Objectives: Liver transplant is emerging as a potential treatment option for patients with isolated colorectal liver metastasis. In this review article, we analyzed the published literature on liver transplant outcomes in such patients.
Materials and Methods: Four prospective studies documenting the clinical outcomes in patients with colorectal liver metastasis who underwent liver transplant were analyzed to study the feasibility of liver transplant in such patients.
Results: The SECA-II trial demonstrated the highest overall survival of 100%, 83%, and 83% at 1, 3, and 5 years, respectively, and disease-free survival of 53%, 44%, and 35%, respectively, with a narrow inclusion criterion. Conversely, extended criteria for selection and donors in arm D of the same trial resulted in median overall survival and disease-free survival of 18 and 4 months, respectively.
Conclusions: Liver transplant provided more prolonged overall survival compared with other therapeutic modalities. Patients with isolated colorectal liver metastasis of less aggressive biology, good performance status, at least 6 weeks of chemotherapy, low clinical risk scores, and negative nodal disease should be considered for patient selection. Moreover, exclusion criteria consisting of patients with the right-sided primary tumor, less than 3 years to liver transplant after diagnosis, and elevation of carbohydrate antigen(CA19-9) in the presence of BRAF mutation should be explored.
Key words : Colorectal cancer, Overall survival, Recurrence
Colorectal cancer is the third most common cause of diagnosed cancers globally and the second most common cause of cancer-related mortality as of 2020.1 Metastasis to the liver occurs in almost half of patients with colorectal cancer.2 Although most patients with colorectal cancer have extrahepatic metastases, some have isolated metastases to the liver.3
Regional treatment approaches alone or with systemic chemotherapy are considered alternative options for such patients.3 These approaches range from surgical resection, thermal ablation, regional hepatic intra-arterial chemotherapy, and chemoembolization to radioembolization and radiation therapy, including stereotactic radiation therapy.3 Despite the emergence of advanced therapeutic options, surgical resection combined with neoadjuvant and adjuvant chemotherapy remains the most successful treatment at curing the disease.4-6 The current literature has reported that the hepatectomy approach for colorectal liver metastasis (CRLM) can achieve a 5-year overall survival (OS) rate of 47% to 60%.7-9 However, 40% to 75% of patients can experience recurrence,10-12 with at least half having recurrent metastases in the liver.11-13 Although repeated hepatectomies can improve survival after recurrence in the liver,14 the 5-year survival drops to less than 10% with palliative chemotherapy.15 Moreover, few candidates with CRLM are eligible for liver resection, with up to 10% of patients with isolated liver disease having unresectable disease.4 This is partly due to the proximity of specific metastases to vital structures and the scant anticipated volume of residual liver after resection.16 Furthermore, several chemotherapeutic agents are known to cause toxicity to the liver, impairing liver function and the process of regeneration.17-19 In particular, irinotecan and oxaliplatin-based regimens are known to cause steatohepatitis and sinusoidal injury, respectively.17,18 The distinct situation has given birth to the idea of total hepatectomy followed by liver transplant (LT) to cure patients with isolated CRLM.16
The initial attempts at transplant for unresectable liver metastases produced discouraging results, closing the doors to this option until 2013.16 The Vienna experience by Mühlbacher and colleagues in 1991,20 in which LT was used to treat unresectable liver metastases, including 17 patients with CRLM, showed a 5-year survival of only 12% and a high recurrence rate of 60%. The North American data from the same year produced equally disappointing results, concluding this chapter for a few years.21 Before 2013, several isolated case reports reported outcomes of LT in CRLM. Table 1 lists these studies and the patient characteristics. These included case reports from Honore (2003), Kocman (2011), and Uskudar (2011) and colleagues of rescue LT following acute liver failure secondary to either chemotherapy or biliary sclerosis due to bile duct damage.22-24 All of these patients survived with no evidence of disease at the end of the follow-up period. On the other hand, Hrehort and colleagues (2013) and Wang and colleagues (2016) reported poor outcomes due to extrahepatic pulmonary metastases following LT, leading to the death of the patient in the latter report.25,26 Between 2013 and the present, a series of trials, albeit few, have shed light on LT as a potential curative option for isolated CRLM.27-30
This review aimed to investigate the feasibility of LT in patients with nonresectable CRLM based on findings available from existing literature to gauge whether the approach can improve outcomes.
