALITY study demonstrated acceptable survival outcomes in a carefully selected cohort [119], these data should be interpreted as feasibility and safety validation, rather than endorsement of routine pre-transplant ICI use, and should be regarded as important proof-of-concept evidence that remains exploratory.

These data support a cautious, study-driven approach to pre-transplant IO, with the overarching aim of improving transplant eligibility and long-term outcomes. Future progress will depend on prospective trials designed to define optimal patient selection, sequencing with LRT, washout duration, and immunosuppressive management. Until such evidence matures, pre-transplant IO should be actively encouraged within clinical trials and registries, rather than adopted as standard practice.

Post-Transplant IO: Current Limitations and Unmet Need. In contrast, post-transplant IO is largely discouraged in routine practice due to the substantial risk of immune-mediated allograft rejection. Importantly, patients with prior LT have been explicitly excluded from all major phase III ICI trials in HCC, resulting in a complete absence of high-level evidence supporting efficacy or safety in this population. As such, post-transplant ICI use remains confined to selected cases of recurrent HCC, where potential oncologic benefit must be carefully weighed against the risk of graft loss [124, 125]. Observational data suggest that rejection risk is highest when ICIs are introduced within 2–3 years post-transplant, whereas longer intervals of 4–6 years may be associated with improved tolerance [126–129]. High intra-graft PD-1/PD-L1 expression has been linked to rejection risk, underscoring the importance of immune profiling and timing [130, 131]. However, the impact of pre-transplant ICI washout duration, persistent immune activation from prior ICI exposure, and post-transplant immunosuppressive regimens on recurrence and graft outcomes remains poorly defined and represents a major unmet research need [116–118].

Regional Perspectives (Divergence)

In Eastern centres, where LDLT predominates, the integration of IO into downstaging, and to a lesser extent bridging, strategies has progressed rapidly. The ability to plan LT on flexible timelines allows dynamic sequencing of ICI with LRT and careful management of washout intervals, facilitating biologically informed patient selection. This approach is supported by high surgical volumes and extensive experience in complex HCC management, enabling early clinical adoption and iterative protocol refinement within LDLT-dominant systems.

To date, true ICI-based downstaging with intent to proceed to LT has been reported primarily from East Asia. Most notably, investigators from Renji Hospital in Shanghai, China, demonstrated the feasibility of pembrolizumab-based therapy as a downstaging strategy prior to LT in selected patients with advanced HCC beyond conventional transplant criteria [132]. This study provides important proof of concept that IO-guided downstaging can expand transplant eligibility while serving as a biological selection tool to mitigate futile transplantation in patients with aggressive disease biology. In contrast, Western centres, which rely predominantly on DDLT, have adopted a more conservative and stepwise approach. Longer and less predictable waiting times necessitate stringent oncologic selection criteria and careful validation of IO washout safety prior to listing. As a result, Western programmes have prioritized prospective, structured studies to characterize safety, immunologic correlates of rejection, and post-transplant outcomes following pre-transplant ICI exposure. Large multicentre efforts such as VITALITY exemplify this approach [119], providing systematic evaluation of pre-transplant IO strategies rather than driving early clinical adoption.

Currently, the practices reflect structural and systemic differences rather than fundamental philosophical divergence. Eastern centres emphasize early, response-guided clinical decision-making enabled by LDLT-dominant systems, while Western centres anchor progress in rigorous validation shaped by deceased-donor allocation constraints. Convergence between these approaches, where real-world innovation is refined and standardized through prospective multicentre validation, will be critical to defining the future role of IO-based downstaging in transplant oncology for HCC.

Pathways towards Integration (Convergence)

Achieving global harmonization in peri-transplant IO will require coordinated efforts to standardize clinical protocols while allowing flexibility across differing transplant systems. Key areas of focus include the optimization of treatment timing, washout strategies, and eligibility criteria, alongside the structured incorporation of immune monitoring and biologic risk stratification frameworks.

Several priorities have emerged. Firstly, refinement of ICI washout intervals through large, multicentre registries is essential to balance antitumour efficacy with graft safety across both living and deceased-donor transplantation platforms. Secondly, tumour burden and biological criteria used to define upper limits for downstaging should be standardized to ensure consistency in patient selection and equitable application across regions [133, 134]. Moreover, immune monitoring strategies, particularly T-cell immune monitoring, are increasingly recognized as promising adjuncts to peri-transplant IO. These approaches may help assess washout adequacy, guide individualized immunosuppressive thresholds [135–138], and identify patients at heightened risk of rejection or chronic graft dysfunction, thereby enabling more tailored post-transplant surveillance and immune management [139]. While not yet ready for routine implementation, such tools represent an important area for prospective validation. The development of predictive and dynamic biomarkers, including ctDNA, immune-related signatures (e.g., PD-L1 expression), and cytokine profiles, may further refine the timing and sequencing of IO relative to transplantation. However, in the near term, these biomarkers are more likely to complement existing peri-transplant decision-making regarding ICI timing, washout, and immunosuppressive management than to replace current clinical frameworks.

Ultimately, multinational collaboration will be critical to ensure equitable access to clinical trials, enable meaningful comparisons across LDLT- and DDLT-dominant systems, and validate outcomes in diverse populations. By integrating the procedural experience and LDLT expertise of Eastern centres with the immunologic precision, regulatory rigour, and prospective trial infrastructure of Western programmes, the global transplant oncology community can establish evidence-based, system-agnostic standards for peri-transplant IO that maximize safety, efficacy, and patient access worldwide.

Management of CCA: East-West Perspectives and Integration

Management of Resectable but “High-Risk” CCA (Statements 32-42)

Shared Principles

Across Eastern and Western hepatobiliary programmes, there is growing recognition that technical resectability does not guarantee oncologic success in CCA. Early recurrence remains common, with more than half of patients relapsing after surgery and 5-year OS rarely exceeding 45% [1, 140–143] (Fig. 3; Table 2). This has driven increasing emphasis on biologic resectability, incorporating tumour behaviour, differentiation, perineural and vascular invasion, and broader indicators of aggressive disease into surgical decision-making.

Fig. 3.

Key areas of convergence and divergence of transplant oncology in the East and West for CCA (Created in BioRender. Kow, A. (2026) https://BioRender.com/p7jvto1). FLR, future liver remnant; LRT, locoregional therapy; DFS/OS, disease-free survival/overall survival; LT, liver transplantation; TACE, transarterial chemoembolization; SBRT, stereotactic body radiation therapy; HAIC, hepatic artery infusion chemotherapy; RFA/MWA, radiofrequency ablation/microwave ablation.

Table 2.

