Criteria for liver transplantation for hepatocellular carcinoma: what is an acceptable outcome?


Jacques Belghiti, Department of Hepato-Pancreato-Biliary and Transplantation Surgery, Beaujon Hospital, [Assistance Publique Hôpitaux de Paris], 92110 Clichy, France
Tel: 33 1 40 87 58 95
Fax: 33 1 40 87 17 24 e-mail:


The incidence of hepato cellular carcinoma (HCC) and the shortage of grafts restrict liver transplantation (LT) in HCC patients with a low risk of recurrence. The risk of recurrence is mainly related to the presence of vascular invasion which increases in parallel with tumour size and number of nodules. A favourable post-transplant outcome has been observed in patients who meet the empirically defined Milan criteria, namely, a single nodule < 5 cm or two or three nodules each < 3 cm in the absence of macroscopic vascular invasion, based on pre-transplant imaging. These criteria were felt to be too restrictive, leading several centers to propose expanded criteria for LT. However, increasing both the size and number of nodules resulted in an increased risk of recurrence. It has not been demonstrated that loco-regional treatment in HCC patients listed for LT (bridging therapies) improve post-transplant survival. More precise predictors of negative prognostic factors including elevated α-feto protein level, poor differentiation and molecular techniques should be considered in order to optimize the use of grafts and achieve zero recurrence.

Liver transplantation (LT), which remains the best treatment of small hepatocellular carcinoma (HCC) from chronic liver disease, is limited by the increasing number of potential candidates and the scarcity of donors (1). Because of organ shortage, the poor results of transplantation in patients with large, unresectable tumours led to stringent selection criteria (2, 3). Indeed, recurrence generally occurs in patients with large tumours within the first months or years following transplantation. Extrahepatic metastases are frequent and tumour growth is markedly accelerated by immunosuppressive therapy. Overall, the prognosis in patients with post-transplant recurrence is still very poor. These stringent criteria were supported by the finding that patients with small incidental tumours on the explant had a low rate of post-transplant recurrence.

The Milan criteria (MC) (a single nodule less than 5 cm or two to three nodules each<3 cm on pretransplant imaging), which were determined empirically, have been shown to select patients with 5-year post-transplant survival exceeding 70% and a rate of tumour recurrence below 15% (4). The MC have been widely adopted since then. With increasing experience, it has become clear that, while the results of LT in patients meeting the MC are good, a substantial number of patients with HCC exceeding these criteria could also derive a good benefit from transplantation (5–9). These findings were an incentive to extend selection criteria.

The risk of recurrence according to tumour size is a continuum. This continuum is clearly illustrated by the notion of the ‘Metro Ticket’. When one travels on the underground, the price of the ticket increases as the distance increases. Similarly, when tumour size increases, the price to pay in terms of recurrence increased proportionally (10). Because of organ shortages, it can be argued that only patients with the lowest risk of recurrence should be selected to optimize organ allocation.

Risk factors associated with recurrence after liver transplantation

In almost all cases, tumour recurrence after transplantation rapidly leads to death and graft loss, even with the use of the most recent anticancer agents. Again, tumour progression is markedly accelerated by immunosuppression (11).

The risk of recurrence is influenced by the presence or the absence of vascular invasion, which can either be macroscopic or microscopic. Macroscopic vascular invasion, corresponding to any invasion of large branches of the portal vein (or less commonly hepatic veins), which can be observed on gross pathological examination, is inevitably associated with early recurrence. Whatever the size and number of nodules, macroscopic vascular invasion represents a definitive contra-indication for LT. However, macroscopic vascular invasion is much more common in patients with large tumours than in those with small tumours (including those who meet the MC). Vascular invasion seen only under microscopic inspection is identified as microvascular. Microscopic vascular invasion is also a risk factor for recurrence although less than microscopic invasion. Again, microscopic vascular invasion is less common in patients with small tumours. However, it remains impossible to detect macroscopic invasion on the basis of pretransplant imaging. The presence of satellite lesions, defined as tumour nodules <2 cm from the primary tumour, carries a high risk of recurrence (12). Although poor differentiated HCC are often associated with vascular invasion, the presence of vascular invasion is significantly correlated to the size and number of nodules. As a result, the size and number of the nodules still represent the most objective and relevant selection criteria for transplantation. Independent of tumour burden, it has been shown recently that a high α-fetoprotein (AFP) level is also predictive of recurrence. Patients with serum AFP over 500 ng/ml or a rapid increase in serum AFP are at a higher risk of recurrence, even if they meet the MC (13).