Materials and Methods
Eligible studies were identified using a PubMed search. Combinations of the following MeSh terms were used to retrieve articles: “liver,” “transplantation,” “metastasis,” “metastases,” “colon,” “rectal,” “colorectal,” “cancer,” and “carcinoma.” Subsequently, the relevant trials were shortlisted manually.
The criteria for inclusion entailed clinical trials and series reporting on LT due to CRLM. Those studies reporting on (1) LT for non-CRLM, (2) involving colorectal and other primary cancer, (3) individual case reports and case series, and (4) animal studies were excluded. Table 2 lists the eligible studies27-30 included in this review. The 4 studies and details of candidate selection are described below.
In 2013, Hagness and colleagues published the results of the first series of trials known as SECA (SEcondary CAncer) to investigate the outcome of LT in patients with unresectable CRLM.27
SECA-I used broad inclusion criteria to evaluate the feasibility, safety, and effectiveness of LT in patients with CRLM as part of a prospective pilot study. The main inclusion criteria for the 25 candidates included in the study were isolated CRLM with radically excised primary tumors, good performance status (Eastern Cooperative Oncology Group score of 0 or 1), and at least 6 weeks of chemotherapy. Moreover, only those patients with at least a 1-year time span between the diagnosis of colorectal cancer and the date of being listed on the transplant list were included. Patients with more than 10% weight loss, standard contraindications to LT, and other concomitant malignancies were excluded from the trial.
The SECA-I trial was followed by SECA-II, which aimed to gauge the impact of narrower inclusion criteria on outcomes.28
SECA-II utilized stricter inclusion criteria and enrolled 15 patients who had isolated CRLM with a minimum of 10% response to chemotherapy (RECIST criteria) and at least 1 year between the diagnosis of the primary tumor and the listing for LT. Moreover, patients in SECA-II had a lower tumor burden and carcinoembryonic antigen (CA19-9) level compared with those in SECA-I.
A noteworthy point to consider for these 2 SECA trials for candidate inclusion was the pre-LT examination with 18-fluorodeoxyglucose positron emission tomography/computed tomography (18-FDG-PET/CT) to identify tumors with less aggressive biology and to better detect extrahepatic metastases, a viable option for patient selection.31 At the time of primary disease resection and low Fong Clinical Risk Score, negative nodal disease yielded better outcomes. These, along with the criteria above, allowed tailoring the criterion for ideal candidate selection.
SECA-II arm D trial
The SECA-II arm D prospective study included patients with nonresectable CRLM with extended criteria to include those excluded from the SECA-I and SECA-II studies.30
In contrast to the earlier SECA trials, arm D extended the criteria to include patients with less than 10% response to chemotherapy, those with resectable pulmonary metastases, and a history of resection of pulmonary metastasis.
Compagnons Hépato-Bilaires from Toso and colleagues was a series of 12 patients who underwent LT for nonresectable CRLM published by a European consortium.29
The Compagnons Hépato-Bilaires trial included patients with isolated liver metastasis but failed to elaborate on the standardized patient selection criteria and intervention protocol.
This is a review article; therefore, this study was exempt from the need for ethical approval.
Table 2 summarizes the primary outcomes of the 4 studies.
Outcomes in SECA-I
The results of SECA-I (n = 25 patients) were characterized by a median follow-up time of 27 months (range, 8-60 mo) and OS of 95%, 68%, and 60% at 1, 3, and 5 years, respectively. Nineteen patients had recurrence, and 6 patients died after a median of 26 months (range, 6-41 mo) from disseminated colorectal cancer. None of the patients with tumor in the ascending colon achieved OS of 5 years. Seven patients had a recurrence of colorectal cancer in the transplanted liver, whereas 7 more developed pulmonary-only metastases. The patients did not receive adjuvant or preoperative chemotherapy as per the standard protocol and were administered a sirolimus-based immunosuppressive regimen posttransplant. The study was limited by the small sample size and the short follow-up. In a follow-up study that graded the recurrence pattern, all patients who were followed beyond 11 months developed recurrence with a median time of 6 months.32 Lungs were the most common first site of recurrence (13/21 patients). Of these patients, 7 had isolated pulmonary metastases and 3 underwent pulmonary resections. All 7 patients were alive after a median follow-up of 27 months. Oddly, the liver was not the first recurrence site in any of these patients, and no patients had isolated liver disease.
