Therapy in Liver Diseases 2003 : Hepatocellular carcinoma

Journal of Hepatology, Vol. 38 (S1) (2003) pp. S136-S149
© 2003 European Association for the Study of the Liver. Published by Elsevier B.V. All rights reserved.
PII: S0168-8278(02)00432-4

Hepatocellular carcinoma (HCC) represents more than 5% of all cancers in the world, and the estimated number of cancer-related deaths exceeds 500,000 per year [1]. Two major epidemiological facts characterize this cancer. It occurs in a previously diseased liver, and the causes of the underlying liver disease differ according to the geographical distribution. Consequently, the mechanisms of hepatocarcinogenesis and the characteristics of the tumour might vary greatly from one part of the world to another. In Africa and Southern Asia, the role of aflatoxin B1 and HBV infection - which is acquired at birth or early in life - is highly predominant. In these patients, HCC develops often at a young age and in the absence of cirrhosis. By contrast, in Japan, Egypt and in Southern Europe, HCV is the main cause of HCC which occurs in older patients, nearly all of them with advanced fibrosis or cirrhosis. In Northern and Central European countries, HCV infection and alcohol are the main causes of cirrhosis. In France, ethanol is still the leading cause of cirrhosis and was responsible for 60% of all HCC cases during the last decade [2]. The incidence of HCC and its clinical presentation are different according to the geographical distribution of risk factors. In parts of Africa or East Asia, where the highest incidence rates are observed, most of the patients are diagnosed with infiltrative or massive tumours. In Western countries, where the underlying liver disease has frequently been detected and the patients are regularly followed, HCC is mostly diagnosed at an asymptomatic stage by routine ultrasonography (US). In a minority of HCV infected or alcoholic patients, HCC appears as a symptomatic disease that decompensates the underlying cirrhosis. Another important epidemiological fact is the rising incidence of HCC in developed countries, where it was a marginal cause of death years ago. In the United States, approximately 15,000 new cases occur each year [3], and in France, where the incidence of HCC has steadily and markedly increased, the estimated number is about 4000 per year [4]. The rising incidence is not only due to a more accurate diagnosis or an increase in the immigration rates from high prevalent countries. Epidemiological data have shown a steady increase in native patients during the last two decades. The reasons advocated for explaining this phenomenon are the increased rate of HCV infection and an improvement of the clinical management of cirrhotic patients. However, the increase incidence of HCC has only been joined by a slight increase in the percentage of cases due to HCV infection. This may favour the hypothesis that enhancing the survival of patients with advanced cirrhosis leads to an increased incidence of HCC. In fact, a decade ago, most of the deaths in cirrhotic patients were due to digestive haemorrhage or bacterial infections, two conditions that are now more efficiently prevented and cured [5]. Therefore, HCC has become the leading cause of death in patients with cirrhosis. Prevention and treatment of this neoplasm are now major concerns.

It is clear that the prevention of the development of liver diseases and of their progression to cirrhosis is the most effective way to prevent HCC. HBV infection, the main cause of HCC in the world, could be efficiently prevented by vaccination. Vaccination of children in Taiwan has led to a drastic decrease in the incidence of HCC [6,7]. This intervention is clearly cost-effective in developing countries with high HBV infection rates. Prevention and cure of HCV infection will also greatly influence the incidence of HCC in many countries, as well as prevention and early detection of alcohol abuse. Primary prevention of HCC in the West concerns patients with advanced fibrosis or cirrhosis. This emphasizes the need for screening for liver diseases in the general population and the validation of non-invasive methods for detecting cirrhosis, which still remains as an underdiagnosed condition. In patients with cirrhosis the risk of developing HCC varies according to simple criteria such as age, sex, the cause of the underlying liver disease and the severity of cirrhosis. The presence of elevated serum alpha-fetoprotein (AFP) values is also predictive. These simple means allow the population of patients with cirrhosis to be split into different subgroups according to the expected incidence of HCC. Histological parameters could help to refine the model. A high proliferate index, the presence of `irregular regeneration' [8,9] or large cell dysplasia in hepatic biopsy [10] have been proposed as predictors of the occurrence of cancer. Obviously, each of these factors and also putative environmental agents such as tobacco smoking have to be further validated by larger studies. Chemoprevention is a tempting option in patients with high risk. Up to now primary chemoprevention has been considered in patients with HCV-related liver disease. A preventive effect of interferon (IFN) has been suggested by many non-randomized studies in long-term responders [11], and seems to be due to the anti-viral effect of IFN or its association with Ribavirin. Up to now, only two randomized controlled trials (RCTs) have addressed this issue in cirrhotic patients, with opposite results [12,13]. The use of differentiating agents such as polyprenoic acid in primary chemoprevention is a promising approach. Polyprenoic acid has shown a chemopreventive effect in animal models and has been claimed to be effective in the prevention of development of new tumours after curative treatments [14,15].

It is clear that HCC in the West occurs mainly in patients with cirrhosis, with a mean annual incidence of around 3-4%. Amongst these patients some are at a higher risk of developing tumours, and, thus, are the ideal target to propose chemoprevention and/or screening. Screening for HCC in patients with cirrhosis has been advocated for years and although its benefit in terms of survival has not been demonstrated (and seems difficult to demonstrate) [16] it is currently highly popular amongst hepatologists in developed countries [17,18]. The purpose of surveillance is to recognize HCC at an early stage, where the tumour could be cured, and consequently tumour-related deaths can be decreased. The identification of small tumours - especially <2 cm in diameter - is important not only because this makes them eligible for curative treatments, but also because at this size they are less prone to invade the surrounding liver parenchyma.

A consensus has been reached concerning screening methods, taking into account the efficacy of each procedure and its cost. Firstly, periodical assessment of serum AFP has a poor sensitivity and a poor specificity and is practically useless as a screening tool [19]. Serum AFP is often intermittently elevated in cirrhotic patients, but is rarely markedly elevated in patients with small tumours. Therefore, if the cut-off value for initiating a recall procedure is the upper limit of normal, the cost will be tremendous with a limited benefit, whereas if the cut-off value is elevated to almost diagnostic levels such as 400 ng/ml, the sensitivity will be extremely poor. Therefore, serum AFP - one of the first and more popular tumour markers - is to be abandoned as a screening tool. It still has a role for defining high-risk patients and also for establishing a non-invasive diagnosis when markedly elevated [19]. Second, among the imaging techniques, CT-scan and MRI have been discarded as screening methods due to their cost and relative invasiveness, and US has emerged as the only usable technology. This technique, however, is observer-dependent, and is also limited because it may be difficult to interpret in some circumstances according to patient characteristics (i.e. obesity), tumour characteristics (isoechoic) or location (subdiaphragmatic area), and characteristics of the non-tumorous parenchyma (heterogeneous pattern in steatosis or macronodular cirrhosis). Despite these limitations, US examination is able to recognize HCC before the critical diameter of 3 cm in more than 85% of cases.

Recall procedures in the presence of a nodule on a cirrhotic liver will vary according to the diameter of the lesion (Table 1). A firm diagnosis of HCC is still difficult to reach when the nodule is less than 1 cm in diameter, and short-term follow-up is suitable in this setting [20]. The nodule could either vanish, particularly in the case of heterogeneous steatosis, persist in the case of monoclonal macronodule (dysplastic nodule) or HCC, or exhibit an even limited growth which is almost synonymous of HCC. Nodules sized 1-2 cm in diameter should be further studied by the determination of serum AFP and additional imaging, either CT-scan or MRI or both, even though diagnosis should rely on histology. Conversely, in nodules above 2 cm the diagnosis of HCC is highly probable, and a firm diagnosis can be established by non-invasive criteria described below [20].