Expert statements on CCA: Eastern and Western perspectives

CCA
(A) Management of resectable but “high-risk” CCA
Statement 32	Technical resectability does not equate to oncological resectability in CCA. Surgery for biologically aggressive tumours may prove futile, as such tumours are prone to early recurrence, negating the potential benefits of resection
Statement 33	A clear distinction must be made between neoadjuvant therapy for upfront resectable tumours and downstaging protocols intended for conversion surgery in borderline-resectable tumours. The intent of treatment should be explicitly documented at the time of initial diagnosis within an MTB discussion
Statement 34	The role of neoadjuvant therapy in resectable and downstaging therapy in borderline-resectable CCA lacks consensus, as some centres, especially in the East, prioritize upfront surgery to attempt to achieve R0 resection and mitigate risks of preoperative therapy dropout due to complications or disease progression
Statement 35	Non-nodal dissemination, such as perineural spread, is a significant negative prognostic marker that remains challenging to evaluate. Future research, including advancements in radiomics, may improve detection, aiding in disease assessment and treatment planning
Statement 36	Future directions in CCA management include the use of biomarkers to predict the futility of resection and identify patients better suited for LT
Statement 37 (pCCA)	Absolute contraindications to pCCA resection include metastatic disease, poor hepatic reserve or performance status, vascular invasion precluding safe reconstruction, and bilateral biliary extension preventing R0 resection (e.g., beyond the limits of a trisectionectomy)
Statement 38 (pCCA)	Relative contraindications to upfront surgery for pCCA include poorly localized PSC-associated tumours, vascular involvement, nodal involvement, and elevated CA19-9 levels post-biliary drainage (e.g., >300), which suggest poor tumour biology despite the lack of a definitive consensus on cut-off values
Statement 39 (pCCA)	Borderline-resectable pCCA cases, particularly those requiring vascular reconstruction, may benefit from downstaging therapy before surgery to improve resectability and outcomes
Statement 40 (pCCA)	The use of neoadjuvant or downstaging therapy in pCCA presents significant challenges due to the high risk of infective complications, including cholangitis and liver abscess, associated with a blocked biliary system
Statement 41 (iCCA)	For iCCA, absolute contraindications to surgery include distant metastatic disease and inadequate hepatic reserve or performance status. Extensive multifocal disease (e.g., more than 3 tumours) and regional nodal involvement are strong adverse prognostic factors that substantially reduce the benefit of resection. However, they are not universally accepted as absolute contraindications, as resection combined with chemotherapy may still benefit carefully selected patients
Statement 42 (iCCA)	For iCCA, relative contraindications to upfront surgery include multifocal lesions (2–3 tumours) and regional nodal involvement, which indicate aggressive biology and warrant consideration of neoadjuvant therapy before resection. There is no consensus on cut-off values for CA19-9 or tumour size, although tumours larger than 5 cm are generally associated with poorer survival outcomes
(B) Patient selection for neoadjuvant trials in CCA
Statement 43	Global trials for peri-operative management of CCA must account for practice heterogeneity between Western and Eastern centres, addressing differences in resectability criteria, transplant considerations, and trial endpoints
Statement 44	The relative importance of strict patient selection criteria vs. the use of neoadjuvant therapy in optimizing outcomes remains a subject of ongoing debate
Statement 45	Neoadjuvant therapy for CCA serves as a test of tumour biology rather than focusing solely on response rates. Strategies include systemic chemotherapy, chemo-IO combinations, and targeted therapies guided by genetic profiling. LRTs, such as HAIC and SIRT, may be incorporated into neoadjuvant protocols in selected patients, but their role remains selective and investigational and requires further validation
Statement 46	A “sandwich” approach, combining neoadjuvant and adjuvant therapy, has proven effective in other malignancies and warrants consideration as a potential strategy for CCA
Statement 47	Liquid biopsy (ctDNA) and radiomics may complement traditional patient selection criteria, providing non-invasive insights into tumour biology and behaviour
Statement 48	Incorporating genomic alterations into selection criteria can refine therapeutic strategies for CCA. FGFR2 fusions and IDH1 mutations predict responsiveness to targeted therapies, although their independent prognostic significance in resectable disease is not well established, while HER2 amplifications and BRAF V600E mutations are under investigation for their therapeutic potential. Conversely, KRAS, TP53, and p15 mutations are linked to resistance and poorer outcomes. Comprehensive molecular profiling using targeted next-generation sequencing should be performed in CCA to identify actionable alterations and personalize treatment. Whole-genome sequencing may provide additional insight but is not currently mandatory and remains exploratory in routine practice
Statement 49	MSI-H status in CCA is associated with a favourable response to IO, particularly pembrolizumab, and should be assessed in treatment planning
Statement 50	Optimal endpoints for peri-operative trials should include progression-free survival, pCR, treatment response rates, and survival metrics specific to disease stage and biology
(C) Downstaging and transplant considerations in CCA
Statement 51	The benefit of LT for CCA is established if the expected 5-year OS exceeds 50%
Statement 52	No consensus exists on whether borderline-resectable cases would potentially benefit more from neoadjuvant therapy followed by transplantation rather than surgery to optimize outcomes
Statement 53	Most CCA transplant protocols require a sustained period of neoadjuvant therapy (commonly at least 6 months) with confirmed disease control prior to LT. However, this remains unstandardized, and both the duration and the response endpoints differ between pCCA and iCCA and between living donor and deceased-donor systems
Statement 54	In the context of LT, it remains unclear whether the focus should be on achieving a positive response vs. stable disease or on specific targets for objective response rates such as reductions in CA19-9 levels and changes in PET avidity as indicators of treatment efficacy. Further investigation is needed to define these benchmarks
Statement 55	Where ICIs are incorporated into transplant protocols, a period of demonstrated disease stability (commonly at least 6 months) together with an adequate washout is generally advised before proceeding with LT, recognizing that optimal thresholds remain undefined and protocol-dependent
Statement 56	Integrating genomic profiling into selection criteria can refine patient selection for LT, with mutations such as FGFR guiding-targeted therapy and predicting tumour behaviour, treatment response, and outcomes
Statement 57	Broader inclusion of CCA as a standard MELD exception should remain contingent upon further evidence demonstrating consistent survival benefit, cost-effectiveness, and outcomes comparable to existing transplant indications, to ensure responsible expansion of access within the constraints of donor organ availability
Statement 58 (pCCA)	Unresectable but transplantable cases of pCCA, such as those meeting Mayo or UNOS criteria, represent a subset where LT may achieve curative outcomes
Statement 59 (pCCA)	The role of radiation in pre-LT neoadjuvant therapy in pCCA remains debated, with concerns about its efficacy and the technical challenges it poses in LDLT, including the frequent need for arterial jump grafts
Statement 60 (pCCA)	Pre-transplant regional lymph node sampling, particularly of station 8 nodes along the common hepatic artery, is crucial for identifying nodal involvement in pCCA, a negative prognostic marker that contraindicates LT. Endoscopic ultrasound-guided sampling of regional lymph nodes is preferred. Direct transperitoneal or percutaneous biopsy of the primary perihilar tumour should be avoided because of the risk of tumour seeding, which can compromise transplant eligibility
Statement 61 (iCCA)	There is consensus that patients with early iCCA who have unresectable disease due to poor liver function or unfavourable tumour location should be considered for LT after stability on systemic treatment +/− LRT (see statements 53 and 55)
Statement 62 (iCCA)	LT for iCCA should be highly selective, with absolute contraindications including lymph node involvement (LN+) and portal vein invasion
Statement 63 (iCCA)	Portal lymph node metastasis and multifocal unilobar iCCA traditionally contraindicate transplantation. Emerging evidence suggests that with effective downstaging protocols, select patients may become eligible for transplantation
Statement 64 (iCCA)	iCCA tumours are classified into four immune subtypes: immunogenic, myeloid, mesenchymal, and immune desert. Their influence on LT suitability and prognosis requires further investigation

MTB, multidisciplinary tumour board; HAIC, hepatic artery infusion chemotherapy.

Established biomarkers such as CA19-9 provide useful prognostic information, while emerging molecular alterations, including FGFR2 fusions, IDH1 mutations, and oncogenic drivers such as KRAS or TP53, are increasingly recognized for their ability to characterize tumour biology and identify potential therapeutic targets. However, these alterations require further validation before they can be reliably applied as predictors of recurrence risk or routinely incorporated into decision-making for neoadjuvant strategies [144–146]. At present, their greatest relevance lies in characterizing tumour biology and identifying potential therapeutic targets that may inform investigational targeted or biomarker-enriched systemic strategies, including in the neoadjuvant or peri-operative setting. However, their role in routine preoperative risk stratification and treatment selection remains evolving and requires prospective validation. Radiomics, together with liquid biopsy platforms such as ctDNA, may further refine detection of occult dissemination and adverse tumour biology beyond what can be identified using conventional preoperative imaging [147–149]. This is important because established pathological features associated with poor prognosis, including perineural invasion, remain difficult to predict reliably before surgery [150, 151]. Collectively, these tools support earlier identification of patients unlikely to derive durable benefit from surgery alone.

Across regions, increasing emphasis is placed on clearly distinguishing neoadjuvant therapy, directed at reducing recurrence risk in upfront resectable disease, from downstaging strategies, which aim to convert borderline-resectable or technically challenging tumours into candidates for R0 resection. Explicit documentation of treatment intent within multidisciplinary deliberations is critical to aligning therapeutic goals with outcome assessment and trial interpretation.

Regional Perspectives (Divergence)

Despite shared biologic principles, East-West differences persist in philosophy and sequencing of therapy.

Perihilar CCA (pCCA). Eastern centres generally favour a surgery-first approach, prioritizing R0 resection whenever technically feasible. This reflects high operative expertise as well as pragmatic concerns that preoperative therapy may increase the risk of biliary sepsis or disease progression that could ultimately preclude curative surgery, particularly in the setting of complex biliary drainage. In contrast, Western programmes more commonly emphasize a biology-informed approach, using preoperative systemic or LRT to refine patient selection and reduce the likelihood of futile resection, while recognizing that robust evidence for routine neoadjuvant therapy in resectable disease remains limited.

Absolute contraindications to resection, including metastatic disease, inadequate hepatic reserve, or vascular/biliary extension precluding an R0 margin, are broadly agreed upon. For example, bilateral biliary extension beyond the limits of a trisectionectomy renders surgery technically unfeasible and oncologically futile [152, 153]. Relative contraindications such as poorly localized PSC-associated tumours [154, 155], nodal disease, and high CA19-9 post-drainage (e.g., >300) [156] signal adverse biology, though specific thresholds vary across regions. Downstaging may be attempted in borderline cases such as those requiring vascular reconstruction but remains limited by high rates of infective complications, especially in the presence of biliary obstruction.

iCCA. Widely recognized absolute contraindications to resection include distant metastases and inadequate liver reserve or performance status. Extensive multifocal disease and regional nodal involvement are strong adverse prognostic factors but are not universally regarded as absolute contraindications [151]. In contrast, relative contraindications such as multifocal (two to three) tumours and regional nodal disease vary in their perceived operability. Although no definitive thresholds exist for CA19-9 levels or tumour size, lesions exceeding 5 cm are consistently associated with inferior survival outcomes [157–160].

Contemporary neoadjuvant trials, including neo-GAP [161] and ZSAB-neoGOLP [25], define high-risk resectable iCCA using largely overlapping anatomic criteria, including large tumour size, multifocality, vascular invasion, and suspected nodal disease, reflecting increasing convergence around clinically meaningful risk stratification. Against this backdrop, Western centres more frequently consider systemic or LRT prior to surgery in selected multifocal or node-positive disease to refine patient selection, whereas Eastern programmes more often proceed directly to resection when an R0 margin appears achievable and hepatic function is preserved.

Pathways towards Integration (Convergence)

Moving forward, convergence between East and West approaches will depend on harmonizing biologically informed selection criteria and peri-operative strategies for high-risk CCA.

