The evolving landscape of colorectal liver metastasis management: from resection to transplantation – a narrative review Authors: Francesco Saverio Papadia, Stefano di Domenico, Matteo Santoliquido, Franco De Cian Published: 2026 ABSTRACT Colorectal cancer remains a leading cause of cancer-related mortality worldwide, with the liver representing the most common site of metastatic spread. The management of colorectal liver metastases (CRLM) has undergone a profound transformation over the past three decades, evolving from a surgical discipline focused solely on resectability to a complex, multimodal enterprise integrating advanced systemic therapies, refined surgical techniques, regional chemotherapy, local ablative procedures, and most recently, liver transplantation for highly selected patients. This narrative review synthesizes contemporary evidence from recent landmark trials and large cohort studies to provide a comprehensive overview of the current treatment landscape. We examine the evolution of surgical resection as the cornerstone of curative-intent therapy, the ongoing debate surrounding perioperative chemotherapy timing and regimens, and the emerging role of thermal ablation as a non-inferior alternative for small metastases. The revival of hepatic arterial infusion pump chemotherapy and the place of targeted therapies are contextualized within modern treatment algorithms. A central focus is the notable introduction of liver transplantation for non-resectable CRLM, with critical appraisal of selection criteria including metabolic tumor volume, the Oslo score, and tumor biology. We conclude by proposing a framework for personalized, biology-driven treatment selection that balances oncologic efficacy with patient quality of life and ethical considerations surrounding organ allocation. Background Colorectal cancer (CRC) is the third most commonly diagnosed malignancy and the second leading cause of cancer death globally, with approximately 1.9 million new cases and 903,000 deaths annually [1]. The liver is the predominant site of metastatic spread, occurring in 20–35% of patients at initial presentation and in a further 40–50% during the disease course [2, 3]. For patients with colorectal liver metastases (CRLM), the presence of hepatic involvement fundamentally alters prognosis and treatment trajectory. In the absence of effective therapy, median survival is measured in months; however, with modern multidisciplinary management, long-term survival and even cure are achievable for a growing subset of patients [4, 5]. The past three decades have witnessed a paradigm shift in the management of CRLM. What was once a binary decision, i.e. resectable versus unresectable, has evolved into a nuanced, personalized approach incorporating systemic chemotherapy, biologic agents, parenchymal-sparing surgical techniques, two-stage hepatectomies, local ablative therapies, regional chemotherapy via hepatic arterial infusion, and most provocatively, liver transplantation. This evolution reflects a deeper understanding of tumor biology, improved patient selection, and the recognition that effective management requires the coordinated expertise of surgeons, medical oncologists, radiologists, and pathologists. Several fundamental advances have driven this transformation. First, the development of more effective chemotherapy regimens, particularly the introduction of oxaliplatin and irinotecan, has improved response rates and enabled conversion of initially unresectable disease [6, 7]. Second, the integration of biologic agents targeting VEGF and EGFR has further extended survival in molecularly defined subgroups [8, 9]. Third, refinements in surgical technique, including parenchymal-sparing resections, associating liver partition and portal vein ligation for staged hepatectomy (ALPPS), and two-stage hepatectomy combined with portal vein embolization, have expanded the pool of patients eligible for curative-intent resection [10–12]. Fourth, the recognition that thermal ablation can achieve oncologic outcomes comparable to surgery for small metastases has added a valuable, less invasive option [13]. Fifth, and most recently, the demonstration that liver transplantation can achieve 5-year survival rates exceeding 70% in highly selected patients with non-resectable CRLM has challenged long-held dogmas and opened new therapeutic frontiers [14, 15]. The concept of oligometastatic disease, an intermediate state between localized and widespread malignancy, has become central to CRLM management. The European Society for Medical Oncology (ESMO) consensus defines oligometastatic disease as up to 5 metastases in up to 2–3 sites [16]. This classification has important prognostic and therapeutic implications, as patients with truly oligometastatic disease may be candidates for more aggressive local therapy, while those with polymetastatic disease require a more systemic approach. This narrative review synthesizes contemporary evidence to provide a comprehensive overview of the current management landscape for CRLM. We examine the evolution of surgical resection, the role and timing of perioperative systemic therapy, the expanding applications of local and regional therapies, and the emerging paradigm of liver transplantation. A central theme throughout is the critical importance of patient selection, i.e. matching treatment intensity to tumor biology and disease burden. Methodology This structured narrative review was conducted following SANRA (Scale for the Assessment of Narrative Review Articles) guidelines. A literature search of PubMed/MEDLINE, Scopus, and Cochrane Library was performed for articles published between January 2000 and December 2025. Search terms included combinations of (“colorectal liver metastases” OR “CRLM”) AND (“hepatectomy” OR “liver resection” OR “ablation” OR “hepatic arterial infusion” OR “liver transplantation”). Inclusion criteria were: landmark prospective trials, large cohort studies (n ≥ 100 patients), systematic reviews and meta-analyses, and consensus guidelines. Exclusion criteria were: case reports, non-English publications, and studies without original data. Reference lists of included articles were hand-searched for additional relevant studies. Given the narrative design, formal quality scoring was not performed; instead, emphasis was placed on methodologically rigorous studies and those with high clinical impact. We acknowledge the inherent limitations of the narrative review format, including the absence of formal quality scoring or risk-of-bias assessment for individual studies, as well as potential selection bias in the inclusion of cited literature despite our systematic search strategy. Accordingly, our synthesis should be interpreted as a qualitative overview intended to highlight key developments and areas of controversy, rather than as a definitive quantitative summary of the evidence. The cornerstone of cure: surgical resection and systemic therapy The evolution of surgical technique Surgical resection remains the only potentially curative treatment for CRLM, offering 5-year overall survival (OS) rates of 40–60% and 10-year OS rates of 25–30% in contemporary series [4, 5, 17]. However, the definition of “resectable” has expanded dramatically over the past three decades, driven by advances in surgical technique, better understanding of liver anatomy and regeneration, and the recognition that parenchymal preservation is paramount. A landmark single-center analysis of 1,736 hepatectomies performed over three decades (1990–2021) illuminates this evolution [18]. Over this period, patient characteristics shifted toward older age, more synchronous disease, and more frequent bilobar involvement, reflecting an increasingly aggressive approach to patients who would previously have been deemed unresectable. Simultaneously, surgical technique evolved toward parenchymal-sparing and non-anatomical resections, with a significant decline in major hepatectomies (from 57.2% in