Table 1. Recall diagnostic procedures in the presence of a liver nodule newly discovered by US screening in a patient with cirrhosis
Finding Diagnosis of HCC Recommended recall procedure
Nodule <1 cm Improbable. <50% are finally HCC Repeat US at 3 months: if growth of the nodule, proceed to next step.
Nodule 1-2 cm Probable Serum AFP, CT-scan and/or MRI. Biopsy required for diagnosis, but 40% false negatives.
Nodule 2-3 cm Highly probable Serum AFP, CT-scan and/or MRI. Apply non-invasive criteria. Biopsy indicated in the absence of arterial hypervascularization.
Nodule >3 cm Almost diagnostic Serum AFP, CT-scan and/or MRI. Apply non-invasive criteria. Biopsy rarely indicated. Search for vascular invasion.

Several cohort studies have shown that surveillance increases the rate of early detection, and also slightly increases the application of curative treatments (Table 2) [21-27]. However, they have failed to demonstrate a benefit in terms of survival. This failure was mainly due to the previously poor performances of US, the absence of clear diagnostic criteria (except histological proof often difficult to reach), and consequently the absence of early curative treatments in most of the cases. Therefore, a cost-effectiveness evaluation of these programs has so far been difficult to perform because its survival benefit has not yet been demonstrated. Nowadays, the wide use of US technology, and the improvement in the application of curative treatments may evolve to a better scenario. The fact that most of these nodules are monoclonal lesions predisposing to cancer [28] - if not in situ HCCs - might help to decide their ablation by minimally invasive procedures.

Table 2. Results of surveillance programs for HCC in cirrhotic patients
Author Year Country No. patients/no. HCC Main cause of cirrhosis Child A patients (%) US periodicity (months) HCC annual incidence (%) Unique nodule <5 cm (%) Patients eligible for curative treatment (%)
Oka et al. [21] 1990 Japan 140/40 Virus NPa 3 6.5 33 (82) 15 (38)
Colombo et al. [22] 1991 Italy 447/56 Virus 86 12 3.2 30 (54) 12 (29)
Pateron et al. [23] 1994 France 118/14 Ethanol 64 6 5.8 6 (42) 2 (14)
Cottone et al. [24] 1994 Italy 147/30 Virus 100 6 4.4 25 (83) 10 (33)
Zoli et al. [25] 1996 Italy 164/34 Virus 72 3-6 6.2 26 (76) 7 (20)
Bolondi et al. [26] 2001 Italy 313/61 Virus 63 6 4.1 49 (80) 39 (69)
Henrion et al. [27] 2000 Belgium 141/6 Ethanol 69 6 2.4 6 (100) 6 (100)a NP, not precise.

Traditionally, and in similarity with other tumours, HCC diagnosis was established on the basis of histological or cytological proof. Limitations to biopsy have been well documented for small and large tumours. In small early tumours, the rate of false negative results may be as high as 40% [29], and some of them are difficult to reach under US or CT-scan guidance due to their location. In larger tumours, the risk of complications - either intraperitoneal or intrahepatic haemorrhage or needle track seeding - has been clearly documented, and may complicate around 3% of cases. These high values could certainly be minimized using fine needles and puncturing the tumour through non-tumorous parenchyma but, even in experienced hands, a small rate of serious complications remains. The establishment of a firm diagnosis of well differentiated HCC with a small histological sample can be difficult, and may require additional samples or cytological smears, each of these procedures carrying a risk. These limitations have led to the use of non-invasive criteria, which chiefly rely on imaging techniques [20].

New imaging technology, either by CT-scan or last generation MRI, allows the studying of tumours with contrast injection at an early arterial phase, as well as at the parenchymatous and portal phase [30]. HCC is characterized by arterial hypervascularization followed by a rapid wash-out, appearing hypovascular in the portal phase. Other signs are the existence of a capsule or pseudocapsule and the presence of fat or haemorrhage inside the tumour. Typically, these tumours are hypointense on T1-weighted phase and hyperintense on T2-weighted phase by MRI imaging but these findings are less sensitive and less reliable than arterial hypervascularization. Among focal liver lesions, few are arterially hypervascularized: some metastasis such as neuroendocrine tumours, renal carcinoma, and others, which are rare in cirrhotic livers, some hemangiomas often typical when studied by MRI, and angiomyolipomas, which are exceptional. More troublesome, highly hypervascularized nodular areas of less than 2 cm in diameter have been recognized in cirrhotic livers with an appearance close to HCC when studied by CT-scan or MRI [31].

Recently a group of experts have proposed non-invasive criteria for the diagnosis of HCC in cirrhotic patients [20]. The rational basis of these criteria is the following: (a) the incidence of HCC is high in cirrhotic patients; (b) any focal hepatic lesion growing in a cirrhotic liver has a high probability of being neoplastic: either a dysplastic macronodule (a monoclonal proliferation that can lead to cancer) or more frequently an established HCC; (c) most HCCs are hypervascularized by an arterial route in contrast with the absent or discrete hypervascularization of other lesions. HCC diagnosis was established by the coincidental findings of two imaging techniques showing a nodule of >2 cm with arterial hypervascularization, or by a single positive imaging technique associated with AFP of >400 ng/ml [20]. These criteria are simple, specific and easy to use, even by a non-specialist. Obviously, they do not summarize all conditions in which a diagnostic of HCC could be made without histology. A focal growing lesion in a cirrhotic liver accompanied by a progressively increasing serum AFP (even below 400 ng/ml) or an arterially hypervascularized tumour in a non-cirrhotic liver accompanied by a markedly elevated serum AFP are also fairly specific. These non-invasive criteria have yet to be validated in a large scale. In addition, new criteria must also be developed for tumours less than 2 cm in diameter.

The prognosis of HCC was considered to be dismal two decades ago. Old series reported survival of less than 1 year in almost all patients [32]. At present, this scenario has completely changed, at least in Western countries, where the 5-year survival rates in surgical patients may exceed 70% [33-35]. This improvement is due to the advancement of the time of diagnosis and the increase in the therapeutic efficacy. Therefore, the prognosis of patients will vary according to the evolutionary stage in which the neoplasm is diagnosed and the treatment received.

The pathological evolution of HCC at initial stages has been accurately described by Japanese authors [36]. Initially HCC is a well-differentiated neoplasm with an ill-defined nodular appearance. Less-differentiated neoplastic tissues with greater proliferate activity evolve within them, when the tumour reaches a size of 1-1.5 cm in diameter. Neoplastic invasion - either as local metastases or portal vein invasion - may appear at the 2 cm cut-off level. Subsequently, moderately differentiated HCC tissues gradually replace the initial surrounding HCC, a phenomenon that is completed when HCC reaches 3 cm in size.

HCC is a multicentric disease in nature. At least 20-60% of HCC individuals diagnosed in the setting of surveillance programs present multicentric small tumours [26]. In addition, intraoperative US in the course of resection identifies 10-15% of additional nodules unrecognized prior to the procedure [34]. Similarly, pathological studies of the explanted liver of HCC patients undergoing transplantation show that 50-60% present multicentric tumours, at least 30% unrecognized by sophisticated pre-operative imaging techniques [34]. Half of these patients present coincidental pre-neoplastic lesions when analyzing the entire liver in 5 mm slices [37]. Multinodular HCC may be the result of intrahepatic metastases of single tumours, or may correspond to multicentric occurrence. Recent data analyzing genomic profiling suggest that multicentric disease involves more than one-third of all cases [38].

Early diagnosis of HCC has become the key issue in increasing the applicability of curative therapies, which currently benefit around 30% of cases [39]. Patients diagnosed at that stage will now receive radical therapies and thus, the natural course of the disease in these cases will remain poorly understood. Few historical series provided 3 year survival rates for untreated early HCC ranging between 13 and 26% [32]. The best estimates showed a 65% 3 year survival, when considering only single HCC on Child-Pugh A cirrhosis [40]. Most of these retrospective data, however, are hampered by the nature of the tumours, which for some unknown reason were left untreated. In referral liver units, a proper selection of candidates for curative therapies provides 5 year survival rates ranging between 50% and 70% [33-35]. It is assumed that active treatments improve the natural course of the neoplasm.