• 1.Standardized definitions and trial design.

There remains a critical need for consensus definitions of “high-risk” and “borderline-resectable” CCA that are applicable across regions. Trial eligibility and endpoint frameworks should integrate established clinicopathologic factors, including tumour burden, nodal status, vascular and perineural invasion, alongside emerging tools such as ctDNA kinetics, radiomic features, and immune profiling. These criteria must be aligned with standardized pathologic response metrics (pCR, MPR) and clinically meaningful survival benchmarks (e.g., 3-year RFS and OS), enabling consistent interpretation across peri-operative and neoadjuvant studies.

• 2.Biomarker-driven personalization.

Serologic markers (CA19-9 kinetics), liquid biopsy approaches (ctDNA-based minimal residual disease detection), and radiomic signatures hold increasing promise for dynamic risk stratification and treatment adaptation. Molecular alterations, including KRAS, TP53, and CDKN2A, have been associated with adverse tumour biology and inferior outcomes in resected CCA, as highlighted in large molecular-pathologic analyses [162], though their independent prognostic value and role in guiding peri-operative decision-making require further prospective validation. In contrast, alterations such as FGFR2 fusions and IDH1 mutations currently appear more relevant for identifying therapeutic vulnerability than for definitive recurrence risk stratification. Collectively, these tools may help identify poor-biology tumours unlikely to benefit from upfront surgery and support earlier consideration of neoadjuvant or transplant-based pathways rather than futile resection [163, 164].

• 3.Collaborative infrastructure.

Progress will depend on strengthening multinational collaborative platforms that integrate genomic sequencing, advanced imaging analytics, and prospective clinical trial networks. Expanding access to molecular and biomarker-driven research infrastructure in Asia, while leveraging established Western precision-oncology systems, will be essential to validate unified, biologically grounded peri-operative strategies and ensure their applicability across diverse healthcare settings.

Patient Selection for Neoadjuvant Trials in CCA (Statements 43-50)

Shared Principles

Across regions, neoadjuvant therapy is increasingly viewed as both a therapeutic strategy and a biologic stress test, revealing tumour chemosensitivity, systemic control, and suitability for definitive surgery or transplantation. A central debate concerns whether improved outcomes are driven primarily by the therapeutic effect of neoadjuvant treatment or by the stringent biological selection embedded within these protocols. This tension directly influences how thresholds for neoadjuvant use are defined, whether it should be reserved for more locally advanced disease or expanded to include earlier stage but biologically high-risk tumours in a more inclusive, real-world framework. Restrictive criteria improve trial homogeneity and internal validity, whereas broader inclusion better captures clinical heterogeneity and enhances external generalizability. Future studies must balance these approaches to produce evidence that is both biologically rigorous and widely applicable.

Systemic gemcitabine-cisplatin over 4–8 cycles across 3–6 months forms the backbone of most protocols, increasingly combined with chemo-IO (durvalumab, pembrolizumab) [165, 166] and locoregional adjuncts such as hepatic artery infusion chemotherapy and selective internal radiation therapy (whose role in neoadjuvant protocols remains selective and investigational), which deliver high-dose therapy directly to the tumour to enhance cytoreduction while minimizing systemic toxicity [165, 167]. Radiologic response assessment using RECIST 1.1, supplemented by mRECIST where LRT is employed, together with biochemical markers (e.g., CA19-9 kinetics) and emerging liquid biopsy measures such as ctDNA clearance, is increasingly used to guide treatment sequencing and assess surgical readiness [163, 164].

The NEO-GAP study evaluating neoadjuvant gemcitabine-cisplatin-nab-paclitaxel demonstrated feasibility and preserved peri-operative safety, encouraging survival outcomes, with a median OS of 24 months in the intention-to-treat population and not reached among patients who proceeded to resection at approximately 16–17 months of follow-up. These findings support the integration of intensified multi-agent neoadjuvant chemotherapy as a biologically selective strategy in patients with high-risk iCCA [161]. Building on the success of chemo-IO in advanced biliary tract cancers, including the phase III TOPAZ-1 [165] and KEYNOTE-966 trials [166], IO is now being explored in an earlier stage. The ZSAB-neoGOLP trial represents one such effort, investigating a neoadjuvant chemo-IO approach in high-risk iCCA. Interim read-outs have demonstrated acceptable tolerability, preservation of surgical feasibility, and encouraging signals of tumour response, supporting the biological rationale for immune priming prior to resection [25]. Final survival and recurrence outcomes from this study are awaited.

Building on these neoadjuvant feasibility signals, peri-operative “sandwich” strategies that integrate neoadjuvant and adjuvant therapy are under active investigation, for example, the DEBATE trial (NCT04624344) [168]. Together, these developments highlight the growing potential of multimodal, biology-guided neoadjuvant strategies to improve resectability, disease control, and long-term outcomes in CCA, while underscoring the need for prospective validation of their oncologic benefit.

Advances in radiomics [147, 149, 169, 170], ctDNA analysis [171, 172], and multi-omics profiling enable more dynamic assessment of tumour biology in CCA, complementing conventional anatomic staging and improving patient stratification. A broad spectrum of genomic alterations, including FGFR2 fusions, IDH1 mutations, KRAS and TP53 alterations, HER2 amplification, BRAF V600E mutations, and microsatellite instability-high (MSI-H) status, has been shown to confer therapeutic sensitivity or resistance in the advanced disease setting, where most evidence for targeted therapies currently resides.

Currently, the clinical utility of these molecular alterations has been established primarily in later lines of therapy, and their integration into neoadjuvant or peri-operative strategies remains challenging. This is exemplified by the difficulty in advancing FGFR inhibition into the first-line setting, with trials such as FIGHT-302 failing to complete accrual [173], underscoring the biological heterogeneity of CCA and the complexity of molecularly driven treatment intensification earlier in the disease course. As such, while molecular profiling provides critical biological insight, its role in guiding peri-operative decision-making remains investigational.

FGFR2 fusions and IDH1 mutations are predictive of response to targeted agents, including FGFR inhibitors such as pemigatinib [174] and IDH1 inhibitors such as ivosidenib [175], but their independent prognostic significance in resectable disease is not well established. Conversely, alterations involving KRAS, TP53, and cell cycle regulators such as CDKN2A/p15 may be associated with more aggressive tumour behaviour and therapeutic resistance, although these associations remain inconsistent and require further validation across disease stages [162].

Additional genomic alterations, particularly HER2 amplification and BRAF V600E mutations, have demonstrated consistently high response rates to matched targeted therapies in advanced CCA, exceeding those historically observed with conventional chemotherapy. These robust activity signals render them especially attractive candidates for investigation in biomarker-enriched neoadjuvant or peri-operative strategies, although prospective validation in these settings is required [176–179]. MSI-H status likewise represents a clinically important biomarker in CCA, reliably predicting responsiveness to immune checkpoint inhibition and offering a critical therapeutic avenue for patients with advanced or biologically aggressive disease where standard options are limited [180]. Beyond these actionable alterations, broader genomic platforms such as whole-genome sequencing may further refine patient selection by providing an integrated view of tumour evolution, clonal architecture, and additional vulnerabilities, although their routine clinical application remains exploratory.

Parallel advances in immune molecular characterization have further refined selection. Immune subtypes of iCCA, immunogenic, myeloid, mesenchymal, and immune desert, demonstrate distinct patterns of microenvironmental activity, treatment responsiveness, and prognosis [181]. Integrating immune profiling with genomic alterations may enable a more precise appraisal of biological fitness for resection and, in selected contexts, transplantation.

Regional Perspectives (Divergence)

Western centres increasingly emphasize tumour biology in surgical decision-making, particularly in selected programmes with access to clinical trials and translational infrastructure. In these settings, neoadjuvant systemic or chemo-IO may be considered to refine candidate selection and reduce futile resections. However, this approach is not uniformly adopted, and many Western hepatobiliary units continue to prioritize upfront surgery in resectable disease in the absence of compelling evidence. Overall, neoadjuvant strategies in the West remain heterogeneous and largely investigational, reflecting ongoing uncertainty regarding patient selection, optimal regimens, and long-term benefit.

Eastern centres, by contrast, more commonly favour a surgery-first strategy, prioritizing R0 resection whenever technically achievable, even when this entails complex or radical surgery. This approach reflects both the high technical expertise of Asian hepatobiliary units and pragmatic concerns that preoperative therapy may precipitate biliary sepsis, cholangitis, or disease progression that ultimately precludes curative resection. In transplant-eligible settings, the predominance of LDLT introduces additional ethical considerations, requiring stringent survival expectations and careful donor-recipient matching when neoadjuvant or downstaging-to-transplant pathways are contemplated [182, 183]. Limited deceased-donor availability further necessitates transparent survival benchmarks (≥50% 5-year OS) to justify oncologic indications while preserving equity relative to non-cancer recipients.

Differences in biomarker access and integration further contribute to regional variation. While some Western centres have broader access to multi-omics platforms and ctDNA-based monitoring, many programmes, both Western and Eastern, continue to rely primarily on clinicopathologic and serologic markers. These disparities influence trial design, risk stratification, and the pace at which biologically driven strategies are incorporated into routine practice, underscoring that gaps in evidence and implementation remain on both sides.