Interestingly, the high recurrence rate did not deteriorate survival, as patients with pulmonary first-site metastases displayed a 5-year survival rate of 72% after the first diagnosis of recurrence. Thus, the authors concluded that the survival outcome was not affected significantly by the development of pulmonary metastases post-LT. Similarly, existing literature documented a 5-year OS of approximately 45% after several pulmonary and hepatic resections of CRLM.33,34
Outcomes in Compagnons Hépato-Bilaires
The Compagnons Hépato-Bilaires29 trial (n = 12 patients) reported a median follow-up of 26 months and OS at 1, 3, and 5 years of 83%, 62%, and 50%, respectively. Six of the 12 patients developed recurrences, of which 5 were to the lungs. Unarguably, the most important finding of this study was that 4 patients did not relapse even ~48 months post-LT. Despite the lack of standardized selection criteria and its small sample size, the encouraging result of this study was the demonstration that a long-term cure could be achieved in patients with LT. Moreover, patients achieved disease-free survival after multiple chemotherapy and liver resection courses and a long time between colorectal surgery and transplant (>24 months).
Outcomes in SECA-II
The SECA-II28 study (n = 15 patients) reported a median follow-up time of 36 months (range, 5-60 mo) and an improved OS at 1, 3, and 5 years of 100%, 83%, and 83%, respectively. Eight patients (53%) developed recurrence following LT, of which 6 patients developed pulmonary metastases as the first or only site of metastatic disease. The median time between LT and detection of lung metastases was months (range, 2.7-24.3 mo). Five patients with pulmonary metastases received 6 resections, whereas 1 patient underwent liver resection for solitary liver metastasis. Moreover, 2 patients underwent resection of lymph nodes followed by postoperative radiation therapy. The 2 patients who developed oligometasta-tic disease after LT received palliative chemotherapy but could only achieve OS of 13 and 17 months from the start of chemotherapy. Seven patients without relapse and 4 patients who had pulmonary resections had no evidence of disease at follow-up, including the patient who underwent liver resection for metastasis. The study demonstrated a good survival rate of 72.9% even 4 years after relapse, and no patient required transplant. Thus, SECA-II was able to prove that the narrower selection criteria for LT significantly improved clinical outcomes. However, the lack of long-term follow-up still limited the credibility of LT outcomes among patients with CRLM. Interestingly, because patients in the SECA-II had lower liver PET uptake values than those in SECA-I, this indicated a more favorable tumor biology than SECA-I.
Outcomes in SECA-II arm D
Patients enrolled in the SECA-II Arm D study had a median follow-up time of 23 months after LT and demonstrated median disease-free survival and OS of 4 months and 18 months, respectively. The disease relapsed in 8 patients, with 6 patients developing metastases to the lungs first while 1 patient had a local pelvic relapse. Of the 6 patients, 1 also had metastases in the donor liver graft, lymph node, and peritoneum. Because of the multiplicity of the lesions, the 6 patients received palliative chemotherapy or radiotherapy instead of undergoing resection. Despite having Fong and Oslo scores similar to those enrolled in SECA-I, patients with Oslo scores of 1 to 2 had a reduced OS compared with patients with similar scores from SECA-I. Two patients with BRAF mutation were included for the first time in this study, with 1 patient having a concomitant elevation of plasma CA19-9 level. Interestingly, the patient with normal CA19-9 levels was alive even after 26 months of follow-up, whereas the other had an OS of 6 months only. Moreover, 5 patients had a tumor originating on the right side, which potentially contributed to poor outcomes.
The current review reported on the use of LT as a curative approach for patients with isolated CRLM based on results from existing studies documenting clinical outcomes.
What parameters need to be investigated to define inclusion criteria for liver transplant candidates?
The surplus of organ donors facilitated the success and feasibility of the SECA trial compared with recipients on the donor list in Norway. This surplus significantly reduced wait times. However, in most other countries, donor grafts are scarce. Several strategies have been proposed to conserve the donor graft pool. These include the use of small grafts (allograft segment 2-3),35 auxiliary liver grafts, implanted either orthotopically or heterotopically, and split liver grafts. The former is followed by second-stage hepatectomy once the graft regenerates to an adequate volume. However, the magnitude of shortages still gives rise to the need for a reasonable selection criterion to avoid the wastage of donor grafts.