the 1990s to 34.7% in the 2010s) and a corresponding increase in the use of ablation as an adjunct to resection. Perioperative outcomes improved markedly despite increasing patient complexity, with 90-day mortality declining from 5.3% to 1.6% and hospital stay decreasing from 12.8 to 10.3 days. Perhaps most importantly, this series demonstrated that 5-year OS improved significantly across the study period, from 40.5% in the 1990s to 51.6% in the 2010s. Notably, this improvement was not attributable to reduced recurrence rates—which remained stable at approximately 70%—but rather to improved survival after recurrence, which nearly doubled from 19 to 31 months. This finding underscores the critical importance of effective post-recurrence therapies, including repeat resection, ablation, and increasingly effective systemic treatments [19]. For patients with extensive bilobar disease not amenable to single-stage resection, two-stage hepatectomy (TSH) has emerged as a valuable strategy. The Memorial Sloan Kettering Cancer Center (MSKCC) experience with TSH combined with perioperative systemic chemotherapy and hepatic arterial infusion (HAI) chemotherapy in 183 patients reported a completion rate of 69% and, among those completing both stages, a median OS of 58.2 months with 5-year OS of 53% [12]. These outcomes, achieved in patients with a median of 10 metastases, underscore the potential of aggressive surgical strategies when supported by effective systemic and regional therapy. Factors associated with failure to complete TSH included extrahepatic disease, tumor size > 5 cm, and lower pathologic response to neoadjuvant chemotherapy, highlighting again the primacy of tumor biology in determining outcomes. The ongoing debate on perioperative chemotherapy The integration of systemic chemotherapy with surgical resection has been a subject of intense investigation and debate. The theoretical rationale is compelling: perioperative chemotherapy may eradicate micrometastatic disease, downstage tumors to facilitate resection, and provide an in vivo test of chemosensitivity. However, the optimal timing, duration, and regimen remain contested. The EORTC 40,983 trial, a landmark randomized controlled trial comparing perioperative FOLFOX (six cycles before and six after surgery) with surgery alone for patients with up to four resectable metastases, demonstrated a significant improvement in progression-free survival (PFS) but failed to show an OS benefit after long-term follow-up [20]. The JCOG0603 trial, which specifically addressed adjuvant chemotherapy after hepatectomy, reported similar findings: while adjuvant mFOLFOX6 improved disease-free survival (DFS), this did not translate into an OS advantage after a median follow-up of 7.7 years (HR 1.07, 95% CI 0.73–1.57) [21]. These and other pivotal trials are summarized in (Table 1.) Table 1Landmark randomized trials in CRLM managementTrialYearPopulationInterventionPrimary EndpointKey FindingEORTC 40,9832013Resectable CRLM (≤ 4 metastases)Perioperative FOLFOX vs. surgery alonePFSImproved PFS (HR 0.77), no OS benefitNew EPOC2020Resectable, KRAS wild-type CRLMChemotherapy ± cetuximabPFSWorse PFS and OS with cetuximabJCOG06032026Resected CRLMAdjuvant mFOLFOX6 vs. observationDFSImproved DFS (HR 0.67), no OS benefit (HR 1.07)COLLISION2025CRLM ≤ 3 cm (≤ 10 lesions)Thermal ablation vs. resectionOS (non-inferiority)Ablation non-inferior for OS; fewer adverse eventsTRANSMET2024Unresectable, BRAF wild-type CRLMLT + chemo vs. chemo aloneOS5-year OS: 56.6% vs. 12.6% (ITT)CAIRO52023Initially unresectable CRLMVarious first-line systemic strategiesConversion rateTriplet + bevacizumab highest conversionAbbreviations: CRLM colorectal liver metastases, PFS progression-free survival, OS overall survival, DFS disease-free survival, LT liver transplantation, ITT intention-to-treat, HR hazard ratio Table 1Landmark randomized trials in CRLM managementTrialYearPopulationInterventionPrimary EndpointKey FindingEORTC 40,9832013Resectable CRLM (≤ 4 metastases)Perioperative FOLFOX vs. surgery alonePFSImproved PFS (HR 0.77), no OS benefitNew EPOC2020Resectable, KRAS wild-type CRLMChemotherapy ± cetuximabPFSWorse PFS and OS with cetuximabJCOG06032026Resected CRLMAdjuvant mFOLFOX6 vs. observationDFSImproved DFS (HR 0.67), no OS benefit (HR 1.07)COLLISION2025CRLM ≤ 3 cm (≤ 10 lesions)Thermal ablation vs. resectionOS (non-inferiority)Ablation non-inferior for OS; fewer adverse eventsTRANSMET2024Unresectable, BRAF wild-type CRLMLT + chemo vs. chemo aloneOS5-year OS: 56.6% vs. 12.6% (ITT)CAIRO52023Initially unresectable CRLMVarious first-line systemic strategiesConversion rateTriplet + bevacizumab highest conversionAbbreviations: CRLM colorectal liver metastases, PFS progression-free survival, OS overall survival, DFS disease-free survival, LT liver transplantation, ITT intention-to-treat, HR hazard ratio Landmark randomized trials in CRLM management Abbreviations: CRLM colorectal liver metastases, PFS progression-free survival, OS overall survival, DFS disease-free survival, LT liver transplantation, ITT intention-to-treat, HR hazard ratio A critical distinction must be drawn between the Level I evidence from randomized trials and the hypothesis-generating nature of retrospective analyses. The EORTC 40,983 and JCOG0603 trials (both prospective randomized studies) failed to demonstrate an overall survival benefit for perioperative or adjuvant chemotherapy, despite improvements in progression-free or disease-free survival. In contrast, large retrospective analyses consistently report OS benefits for perioperative chemotherapy. This discrepancy likely reflects selection bias inherent to observational data: patients receiving perioperative chemotherapy may differ systematically from those undergoing resection alone in ways that independently predict survival (e.g., performance status, access to care, treating institution volume). Accordingly, retrospective data should be viewed as hypothesis-generating rather than practice-defining. Several explanations for the DFS-OS disconnect have been proposed. First, oxaliplatin-induced sinusoidal obstruction syndrome may alter liver parenchyma and CT appearance, potentially leading to overdiagnosis of intrahepatic recurrence and artifactually improving DFS in the chemotherapy arm. Second, effective post-recurrence therapy, particularly re-exposure to oxaliplatin-based regimens in patients who received it adjuvantly, may “rescue” patients who recur, attenuating any OS difference. Third, the EORTC 40,983 trial was underpowered for OS, and the JCOG0603 adjuvant trial may have been confounded by differential use of post-recurrence therapy. Fourth, perioperative chemotherapy may benefit only specific subgroups, and pooling all-comers dilutes the signal. Despite these negative trials, large retrospective analyses suggest a benefit for perioperative chemotherapy in appropriately selected patients. An analysis of 43,039 patients with synchronous liver-only metastases from the National Cancer Database (NCDB) found that perioperative chemotherapy (administered both before and after resection) was associated with the highest median OS (44.7 months), compared with preoperative chemotherapy alone (38.4 months), postoperative chemotherapy alone (27.9 months), or resection alone (9.5 months) [22]. On multivariable analysis, perioperative chemotherapy with metastasectomy was associated with a 67% reduction in hazard of death compared with chemotherapy alone. However, the authors acknowledge that unmeasured confounding (e.g., KRAS status, detailed comorbidity assessment) likely persists despite propensity score methods. The choice between doublet (FOLFOX/FOLFIRI) and triplet (FOLFOXIRI) regimens is increasingly informed by disease burden. A multicenter retrospective study comparing outcomes after hepatectomy in patients receiving preoperative FOLFOXIRI versus doublet therapy found that, although triplet-treated patients had more advanced disease (higher tumor