The prognosis of HCC was reported to be less than 1 year in patients at advanced evolutionary stages, when curative procedures were unfeasible [32]. Survival estimates gathered from population-based cancer registries have recently confirmed these figures [41]. However, these patients were recruited retrospectively more than 10 years ago. The management of cirrhotic patients and the imaging techniques have improved since then, and consequently, the outcome of untreated HCC has dramatically changed. During the last years, more than 20 RCTs have been published including an untreated arm of conservative management [39]. The 1 and 2 year control rates in these trials were 10-72% and 8-50%, respectively. The wide disparity of these figures reflects the heterogeneity of the population considered as merely with `non-surgical HCC'. Recently, the natural history of untreated HCC in a cohort of 102 patients randomly assigned to a control arm in the setting of two RCTs has been revised [42]. Overall, the 1, 2 and 3 year survival rates were 54%, 40% and 28%. The best predictors of survival were the presence of cancer-related symptoms (performance status test=1-2) and the identification of an invasive pattern evidenced by the presence of vascular invasion or extrahepatic spread. Thereby, two subgroups with a markedly different life expectancy were identified: patients in a truly intermediate stage (asymptomatic patients without any tumoral invasive pattern) had 1, 2 and 3 year survival rates of 80%, 65% and 50%, whereas patients at an advanced stage (at least one adverse prognostic factor) had survival rates of 29%, 16% and 8%, respectively. As is described below, several HCC staging systems share similar predictors of survival in this subset of individuals [2,43-46]. In addition, other prognostic factors have been included in these models such as AFP and alkaline phosphatase, Child-Pugh score and ascitis.

End-stage HCC is characterized by very advanced tumours associated with tumor-related symptoms and an overall survival of less than 3 months [32,44]. The prognostic factors identified in old series are reflected by both the Okuda stage III and a performance status of 3-4. Similarly, advanced tumours in Child-Pugh C patients also account for a very poor prognosis. A 5% 6-month survival in Child-Pugh C patients presenting spontaneous bacterial peritonitis and advanced tumours has been recently reported [47]. Thereby, these terminal patients should not be included in RCTs as no survival benefit can be reliably obtained. The inclusion of these patients in some RCTs may account for the poor control rate obtained.

Cancer staging should serve to select the appropriate primary and adjuvant therapy, to estimate the prognosis, and also to assist in the evaluation of the results of treatment, and to exchange information without ambiguity [48]. In oncology, the prognosis of patients with solid tumours is solely related to tumour stage, and other co-factors such as age or histological grade are only seldom considered. However, HCC patients constitute a particular neoplasm. Cirrhosis underlies HCC in most individuals and thus, their outcome is also related to this entity that simultaneously determines the applicability and efficacy of treatments. Accordingly, prognostic modelling in HCC has a high complexity. The EASL Panel of Experts recommended considering four related aspects: tumour stage, degree of liver function impairment, the patient's general condition, and treatment efficacy [20]. Prognostic systems assessing just one of these aspects (Child-Pugh, TNM, performance status) have a marginal usefulness. The Okuda staging [32] is unable to distinguish between early and advanced HCC and serves mostly to identify end-stage individuals.

Due to the lack of a common universally-used staging system, the outcomes reported in cohort studies describing the natural history of the disease and survival rates detailed in the setting of phase II trials or RCTs are heterogeneous. These discrepancies reflect the fact that the tools used to stage HCC have not been able to differentiate patients at a distinct evolutionary stage. Recently, five new systems have attempted to overcome these difficulties and aim to establish a useful tumour classification [2,43-46] (Table 3).

Table 3. New HCC staging systems
Classification (type) Variables Actuarial survival (%)
1-year 3-year 5-year
French classification (score) [2] Tumour stage: portal invasion, AFP, alkaline phosphatase; liver function: bilirubin; health status: Karnofsky index
   Stage A (score 0) 71 38 -
   Stage B (score 1-5) 33 12 -
   Stage C (score 6) 7 3 -
CLIP classification (score) [43] Tumour stage: portal invasion, morphology, AFP; liver function: Child-Pugh
   Score 0 84 50 -
   Score 1 66 35 -
   Score 2 45 12 -
   Score 3 36 5 -
   Score 4-6 9 0 -
BCLC staging (staging system) [44] Tumour stage: portal invasion/metastases, morphology, Okuda; liver function: Child-Pugh, portal hypertension, bilirubin; health status: PST
   Stage A (early) 85 70 50-70
   Stage B (intermediate) 80 50 16
   Stage C (advanced) 29 8 0
   Stage D (end-stage) 10 0 0
CUPI Index (score) [45] Tumour stage: TNM, AFP, alkaline phosphatase; liver function: ascites, bilirubin; health status: presence of symptoms
   Low risk (score 1) 47 25 -
   Intermediate risk (score 2-7) 17 5 -
   High risk (score 8) 5 0 -
TNM staging (staging system) [46] Tumour stage: morphology, portal invasion/metastases; liver function: fibrosis
   Stage I 90 75 55
   Stage II 80 65 33
   Stage III 60 25 15

The French classification was constructed with the analysis of 761 HCC patients, amongst which 47% received specific treatments [2]. This classification combines five variables in a score system that stratifies patients in three stages. Survival of these stages at 2 years was 51%, 16%, and 3%, respectively, and reflects the fact that this cohort mostly included patients at advanced stages. Its application to all HCC patients remains to be elucidated.

The Cancer of the Liver Italian Program (CLIP) classification [43] was constructed in a retrospective study and validated by the authors and other groups in prospective studies. This score combines four variables that provide a seven-stage classification system. It has been compared with Okuda staging and TNM staging with better discriminatory power. Asian groups have reported survival rates clearly different than the original authors, thus compromising their external validation [49]. It is also limited by the fact that it does not serve to select the appropriate therapy for each patient.

The Barcelona-Clínic Liver Cancer (BCLC) staging system [44] was constructed based on the results obtained in the setting of several cohort studies and RCTs by the Barcelona group. This proposal is not a scoring system since it derives from the identification of independent prognostic factors in the setting of several studies, conforming to a staging classification. This classification uses variables related to tumour stage, liver functional status, physical status and cancer-related symptoms, and links the four stages described with a treatment algorithm [39,44]. It has been suggested that it is best suited for treatment guidance, and particularly to select early-stage patients that could benefit from curative therapies [50]. However, at present, it has no external validation.

Investigators in Hong Kong described a staging system analyzing their experience in 926 patients, most of them with HBV-related cirrhosis [45]. The Chinese University Prognostic Index (CUPI) considers six predictive variables, and divides patients in three stages. The authors estimate that this classification has better estimation of survival than the CLIP score and Okuda stage, although its discriminatory power in early stages is questionable, since the best 1 year survival was around 50%.

The conventional TNM system, which only contained variables related to tumour stage, has been mostly tested in the surgical setting, and showed poor prognostic prediction in HCC patients undergoing either resection [51] or transplantation [52]. Based on the results of a series of 557 patients who underwent resection [46,53], a recent modification has been proposed, including tumour stage and presence of fibrosis. The new four-stage system may improve the stratification of resected tumours, even though it is controversial whether they will apply for non-surgical patients.