Pathways towards Integration (Convergence)

Bridging Eastern pragmatism and Western precision requires harmonization across trial design, biomarker frameworks, and ethical governance:

• 1.Unified eligibility and endpoints.

Global harmonization of trial design requires consensus definitions for “high-risk,” “borderline-resectable,” and “transplant-eligible” CCA. Multinational protocols should incorporate unified biologic criteria, including FGFR2 and IDH1 mutations, KRAS/TP53 drivers, ctDNA dynamics, CA19-9 kinetics, and radiomic features, into eligibility thresholds and endpoint definitions [163, 164]. Standardized reporting of pCR, MPR, PFS, and 3-year RFS/OS will enable valid cross-regional comparison of endpoints.

• 2.Biomarker-driven stratification.

Integration of multi-omics, immune profiling, CA19-9 kinetics [184], and ctDNA-MRD [171, 172] should guide real-time therapeutic adaptation, allowing early identification of non-responders and timely redirection to alternative or transplant-based therapies. Such biologically anchored stratification transforms CCA management from a traditionally anatomy-driven paradigm into one centred on tumour biology and personalized therapy.

• 3.Collaborative infrastructure and ethical alignment.

Expanding genomic and clinical trial capacity in Asia, supported by Western precision-oncology infrastructure, will facilitate biologically coherent, ethically robust implementation of neoadjuvant and transplant-inclusive protocols.

• 4.Innovative trial models.

Embedding translational endpoints such as ctDNA clearance, MSI status, radiomic progression signatures, and immune microenvironment profiling into peri-operative trials that include multimodal therapies (e.g., NEO-GAP, DEBATE) will accelerate biologic validation and global uptake.

Downstaging Strategies in CCA (Statements 51-64)

Shared Principles

Across regions, downstaging strategies in CCA are increasingly recognized as tools not only for cytoreduction, but for biological selection. The overarching aim is to identify tumours with sufficiently favourable behaviour to justify curative-intent intervention, either surgical resection or, in highly selected contexts, LT. Durable disease control, rather than radiographic response alone, has emerged as the critical determinant of benefit. A clear distinction must therefore be made between downstaging to resection and downstaging to transplantation, as these strategies differ fundamentally in intent, selection thresholds, and acceptable risk.

Downstaging to Surgical Resection. Downstaging to resection targets patients with initially unresectable CCA due to tumour burden, vascular involvement, or unfavourable anatomic features, with the goal of achieving an R0 resection following demonstrable biological control. In this setting, systemic therapy forms the backbone of treatment, often supplemented by locoregional approaches. The ZSAB-TransGOLP study represents a contemporary example of this approach [23]. By integrating intensive systemic therapy with response-adapted decision-making, the study demonstrated that a subset of patients with locally advanced CCA could be successfully downstaged to resection with acceptable peri-operative outcomes and encouraging survival. Importantly, progression to surgery was guided by sustained disease stability, rather than short-term radiologic response, reinforcing the principle that downstaging functions as a biological filter rather than a purely cytoreductive manoeuvre.

At present, these findings remain relatively isolated. Beyond ZSAB-TransGOLP, evidence for downstaging to resection in CCA is limited to small retrospective series and heterogeneous conversion experiences derived from multi-agent chemotherapy, occasionally combined with IO or LRT. These reports lack standardized selection criteria, treatment duration, and response thresholds, making outcomes difficult to compare and precluding consensus regarding optimal regimens or benchmarks for proceeding to surgery. Accordingly, downstaging-to-resection strategies should currently be pursued within protocolized frameworks and high-volume centres, where careful multidisciplinary oversight can mitigate the risk of futile intervention.

Downstaging and Transplantation Protocols. In contrast, downstaging to LT is reserved for a much narrower population, reflecting the ethical imperative to align oncologic indications with graft scarcity. A unifying principle is that LT should be considered only when expected 5-year OS exceeds approximately 50%, with RFS comparable to existing UNOS standards [185–187], aligning oncologic indications with established transplant benchmarks and ensuring ethical allocation of scarce grafts [188]. Most protocols require ≥6 months of neoadjuvant therapy and confirmed disease stability to ensure biologically favourable tumour behaviour prior to LT.

For pCCA, protocols such as the Mayo Clinic and UNOS pathways [189, 190] have demonstrated that neoadjuvant therapy followed by LT can achieve 3-year survival of 82% after transplantation, compared with 48% after resection [163], with intention-to-treat survival of 68% and 53% at 2 and 5 years across 12 US centres [188]. These protocols mandate prolonged neoadjuvant therapy (typically ≥6 months), negative lymph node staging, controlled CA19-9 kinetics, and documented disease stability as surrogates of favourable tumour biology.

For iCCA, transplantation remains highly selective and largely investigational. Current indications are generally limited to anatomically unresectable early-stage tumours, cirrhotic patients, or cases with marginal liver function, provided there is sustained biological control on systemic or LRT and favourable biological features [164, 191, 192]. Lymph node metastasis and major vascular invasion remain absolute contraindications. Although emerging data suggest that effective downstaging may expand eligibility in select cases, including patients with portal lymph node involvement or multifocal unilobar disease, recurrence risk remains substantial, underscoring the need for stringent selection and clearly defined benchmarks [151, 182–184].

Biologic selection for LT increasingly incorporates CA19-9 kinetics, PET avidity [193–195], and emerging biomarkers such as ctDNA dynamics, although none are yet validated as standalone decision tools [175–179]. Molecular profiling (e.g., FGFR2, IDH1, KRAS, TP53, p15) may further refine risk stratification, but its role remains adjunctive rather than determinative [1, 4, 167, 176, 177, 196]. Collectively, the use of biomarkers to guide transplant eligibility and reliably predict post-transplant recurrence remains an important area of unmet need.

Existing transplant protocols illustrate this cautious approach. The Mayo protocol differentiates PSC-associated from de novo tumours, while UNOS criteria require negative lymph nodes, tumour size <3 cm, acceptable CA19-9 levels, and confirmatory diagnostics such as FISH and brush cytology. Expansion of these criteria warrants further investigation to improve access while maintaining acceptable outcomes. OPTN guidance recommends that carefully selected candidates with biopsy-proven, unresectable iCCA or mixed HCC-iCCA ≤3 cm in the setting of cirrhosis, and with documented disease stability for at least 6 months from diagnosis or last treatment, be considered for a MELD exception score of MMaT-3 [197]. Such requests are non-standardized and are adjudicated case by case by the Adult Transplant Oncology Review Board, the broadened and renamed successor to the Adult HCC Review Board within the National Liver Review Board. Because this guidance is explicitly not OPTN Policy, the pathway is distinct from the standardized MELD exception established in OPTN Policy 9.5.A for pCCA, applies highly selective criteria, and should not be presented as a universal standard. Broader inclusion of CCA as a standard MELD exception for both early and advanced diseases will require robust evidence demonstrating survival benefit and cost-effectiveness compared to continued systemic therapy. Additionally, comparative outcome data from existing transplant indications are essential to justify expanding transplant eligibility and listing CCA patients on the DDLT waitlist. Studies such as the Houston Methodist series highlight acceptable OS rates for unresectable iCCA following LT but emphasize high recurrence rates, underscoring the need for stringent selection criteria and well-defined benchmarks for patient inclusion [198].

Areas of Ongoing Uncertainty. No consensus exists regarding the optimal management of borderline-resectable CCA, including whether patients are best served by upfront surgery, neoadjuvant therapy followed by resection, or escalation to transplant-based pathways [163, 199]. Similarly, in upfront unresectable disease, the choice between downstaging with intent for resection versus LT remains unresolved and likely dependent on biological response and institutional context.

The role of radiation therapy in pre-transplant neoadjuvant protocols remains an area of active debate, both regarding its necessity and its potential to complicate subsequent LT, as well as with respect to radiation modality, dose, and target volume. Conventional fractionated external beam radiotherapy, stereotactic body radiotherapy, and particle therapies may enhance local control and contribute to biological selection, but all carry risks of vascular and biliary injury that can complicate transplantation, particularly in LDLT, often necessitating complex vascular reconstruction and increasing procedural risk [200, 201]. In the absence of comparative, transplant-specific outcome data, radiation modality selection remains institution-dependent. While RT-containing neoadjuvant protocols have been associated with favourable outcomes in rigorously selected cohorts and observational data suggest reduced recurrence at the cost of increased vascular complications, there is currently no definitive randomized evidence demonstrating that radiation is necessary or superior to non-radiation neoadjuvant approaches prior to transplantation.

Accurate staging also remains central to treatment allocation. Pre-transplant lymph node sampling, particularly of station 8 nodes along the common hepatic artery, is essential, as nodal involvement predicts poor post-transplant outcomes. Minimally invasive techniques such as endoscopic ultrasound-guided fine-needle aspiration improve staging accuracy and are increasingly incorporated into clinical practice [202, 203]. Direct transperitoneal or percutaneous biopsy of the primary perihilar tumour should be avoided because of the risk of tumour seeding, which can compromise transplant eligibility. However, the optimal extent and standardization of nodal staging within pre-transplant algorithms remain uncertain.