burden score, more synchronous metastases), their 5-year OS was similar to that of doublet-treated patients (66.6% vs. 56.4%, p = 0.072) [6]. Importantly, among patients with medium-high tumor burden, FOLFOXIRI was associated with significantly improved 5-year OS (67.6% vs. 50.1%, p = 0.004). These findings suggest that treatment intensification with triplet chemotherapy should be considered for patients with higher disease burden, with the caveat that perioperative morbidity, particularly bile leak, was increased in the FOLFOXIRI group. Based on available evidence, patients who may benefit from perioperative chemotherapy include, in our interpretation, those with: (1) high tumor burden (≥ 5 metastases or tumor burden score > 8); (2) synchronous presentation; (3) elevated CEA (> 30 µg/L); (4) RAS/BRAF mutant tumors (where biologic agents add less value); and (5) borderline resectability where downstaging may facilitate parenchymal-sparing resection. Patients with low-risk features (solitary metastasis ≤ 3 cm, metachronous presentation > 12 months from primary, normal CEA) may reasonably proceed directly to resection, as the absolute benefit of perioperative chemotherapy is likely small and must be balanced against toxicity. Microsatellite instability (MSI) status has emerged as an important consideration in perioperative decision-making. MSI-high (MSI-H) tumors, comprising approximately 3–5% of metastatic CRC, have excellent responses to immune checkpoint inhibitors and a more favorable prognosis independent of treatment [23]. However, MSI-H tumors are relatively chemoresistant, and perioperative chemotherapy provides less benefit. For MSI-H patients with resectable CRLM, perioperative immunotherapy rather than conventional chemotherapy may be preferred, though prospective data are limited. The role of targeted therapies in the perioperative setting is more nuanced. The New EPOC trial, which evaluated the addition of cetuximab to perioperative chemotherapy in patients with KRAS wild-type resectable CRLM, unexpectedly demonstrated worse outcomes in the cetuximab arm, with shorter PFS and OS [24]. This finding has been attributed to the development of acquired resistance clones and underscores the caution needed when combining targeted agents with perioperative therapy in patients with low-volume disease. In contrast, for patients with unresectable disease, the addition of anti-EGFR therapy in RAS wild-type left-sided tumors or anti-VEGF therapy in appropriate candidates improves response rates and conversion to resectability [9]. Challenging the gold standard: ablation for small metastases For decades, surgical resection has been the unquestioned standard for patients with resectable CRLM. The recent COLLISION trial has prompted reconsideration of this paradigm, providing high-level evidence that thermal ablation may be non-inferior to resection for small metastases [13, 14]. This international, randomized phase III trial enrolled 300 patients with up to 10 CRLM ≤ 3 cm in diameter, comparing thermal ablation (primarily microwave) with surgical resection. At a planned interim analysis, the trial met predefined stopping rules for early benefit, demonstrating that the conditional likelihood of proving non-inferiority for OS exceeded 90%. With a median follow-up of 28.9 months, OS was similar between groups (HR 1.05, 95% CI 0.69–1.58), with 5-year OS rates of 51.2% for ablation and 58.0% for resection. Local tumor control was non-inferior, and for per-tumor analysis, ablation was actually superior to resection (HR 0.09 for loss of local control). Crucially, the safety profile strongly favored ablation. Adverse events occurred in only 19% of ablation patients versus 46% of resection patients (p < 0.001), with serious adverse events in 7% versus 20%, respectively. There were no treatment-related deaths in the ablation group versus three (2%) in the resection group. Hospital stay was significantly shorter with ablation (median 1 vs. 4 days). However, several important caveats warrant emphasis. The median follow-up was 28.9 months, which is relatively short for oncologic outcomes in CRLM. The median number of metastases was only two (IQR 1–4), limiting generalizability to patients with more numerous lesions. Additionally, 43% of ablation patients underwent laparoscopy or laparotomy rather than percutaneous ablation, highlighting that optimal outcomes require multidisciplinary expertise in technique selection rather than universal applicability of percutaneous approaches. The non-inferiority finding should not be interpreted as superiority of ablation, nor as evidence that all small metastases should be ablated. Rather, these data support ablation as an equivalent option in appropriately selected patients when performed by experienced operators. These findings have implications for clinical practice. For patients with small CRLM (≤ 3 cm) that are technically amenable to ablation, this approach may offer comparable oncologic outcomes with markedly reduced morbidity and faster recovery. The results align with those of the MAVERRIC trial, which similarly reported comparable OS between ablation and resection [25]. As noted in an accompanying editorial, the assumption that thermal ablation should only be used for unresectable colorectal liver metastases needs to be reconsidered [26]. Expanding the arsenal: regional and locoregional therapies The revival of hepatic arterial infusion pump chemotherapy Hepatic arterial infusion pump (HAIP) chemotherapy represents a rational approach to liver-dominant disease, exploiting the dual blood supply of liver metastases (predominantly arterial) and normal parenchyma (predominantly portal). Floxuridine (FUDR), the agent most commonly used, has a 95% first-pass hepatic extraction, enabling high intra-tumoral concentrations with minimal systemic toxicity [27]. The MSKCC experience with HAIP combined with systemic chemotherapy (HAIP-SYS) for unresectable CRLM has been reported over several decades, demonstrating conversion-to-resection rates of up to 57% in chemotherapy-naive patients and median OS of 50–76 months [28, 29]. However, adoption outside MSKCC has been limited by technical complexity, the need for multidisciplinary expertise, and regulatory barriers (FUDR is not approved in the European Union). A recent phase II feasibility study from the Netherlands has addressed these barriers, demonstrating that HAIP-SYS can be successfully implemented in a European setting [30]. Among 31 patients with borderline resectable or unresectable CRLM (median 13 metastases), 90% completed at least two cycles of HAIP-SYS. Surgical and pump-related complications occurred in 16%, and grade ≥ 3 chemotherapy toxicity in 38%, rates comparable to the MSKCC experience. At first response evaluation, disease control rate was 83%, with hepatic disease control rate of 93%. Quality of life, initially reduced by surgery, returned to baseline by the time of first response evaluation. These findings have paved the way for a multicenter phase III trial (PUMP-IT) comparing HAIP-SYS with standard-of-care systemic therapy in patients with unresectable CRLM. If positive, this trial could establish HAIP therapy as a standard option for appropriately selected patients, particularly those with high-volume liver-only disease who face poor outcomes with systemic therapy alone. The role of targeted therapy in the metastatic setting The integration of targeted therapies has improved outcomes for molecularly defined subgroups of patients with metastatic CRC. A large real-world study using the Flatiron Health database (9,134 patients) provides valuable insights into patterns of targeted therapy use and associated outcomes [8]. Among patients with RAS wild-type tumors, 15.5% received EGFR inhibitors (cetuximab or panitumumab) as part of first-line therapy, and this was associated with improved OS (adjusted HR 0.85, 95% CI 0.74–0.98). In the Flatiron analysis, patients receiving first-line EGFR