All these attempts to improve the classification and prognosis prediction of HCC are still evolving, and there is no agreement on the best staging that can be recommended world-wide. Also to be considered is that survival of patients with early tumours is modified by treatment and thus, prognostic prediction has to include treatment-related variables. By contrast, the impact of treatment in more advanced stages is negligible. The heterogeneous survival figures described for the best stages (3-year survival from 80% to 25%) in Table 3 reflect the fact that some studies include mostly advanced cases with a minor number of effectively treated patients. In these studies, treatment-related variables might not be identified as a relevant survival predictor and the use of the same set of variables for all patients may appear adequate [2,43,45]. However, for clinical purposes, it is more informative to divide patients into the relevant strata in which they may roughly be classified and then develop a prognostic system for each category [20].

Before any decision is made for treating a patient with HCC, the following questions have to be addressed. What is the status of the non-tumorous liver? What is the size and extension of the tumour? What is the general condition of the patient, their age and their expected life expectancy? The status of the non-tumorous liver is essential. In Europe, about 85-90% of patients with HCC have a cirrhotic liver [2,40,42-44] and most of the remaining cases have an underlying liver disease. Patients with non-cirrhotic liver are the best candidates for resection or other therapeutic means that could damage liver function, such as chemoembolization. The case of cirrhotic patients is more complex. It is clear that HCC patients with decompensated cirrhosis are not eligible for any invasive method, either surgical or radiological. In these cases there is agreement that the first-line option is liver transplantation (LT) [20]. Decision-making in Child-Pugh's A class patients with single tumours requires more sophisticated assessment. Aggressive treatments should be avoided in patients that will decompensate, and whose prognosis afterwards is usually bleak. Many selection criteria have been proposed such as indocyanin green clearance [54], serum bilirubin and portal hypertension (PHT) [34,55], or transaminases level. These criteria have been proposed for resection, but could presumably also apply for chemoembolization which induces ischemic damage to the non-tumoral liver.

Evaluation of tumour extension, with a search for the presence of daughter nodules, and/or portal vein thrombosis - even segmental - is necessary in all cases. This could be done by US, CT-scan and/or MRI. Angiography is nowadays considered a complementary method due to its invasiveness and the improvement of other radiological techniques. In selected cases, extrahepatic spread should be searched for by chest CT-scan or bone scintigraphy. The presence of tumoral portal invasion or metastasis leads to a bleak prognosis and precludes most of the potentially curative interventions [2,42]. Finally, the patient's general condition must be assessed before taking any therapeutic decision. In all cases, this issue should be analyzed by means of the Karnofsky index or PS test, which will identify patients with advanced or terminal disease, that are not eligible for most of the procedures [39]. Selection of candidates for transplantation is made according to different criteria, because variables dependent on liver failure (even involving the general status) are less relevant, although they might increase the perioperative mortality. In this case, assessment of cardiovascular and pulmonary function and other systemic conditions is crucial. In the light of all these elements, the clinician has to consider as the first step potentially curative treatments, resection, transplantation, and percutaneous destruction.

Few therapies are currently considered to be able to achieve complete tumour ablation. Surgical treatments are assumed to eliminate the total bulk of the tumour when properly applied. Percutaneous treatments have also proved to completely eliminate a significant proportion of tumours, mainly if sized below 3 cm. Selected candidates receiving any of these therapies are expected to have improved survival, although this assumption derives from observational studies (Table 4). There are no RCTs comparing these interventions, and thus there is no robust evidence of which of them provides the best outcome and cost-effective scenario. Consequently, there is no agreement to establish the first-line option in patients with single HCC in compensated cirrhosis, and each group should establish the preferred schedule taking into account their available resources and technical skills [20].

Table 4. Sources of evidence for primary treatment and prevention of recurrence of HCC
Treatment Source Evidence of benefitsa
Curative treatments
Resection Cohort studies Increased survival in selected candidates; CR: 100%
Liver transplantation (CLT/LDLT)a Cohort studies Increased survival in selected candidates; CR: 100%
Percutaneous treatments 5 RCTs; cohort studies Increased survival in selected candidates; CR: 60-80%
Palliative treatments
Arterial embolization/chemoembolization 17 RCTs; meta-analysis May increase survival in selected candidates; CR: 5%, PR: 20-50%
Arterial lipiodolization/chemotherapy 10 RCTs No survival benefit; PR: 20-30%
Internal radiation 3 RCTs No survival benefit; PR: 20-30%
Hormonal compounds
Tamoxifen 8 RCTs; meta-analysis No survival benefit; PR: 0%
Other compounds 4 RCTs No survival benefit
Systemic chemotherapy 9 RCTs No survival benefit; PR: 5-15%
Immunotherapy 4 RCTs No survival benefit; PR: 10%
Prevention of recurrence after curative treatments
Arterial embolization 1 RCT No effect
Chemotherapy 3 RCTs No effect
Intra-arterial I131; adoptive immunotherapy; polyprenoic acid; IFN 4 RCTs May reduce recurrence, and more data are neededa CR, complete response; PR, partial response; CLT, cadaveric liver transplantation; LDLT, living donor liver transplantation.

Hepatic resection is the first-line approach in non-cirrhotic patients with HCC. This only applies to 5% of cases in the West [39]. In cirrhotic patients, the poor selection of candidates for resection in the eighties led to unacceptable survival rates below 50% at 3 years [56]. Nowadays, this selection has been refined to avoid treatment-related complications, e.g. liver failure. Some groups consider Child-Pugh's A patients with single HCC as the target population, but this rough criteria is still associated with unacceptable morbi-mortality rates. Experienced groups have refined the patient's selection, and recommended this procedure for single asymptomatic HCC in patients with extremely well-preserved liver function (so called Child-Pugh's `hyper A') [34,35,54]. Japanese authors use the indocyanin green retention rate to identify the best candidates [54], a technique that is not routinely performed by Western groups. Portal pressure and bilirubin have been established as the best parameters to select the optimal surgical candidates in Europe [34,55]. Clinically relevant PHT was defined as the presence of either a hepatic vein pressure gradient of >10 mmHg [55], esophageal varices, or splenomegaly with a platelet count of <100,000/mm3. Thereby, subjects without relevant PHT and normal bilirubin achieved 70% survival at 5 years, whereas this decreased to 50% in patients with PHT, and to 25% in the worst candidates (with relevant PHT and abnormal bilirubin) [34]. A 50% survival expectancy at 5 years, considering a perioperative mortality below 3%, has been established as the minimal cut-off value to propose this procedure [39].

Tumour recurrence complicates 50% of cases at 3 years, and deteriorates long-term survival [34,35]. This occurs despite the fact that resection is nowadays restricted to excellent candidates [34,35]. In these cases, microvascular invasion, poor differentiation degree and satellites are considered the best predictors of recurrence [34]. These factors may be eventually associated with true recurrence rather than with new HCC development. Chen et al. [38] have proposed a comparative genomic hybridization (CGH) technique to document the chromosomal aberration profile that distinguishes between true recurrences (tumours that arise from the spread of the original) and second de novo HCC (a new focus independent of the previous cancer). These genetic changes were able to differentiate intrahepatic metastases after resection (35%) from second primary tumours (35%), but were non-specific in the remaining cases. Further molecular studies are needed to discriminate the incidence of each entity.

The high recurrence rate makes these patients the best target for the evaluation of preventive agents after resection. These strategies include agents to prevent true recurrences, such as adjuvant chemotherapy/chemoembolization, intra-arterial lipiodolized I131 and adoptive immunotherapy, or others aimed at preventing metachronic tumours, such as retinoids or IFNs (Table 4) [14,57-59]. Adjuvant chemotherapy or chemoembolization prior to or after surgical resection for HCC have failed to show any effectiveness in four RCTs. Pre-operative arterial embolization with epirubicin was assessed in 97 patients with solitary 2-5 cm HCC without significant benefit in terms of recurrence and survival [57]. Conversely, encouraging results were published with internal radiation I131, adoptive immunotherapy and acyclic retinoids. A unique RCT assessing adjuvant internal radiation with I131 after resection reported positive results in patients with hepatitis-B-related HCC [58]. Similarly, adoptive immunotherapy by the injection of activated lymphocytes was assessed in 150 Japanese patients, with positive benefits in terms of prevention of 2-year recurrence after surgery [59]. This is an expensive and complex therapy which is difficult to apply as a routine practice in the clinical setting. Finally, retinoids administration was shown to prevent secondary HCC tumours in 89 patients after curative treatments (resection or percutaneous ethanol injection) [14]. However, the low number of patients included in single studies raised some concerns about the findings and thus, more trials are needed. Therefore, these interventions are not accepted world-wide as the standard preventive agents after resection.