Regional Perspectives (Divergence)

Despite broad agreement on overarching principles, meaningful variation persists both between and within Eastern and Western centres, reflecting differences in donor availability, regulatory frameworks, and institutional philosophy. Many Western programmes operate within DDLT-dominant systems and place emphasis on disease biology and durability of control in transplant selection. However, this is most commonly implemented through structured clinical and radiologic criteria, such as mandated periods of disease stability, histologic confirmation, and predefined listing pathways, rather than routine use of advanced molecular profiling or ctDNA monitoring, which remain largely investigational. Moreover, MELD exception-based pathways are not uniformly applied across Western jurisdictions and represent only one mechanism for operationalizing biological selection. For example, selected oncologic indications, including unresectable small iCCA in cirrhotic patients, may be considered under formal MELD exception frameworks requiring strict tumour size limits (≤3 cm), biopsy confirmation, and a minimum period of documented disease stability (typically ≥6 months), illustrating how tumour behaviour is translated into allocation policy within specific regulatory contexts rather than reflecting a universal Western approach.

Eastern programmes, in contrast, function primarily within LDLT-dominant systems, where the principle of double equipoise necessitates particularly stringent survival expectations for oncologic indications [182, 183]. As a result, many Asian centres continue to favour a surgery-first strategy, reserving transplantation for rigorously vetted cases in which the anticipated survival benefit clearly justifies donor risk. The relative scarcity of deceased-donor grafts further amplifies the need for transparent survival benchmarks and careful consideration of tumour burden, biological behaviour, treatment response, and waiting time, ensuring that expanded oncologic indications do not disadvantage non-cancer recipients. Given the intrinsically higher recurrence risk associated with malignancy, transplant candidacy is typically limited to scenarios where transplantation is expected to confer a predictable and meaningful advantage over the best available non-transplant alternatives.

Differences in access to, and readiness for, biomarker integration also contribute to regional variation, though this remains heterogeneous on both sides. While selected centres, both Western and Asian, are actively evaluating genomic profiling, radiomics, and ctDNA dynamics within research or trial settings, most transplant programmes worldwide continue to rely primarily on clinicopathologic, serologic, and radiologic parameters for routine decision-making. These structural, ethical, and resource-driven differences underscore the need for regionally contextualized yet biologically informed frameworks that balance donor availability, equity, and oncologic benefit, while progressively moving towards shared standards as evidence matures in CCA transplantation.

Pathways towards Integration (Convergence)

Harmonization of global LT strategies for CCA will require the following:

• 1.Consensus on biologically grounded eligibility.

Standardized definitions for LT-eligible pCCA and early iCCA should increasingly incorporate molecular alterations (FGFR2, IDH1, KRAS, TP53), ctDNA kinetics, PET avidity [193–195], and CA19-9 dynamics [1, 4, 167, 176, 177, 196]. Comparable biologic frameworks are needed across DDLT and LDLT systems.

• 2.Clarification of pre-LT downstaging endpoints.

Uniform definitions of treatment success, radiologic response, CA19-9 improvement, ctDNA clearance, or durable stable disease, should guide timing of LT and define minimal acceptable tumour control prior to transplantation. Clear, standardized criteria are essential to determine when radiographic regression, biochemical improvement, or disease stability alone is sufficient to justify transplantation.

• 3.Ethical and logistical adaptation.

DDLT-based Western protocols (e.g., Mayo, UNOS) must be adapted with sensitivity to LDLT contexts in Asia, incorporating donor safety, regional organ availability, and culturally informed ethical thresholds. Clear survival expectations, transparent prioritization schemes, and integration of advanced staging tools such as EUS-FNA for lymph node evaluation will be essential for equitable and safe expansion of transplant indications.

• 4.Infrastructure and data sharing.

Expanding genomic, ctDNA, and radiomic capacity in Asia, supported by Western precision-oncology systems, will enable global validation of biologically driven pathways. Unified registries and cross-regional datasets will define benchmarks for survival, recurrence, and donor risk.

Through global collaboration, transplant programmes can establish coherent, biologically rigorous, and ethically balanced standards for downstaging and transplantation in CCA, transforming a highly heterogeneous landscape into a unified, precision-driven paradigm.

Management of CRLM: East-West Perspectives and Integration

Management of Resectable and Transplantable CRLMs (Statements 65-73)

Shared Principles

A unifying global principle is that technical resectability alone does not determine oncologic success in CRLM [204, 205] (Fig. 4; Table 3). Long-term outcomes depend far more on tumour biology than operative complexity. Landmark analyses have shown that curative-intent hepatic resection yields 10-year OS of ∼20–25%, with a “cure plateau” near 20% [206, 207], a benchmark that defines surgical cure. In contrast, systemic therapy alone achieves a median OS of approximately 32 months in contemporary cohorts, with outcomes substantially better in left-sided, RAS wild-type disease [208, 209]. These distinctions have driven a shift towards biological resectability, guided by growth kinetics, systemic therapy response, metabolic tumour volume (MTV), RAS/BRAF status, and the presence or absence of extrahepatic disease [205, 207, 210].

Fig. 4.

Key areas of convergence and divergence of transplant oncology in the East and West for CRLM (Created in BioRender. Kow, A. (2026) https://BioRender.com/3olk8aa). FLR, future liver remnant; LRT, locoregional therapy; CRLM, colorectal liver metastasis; DFS/OS, disease-free survival/overall survival; LT, liver transplantation; TACE, transarterial chemoembolization; SBRT, stereotactic body radiation therapy; HAIC, hepatic artery infusion chemotherapy; RFA/MWA, radiofrequency ablation/microwave ablation; LDLT, living donor liver transplantation; MTV, metabolic tumour volume; ctDNA, circulating tumour DNA.

Table 3.

Expert statements on CRLMs: Eastern and Western perspectives

CRLM
(A) Management of resectable and transplantable CRLMs
Statement 65	Technical resectability does not equate to oncological success in CRLM. Biological resectability must also be considered. After curative-intent hepatic resection, landmark data indicate an actual 10-year OS of approximately 20–25% (with a cure plateau near 20%), compared with a median OS of approximately 32 months in contemporary cohorts, with outcomes substantially better in left-sided, RAS wild-type disease
Statement 66	LT may be considered in select CRLM cases with favourable tumour biology, demonstrated by durable systemic disease control, including patients with more than 3 lesions, high tumour burden, or upfront unresectable disease showing a strong radiological response to therapy or prolonged stable disease (>1 year)
Statement 67	Absolute contraindications for LT in CRLM include poor prognostic factors such as progressive disease while on systemic therapy, signet-ring histology, undifferentiated tumours, BRAF V600E mutations, and RAS/TP53 co-mutations. A right-sided primary tumour is a strong adverse prognostic factor but is generally regarded as a relative rather than an absolute contraindication. MSI-H/dMMR disease is generally not an indication for LT, as durable control is usually achievable with ICI, and is therefore considered a contraindication except in the rare, genuinely IO-refractory case, which may be considered only within investigational, experimental pathways (see statement 70)
Statement 68	Relative contraindications for LT in CRLM include tumour size >5.5 cm, MTV >70 cm3 during a chemotherapy break, tumour burden score >9, >9 liver lesions, Oslo or Fong score >2, poor treatment response evidenced by RECIST 1.1 <10% change, and N2 nodal involvement of primary tumour
Statement 69	Young patients with bilobar CRLM and no extrahepatic disease should be referred early in their treatment course by medical oncologists to surgeons for evaluation of their suitability for LT
Statement 70	The role of LT in MSI-H tumours unresponsive to IO warrants further investigation. MSI-H tumours that do not respond to pembrolizumab, including immunologically “cold” tumours, may represent a subset that could benefit from LT. This should be regarded as experimental and pursued only within carefully defined, investigational protocols, as evidence supporting transplant benefit in this subgroup remains extremely limited
Statement 71	Portal lymphadenectomy has a role in staging CRLM and can provide valuable information for treatment planning. However, it should be selectively performed pre-transplant to avoid unnecessary procedures that may not impact transplant eligibility or outcomes
Statement 72	HAI therapy is a valuable treatment modality that can be implemented during colon surgery for CRLM. However, HAI can complicate LT, particularly in LDLT; thus, careful planning is required to balance the benefits of HAI against its potential impact on LT feasibility
Statement 73	The impact of ABO-incompatible (ABO-I) LT on RFS in CRLM is unclear. While ABO-I has not been shown to affect RFS in HCC, its role in CRLM requires further investigation, particularly regarding the risk of biliary complications
(B) Transplant vs. resection in borderline-resectable CRLM
Statement 74	In carefully selected patients with favourable tumour biology, LT may offer outcomes comparable or superior to resection, and a high tumour load that is technically resectable only through extensive or high-risk hepatectomy should not be regarded as an absolute exclusion from LT consideration. It should be recognized, however, that the strongest direct evidence (the SECA studies and the randomized TransMet trial) is derived from permanently unresectable, liver-only disease. Extrapolation of these principles to borderline-resectable or technically resectable CRLM is not yet supported by comparative data and should currently be pursued cautiously and, where possible, within clinical trials
Statement 75	There is no consensus on whether conversion to surgery offers better survival than LT for patients with the same malignancy grade. While some advocate for attempting conversion surgery before considering LT in unresectable cases, this approach remains debated
Statement 76	Extending LT indications for CRLM in Asia requires dedicated transplant oncologists in MDTs, robust cost-effectiveness studies, and an Asian registry for potentially transplantable patients. This registry could support MELD exception points and organ allocation policies
Statement 77	While MELD exception points are now granted for highly selected CRLM cases in some transplant programmes, broader inclusion of CRLM as a standard MELD exception will require robust evidence of survival benefit and cost-effectiveness compared to continued systemic therapy, as well as comparative data from existing transplant indications to justify listing such patients on the DDLT waitlist
(C) Integration of chemotherapy and IO in CRLM LT
Statement 78	Targeted therapies may be integrated into transplant protocols to optimize outcomes in selected patients. Anti-EGFR agents (e.g., cetuximab) are restricted to extended RAS (KRAS and NRAS) and BRAF wild-type tumours and are most beneficial in left-sided primaries, whereas bevacizumab is used across molecular subgroups irrespective of RAS status and is not specifically indicated by KRAS-mutant status
Statement 79	The optimal duration of chemotherapy before assessing suitability for LT is undefined, but current practices recommend at least 3–6 months to evaluate tumour biology and treatment response
Statement 80	Chemotherapy cessation prior to LT requires careful consideration. A chemo-free period of 4–6 weeks is recommended for assessment and optimal outcomes, though prolonged observation may be warranted in select cases
Statement 81	Post-transplantation chemotherapy is generally avoided in LT for CRLM, as the primary goal of LT is to provide a chemotherapy-free treatment pathway for patients
(D) Assessment of treatment efficacy in CRLM
Statement 82	Endpoints such as SAR and treatment response may better reflect the efficacy of CRLM treatment than traditional RFS, which does not reliably correlate with OS
Statement 83	ctDNA may have prognostic value in identifying minimal residual disease pre-transplant and may serve as a tool for dynamic monitoring of treatment response and recurrence risk