inhibitors had a median OS of 32.1 months (95% CI 29.4–34.8) compared with 27.3 months (95% CI 25.1–29.6) in those who did not receive targeted therapy. The adjusted hazard ratio of 0.85 corresponds to an absolute gain of approximately 4.8 months in median survival. In contrast, and diverging from clinical trial data, VEGF inhibitor use was not associated with OS benefit in either RAS wild-type (HR 1.00) or RAS mutant (HR 1.01) populations. Several factors may explain this discrepancy: the real-world population is older and has more comorbidities than trial populations; VEGF inhibitor therapy may be preferentially used in patients with higher-risk features; and the analysis could not fully account for tumor sidedness or the specific chemotherapy backbone. Several limitations of this real-world analysis warrant acknowledgment. The Flatiron database captures structured data from electronic health records but lacks granular information on treatment line, dose reductions, toxicity management, and reasons for therapy discontinuation. Unmeasured confounding is likely: patients receiving targeted therapy may differ systematically from those who do not in ways that correlate with prognosis (e.g., frailty, social support, access to subspecialty care). Additionally, the analysis could not fully account for tumor sidedness (left vs. right primary) or the specific chemotherapy backbone, both of which modify targeted therapy efficacy. Finally, real-world survival estimates are susceptible to immortal time bias and informative censoring. Importantly, this study also documented that more than one-third of patients (36.6%) received no targeted therapy at all. Older age (≥ 80 years) and higher ECOG scores were associated with lower odds of receiving targeted treatment, likely reflecting concerns about toxicity in vulnerable patients and appropriate caution given the uncertain cost-effectiveness of these agents in patients with limited life expectancy. Primary tumor sidedness is a critical determinant of prognosis and treatment response in metastatic CRC. Left-sided primary tumors (splenic flexure to rectum) have significantly better prognosis than right-sided tumors (cecum to transverse colon), independent of RAS/BRAF status [31]. This difference is driven by embryologic, molecular, and immunologic distinctions: right-sided tumors more frequently harbor BRAF mutations, CpG island methylator phenotype (CIMP), and MSI-H, while left-sided tumors are enriched for RAS wild-type and HER2 amplifications. Clinically, anti-EGFR therapies (cetuximab, panitumumab) are effective only in left-sided RAS/RAF wild-type tumors; right-sided tumors derive no benefit and potentially harm from EGFR inhibition [32]. Sidedness should be incorporated into treatment algorithms, particularly when considering biologic agents in the perioperative or first-line metastatic setting. Pushing the boundaries: liver transplantation for non-resectable disease From contraindication to curative option Liver transplantation for CRLM was historically abandoned after poor outcomes in the pre-1995 era, when 5-year OS was only 18% and most deaths were due to transplant-related complications rather than tumor recurrence [33]. However, improvements in surgical technique, immunosuppression, and patient selection have prompted reconsideration. The Norwegian SECA (SEcondary CAncer) studies provided the first modern evidence that transplantation could achieve meaningful long-term survival. In the SECA-I pilot study (21 patients), 5-year OS was 60% despite a recurrence rate of 90% [34]. The subsequent SECA-II study (15 patients) employed stricter selection criteria, excluding patients with progressive disease on chemotherapy, CEA >80 μg/L, or time from primary diagnosis <1 year—and achieved 5-year OS of 83% [35]. These outcomes, far exceeding the 9-10% 5-year survival expected with palliative chemotherapy, demonstrated that transplantation could achieve long-term disease control in a subset of patients [36]. The recently reported TRANSMET trial represents the first randomized comparison of liver transplantation plus chemotherapy versus chemotherapy alone in patients with definitively unresectable, BRAF wild-type CRLM [14, 15]. In this French multicenter trial, 94 patients were randomized 1:1. In the intention-to-treat analysis, 5-year OS was 56.6% in the transplantation group versus 12.6% in the chemotherapy-alone group; in the per-protocol analysis (patients actually receiving allocated treatment), 5-year OS was 73.2% versus 9.3%. Median PFS was 17.4 versus 6.4 months. Notably, although 74% of transplanted patients developed recurrence, 46% of these underwent curative-intent local treatment (surgery or ablation), and 42% of transplanted patients remained disease-free at 5-year follow-up. The clinical implications of a 70-90% recurrence rate warrant careful consideration. While 5-year OS of 56-83% is remarkable for patients with otherwise terminal disease, the majority of transplanted patients will experience recurrence, often within the first 2 years. Post-transplant immunosuppression may theoretically accelerate tumor growth, and recurrent disease can be challenging to treat. However, the TRANSMET experience demonstrated that 46% of recurrences were amenable to curative-intent local therapy (surgery or ablation), suggesting that transplantation converts diffuse, untreatable liver metastases into a state of oligometastatic disease that can be managed with conventional local therapies. These results have been met with both enthusiasm and caution. It is essential to recognize, however, that these outcomes derive from highly selected trial populations (e.g., TRANSMET required BRAF wild-type, age <65 years, and non-progressive disease after induction chemotherapy) and are not generalizable to most patients with unresectable CRLM, particularly those with progressive disease or unfavorable molecular profiles. A systematic review and meta-analysis pooling data from 20 studies (377 unique patients) confirmed these findings, reporting 5-year OS of 54% overall, with low MTV (<70 cm³) and favorable Oslo score (<2) identifying patients with particularly good outcomes [37]. The meta-analysis also highlighted that surgical resection of post-transplant recurrence, particularly pulmonary metastases, was associated with improved survival, supporting aggressive surveillance and intervention. Refining patient selection: the critical role of tumor biology The success of transplantation for CRLM hinges entirely on rigorous patient selection. The Oslo score, derived from the SECA-I experience, incorporates four negative prognostic factors: largest tumor diameter > 5.5 cm, time from primary surgery to transplantation < 2 years, pre-transplant CEA > 80 µg/L, and progressive disease on chemotherapy [29]. Patients with scores of 0–1 have 5-year OS exceeding 75%, compared with those with scores ≥ 2 who have poor outcomes. A comprehensive summary of selection criteria, including the Oslo score components and their evidence base, is provided in (Table 2). Table 2Patient selection criteria for liver transplantation in CRLMCriterionFavorableUnfavorable (Exclusion)EvidenceOslo Score components Largest tumor diameter≤ 5.5 cm> 5.5 cmSECA-I [34] Time from primary to LT≥ 2 years< 2 yearsSECA-I [34] Pre-transplant CEA≤ 80 µg/L> 80 µg/L (especially if rising)SECA-I [34] Disease progression on chemoStable/respondingProgressive diseaseSECA-I [34]PET/CT parameters Metabolic tumor volume (MTV)< 70 cm³≥ 70 cm³SECA 15-year [38] Total lesion glycolysis (TLG)< 260 g≥ 260 gIHPBA consensus [39]Molecular/histologic BRAF statusWild-typeV600E mutationIHPBA consensus [39] MSI statusMSSMSI-HIHPBA consensus [39] HistologyAdenocarcinoma, NOSSignet ring, undifferentiatedIHPBA consensus [39]Disease extent Extrahepatic diseaseNoneAny (relative contraindication)TRANSMET [14] Peritoneal involvementNonePrior or currentTRANSMET [14] Nodal status (extrahepatic)NegativePositive (relative)IHPBA consensus [39]Abbreviations: CRLM colorectal liver metastases, LT liver transplantation, CEA carcinoembryonic antigen, MTV metabolic tumor volume, TLG total lesion glycolysis, MSI microsatellite instability, MSS microsatellite stable, MSI-H microsatellite instability-high, NOS not otherwise specified Table 2Patient selection criteria for liver transplantation in CRLMCriterionFavorableUnfavorable (Exclusion)EvidenceOslo Score components Largest tumor diameter≤ 5.5 cm> 5.5 cmSECA-I [34] Time from primary to LT≥ 2 years< 2 yearsSECA-I [34] Pre-transplant CEA≤ 80 µg/L> 80 µg/L (especially if rising)SECA-I [34] Disease progression on chemoStable/respondingProgressive diseaseSECA-I [34]PET/CT parameters Metabolic tumor volume (MTV)< 70 cm³≥ 70 cm³SECA 15-year [38] Total lesion glycolysis (TLG)< 260 g≥ 260 gIHPBA consensus [39]Molecular/histologic BRAF statusWild-typeV600E mutationIHPBA consensus [39] MSI statusMSSMSI-HIHPBA consensus [39] HistologyAdenocarcinoma, NOSSignet ring, undifferentiatedIHPBA consensus [39]Disease extent Extrahepatic diseaseNoneAny (relative contraindication)TRANSMET [14] Peritoneal involvementNonePrior or currentTRANSMET [14] Nodal status (extrahepatic)NegativePositive (relative)IHPBA consensus [39]Abbreviations: CRLM colorectal liver metastases, LT liver transplantation, CEA carcinoembryonic antigen, MTV metabolic tumor volume, TLG total lesion glycolysis, MSI microsatellite instability, MSS microsatellite stable, MSI-H microsatellite instability-high, NOS not otherwise specified Patient selection criteria for liver transplantation in CRLM Abbreviations: CRLM colorectal liver metastases, LT liver transplantation, CEA carcinoembryonic antigen, MTV metabolic tumor volume, TLG total lesion glycolysis, MSI microsatellite instability, MSS microsatellite stable, MSI-H microsatellite instability-high, NOS not otherwise specified Metabolic tumor volume (MTV) measured on pre-transplant 18 F-FDG PET/CT has emerged as a powerful prognostic tool. In an analysis pooling 15 years of SECA experience (40 patients), MTV < 70 cm³ was associated with dramatically improved OS (5-year OS 76% vs. 21%, p < 0.001), DFS, and post-recurrence survival [38]. Importantly, the predictive value of MTV was strongest when PET/CT was performed within 90 days of transplantation, emphasizing the need for current imaging in patients on the waiting list. MTV outperformed other PET parameters (SUVmax, SUVmean, T/B ratio) and correlated with other adverse features including higher CEA, more metastases, larger tumor size, higher N-stage, and progressive disease at transplantation. These findings have been incorporated into the International Hepato-Pancreato-Biliary Association (IHPBA) consensus guidelines, which recommend exclusion of patients with progressive disease, CEA > 80 µg/L with increasing trend, MTV > 70 cm³, and TLG > 260 g [39]. The guidelines also recommend exclusion of high-risk histologies (undifferentiated adenocarcinoma, signet cell carcinoma) and BRAF V600E-mutated tumors, while noting that MSI-high tumors should be excluded due to their excellent prognosis with immunotherapy and potential risk of rejection. LT vs. staged resection (ALPPS): a new dilemma The emergence of transplantation as an option for patients with extensive CRLM has created a new therapeutic dilemma: for patients with extensive but potentially resectable disease, should aggressive surgical resection (including two-stage procedures and ALPPS) be preferred, or is transplantation a better option? Two recent studies have reached apparently opposite conclusions. A single-center French study comparing TSH (n = 32) with LT (n = 19) for extensive bilobar CRLM reported dramatically better outcomes for transplantation, with 5-year OS of 84.4% versus 32.2% (p = 0.003) [40]. However, baseline characteristics differed substantially: LT patients had more tumors (median 15 vs. 11) and more preoperative chemotherapy cycles (20 vs. 14), suggesting that transplantation was offered to patients with more advanced, but perhaps more chemotherapy-responsive disease. In contrast, a European multicenter study comparing LT (n = 54, Oslo experience) with ALPPS (n = 46, Italian registry) using entropy balancing to adjust for confounders found no difference in median OS (76 vs. 73 months) [41]. When ALPPS patients were further selected according to more stringent SECA-II criteria, LT appeared superior, suggesting that any survival benefit of transplantation depends on rigorous patient selection rather than the procedure itself. As noted in an accompanying commentary, “at this stage, ALPPS or other TSH must be preferred in patients with resectable disease, if feasible” [42]. Transplantation should be reserved for patients with truly unresectable disease who meet stringent selection criteria, and even then, must be viewed as a potentially curative but still palliative approach given recurrence rates exceeding 60%. The ethical and logistical challenges of organ allocation The extension of liver transplantation to patients with metastatic cancer raises ethical and logistical questions regarding organ allocation. A discrete event simulation model from Norway examined the potential impact of expanding eligibility to include CRLM patients on waiting times and life expectancy [43]. With the current Norwegian donor liver availability (approximately 115 annually), adding 2–4 CRLM patients per year increased median waiting time from 52 days to 61–73 days, with variations by blood type. However, the life-years gained for CRLM patients (10.6 years per patient under SECA-II criteria) outweighed the losses incurred by currently listed patients (0.05 years per patient), resulting in a net gain of 96–150 life-years over a 10-year programmatic period. These findings suggest that, in settings with relatively short waiting times, expanding eligibility to include highly selected CRLM patients can improve overall population health. However, generalizability to other countries depends critically on local organ availability, waiting list dynamics, and allocation policies. In the United States, where waiting times are longer and waitlist mortality higher, the calculation may differ substantially. Additional ethical considerations include the use of MELD exception points for malignant indications, which remains controversial, and the principle of utility versus justice in allocating a scarce resource to patients with metastatic cancer when patients with benign disease face mortality on the waiting list. Several strategies may mitigate the impact on waiting lists: use of extended criteria donors, living donor transplantation (which also enables elective scheduling and avoids allocation conflicts entirely), and novel surgical techniques such as the RAPID procedure (Resection And Partial Liver Segment 2–3 Transplantation With Delayed Total Hepatectomy), which allows use of small auxiliary grafts [44, 45]. The IHPBA consensus guidelines recommend consideration of these approaches and emphasize the importance of rigorous pre-transplant assessment, including intra-operative nodal sampling when indicated [39]. These criteria represent expert opinion rather than prospective validation, and local protocols may reasonably differ. It is essential to recognize that liver transplantation for CRLM remains available to only a fraction of patients with unresectable disease. Even in the TRANSMET trial, which enrolled patients at high-volume European transplant centers, only 94 patients were randomized over several years. For the vast majority of patients with unresectable CRLM, particularly those in regions with long waiting times or limited transplant infrastructure, transplantation may not be a realistic option. The enthusiasm for transplantation as a curative option must be tempered by recognition that, for most patients with unresectable CRLM, this is not, and probably will not become, a realistic option for many patients. Rigorous selection is not merely advisable but essential to avoid both