The clinical impact of transplantation has changed the treatment strategy for HCC. This therapy is claimed to cure the tumour and the underlying cirrhosis at the same time, in well-selected patients. The high recurrence rate (32-54%) and poor outcomes (5-year survival of <40%) obtained from transplant registries during the eighties were related to the wide selection criteria applied, including patients with macroscopic vascular invasion, lymph node involvement and extrahepatic spread [56,60]. These devastating figures questioned the inclusion of HCC patients in some programs. During the nineties a second era was faced. Some groups selecting the `optimal candidates' (patients with single HCC 5 cm or up to three nodules 3 cm) reported a 70% 5-year survival with a recurrence rate below 15% [33,34,61]. These studies shared not only strict selection criteria, but also a short waiting time of less than 6 months. The TNM classification was questioned as an efficient tool for HCC staging prior to LT [52]. During these years, this procedure was considered in most centres as the first-line option for early HCC. Nevertheless, the ideal candidates were already recognized, and a common agreement was that patients should only be listed when the probability of living 5 years after the procedure exceeded 50% [62].

At present, a new situation has arisen. The shortage of donors has led to extremely large waiting times in almost all countries distorting the encouraging outcomes reported. Leading programs in the United States and Europe face a waiting time of over 12 months, and the United Network of Organ Sharing (UNOS) data indicate that patients excluded while waiting almost equal those that are effectively transplanted [63]. The drop-out rate may account for 20-50% of cases, and is a major clinical problem for most programs. This event is also common, but less frequent, in patients listed without HCC, and has questioned the allocation of livers. In the United States, the UNOS have adopted the Model for End-Stage Liver Disease (MELD) system to allocate patients according to a composite score, which includes bilirubin, prothrombin activity, creatinine levels and aetiology of liver disease, for non-cancer patients, and a variable score between 24 and 29 points for HCC patients [64]. In the clinical practice this policy may represent a prioritization of HCC patients, although its real impact is unknown.

Several strategies have been proposed to ameliorate the impact of tumour progression while waiting for a donor. Adjuvant therapies - percutaneous ablation, chemoembolization or even chemotherapy - have been assessed in the setting of observational studies. At present, there is no evidence of survival benefit for these procedures, and thus randomized studies are needed [39]. Only cost-effectiveness analysis suggested survival advantages of 6 months for adjuvant therapies using resection or percutaneous treatment in comparison to conservative treatments [65]. Marginal livers, domino and split transplantation represent alternatives that only solve a minor proportion of cases.

Living donor liver transplantation (LDLT) has emerged as the most feasible alternative to cadaveric LT [66]. Over 2000 adult LDLTs using the right hepatic lobe have been performed throughout the world. LDLT was expected to change the therapeutical scenario of early HCC, particularly in countries with extremely large waiting times, or with local problems in the development of cadaveric transplantation. However, the absence of donor constraints is counteracted by the fact that the procedure is associated with a 0.5% mortality rate of the donor, and around 20% of morbidity. At present, only short series including HCC patients have been published, and their long-term outcome is still unknown [67]. Decision analyses have shown that LDLT is cost-effective for early HCC in comparison to cadaveric transplantation when waiting times exceed 7 months [68].

The standard selection criteria for cadaveric transplantation rely upon a strict size limitation. These criteria have been recently questioned. Five year survival above 70% was reported in patients with tumours up to 6.5 cm, or three less than 4.5 cm at pathological examination of the explanted liver [69]. Almost simultaneously, the Mount Sinai group reported results applying pre-operative expanded criteria (HCC >5 cm, intrahepatic portal involvement) [70]. Amongst 80 candidates, only 43 patients were effectively transplanted, with an overall 5 year survival of 44% (55% for 5-7 cm HCCs) [70]. These findings reinforce the idea that selection criteria for candidates should rely upon pre-established pre-operative staging, rather than on pathological findings. This may avoid unfair situations, particularly in the era of a shortage of donors. On the contrary, the principle of expanded criteria may apply to LDLT, considering the potential unlimited availability of this procedure as well as the almost non-existent waiting time. A set of new expanded criteria have been proposed and are expected to achieve 50% 5-year survival rates in patients with either single tumours 7 cm, three nodules 5 cm, five nodules 3 cm, or downstaging to conventional criteria after any therapy lasting more than 6 months [39].

Percutaneous treatments are considered as minimally invasive procedures that constitute the best medical option for non-surgical HCC [39,71-74] (Table 4). Tumour ablation is achieved by using chemical substances (alcohol, acetic acid) or by modifying the temperature of neoplastic cells (radiofrequency (RF), microwave, laser, and cryoablation). Percutaneous ethanol injection (PEI) achieves complete responses of 70% in HCC 3 cm, and is considered the gold standard treatment [39,71]. Selected Child-Pugh's A candidates with complete responses achieve 5 year survival rates of 50% [35,71]. RF ablation constitutes the most extensively used alternative to PEI, either through single or multiple cooled-tip electrode, or single electrode associated with J-hook needles, according to different engines available. It can be applied percutaneously, laparoscopically or during laparotomy, and is claimed to achieve at least the same objective responses as PEI, in significantly few sessions [72]. RF may provide marginal anti-tumoral benefits in tumours larger than 3 cm in diameter, where PEI is unable to disrupt the intratumoral septa. At present, the 5 year survival estimate from a cohort of 88 patients with tumours below 3.5 cm treated by RF is 33% [73]. Nevertheless, controversies regarding the beneficial role of RF ablation in comparison to PEI should be evaluated in the setting of a large RCT. The remaining treatments are either associated with a high rate of complications (cryoablation), have not been proven to represent any advantage (microwave coagulation) or are still experimental.

Predictors of treatment response are both the size (3 cm) [74] and the morphology of the lesion (well encapsulated vs. invasive) [71]. Kojiro [36] has recently described the pathological events that appeared in 106 resected small HCC of less than 2 cm in diameter. Among the most frequent type of tumour (the so-called distinctly nodular type), local metastases (located 10 mm from the nodule) were observed in 10% of cases, and microscopic portal invasion in up to 25% of cases. This provides the rationale to produce a 1 cm safety ring surrounding the tumour, particularly in tumours larger than 2 cm, regardless of the percutaneous therapy used.

The assessment of treatment response in HCC, particularly after percutaneous ablation, has been cumbersome. Definitions of responses are crucial to have common criteria for evaluation that enables comparison of different regimens in different trials. Response rates vary according to the criteria applied and the definitions established for late treatment failure and local recurrences. The conventional World Health Organization (WHO) criteria have been heterogeneously used [75]. Recently, the National Cancer Institute has adopted standardized criteria called Response Evaluation Criteria in Solid Tumours (RECIST) [76]. Simultaneously, the EASL Panel of Experts of HCC proposed using a modification of the WHO criteria [20] (Table 5). In brief, three considerations were adopted by this panel: (1) assessment of response by spiral CT or MRI 4 weeks after the treatment is considered the gold standard; (2) measurements of the diameter refer to the viable tumour (identified by the area enhanced during spiral CT), rather than changes in the overall diameter; (3) local recurrences should be defined as treatment failure. These guidelines are aimed at homogenizing the criteria used in the HCC field.