MDT, multidisciplinary team.

Within this framework, LT has emerged as a potential option for a highly selected subgroup of patients. Candidates for LT include patients with high hepatic tumour burden or upfront unresectable, liver-only disease who demonstrate favourable tumour biology through durable systemic disease control, typically reflected by a strong radiological response to therapy or prolonged stable disease for more than 1 year, together with absence of extrahepatic spread and favourable molecular features. This “test of time” ensures LT is reserved for those with controlled systemic disease who stand to benefit from transplantation [211].

Universal contraindications include progressive disease on optimal systemic therapy, signet-ring or undifferentiated histology, and high-risk molecular alterations such as BRAF-V600E and RAS/TP53 co-mutations, all of which consistently predict dismal post-transplant survival and are therefore excluded in contemporary series [212]. MSI-H/dMMR tumours, although generally associated with favourable prognosis due to heightened immunogenicity, respond poorly to standard chemotherapy regimens such as FOLFOX or XELOX but typically demonstrate strong and durable responses to immune checkpoint inhibition [213]. For this reason, MSI-H disease is not routinely considered a suitable indication for LT, as the survival benefit achievable with IO alone usually exceeds that of transplantation. LT may be considered only in the rare circumstance of genuine IO-refractory MSI-H disease, often characterized by an “immunologically cold” phenotype with poor T-cell infiltration, but even then, only within highly individualized or investigational pathways as evidence supporting transplant benefit in this subgroup remains extremely limited.

Relative contraindications, including right-sided primary tumours, tumour size >5.5 cm, MTV >70 cm³ (measured during a chemotherapy-free interval of 4–6 weeks and within 3 months of LT), >9 liver lesions, tumour burden score >9, Oslo or Fong score >2, poor treatment response (<10% change on RECIST 1.1), and N2 nodal involvement, serve as surrogate markers of aggressive biology and guide cautious selection [214–217]. Although these thresholds are not uniformly standardized across centres, they remain useful clinical anchors. Of these factors, MTV has emerged as one of the strongest prognosticators and should be routinely incorporated into the pre-transplant evaluation.

An essential shared principle is early referral. Young patients with bilobar CRLM without extrahepatic disease should be referred early for hepatobiliary/transplant evaluation to ensure adequate time for biological assessment, multidisciplinary planning, and sequencing of systemic therapy. Accurate staging is equally fundamental. Any regional lymph node involvement at the time of transplant evaluation constitutes a contraindication to LT, reflecting its strong association with systemic disease. Accordingly, targeted sampling of clinically relevant nodal basins, commonly including portal lymph nodes such as station 8 along the common hepatic artery, plays a critical role in confirming nodal negativity. A selective, risk-adapted approach is preferred, as routine lymphadenectomy increases morbidity without improving decision-making or transplant eligibility. Targeted sampling therefore optimizes staging accuracy while minimizing risk.

Regional Perspectives (Divergence)

Western programmes have played a foundational role in the development of transplant oncology for CRLM, supported by high-volume DDLT, established trial infrastructure, and an early emphasis on biologic selection. The Oslo experience demonstrated the feasibility of LT for unresectable CRLM using stringent criteria, including sustained response or stability to systemic therapy (typically ≥1 year), limited hepatic tumour burden, absence of extrahepatic disease, and favourable molecular characteristics. Outcomes from the SECA-I and SECA-II studies catalysed broader international interest in transplantation for CRLM and informed the development of structured frameworks, including MELD exception-based pathways in parts of North America and Europe [218, 219].

Contemporary Western practice increasingly incorporates biologic and functional markers, such as FDG-PET avidity, MTV thresholds [215], and molecular profiling (particularly RAS and BRAF status), to refine risk stratification and guide transplant candidacy. Broader molecular classifications, including consensus molecular subtypes and extended genomic profiling, are conceptually attractive and increasingly explored in research settings, but are not yet formalized components of LT eligibility criteria [220]. Nevertheless, these tools reflect a broader trajectory towards biologically grounded selection, with transplantation reserved for patients demonstrating durable systemic disease control and favourable tumour behaviour over time.

Eastern centres, by contrast, practise within LDLT-dominant systems where donor availability, ethics, and risk-benefit considerations differ substantially. The double equipoise principle mandates stricter survival expectations than in DDLT. Because cancer recipients face elevated recurrence risk, Eastern programmes often require exceptionally favourable tumour biology, longer observation periods, and tighter selection margins before proceeding with LT. Biomarker and genomic profiling access may be more variable, and selection often relies more heavily on clinical/serologic markers (CEA kinetics, radiologic response, absence of progression), although precision-oncology platforms are rapidly expanding.

Western centres, operating within DDLT-dominant systems, follow strict allocation frameworks and conservative governance of graft usage. LT for CRLM is often done under clinical trials setting in many centres in the West. Access to deceased-donor grafts reduces reliance on ABO-I LDLT, diminishing the need to broaden donor compatibility for investigational oncologic indications. Consequently, Western programmes tend to prioritize stringent biological selection, radiologic response, durable disease stability, and long-term survival benchmarks rather than expanding compatibility thresholds through ABO-I strategies.

ABO-I LT for CRLM should currently be regarded as investigational, appropriate only in highly experienced LDLT centres where desensitization, biliary intervention, and multidisciplinary oncologic care are optimized. Further research, including comparative analyses of graft outcomes, biliary morbidity, RFS, and OS, will be essential before ABO-I LT can be rationally integrated into broader CRLM transplant pathways [221]. Until such data exist, the East-West disparity in ABO-I use is likely to persist, reflecting not only differences in donor availability but also divergent ethical thresholds and tolerance for graft-related risk in oncologic transplantation.

Pathways towards Integration (Convergence)

Global integration requires harmonization of biologic selection criteria, molecular profiling, and minimum benchmarks for LT justification. The broad criteria for transplant-eligible CRLM should incorporate the following:

• Durable systemic control (≥6–12 months)

• Absence of extrahepatic disease

• Favourable molecular features (RAS-WT or limited RAS mutations, absence of BRAF-V600E, absence of RAS/TP53 co-mutations)

• MTV <70 cm³ and limited tumour burden

• ctDNA negativity or clearance pre-LT

• Radiologic response by RECIST 1.1.

A collaborative East-West framework should also establish minimum survival expectations to ethically justify LT in oncology, likely requiring 5-year OS ≥50%, comparable to other extended criteria indications. To reduce regional disparities, Asia-focused initiatives should expand access to next-generation sequencing, PET-based tumour volume assessment, and ctDNA platforms. Meanwhile, Western programmes must generate stronger prospective data on cost-effectiveness, durability of benefit, and long-term recurrence patterns which are essential for global adoption. Finally, international registries and harmonized MELD exception policies (with transplant oncology-specific review boards) are critical for generating high-quality evidence and ensuring ethical, equitable allocation.

Transplant versus Resection in Borderline-Resectable CRLM (Statements 74-77)

Shared Principles

Across global programmes, there is consensus that borderline resectability in CRLM represents a biologically heterogeneous space in which technical feasibility alone does not define optimal therapy. Patients with high tumour burden who are technically resectable may still have unfavourable long-term outcomes after major hepatectomy, whereas others with unresectable disease but indolent tumour biology may derive substantial benefit from LT. Thus, the decision between LT and resection increasingly hinges on tumour biology rather than anatomy, with identical biological phenotypes potentially achieving superior long-term survival after transplantation compared with extensive resection.