poor outcomes for recipients and ethical breaches in organ allocation. Prognosis and surveillance in the modern era Defining long-term survival and cure As outcomes after resection of CRLM continue to improve, understanding the patterns and determinants of long-term survival becomes increasingly important. The previously described 30-year single-center series documented that while OS has improved, recurrence rates have remained stable at approximately 70% [18]. This apparent paradox is explained by improved survival after recurrence, which nearly doubled over the study period. Patients now live longer after recurrence, likely due to more effective systemic therapies and aggressive local treatment of recurrent disease when feasible. Importantly, the study identified 18 patients who recurred after 5 years of follow-up, supporting the recommendation for extended surveillance. The probability of being recurrence-free at 5 and 10 years was 34% and 33%, respectively, suggesting that patients who remain disease-free at 5 years have a high likelihood of long-term cure. Oligometastatic vs. polymetastatic disease The concept of oligometastatic disease has gained traction in CRLM management. A large Japanese single-center study of 1,123 patients undergoing resection of hepatic or pulmonary metastases classified patients as oligometastatic (1–5 metastases in 1–2 organs) or polymetastatic (≥ 6 metastases) [46]. Oligometastatic patients had significantly better 5-year OS (59.0% vs. 35.3%) and 5-year RFS (35.3% vs. 11.6%). On multivariable analysis, polymetastatic status was an independent predictor of both poorer OS (HR 1.59) and RFS (HR 1.67), along with elevated CEA, largest metastasis ≥ 2 cm, and synchronous presentation. These findings support the ESMO consensus definition of oligometastatic disease as up to 5 metastases in up to 2–3 sites [16]. Importantly, the prognostic impact of metastasis number was independent of metastatic site, suggesting that tumor burden—rather than specific organ involvement—should guide treatment intensity. Extrahepatic disease: site matters While this review focuses on liver-directed therapies, the presence and location of extrahepatic metastases critically influences treatment decisions. Pulmonary metastases are generally considered the most favorable extrahepatic site: patients with resectable lung metastases following or concurrent with CRLM resection achieve 5-year OS of 30–50%, comparable to liver-only disease in some series. The favorable biology of isolated lung metastases, often indolent and amenable to repeat resections, has led many centers to consider these patients eligible for aggressive local therapy. In contrast, peritoneal metastases portend a substantially worse prognosis. Even with complete cytoreduction and hyperthermic intraperitoneal chemotherapy (HIPEC), 5-year OS for colorectal peritoneal metastases ranges from 20 to 40%, and concomitant liver metastases further worsen outcomes. Most transplantation protocols explicitly exclude patients with prior or current peritoneal disease, and aggressive surgical strategies for CRLM are generally reconsidered in the presence of peritoneal spread. Nodal metastases outside the hepatic pedicle (e.g., para-aortic, celiac) are similarly unfavorable, though isolated hilar lymphadenopathy may be resectable. The IHPBA consensus guidelines recommend intraoperative sampling of suspicious nodes at the time of transplantation, with positive extrahepatic nodes constituting a relative contraindication. Optimizing follow-up The optimal duration and intensity of follow-up after CRLM resection remains debated. A UK single-center study of 486 liver resections with prospective follow-up (initially 10 years) examined recurrence patterns to inform surveillance protocols [47]. Recurrence occurred in 56% of patients, with a median time to recurrence of 1.6 years and longest time to recurrence of 4.8 years. Notably, no recurrences occurred beyond 4.8 years, even among patients surviving longer than 5 years. Based on these findings, the authors reduced their surveillance protocol from 10 to 6 years. However, they caution that the 95% confidence interval for recurrence in the 6th year (based on 66 patients entering that year with no events) was 0-5.4%, indicating that a small risk of late recurrence cannot be definitively excluded. For patients undergoing repeat resection for recurrence, the clock resets, and a new 6-year surveillance period is initiated. These data align with other long-term series suggesting that most recurrences occur within the first 5 years, and that patients who remain disease-free at 5 years have a high likelihood of cure [5]. However, given the significant survival benefit associated with aggressive treatment of recurrent disease, particularly when limited and amenable to resection or ablation, surveillance should be sufficiently intensive to detect recurrence at a treatable stage. Most guidelines recommend CT imaging every 3–6 months for the first 2–3 years, then every 6–12 months for up to 5 years [48]. A practical framework for clinical decision-making The following framework represents the authors’ interpretative synthesis of the evidence reviewed above and is not intended as a prescriptive clinical guideline. Final treatment decisions must be individualized and made through multidisciplinary discussion. The preceding sections have reviewed the expanding armamentarium for CRLM management. Translating this evidence into clinical practice requires a structured approach that integrates disease characteristics, tumor biology, patient factors, and local expertise. Figure 1 presents a treatment algorithm that operationalizes these principles, Figure 2 provides a historical timeline of the major paradigm shifts in CRLM management from 1990 to 2025, and Table 3 provides a framework for biology-driven treatment selection. Fig. 1Treatment algorithm for colorectal liver metastases. A decision tree stratifying patients by resectability, metastasis size and number, tumor biology, and response to systemic therapy. The algorithm integrates surgical resection, thermal ablation, perioperative chemotherapy, HAIP, and liver transplantation. Importantly, the applicability of this framework depends on local institutional expertise (e.g., availability of percutaneous ablation, HAIP pump placement, or liver transplantation programs) and regional variability in access to systemic therapies, organ allocation policies, and reimbursement structures. Treatment decisions should be tailored accordingly, with multidisciplinary input Treatment algorithm for colorectal liver metastases. A decision tree stratifying patients by resectability, metastasis size and number, tumor biology, and response to systemic therapy. The algorithm integrates surgical resection, thermal ablation, perioperative chemotherapy, HAIP, and liver transplantation. Importantly, the applicability of this framework depends on local institutional expertise (e.g., availability of percutaneous ablation, HAIP pump placement, or liver transplantation programs) and regional variability in access to systemic therapies, organ allocation policies, and reimbursement structures. Treatment decisions should be tailored accordingly, with multidisciplinary input Fig. 2Evolution of CRLM Management: 1990–2025. A timeline schematic showing paradigm shifts from the resection-only era (1990–2000) through the multimodal era (2000–2015) to the transplantation era (2015-present), with key landmark trials and technical innovations annotated Evolution of CRLM Management: 1990–2025. A timeline schematic showing paradigm shifts from the resection-only era (1990–2000) through the multimodal era (2000–2015) to the transplantation era (2015-present), with key landmark trials and technical innovations annotated Table 3Framework for biology-driven treatment selection in CRLM Importantly, the applicability of this framework depends on local institutional expertise (e.g., availability of