Comparison of results from SECA-I versus SECA-II arm D suggested that, in addition to the Fong score, Oslo score, and PET metabolic tumor volume liver uptake values, other factors such as tumor location, histological differentiation, and lymph node status of the primary should be considered while selecting candidates for LT among patients with nonresectable CRLM. With regard to candidate inclusion, there are some noteworthy points to consider from these trials. All 4 studies discussed in our review utilized traditional clinical risk scores such as Fong and Oslo scores for risk stratification. These clinical risk scores use clinicopathological indicators such as the size of the most significant lesion, CEA levels, time from primary tumor resection to LT, characteristics of the disease (metachronous vs synchronous, progressive disease), and lymph node positivity for the calculation.
Existing literature has suggested poorer outcomes in patients with right-sided tumors following liver resection, with reduced OS36-38 from the start of first-line chemotherapy.39 None of the patients in the SECA-I study with right-sided disease survived more than 5 years after LT, and 5 patients in SECA-II arm D study had disease in the ascending colon, potentially contributing to their poorer outcomes. Furthermore, Toso and colleagues reported better OS in those who received LT at least 3 years after diagnosis of colorectal cancer. This finding was corroborated by SECA-II arm D, where 1 of the 2 patients without relapse was transplanted more than 14 years after the initial diagnosis. Another interesting finding from the arm D study was the poor prognosis in the patient with both BRAF mutation and raised plasma CA19-9 level. Although only 2 patients with BRAF mutations had been included in the trials, previous investigations have reported reduced OS following first-line chemotherapy when both of these markers were elevated.38 Hence, a stringent exclusion criterion consisting of patients with the right-sided primary tumor, less than 3 years to LT after theinitial diagnosis, and concomitant elevation of CA19-9 in the presence of BRAF mutation should be validated. This may help to combat the issue of global organ scarcity by selecting the most eligible patients.
The most common extrahepatic recurrence site from all 4 included studies was the lung. The existing literature demonstrated better survival among patients with pulmonary metastases compared with liver-only metastases.40 Similarly, data from Dave and colleagues showed a 5-year survival of 30% in a cohort of patients with liver and lung metastases who only underwent liver resection due to tumor progression.41 These findings sparked a debate on whether patients with adequate pulmonary function should be selected for LT, given that a significant number of them will develop pulmonary metastases.
Patients with nonresectable CRLM lesions and low Fong and Oslo scores or PET-derived metabolic tumor volume may achieve a 5-year OS comparable to LT using the Milan criteria.42 More recently, the Genetic and Morphological Evaluation (GAME) score was developed, which uses clinicopathological and biological indicators (the KRAS biomarker) to guide preoperative selection for resection among patients with CRLM.43 The KRAS biomarker can provide a snapshot of tumor biology, while the composite variable, “tumor burden size,” may serve as a proxy for the morphological features of the tumor.44 The GAME score has already outperformed the Fong and Oslo scores in large heterogenous populations to predict liver resection outcomes.45 Thus, future trials should incorporate this score to ascertain its utility in candidate selection for resection versus LT in nonresectable CRLM.
Are liver transplant outcomes superior to other forms of treatment such as chemotherapy and portal vein embolization?
Dueland and colleagues46 compared the results of LT from the SECA-I study with a much wider cohort of patients who underwent chemotherapy in the NORDIC VII study.47 The results favored LT over standard chemotherapy based on 5-year OS of 9% in the cohort receiving first-line chemotherapy only. However, these data have limited utility as newer chemotherapeutic regimens (such as FOLFOXIRI and bevacizumab) have shown an 80% response, with 61% of patients being responsive to resection following treatment.48 Hence, the outcomes of LT need to be compared with those of modern alternate systemic and locoregional treatment options before any conclusion can be drawn. Moreover, the comparison between these 2 cohorts can be misleading as the difference in selection criteria (inclusion of extrahepatic metastases), changes in the imaging technology, tumor burden, and response to chemotherapy were not accounted for in the study.
Another study compared OS between patients with high liver metastasis tumor load treated with LT with portal vein embolization (PVE) followed by liver resection.49 The results favored LT over the latter for patients with isolated liver metastasis except for primary tumors in the ascending colon. The studies differed in imaging methods (FDG-PET/CT vs CT/magnetic resonance imaging), the degree of pretreatment with chemotherapy, and the number of CRLM lesions. Furthermore, patients in the LT cohort had unresectable disease if the tumor board labeled it, whereas those in the PVE cohort were considered to have a resectable disease if adequate future liver remnant was achieved after PVE. Hence, prospective trials accounting for such differences for both chemotherapy and PVE should be conducted to allow fairer comparison of outcomes with LT. Nevertheless, LT does offer a promising approach for a better survival outcome.