Table 5. Definitions of treatment response in solid tumors according to the WHO [75] and RECIST [76] guidelines, and amendments proposed by the EASL Panel of Experts for HCC [20]
Variables WHO RECIST EASL
Characteristics
Lesions assessed Measurable lesions (all lesions) Target lesions (five lesions or up to ten if more than one organ involved) Measurable lesions (all lesions)
Type of assessment Change in sum of the products of largest diameter and greatest perpendicular diameter Change in sum of largest diameter WHO criteria, but change in diameter of viable tumour volume (enhanced areas within the tumour by CT-scan) instead of diameter of the lesion
Type of response
Complete response (CR) Disappearance of all disease, confirmed at 4 weeks Disappearance of all target lesions, confirmed at 4 weeks WHO criteria, by spiral CT-scan or MRI 4 weeks after treatment*
Partial response (PR) 50% decrease from baseline, confirmed at 4 weeks 30% decrease from baseline, confirmed at 4 weeks *
Stable disease (SD) All other cases All other cases *
Progressive disease (PD) 25% increase in any lesion, or appearance of new lesions 20% increase over smallest sum observed, or new lesions *
Objective responses CR+PR CR+PR CR+PR

The majority of HCC patients in the West are diagnosed at advanced stages, and are treated by a heterogeneous armamentarium of procedures, which scarcely achieve complete responses [39,50]. The absence of any standard therapy for those individuals prompted the design of several phase II and prospective trials during the last 25 years. A recent systematic review of randomized trials for the treatment of unresectable HCC has identified 61 studies assessing all types of therapies [77]. This study shows that there are no large randomized trials including more than 1000 individuals, which are considered the main source of evidence in medicine, and only few of them include more than 100 patients, a sample size considered only acceptable for rare diseases. Amongst all the studies, only 26 included a control arm of conservative management, a request that has been considered essential in identifying survival benefits [20]. These studies have analyzed the effectiveness of embolization, chemoembolization, arterial or systemic chemotherapy, internal radiation with I131, proton beam radiation, hormonal compounds, immunotherapy and others (Table 4).

In the absence of large RCTs, solid evidence of survival benefits from a particular intervention may derive from meta-analysis of small RCTs [78]. This technique enables the integration of valid information and provides estimates of treatment effects when single studies themselves are not of sufficient size. The conditions needed to perform a meta-analysis should also be applied in the HCC field. First, there should be enough patients recruited in the setting of well-design RCTs assessing a particular intervention. This is the case for two treatments, embolization/chemoembolization and tamoxifen. All other approaches have been assessed in the setting of few small studies that do not have enough statistical power to provide solid conclusions. This only means that there are not enough data to establish the effectiveness of these therapies, rather than to conclude that they have no effect. This is the case for internal radiation with I131, octreotide or immunotherapy with IFN. The encouraging results of initial trials with IFN [79] and octreotide [80] have been counteracted by the negative results recently published [81,82]. Second, as stated before, only some studies have included an untreated control arm, which is mandatory to demonstrate survival benefits. Intra-arterial chemotherapy has been assessed in the setting of ten RCTs, but only two of them including 118 patients compare active vs. non-active therapy [83,84]. Despite this treatment achieving 20-30% partial response rates, no unquestionable conclusion regarding survival advantages can be derived, and further investigations are needed. Similarly, systemic chemotherapy has been assessed in nine RCTs, but only two include an untreated controlled arm [85,86]. Survival benefits are lacking with this treatment which only achieves partial responses in 5-15% of cases, and is associated with relevant toxicity.

Another issue that has to be taken into account is that treatments analyzed and the target population studied should be homogeneous. For instance, several treatments can be administered through the hepatic artery. Embolization/chemoembolization aims to disrupt the arterial blood flow to achieve necrosis through hypoxia (and deliver or not chemotherapy), whereas arterial chemotherapy - also called lipiodolization - only aims to deliver chemotherapeutic agents without blood flow obstruction. These two treatment approaches are too heterogeneous to be analyzed together, as has been proposed in previous meta-analysis [87].

Arterial embolization is the most widely used treatment for patients unsuitable for radical therapies. Embolization agents - usually gelatin - may be administered in association with selective intra-arterial chemotherapy mixed with lipiodol (chemoembolization). Doxorubicin, mitomycin and cisplatin are the most frequently used anti-tumoral drugs. Phase II studies have estimated partial responses ranging from 20% to 70%, either with or without associated chemotherapy. This brilliant anti-tumoral effect has stimulated several groups to conduct prospective trials. At present, there are seven RCTs comparing embolization vs. conservative management/suboptimal therapies, including a total of 516 patients [88-94]. Five of these studies assess chemoembolization with doxorubicin or cisplatin. Arterial embolization is an effective anti-tumoral therapy, which achieved partial responses in 15-55% of patients, and significantly delayed tumour progression and prevented neoplastic vascular involvement [88-94]. Survival benefits favouring chemoembolization were identified in the two most recent studies [92,93], although a trend had been previously recognized [90]. In one of these studies treatment response was the best independent predictor of survival [93]. Meta-analysis of all these data showed a beneficial survival effect of embolization/chemoembolization in comparison to the control group [77] (Table 6). The 2 year survival of the treated group was 41% vs. 27% in the control group. Significant differences in survival were identified with chemoembolization, but not with embolization alone [77].

Table 6. Meta-analysis of seven RCTs assessing embolization/chemoembolization vs. conservative management or suboptimal treatments as a primary treatment of non-surgical HCC [77] (random effects model (DerSimonian&Laird))
Authors (journal, year) Treatment arms (n)
Lin (Gastroenterology, 1988) [88] Embolization (21) vs. embolization+i.v. 5-FU (21) vs. i.v. 5-FU (21)
Pelletier (J Hepatol, 1990) [89] Chemoembolization (doxorubicin) (21) vs. control (21)
GETCH (N Engl J Med, 1995) [90] Chemoembolization (cisplatin) (50) vs. control (46)
Bruix (Hepatology, 1998) [91] Embolization+coils (40) vs. control (40)
Pelletier (J Hepatol, 1998) [92] Chemoembolization (cisplatin)+tamoxifen (37) vs. tamoxifen (36)
Lo (Hepatology, 2002) [93] Chemoembolization (cisplatin) (40) vs. control (39)
Llovet (Lancet, 2002) [94] Embolization (37) vs. chemoembolization (doxorubicin) (40) vs. control (35)
Meta-analysis [77]
Treatment RCTs Patients Odds ratio (95% CI) P value
Arterial embolization/chemoembolization
   2 year survival 6 503 0.53 (0.32-0.89) 0.017
Sensitivity analysis
   Chemoembolization 4 323 0.42 (0.20-0.88) 0.021
   Embolization 3 215 0.59 (0.29-1.2) 0.14
   High quality RCTs 5 440 0.53 (0.29-0.97) 0.039
   1 year survival 7 545 0.64 (0.41-1.01) 0.051

Survival benefits may be derived from a maintained response as time goes by. However, the benefits obtained in treatment responders should not be counteracted by treatment-induced liver failure. The most important severe adverse effect is related to ischemic damage to the non-tumoral liver. This is particularly striking in patients with liver decompensation or hepatic failure (Child-Pugh's B-C), in which this procedure is not recommended. Whether aetiology (viral hepatitis vs. alcoholic cirrhosis) influences treatment response is an issue that should be further explored. From all these data, it could be concluded that chemoembolization is recommended for a subgroup of patients with unresectable HCC, since it may benefit in terms of survival. The best candidates are those patients with preserved liver function and with asymptomatic multinodular tumours without vascular invasion or extrahepatic spread.