percutaneous ablation, HAIP pump placement, or liver transplantation programs) and regional variability in access to systemic therapies, organ allocation policies, and reimbursement structures. Treatment decisions should be tailored accordingly, with multidisciplinary inputDisease CategoryTumor Biology CriteriaRecommended StrategyEvidence LevelRationaleOligometastatic, resectable (1–3 metastases, any size)Any favorable biology (RAS/BRAF wild-type, left-sided, normal CEA)Resection or ablation alone (no perioperative chemo)Level I (EORTC 40983, JCOG0603)No OS benefit from perioperative chemo; avoid toxicityOligometastatic, resectable (1–3 metastases, any size)Unfavorable biology (RAS/BRAF mutant, high CEA)Resection + perioperative chemo (doublet)Level II (retrospective)Higher recurrence risk; chemo may improve DFSHigh burden, resectable (≥ 4 metastases or bilobar)Any biologyResection + perioperative chemo (triplet if high burden)Level II (De Rose et al. [6])Triplet improves OS in medium-high burdenBorderline resectableChemotherapy-responsiveTSH or ALPPS + perioperative chemoLevel II (MSKCC series [12])Achieves resection in 69%; 5-year OS 53%Unresectable, liver-onlyFavorable (responder, low MTV, Oslo 0–1)Consider LT (if meets IHPBA criteria)Level I (TRANSMET [14])5-year OS 56–73% vs. 9–13% with chemo aloneUnresectable, liver-onlyUnfavorable (progression on chemo, high MTV)Palliative systemic therapy or HAIPLevel IIPoor LT outcomes; transplant contraindicatedUnresectable, liver-dominantChemotherapy-naive or limited progressionHAIP-SYS (if available)Level II (MSKCC, Dutch feasibility)Conversion rates 57%; median OS 50–76 monthsSmall metastases (≤ 3 cm)Any biology, technically amenable to ablationAblation or resection (equivalent options)Level I (COLLISION [13])Non-inferior OS, fewer complications, shorter stayExtrahepatic disease (lung only)Resectable, favorable biologyAggressive local therapy (resection/ablation of all sites)Level III (expert consensus)5-year OS 30–50%; comparable to liver-onlyExtrahepatic disease (peritoneum)AnySystemic therapy; surgery only if limited and specialized centerLevel IIIPoor prognosis with CRLM+peritoneal diseaseAbbreviations: CRLM colorectal liver metastases, CEA carcinoembryonic antigen, DFS disease-free survival, OS overall survival, TSH two-stage hepatectomy, ALPPS associating liver partition and portal vein ligation for staged hepatectomy, LT liver transplantation, HAIP-SYS hepatic arterial infusion pump with systemic chemotherapy, MTV metabolic tumor volume, IHPBA International Hepato-Pancreato-Biliary AssociationEvidence level definitions: Level I prospective randomized controlled trial, Level II prospective cohort or high-quality retrospective study, Level III expert consensus or lower-quality retrospective study Table 3Framework for biology-driven treatment selection in CRLM Importantly, the applicability of this framework depends on local institutional expertise (e.g., availability of percutaneous ablation, HAIP pump placement, or liver transplantation programs) and regional variability in access to systemic therapies, organ allocation policies, and reimbursement structures. Treatment decisions should be tailored accordingly, with multidisciplinary inputDisease CategoryTumor Biology CriteriaRecommended StrategyEvidence LevelRationaleOligometastatic, resectable (1–3 metastases, any size)Any favorable biology (RAS/BRAF wild-type, left-sided, normal CEA)Resection or ablation alone (no perioperative chemo)Level I (EORTC 40983, JCOG0603)No OS benefit from perioperative chemo; avoid toxicityOligometastatic, resectable (1–3 metastases, any size)Unfavorable biology (RAS/BRAF mutant, high CEA)Resection + perioperative chemo (doublet)Level II (retrospective)Higher recurrence risk; chemo may improve DFSHigh burden, resectable (≥ 4 metastases or bilobar)Any biologyResection + perioperative chemo (triplet if high burden)Level II (De Rose et al. [6])Triplet improves OS in medium-high burdenBorderline resectableChemotherapy-responsiveTSH or ALPPS + perioperative chemoLevel II (MSKCC series [12])Achieves resection in 69%; 5-year OS 53%Unresectable, liver-onlyFavorable (responder, low MTV, Oslo 0–1)Consider LT (if meets IHPBA criteria)Level I (TRANSMET [14])5-year OS 56–73% vs. 9–13% with chemo aloneUnresectable, liver-onlyUnfavorable (progression on chemo, high MTV)Palliative systemic therapy or HAIPLevel IIPoor LT outcomes; transplant contraindicatedUnresectable, liver-dominantChemotherapy-naive or limited progressionHAIP-SYS (if available)Level II (MSKCC, Dutch feasibility)Conversion rates 57%; median OS 50–76 monthsSmall metastases (≤ 3 cm)Any biology, technically amenable to ablationAblation or resection (equivalent options)Level I (COLLISION [13])Non-inferior OS, fewer complications, shorter stayExtrahepatic disease (lung only)Resectable, favorable biologyAggressive local therapy (resection/ablation of all sites)Level III (expert consensus)5-year OS 30–50%; comparable to liver-onlyExtrahepatic disease (peritoneum)AnySystemic therapy; surgery only if limited and specialized centerLevel IIIPoor prognosis with CRLM+peritoneal diseaseAbbreviations: CRLM colorectal liver metastases, CEA carcinoembryonic antigen, DFS disease-free survival, OS overall survival, TSH two-stage hepatectomy, ALPPS associating liver partition and portal vein ligation for staged hepatectomy, LT liver transplantation, HAIP-SYS hepatic arterial infusion pump with systemic chemotherapy, MTV metabolic tumor volume, IHPBA International Hepato-Pancreato-Biliary AssociationEvidence level definitions: Level I prospective randomized controlled trial, Level II prospective cohort or high-quality retrospective study, Level III expert consensus or lower-quality retrospective study Framework for biology-driven treatment selection in CRLM Importantly, the applicability of this framework depends on local institutional expertise (e.g., availability of percutaneous ablation, HAIP pump placement, or liver transplantation programs) and regional variability in access to systemic therapies, organ allocation policies, and reimbursement structures. Treatment decisions should be tailored accordingly, with multidisciplinary input Abbreviations: CRLM colorectal liver metastases, CEA carcinoembryonic antigen, DFS disease-free survival, OS overall survival, TSH two-stage hepatectomy, ALPPS associating liver partition and portal vein ligation for staged hepatectomy, LT liver transplantation, HAIP-SYS hepatic arterial infusion pump with systemic chemotherapy, MTV metabolic tumor volume, IHPBA International Hepato-Pancreato-Biliary Association Evidence level definitions: Level I prospective randomized controlled trial, Level II prospective cohort or high-quality retrospective study, Level III expert consensus or lower-quality retrospective study The following principles guide decision-making: First, determine resectability. Resectability is defined not by arbitrary number or size cutoffs but by the ability to achieve an R0 resection (negative margins) while preserving sufficient functional liver remnant (≥ 30% in healthy liver, ≥ 40% after chemotherapy). The evolution toward parenchymal-sparing techniques has expanded resectability; patients previously considered unresectable may now be candidates for TSH or ALPPS.Second, assess tumor biology. Regardless of resectability, tumor biology drives outcomes. Favorable biology includes: RAS/RAF wild-type, left-sided primary, normal CEA, response to neoadjuvant chemotherapy, and metachronous presentation >12 months from primary. Unfavorable biology includes: RAS/RAF mutation, right-sided primary, elevated/rising CEA, progression on chemotherapy, and synchronous presentation with short interval.Third, match treatment intensity to biology and burden. For patients with favorable biology and low burden (e.g., solitary metastasis ≤3 cm), resection or ablation alone may suffice, with chemotherapy reserved for recurrence. For patients with unfavorable biology or high burden, perioperative