In situations where parenchymal-preserving hepatectomy is not feasible and complex liver resection would otherwise be required, such as ALPPS, two-stage hepatectomy with portal vein embolization, or major parenchymal sacrifice for bilobar disease, LT may offer a more durable oncologic solution in tumours with favourable biology [222]. Evidence from the Oslo series, subsequent SECA trials [218, 219], and the randomized TransMet trial [211] suggests that carefully selected patients with permanently unresectable, liver-only disease can achieve 5-year OS of 60–83% across the Oslo/SECA series [218, 219] and 56.6% by intention-to-treat (73.3% per protocol) in TransMet [211], which compares favourably with the outcomes of extended hepatectomy in similar biological contexts, although the available comparisons remain non-randomised and single-centre [222, 223]. However, a fundamentally unresolved question remains: does conversion surgery outperform LT in patients with the same biological grade of malignancy? No controlled studies currently compare survival between conversion hepatectomy and LT in biologically matched cohorts. This lack of comparative evidence fuels ongoing debate and underscores the need for standards that evaluate the relative benefit of each approach based on tumour biology rather than solely technical considerations.

Regional Perspectives (Divergence)

Western programmes have led most early evaluations of LT for borderline-resectable or borderline-unresectable CRLM, driven by the availability of DDLT, well-established allocation frameworks, and have been early adopters of biologic selection tools such as PET-derived MTV and comprehensive genomic profiling. In these DDLT-dominant systems, structured transplant pathways have emerged for highly selected CRLM candidates, reflected in evolving MELD exception mechanisms within UNOS/OPTN and several Canadian provincial review boards. These pathways typically require biopsy confirmation, durable radiologic stability, and a minimum listing MELD (e.g., MELD 15 with MMaT-20 assignment), thereby embedding tumour biology and treatment response directly into organ allocation policy. Despite these advances, significant challenges remain. Access to MELD exceptions is highly variable, donor organ scarcity persists, and graft competition with established indications limits broader adoption. For CRLM to be accepted as a standard oncologic LT indication, robust evidence must demonstrate clear survival benefit, cost-effectiveness, and outcome comparability with existing transplant indications. Comparative data against both non-transplant therapies and other accepted indications will be essential to justify waitlist prioritization in DDLT systems.

Eastern programmes, operating primarily within LDLT frameworks, adopt a more conservative stance. The double equipoise principle demands strong justification before exposing a healthy donor to risk, meaning that borderline-resectable CRLM is approached first with surgery if an R0 resection is technically feasible. LT is therefore reserved for highly vetted CRLM cases demonstrating exceptional disease control, preserved performance status, and durable stability on systemic therapy, with these stringent thresholds reflecting ethical considerations, heterogeneous access to precision-oncology tools, and the scarcity of deceased-donor grafts in many Asian systems, where expanding indications to compete for DDLT organs remains challenging.

Expanding LT indications for CRLM in Asia will require substantial structural support. Dedicated transplant oncologists and multidisciplinary team frameworks integrating transplant hepatology, oncology, and HPB surgery are essential to ensure appropriate biological and technical assessment of potential candidates. A major barrier is the absence of a unified Asian transplant oncology registry, which limits longitudinal data collection, MELD exception advocacy, and region-specific policy development. Establishing such a registry would enable systematic tracking of outcomes, support evidence-based allocation models, and facilitate robust cost-effectiveness analyses comparing LT with continued systemic therapy, analyses that are critical for justifying LT as a viable option for carefully selected CRLM patients.

Pathways towards Integration (Convergence)

Harmonizing global practice requires a unified biologic framework for borderline-resectable CRLM, with clear definitions of the following:

• What constitutes “biologically favourable” borderline disease

• When LT should supersede high-risk resections

• Minimum survival expectations required to ethically justify LT.

International consensus should prioritize biologically anchored metrics such as MTV, ctDNA status, RAS/BRAF/TP53 profiles, radiologic response, and sustained disease stability. To resolve the central uncertainty of conversion surgery versus LT, prospective comparative studies or registry-based matched analyses are urgently needed.

Regionally tailored policies are also essential. Western programmes must continue validating cost-effectiveness, durability of response, and risk-benefit profiles of LT. Asian programmes require expanded precision-oncology infrastructure, standardized LDLT-specific criteria, and dedicated registries to support MELD exception pathways in select cases. Ultimately, integrating Eastern technical expertise with Western biologic frameworks will enable a coherent, ethically grounded global strategy for managing borderline-resectable CRLM.

Integration of Chemotherapy and IO in CRLM LT (Statements 78–81)

Shared Principles

Neoadjuvant systemic therapy serves two core purposes in transplant oncology: cytoreduction and biological selection. For CRLM, systemic chemotherapy, typically oxaliplatin- or irinotecan-based doublets, remains the backbone of pre-LT management. A minimum duration of 3–6 months is widely adopted to allow assessment of tumour biology, therapeutic responsiveness, and systemic disease behaviour [211]. Tumour regression, response by RECIST 1.1, or prolonged stable disease during this interval offers valuable prognostic insights and helps identify candidates with sufficiently controlled biology for LT.

IO is not part of standard downstaging or LT selection strategies for CRLM. The vast majority of CRCs are MSS, and ICIs have not demonstrated meaningful clinical activity in this population [224]. As a result, IO is not incorporated into contemporary neoadjuvant, downstaging, or peri-transplant protocols for CRLM. Although MSI-H disease represents a biologically distinct subgroup with established sensitivity to IO, cases of genuine IO-refractory MSI-H CRC that might be considered for LT are exceedingly rare. Targeted therapies may complement chemotherapy in selected molecular subgroups. Anti-EGFR agents such as cetuximab are restricted to extended RAS (KRAS and NRAS) and BRAF wild-type tumours and derive their greatest benefit in left-sided primaries, whereas anti-angiogenic therapy with bevacizumab is used across molecular subgroups irrespective of RAS status rather than being specifically indicated by KRAS-mutant status. These agents require careful peri-operative planning: bevacizumab should be discontinued at least 6–8 weeks prior to transplantation to minimize risks of impaired wound healing and bleeding, whereas cetuximab can generally be stopped closer to the time of LT (approximately 4 weeks) without compromising peri-operative safety [225].

A chemotherapy-free interval of 4–6 weeks is generally recommended before LT to reduce treatment-related toxicities and allow reassessment of tumour biology, though this interval should be individualized based on regimen-specific risks. Agents such as 5-FU pose minimal hepatotoxicity and typically do not require prolonged cessation, while irinotecan and oxaliplatin may induce steatohepatitis or sinusoidal injury, necessitating longer observation and more cautious progression towards LT [226, 227]. In selected cases, particularly those with equivocal initial treatment response, extended observation may be warranted to confirm durable disease stability and ensure optimal patient selection. Prospective studies are needed to define agent-specific cessation timelines and harmonize practice across transplant centres.

Post-transplant systemic therapy is typically avoided. LT is intended to provide a chemotherapy- and targeted therapy-free curative pathway, and immunosuppression may limit the efficacy and safety of cytotoxic or targeted agents after graft implantation. This further underscores the need for optimal tumour control and careful biological selection before LT.

Regional Perspectives (Divergence)

Western programmes more routinely integrate systemic therapy, targeted agents, PET-based metabolic assessment, and immunologic biomarkers into transplant decision-making. Broad access to cetuximab, bevacizumab, and comprehensive molecular profiling, together with well-established clinical trial pathways, better facilitates precision sequencing of therapy. As a result, cessation intervals for bevacizumab and irinotecan-associated hepatotoxicity are better characterized, and PET-based MTV assessment is increasingly incorporated into biologic risk stratification.

In contrast, Eastern programmes face different practical constraints. Access to biologics (e.g., cetuximab, bevacizumab), broad genomic profiling, or PET-based MTV assessment is more variable, and the predominance of LDLT substantially influences how systemic therapy and LRT are deployed. Prolonged pre-LT chemotherapy may increase operative complexity in LDLT due to chemotherapy-associated parenchymal injury or vascular fragility, while anti-angiogenic therapy (e.g., bevacizumab) can heighten the risk of bleeding complications or anastomotic dehiscence for both donor and recipient.

Regional differences are most pronounced with hepatic arterial infusion (HAI) therapy. In Western DDLT settings, particularly in North America, HAI is a well-established modality that offers potent locoregional tumour control with limited systemic toxicity [228]. However, HAI can distort the hepatic artery and perihilar structures, complicating arterial reconstruction at transplant. While such challenges are manageable in DDLT pathways, they pose significant hazards in LDLT, where precise hilar dissection and arterial integrity are critical. Consequently, Eastern programmes use HAI far more selectively or avoid it altogether when future LT is contemplated. These vascular considerations reflect why integration of HAI into LDLT pathways requires substantial caution, as altered anatomy can increase thrombosis risk, compromise graft inflow, and impair post-transplant outcomes. Together, these structural and therapeutic differences shape how chemotherapy, targeted therapy, and locoregional treatments are sequenced before LT in CRLM across regions, underscoring the need for regionally adaptable but biologically consistent integration frameworks.

Pathways towards Integration (Convergence)

A harmonized global approach should define the following:

• The optimal duration and sequencing of chemotherapy before LT

• Regimen-specific cessation protocols based on hepatotoxicity, wound-healing risk, and vasculopathy

• The role of targeted therapy integration (KRAS-WT vs. KRAS-mutant paradigms)

• Approaches for chemotherapy-free intervals to ensure safe operative windows.