The role of anti-estrogen agents in the treatment of advanced HCC has also been controversial. At present, seven RCTs have been published comparing tamoxifen vs. conservative management, including overall 898 patients. Meta-analysis of these studies shows the lack of effectiveness of tamoxifen in terms of anti-tumoral effect and survival [77]. Sensitivity analysis suggests that the positive benefits of tamoxifen claimed during the early nineties [95,96] were the result of low quality scale trials, and have never been confirmed by further well-designed placebo-controlled RCTs [97].

Despite the improvement in the overall survival of the HCC population during the last years, the prognosis of the disease is still discouraging. Transplantation offers a perspective of long-term survival, but can only be applied to a minority of patients in developed countries. Other curative options such as resection or percutaneous ablation do not preclude the occurrence of recurrences after 3 years. This scenario might change in the coming years for the following reasons. Improved anti-viral treatments of HCV and HBV infections will decrease the proportion of patients developing cirrhosis, leading to a reduction of long-term HCC development. In established cirrhosis, improvements might come from chemoprevention and early molecular diagnosis, and will mostly benefit high risk individuals. Nowadays, it is already possible to identify those cirrhotic patients at high risk of HCC. In the future, more accurate knowledge of risk factors will help to define subgroups of patients according to their expected incidence of the disease. In this ideal scenario, chemoprevention and surveillance could be tailored according to the expected risk. Prospective trials should clarify whether chemopreventive or anti-viral treatments are effective in inducing a deletion of minute malignant clones (clonal deletion), as is claimed to happen with retinoids. This might effectively prevent HCC. Finally, recognition of tumours of 1 cm in size may be possible with the improvements in surveillance, recall policies and imaging technology. This is essential to establish the molecular profile of preneoplastic and carcinoma-in-situ lesions, which may provide a molecular diagnosis of HCC. Molecular biology applied to liver samples and proteomic studies in the peripheral blood are promising tools for such an approach.

Treatments of HCC will also improve. It is expected that the advancement of HCC diagnosis together with the application of less invasive surgical approaches, such as laparoscopic resection, may increase the application of this procedure. Similarly, the number of patients transplanted will be slightly expanded probably by split LT and LDLT. Neoadjuvant procedures used to reduce tumoral growth during waiting time must be more objectively evaluated and compared. There are presently some indications that percutaneous treatment could be helpful in that setting, at least for small tumours. Advancements in percutaneous treatments and extension of their indications are widely expected. Despite new technical variants such as intra-arterial ethanol injection, PEI has reached its limits. RF may provide additional benefits in tumours larger than 3 cm in size, where PEI is unable to disrupt the intratumoral septa. It will also be possible to measure the temperature inside the liver using MRI and presumably future procedures will be performed under open MRI guidance with real time thermometric monitoring. In a longer term, extracorporeal devices could allow thermal destruction by ultrasound.

Despite extended screening some patients will still be diagnosed at an advanced stage, and a search for effective treatments in these situations is needed. We must remember that HCC usually occurs in a cirrhotic liver, and both diseases may respond differently to treatments. For instance, agents favouring hepatocyte apoptosis could be detrimental in patients with active inflammatory disease. Increased angiogenesis, even if pre-eminent in most tumours, is also present and presumably beneficial in the hypoxic liver of cirrhotic patients. Thus, results with new anti-angiogenic agents aimed at blocking vascular endothelial growth factor should be tested with caution. Immunotherapy using T cells sensitized to tumoral antigens has been used to prevent recurrences after resection [59], and could presumably be extended to other situations. Immuno-gene therapy through gene transfer into stem cells to induce tumour specific effector T cells has shown promising results in solid neoplasms [98]. Other types of gene therapy using intratumoral immunostimulatory cytokines in humans give the rationale to investigate vector management and toxicity. The need for large RCTs with survival as the primary end-point is more than ever mandatory.

The authors who have taken part in this study have not a relationship with the manufacturers of the drugs involved either in the past or present and did not receive funding from the manufacturers to carry out their research.
------------------------------------------------------------------------

Josep M. Llovet is a recipient of a contract from Programa `Ramon y Cajal' (IDIBAPS, Ministerio de Ciencia y Tecnología). This review was partially supported by a grant from the Fundació Marató TV3 (Catalonia, Spain).