The Milan criteria

Results of transplantation in patients with a single nodule<5 cm or two or three nodules each <3 cm, in the absence of macroscopic vascular invasion, provides excellent results, with a 3-year survival rate exceeding 80% and a recurrence rate lower than 10%. These selection criteria, frequently termed as the ‘Milan criteria’, are currently used as a reference (4). However, there is a risk that the morphological characteristics may result in misdiagnosis due to the limits of imaging technology (13). Tumours may either be under-staged, allowing LT to proceed in patients with a higher risk of recurrence, or over-staged, excluding patients from LT who have a low risk of post-LT recurrence. Because the MC were felt to be too restrictive, expanded criteria have been proposed.

Beyond the Milan criteria and ‘metro-ticket’

According to the frequently termed ‘UCSF criteria’, patients should meet the following criteria to be eligible for transplantation: a solitary tumour <6.5 cm or two or three nodules with the largest lesion <4.5 cm and a total of 8 cm (9). This relatively modest expansion in the selection criteria resulted in survival rates of 90 and 75% at 1 and 5 years respectively. Until now, the UCSF criteria have not been widely accepted. The concept of the ‘metro ticket’ has been used to demonstrate this point, showing that expansion of increasing both the size and the number of nodules resulted in an increasing risk of recurrence. The further the criteria are pushed, the higher the price in terms of survival. Another potentially confounding issue when expanding the MC is the extreme variability of the time to transplantation of patients with HCC.

Preliver transplantation adjuvant treatment

Local-regional therapy (LRT) for HCC, including various transarterial and ablative techniques, is increasingly being used to prevent list dropout, improve post-transplant survival and downstage advanced disease. Pre-LT treatments included transarterial chemo-embolization, radiofrequency ablation and partial liver resection. Although there is no controlled study showing a beneficial effect of LRT before LT, evidence indicates that pretransplant LRT of HCC prevents list dropout (14). Patients with advanced HCC, especially beyond the MC, should typically undergo LRT, followed by a surveillance wait period, even where there is a short waiting time or when a living donor is available. Such an approach may facilitate the identification of patients who have HCC with poor biological behaviour that is more likely to recur post-transplant.

Role of resection

Advances in liver surgery have significantly improved the safety of resection and prior resection neither increases operative morbidity nor impairs survival following cadaveric donor transplantation (15). Resection can be used as a treatment for HCC before LT as a primary therapy considering LT in the case of recurrence or as a selection of good candidates of LT. Resection as the first-line treatment for patients with small HCC with preserved liver function, followed by salvage transplantation in the case of recurrence, has been shown to be a feasible strategy providing rapid access to an effective therapy. The main obstacle to this strategy is the risk of ‘loss of chance’ in the case of rapid and extensive recurrence not amenable to salvage LT. Another justification for resection before transplantation is that it provides access to detailed pathological examination of the tumour and the surrounding liver parenchyma, providing important prognostic information. As a result, resection may help deny transplantation in patients with tumours apparently within the MC but with the histological features of an especially poor prognosis (undetected macrovascular invasion in particular). In contrast, resection may help decide transplantation in patients with tumours slightly outside the MC but with the histological features of a good prognosis.

Which outcome is acceptable?

Liver transplantation is an effective treatment of HCC with results equal to or better than treatments for most other gastrointestinal malignancies. Importantly, for the donor pool, properly selected patients with HCC achieve equivalent post-transplantation results compared with candidates without malignancy. Many questions remain regarding the selection of candidates who are the most likely to achieve long-term success, which, if any, pretransplantation treatments are efficacious for long-term survival and how to fairly rank candidates with HCC among all candidates waiting for transplants.

Towards zero recurrence

In a context of graft storage, a policy to select HCC candidates with post-transplant outcomes similar to non-HCC recipients with elimination of all available and detectable recurrence risk factors should be considered. In order to optimize the use of grafts achieving zero recurrence, a retrospective study from our group showed that among patients within MC, those with pre-LT AFP rate over 400 ng/ml before transplantation, poor differentiation and/or CCK component should be excluded from transplantation (16). These two histological factors can be selected with a preoperative biopsy or resection.

Conflicts of interest

The authors have declared no potential conflicts.