chemotherapy (doublet or triplet based on burden) is warranted, and transplantation may be considered for those with truly unresectable, chemotherapy-responsive, liver-only disease.Fourth, consider patient factors. Age, performance status, comorbidities, and liver function (including chemotherapy-associated liver injury) influence treatment tolerance and outcomes. Frail patients or those with significant comorbidities may benefit from less invasive approaches (ablation over resection) when oncologically equivalent.Fifth, incorporate local expertise. The evidence base supports multiple equivalent strategies (e.g., resection vs. ablation for small metastases; doublet vs. triplet chemotherapy). The optimal choice depends on local availability of expertise in percutaneous ablation, HAIP pump placement, transplantation, and molecular profiling. First, determine resectability. Resectability is defined not by arbitrary number or size cutoffs but by the ability to achieve an R0 resection (negative margins) while preserving sufficient functional liver remnant (≥ 30% in healthy liver, ≥ 40% after chemotherapy). The evolution toward parenchymal-sparing techniques has expanded resectability; patients previously considered unresectable may now be candidates for TSH or ALPPS. First, determine resectability. Resectability is defined not by arbitrary number or size cutoffs but by the ability to achieve an R0 resection (negative margins) while preserving sufficient functional liver remnant (≥ 30% in healthy liver, ≥ 40% after chemotherapy). The evolution toward parenchymal-sparing techniques has expanded resectability; patients previously considered unresectable may now be candidates for TSH or ALPPS. Second, assess tumor biology. Regardless of resectability, tumor biology drives outcomes. Favorable biology includes: RAS/RAF wild-type, left-sided primary, normal CEA, response to neoadjuvant chemotherapy, and metachronous presentation >12 months from primary. Unfavorable biology includes: RAS/RAF mutation, right-sided primary, elevated/rising CEA, progression on chemotherapy, and synchronous presentation with short interval. Second, assess tumor biology. Regardless of resectability, tumor biology drives outcomes. Favorable biology includes: RAS/RAF wild-type, left-sided primary, normal CEA, response to neoadjuvant chemotherapy, and metachronous presentation >12 months from primary. Unfavorable biology includes: RAS/RAF mutation, right-sided primary, elevated/rising CEA, progression on chemotherapy, and synchronous presentation with short interval. Third, match treatment intensity to biology and burden. For patients with favorable biology and low burden (e.g., solitary metastasis ≤3 cm), resection or ablation alone may suffice, with chemotherapy reserved for recurrence. For patients with unfavorable biology or high burden, perioperative chemotherapy (doublet or triplet based on burden) is warranted, and transplantation may be considered for those with truly unresectable, chemotherapy-responsive, liver-only disease. Third, match treatment intensity to biology and burden. For patients with favorable biology and low burden (e.g., solitary metastasis ≤3 cm), resection or ablation alone may suffice, with chemotherapy reserved for recurrence. For patients with unfavorable biology or high burden, perioperative chemotherapy (doublet or triplet based on burden) is warranted, and transplantation may be considered for those with truly unresectable, chemotherapy-responsive, liver-only disease. Fourth, consider patient factors. Age, performance status, comorbidities, and liver function (including chemotherapy-associated liver injury) influence treatment tolerance and outcomes. Frail patients or those with significant comorbidities may benefit from less invasive approaches (ablation over resection) when oncologically equivalent. Fourth, consider patient factors. Age, performance status, comorbidities, and liver function (including chemotherapy-associated liver injury) influence treatment tolerance and outcomes. Frail patients or those with significant comorbidities may benefit from less invasive approaches (ablation over resection) when oncologically equivalent. Fifth, incorporate local expertise. The evidence base supports multiple equivalent strategies (e.g., resection vs. ablation for small metastases; doublet vs. triplet chemotherapy). The optimal choice depends on local availability of expertise in percutaneous ablation, HAIP pump placement, transplantation, and molecular profiling. Fifth, incorporate local expertise. The evidence base supports multiple equivalent strategies (e.g., resection vs. ablation for small metastases; doublet vs. triplet chemotherapy). The optimal choice depends on local availability of expertise in percutaneous ablation, HAIP pump placement, transplantation, and molecular profiling. Conclusion: toward personalized, biology-driven treatment The recommendations offered in this review should be viewed as interpretative syntheses of available evidence, not as binding clinical guidelines. Final treatment decisions must be individualized and made through multidisciplinary discussion, incorporating patient preferences and local resources. The management of colorectal liver metastases has evolved from a binary decision based on resectability to a nuanced, personalized approach that integrates surgical resection, systemic therapy, regional chemotherapy, local ablation, and—for a highly selected subset—liver transplantation. The unifying theme across this therapeutic landscape is the primacy of tumor biology: outcomes are determined less by the specific treatment modality than by the underlying behavior of the disease. For patients with resectable disease, surgical resection remains the cornerstone of curative-intent therapy. The evolution toward parenchymal-sparing techniques, combined with effective perioperative systemic therapy, has enabled aggressive treatment of increasingly complex disease while improving perioperative outcomes. For patients with high disease burden, intensification with triplet chemotherapy (FOLFOXIRI) may be warranted, while those with favorable biology and small metastases may achieve equivalent outcomes with thermal ablation, sparing the morbidity of major surgery. For patients with unresectable but liver-only disease, regional therapy with HAIP chemotherapy offers a bridge to resection for some, while liver transplantation provides a potentially curative option for a highly selected few. The success of transplantation depends entirely on rigorous patient selection using validated tools including the Oslo score, metabolic tumor volume, and response to chemotherapy. The ethical challenges of organ allocation in this context require careful consideration of local organ availability, waiting list dynamics, and the balance of benefit between patients with malignant and non-malignant indications. Looking forward, several priorities emerge. First, the integration of molecular biomarkers, including RAS, BRAF, and MSI status, as well as emerging markers such as circulating tumor DNA, into prognostic and predictive algorithms will enable more precise patient selection [49, 50]. Second, randomized trials such as PUMP-IT and ongoing studies of transplantation will provide higher-level evidence to guide treatment decisions. Third, the development of novel surgical techniques and expanded use of living donation may mitigate organ scarcity and expand access to transplantation for appropriately selected patients. Ultimately, the management of CRLM exemplifies the principles of modern oncologic care: multidisciplinary collaboration, personalization based on tumor biology and patient factors, and a commitment to continuously reevaluating and refining treatment paradigms as evidence accumulates. The goal is no longer simply to determine whether disease is resectable, but to select the right treatment, or combination of treatments, for the right patient at the right time.