Andersen and colleagues assessed the quality of life of the cohorts included in the SECA study using prospective nonrandomized surveys at 3, 6, and 12 months after LT.50 The results indicated good functional scores overall, with some patients reporting improved quality of life. Thus, in addition to high OS, good quality of life is another indication for LT to be explored as a treatment option for nonresectable CLRM.
Several prospective trials are currently recruiting patients. A trial from Toronto is currently investigating the OS, disease-free survival, and quality of life in patients with isolated CRLM undergoing a combination of chemotherapy and living donor LT.51 Another ongoing randomized control trial from Hôpitaux de Paris is aiming to compare LT outcomes in combination with chemot-herapy versus outcomes with standard chemotherapy alone among patients with unresectable CRLM.52 Similarly, the Oslo group of the SECA studies are recruiting patients to compare outcomes between those receiving chemotherapy and LT in a randomized trial (SECA-III).53 More recently, the same group just started a trial to evaluate outcomes between a combination of second-line chemotherapy and LT versus second-line chemotherapy alone in patients with nonresectable CRLM and high tumor burden.54 Other ongoing trials and prospective studies investigating similar outcomes include 2 trials described in ClinicalTrials.gov identifier numbers NCT03803436 and NCT04742621.
In addition, several trials have shifted their focus toward investigating the outcomes of using different surgical approaches and types of donor grafts for transplant. One such study is the MELODIC trial from Italy, which is evaluating the efficacy of using grafts from deceased donors.55 Another trial from Germany is investigating a novel approach of using 2-stage hepatectomy followed by transplant of a living donor graft for such patients.56 Two other ongoing trials evaluating such approaches are described in ClinicalTrials.gov identifier numbers NCT04161092 and NCT02215889. The results of these studies are expected to tackle the global scarcity of organs in offering LT as a treatment option for CRLM. The details of these studies are summarized in Table 3.
It is premature to extend the LT approach to patients with unresectable CRLM, given the small number of trials documenting clinical outcomes and the lack of credible comparisons with existing treatment modalities. The first step would entail refining the selection criteria to include optimal tumor biology with a predisposition for isolated liver disease. This may require genetic profiling as well as validation and development of preoperative clinical risk scores. In addition, the application of LT in nonresectable CRLM would lead to increased competition for donor grafts, which are already scarce globally even for the current clinical indications. Hence, novel strategies need to be investigated to conserve the pool of grafts as the outcomes of LT for CRLM reach acceptable levels.
The most common recurrence site appears to be the lungs. Therefore, patients with adequate lung function should be screened since they may need a resection of the pulmonary metastases in the future. It is also possible that the benefit of transplantation could be higher for patients with a more significant tumor burden but more certainly limited to the liver. Advances in imaging could help select such patients, especially those most likely to die due to isolated hepatic metastasis. Moreover, a stricter exclusion criterion will ensure organ transplantation to the neediest and to patients with the best prognosis.
In summary, patients with liver-only colorectal metastasis with good response to chemotherapy meeting the inclusion mentioned earlier and exclusion criteria based on the parameters discussed in this article should be considered for future trials. Large-scale prospective randomized control trials spanning multiple centers are required to assess the generalizability of LT to all patients meeting the developed selection criteria. The difference in tumor biology and spread pattern means LT may not achieve the same disease-free survival rate regardless of how stringent the selection criteria are. However, the promising results of the SECA-II trials indicated that a good OS can be achieved even for CRLM and that LT may be a feasible treatment for nonresectable CRLM in the not-so-distant future.
Volume : 20
Issue : 2
Pages : 113 - 121
DOI : 10.6002/ect.2021.0421
From the 1Department of Surgery, Division of Surgical Oncology, University Hospitals Cleveland Medical Center, Cleveland, Ohio; the 2Keck School of Medicine, University of Southern California, Los Angeles, California; 3Transplant Surgery, Boston University School of Medicine, Medford, Massachusetts; and the 4Tufts University School of Medicine, Boston, Massachusetts, USA
Acknowledgements: This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. The authors have no conflicts of interest to declare.
Corresponding author: Hassan Aziz, South Building, Floor 4, 800 Washington Street, Boston, MA 02111, USA
Phone: +1 617 636 8228
Table 1. Studies Documenting Outcomes of Liver Transplant in Colorectal Liver Metastases (1972-2020)
Table 2. Shortlisted Studies Analyzed in This Review
Table 3. Ongoing Clinical Trials