International platforms should generate prospective data evaluating therapy duration, predictive biomarkers of response, and the impact of specific chemotherapeutic agents on LT outcomes. Integration of IO remains investigational in CRLM LT because of risk of acute rejection, but MSI-H non-responders represent a biologically distinct cohort that may require tailored pathways.

Assessment of Treatment Efficacy in CRLM (Statements 82–83)

Shared Principles

As LT for CRLM gains acceptance, there is a growing recognition that traditional oncologic endpoints such as RFS inadequately reflect treatment benefit as they do not reliably correlate with OS. Studies from the Oslo group and the TRANSMET trial demonstrate that many LT recipients recur, often in the lungs, yet maintain excellent long-term survival with treatable, indolent recurrences [211, 218, 219]. Thus, survival after recurrence (SAR) and patterns of recurrence may be more meaningful metrics of LT efficacy and long-term patient outcomes than RFS alone. SAR reflects the unique biology of CRLM after transplantation, where recurrence does not necessarily compromise long-term survival and may be controllable with metastasectomy or stereotactic radiotherapy.

ctDNA has emerged as a promising dynamic biomarker of MRD and tumour biology in colorectal cancer, reflecting tumour burden and treatment responsiveness. ctDNA positivity after systemic therapy is associated with a higher risk of recurrence and poorer survival, whereas ctDNA clearance during treatment correlates with more favourable biology and improved oncologic outcomes in non-transplant surgical cohorts. These properties make ctDNA an attractive tool for assessing biological stability over time and for monitoring recurrence risk, with post-treatment ctDNA surveillance capable of detecting molecular relapse earlier than conventional imaging [229–231]. However, its role in transplant oncology remains investigational, as transplant-specific evidence demonstrating that pre-LT ctDNA status independently predicts post-transplant recurrence or survival is lacking. Therefore, ctDNA should currently be viewed as an adjunctive marker of response and disease control rather than a standalone determinant of transplant eligibility.

Regional Perspectives (Divergence)

Western programmes are increasingly incorporating ctDNA assays into CRLM research and increasingly into clinical decision-making. These tools assist in selecting transplant candidates, predicting outcomes, and refining surveillance strategies after LT.

Eastern programmes face variable access to high-sensitivity ctDNA platforms and may still rely primarily on CEA kinetics, imaging, and radiologic response for selection. As precision-oncology infrastructure expands, integration of ctDNA into pre-LT and post-LT protocols is expected to grow.

Pathways towards Integration (Convergence)

Global pathways should prioritize the following:

• Standardized use of ctDNA for pre-LT risk stratification and post-LT surveillance protocols

• Adoption of SAR, recurrence pattern profiling, and conditional survival as routine endpoints in CRLM transplant trials

• Cross-regional registries collecting ctDNA kinetics, recurrence patterns, and SAR outcomes

• Prospective validation of ctDNA as a surrogate endpoint for biological selection in LT.

By incorporating dynamic biomarkers and redefining meaningful endpoints, the field can more accurately evaluate treatment efficacy, optimize patient selection, and reshape expectations for survival in CRLM managed with transplantation.

Conclusion

Across hepatobiliary malignancies, it is now clear that technical success alone does not guarantee oncologic benefit. Major hepatectomy or even anatomically flawless resection may be futile in the setting of adverse tumour biology, leading to early recurrence and limited survival. Likewise, LT, which is arguably the most powerful tool in HPB oncology, cannot be justified on anatomical feasibility or operative expertise alone. Both resection and transplantation must be anchored in robust, biology-driven criteria that prioritize durable oncologic outcomes over what is surgically possible.

Rapid advances in biomarkers, radiomics, multi-omics profiling, and ctDNA have begun to reshape how “resectability,” “transplantability,” and surgical futility are conceptualized. Tumour behaviour, treatment response, and biological stability are increasingly recognized as central determinants of outcome. Yet, despite this conceptual convergence, the field still lacks harmonized definitions of high-risk disease, borderline resectability, transplant eligibility, and acceptable thresholds for curative-intent intervention. In the absence of shared benchmarks, practice remains fragmented and meaningful comparison across centres and regions remains challenging.

Importantly, this statement goes beyond acknowledging the existence of East-West differences and highlights the need to understand how distinct strategies translate into patient outcomes. Broadly, Eastern programmes have tended to favour interventional, response-driven pathways, leveraging aggressive resection, downstaging, and LDLT within ethically constrained donor frameworks, whereas Western programmes have more often emphasized biologic selection, prolonged disease stability, and structured allocation systems within DDLT-dominant environments. These approaches are frequently portrayed as divergent. However, whether they result in different long-term outcomes, or instead represent alternative routes to comparable benefit, remains largely untested.

This gap represents a critical opportunity. Comparative evaluation of these strategies, rather than their isolated description, could provide unprecedented insight into which elements truly drive survival, recurrence control, and cost-effectiveness. Addressing such questions now is especially timely, as rapid advances in systemic therapy may soon blur these distinctions, limiting the ability to disentangle the impact of strategy from that of treatment efficacy.

Equally intriguing is the possibility that these paradigms are not mutually exclusive. If outcomes are similar despite different philosophies, this would support the safe, context-dependent application of either strategy. Alternatively, sequential or combined approaches, such as initial surgical debulking followed by biology-guided systemic or immune modulation, may ultimately offer superior outcomes. Such a bidirectional treatment-planning framework, grounded in tumour biology and treatment response rather than rigid sequencing, could provide a unifying rationale across primary and metastatic liver malignancies.

The discussions catalysed at APTOS 2024 highlight both the urgency and opportunity to move beyond descriptive East-West comparisons towards outcome-focused integration in hepatobiliary oncology. By aligning Eastern procedural expertise with Western biology-driven frameworks, and by placing tumour biology, treatment response, and ethical stewardship at the centre of decision-making, the global community can define evidence-based standards that transcend regional boundaries. This statement provides a conceptual foundation for the next phase of the field, in which comparative analyses, shared registries, and harmonized trial platforms will be essential to determine not only how patients are treated, but which strategies most reliably deliver durable benefit. Through this convergence, resection and LT can expand their legitimate roles as curative modalities, applied rationally, ethically, and consistently, offering meaningful survival gains to patients whose diseases were once considered beyond effective intervention.

Conflict of Interest Statement

Gonzalo Sapisochin reports grant support from Bayer^® and Roche^®; consulting relationships with Novartis^®, Integra^®, Roche^®, AstraZeneca^®, Chiesi^®, Eisai^®, HepaRegeniX^®, and Stryker^®; and research collaborations with AstraZeneca^®, Roche^®, and Natera^®. Alfred Wei Chieh Kow reports consulting relationships for Roche^® and Stryker^®; and research collaborations with Roche^®. Stephen L Chan serves as an advisory member for AstraZeneca, MSD, Eisai, BMS, Ipsen, and Hengrui; received research funds from MSD, Eisai, Ipsen, Sirtex, and Zailab; and received honoraria from AstraZeneca, Eisai, Roche, Ipsen, and MSD. Arndt Vogel reports consultancy and advisory roles with Roche, AstraZeneca, Boehringer Ingelheim, Ipsen, Incyte, Cogent, Eisai, Zymeworks, Biologix, BMS, Terumo, Elevar, Servier, MSD, Taiho, Jazz Pharmaceuticals, Medivir, AbbVie, and Tyra. Sherrie Bhoori reports consulting and scientific honoraria from Roche, AstraZeneca, Terumo, and Boston Scientific. Teresa Macarulla Mercade reports consultant or advisory roles with Amgen, Servier, Incyte, Sanofi, AstraZeneca, Taiho, Celgene, and Eisai; research funding from Celgene, AstraZeneca, BeiGene, and Incyte; and speaking engagements for AstraZeneca, Incyte, Servier, Roche, and Eisai. Masatoshi Kudo, Etsuro Hatano, Stephen L Chan, Arndt Vogel, Tim Meyer, and Richard Finn were members of the journal’s Editorial Board at the time of submission. The remaining authors have no conflict of interest to declare.

Funding Sources

This work received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors. No funder or sponsor had any role in the conception, preparation, or writing of this manuscript, or in the decision to submit it for publication.

Author Contributions

Jia-Hao Law, Kunlin Xie, Hong Wu, Gonzalo Sapisochin, and Alfred Wei Chieh Kow were involved in conceptualization, methodology, investigation, project administration, resources, supervision, validation, and writing (original draft, review, and editing). Jiayin Yang, Masatoshi Kudo, Mark Dhinesh Muthiah, Neil Mehta, Sherrie Bhoori, David J. Pinato, Cheng Ean Chee, Etsuro Hatano, Pål-Dag Line, Stephen L. Chan, Majella Doyle, Teresa Macarulla Mercade, Adam C. Yopp, Arndt Vogel, Masayuki Ohtsuka, Tomoki Ebata, Mahmoud El-Meteini, Mohamed Fathy Abdelghafar, Grainne O’Kane, Ahmed El-Gendi, Tim Meyer, Suk Kyun Hong, Prashant Bhangui, Richard S. Finn, Albert Chi Yan Chan, Vincenzo Mazzaferro, Dong-Sik Kim, Jian Zhou, Jiahong Dong, and Taizo Hibi were involved in conceptualization, methodology, investigation, and writing (review and editing).

Funding Statement

This work received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors. No funder or sponsor had any role in the conception, preparation, or writing of this manuscript, or in the decision to submit it for publication.