[1]
F.X. Bosch, J. Ribes, J. Borras, Epidemiology of primary liver cancer, Semin Liver Dis 19 (1999) 271-285
[2]
S. Chevret, J.C. Trinchet, D. Mathieu, A.A. Rached, M. Beaugrand, C. Chastang, A new prognostic classification for predicting survival in patients with hepatocellular carcinoma, J Hepatol 31 (1999) 133-141
[3]
H.B. El Serag, A.C. Mason, Rising incidence of hepatocellular carcinoma in the United States, N Engl J Med 340 (1999) 745-750
[4]
S. Deuffic, L. Buffat, T. Poynard, A.J. Valleron, Modeling the hepatitis C virus epidemic in France, Hepatology 29 (1999) 1596-1601
[5]
G. Garcia-Tsao, Current management of the complications of cirrhosis and portal hypertension: variceal hemorrhage, ascites, and spontaneous bacterial peritonitis, Gastroenterology 120 (2001) 726-748
[6]
M.H. Chang, C.J. Chen, M.S. Lai, H. Hsu, T.C. Wu, M.S. Kong et al., Universal hepatitis B vaccination in Taiwan and the incidence of hepatocellular carcinoma in children, N Engl J Med 336 (1997) 1855-1859
[7]
M.H. Chang, W.Y. Shau, C.J. Chen, T.C. Wu, M.S. Kong, D.C. Liang et al., The Taiwan Childhood Hepatoma Study Group. Hepatitis B vaccination and hepatocellular carcinoma rates in boys and girls, J Am Med Assoc 284 (2000) 3040-3042
[8]
M. Shibata, T. Morizane, T. Uchida, T. Yamagami, Y. Onozuka, M. Nakano et al., Irregular regeneration of hepatocytes and risk of hepatocellular carcinoma in chronic hepatitis and cirrhosis with hepatitis-C-virus infection, Lancet 351 (1998) 1773-1777
[9]
Y. Ueno, M. Moriyama, T. Uchida, Y. Arakawa, Irregular regeneration of hepatocytes is an important factor in the hepatocarcinogenesis of liver disease, Hepatology 33 (2001) 357-362
[10]
N. Ganne-Carrie, C. Chastang, F. Chapel, C. Munz, D. Pateron, M. Sibony et al., Predictive score for the development of hepatocellular carcinoma and additional value of liver large cell dysplasia in Western patients with cirrhosis, Hepatology 23 (1996) 1112-1118
[11]
C. Camma, M. Giunta, P. Andreone, A. Craxi, Interferon and prevention of hepatocellular carcinoma in viral cirrhosis: an evidence-based approach, J Hepatol 34 (2001) 606-609
[12]
S. Nishiguchi, S. Shiomi, S. Nakatani, T. Takeda, K. Fukuda, A. Tamori et al., Prevention of hepatocellular carcinoma in patients with chronic active hepatitis C and cirrhosis, Lancet 357 (2001) 196-197
[13]
D.C. Valla, M. Chevallier, P. Marcellin, J.L. Payen, C. Trepo, M. Fonck et al., Treatment of hepatitis C virus-related cirrhosis: a randomized, controlled trial of interferon alfa-2b versus no treatment, Hepatology 29 (1999) 1870-1875
[14]
Y. Muto, H. Moriwaki, M. Ninomiya, S. Adachi, A. Saito, K.T. Takasaki, Prevention of second primary tumors by an acyclic retinoid, polyprenoic acid, in patients with hepatocellular carcinoma. Hepatoma Prevention Study Group, N Engl J Med 334 (1996) 1561-1567
[15]
Y. Muto, H. Moriwaki, A. Saito, Prevention of second primary tumors by an acyclic retinoid in patients with hepatocellular carcinoma, N Engl J Med 340 (1999)1046-1047
[16]
M. Sherman, Screening for hepatocellular carcinoma, Bailliere's Best Pract Res Clin Gastroenterol 13 (1999) 623-635
[17]
N. Ganne-Carrie, S. Chevret, J.C. Barbare, C. Chaffaud, V. Grando, A.M. Vogt et al., Practical screening and early treatment of hepatocellular carcinoma. Results of a French survey, Gastroenterol Clin Biol 26 (2002) 570-577
[18]
N. Chalasani, J.C. Horlander, Sr., A. Said, H. Hoen, K.K. Kopecky, S.M. Stockberger, Jr et al., Screening for hepatocellular carcinoma in patients with advanced cirrhosis, Am J Gastroenterol 94 (1999) 2988-2993
[19]
M. Sherman, Alphafetoprotein: an obituary, J Hepatol 34 (2001) 603-605
[20]
J. Bruix, M. Sherman, J.M. Llovet, M. Beaugrand, R. Lencioni, E. Christensen et al., Clinical management on hepatocellular carcinoma. Conclusions of the Barcelona-2000 EASL Conference, J Hepatol 35 (2001) 421-430
[21]
H. Oka, N. Kurioka, K. Kim, T. Kanno, T. Kuroki, Y. Mizoguchi et al., Prospective study of early detection of hepatocellular carcinoma in patients with cirrhosis, Hepatology 12 (1990) 680-687
[22]
M. Colombo, R. De Franchis, E. Del Ninno, A. Sangiovanni, C. De Fazio, M. Tommasini et al., Hepatocellular carcinoma in Italian patients with cirrhosis, N Engl J Med 325 (1991) 675-680
[23]
D. Pateron, N. Ganne-Carrie, J.C. Trinchet, M.H. Aurousseau, F. Mal, C. Meicler et al., Prospective study of screening for hepatocellular carcinoma in Caucasian patients with cirrhosis, J Hepatol 20 (1994) 65-71
[24]
M. Cottone, M. Turri, M. Caltagirone, P. Parisi, A. Orlando, G. Fiorentino et al., Screening for hepatocellular carcinoma in patients with Child's A cirrhosis: an 8-year prospective study by ultrasound and alphafetoprotein, J Hepatol 21 (1994) 1029-1034
[25]
M. Zoli, D. Magalotti, G. Bianchi, C. Gueli, G. Marchesini, E. Pisi et al., Efficacy of a surveillance program for early detection of hepatocellular carcinoma, Cancer 78 (1996) 977-985
[26]
L. Bolondi, S. Sofia, S. Siringo, S. Gaiani, A. Casali, G. Zironi et al., Surveillance programme of cirrhotic patients for early diagnosis and treatment of hepatocellular carcinoma: a cost-effectiveness analysis, Gut 48 (2001) 251-259
[27]
J. Henrion, E. Libon, S. De Maeght, M. Schapira, J.M. Ghilain, J.M. Maisin et al., Surveillance for hepatocellular carcinoma: compliance and results according to the aetiology of cirrhosis in a cohort of 141 patients, Acta Gastroenterol Belg 63 (2000) 5-9
[28]
V. Paradis, I. Laurendeau, M. Vidaud, P. Bedossa, Clonal analysis of macronodules in cirrhosis, Hepatology 28 (1998) 953-958
[29]
F. Durand, F. Durand, J.M. Regimbeau, J. Belghiti, A. Sauvanet, V. Vilgrain et al., Assessment of the benefits and risks of percutaneous biopsy before surgical resection of hepatocellular carcinoma, J Hepatol 35 (2001) 254-258
[30]
T. Murakami, T. Kim, S. Takahashi, H. Nakamura, Hepatocellular carcinoma: multidetector row helical CT, Abdom Imaging 27 (2002) 139-146
[31]
T. Kamura, M. Kimura, K. Sakai, T. Ichida, H. Seki, S. Yamamoto et al., Small hypervascular hepatocellular carcinoma versus hypervascular pseudolesions: differential diagnosis on MRI, Abdom Imaging 27 (2002) 315-324
[32]
K. Okuda, T. Ohtsuki, H. Obata, M. Tomimatsu, N. Okazaki, H. Haregawwa et al., Natural history of hepatocellular carcinoma and prognosis in relation to treatment, Cancer 56 (1985) 918-928
[33]
V. Mazzaferro, E. Regalia, R. Doci, S. Andreola, A. Pulvirenti, Liver transplantation for treatment of small hepatocellular carcinomas in patients with cirrhosis, N Engl J Med 334 (1996) 693-699
[34]
J.M. Llovet, J. Fuster, J. Bruix, Intention-to-treat analysis of surgical treatment for early hepatocellular carcinoma: resection versus transplantation, Hepatology 39 (1999) 1434-1440
[35]
S. Arii, Y. Yamaoka, S. Futagawa, K. Inoue, K. Kobayashi, M. Kojiro et al., Results of surgical and nonsurgical treatment for small-sized hepatocellular carcinomas: a retrospective and nationwide survey in Japan, Hepatology 32 (2000) 1224-1229
[36]
M. Kojiro, The evolution of pathologic features of hepatocellular carcinoma, In: E. Tabor, ed., Viruses and liver cancer, (2002) pages 113-122 (Elsevier, Amsterdam)
[37]
F. Mion, L. Grozel, O. Boillot, P. Paliard, F. Berger, Adult cirrhotic liver explants: precancerous lesions and undetected small hepatocellular carcinomas, Gastroenterology 11 (1996) 1587-1592
[38]
Y.J. Chen, S.H. Yeh, J.T. Chen, C.C. Wu, M.T. Hsu, S.F. Tsai et al., Chromosomal changes and clonality relationship between primary and recurrent hepatocellular carcinoma, Gastroenterology 119 (2000) 431-440
[39]
J. Bruix, J.M. Llovet, Prognostic prediction and treatment strategy in hepatocellular carcinoma, Hepatology 35 (2002) 519-524
[40]
L. Barbara, G. Benzi, S. Gaiani, F. Fusconi, G. Zironi, S. Siringo et al., Natural history of small untreated hepatocellular carcinoma in cirrhosis: a multivariate analysis of prognostic factors of tumor growth rate and patient survival, Hepatology 16 (1992) 132-137
[41]
H.B. El Serag, A. Mason, C. Key, Trends in survival of patients with hepatocellular carcinoma between 1997 and 1996 in United States, Hepatology 33 (2001) 62-65
[42]
J.M. Llovet, J. Bustamante, A. Castells, R. Vilana, M.C. Ayuso, M. Sala et al., Natural history of untreated nonsurgical hepatocellular carcinoma: rationale for the design and evaluation of therapeutic trials, Hepatology 29 (1999) 62-67
[43]
CLIP, Prospective validation of the CLIP score: a new prognostic system for patients with cirrhosis and hepatocellular carcinoma, Hepatology 31 (2000) 840-845
[44]
J.M. Llovet, C. Bru, J. Bruix, Prognosis of hepatocellular carcinoma: the BCLC staging classification, Semin Liver Dis 19 (1999) 329-338
[45]
T.W. Leung, A.M. Tang, B. Zee, W.Y. Lau, P.B. Lai, K.L. Leung et al., Construction of the Chinese University Prognostic Index for hepatocellular carcinoma and comparison with the TNM staging system, the Okuda staging system, and the Cancer of the Liver Italian Program staging system: a study based on 926 patients, Cancer 94 (2002) 1760-1769
[46]
F. Greene, D. Page, I. Fleming, A. Fritz, C. Balch, D. Haller, M. Morrow, In: AJCC cancer staging handbook, (6th ed., edition), (2002) pages 145-153 (Springer-Verlag, New York)
[47]
J.M. Llovet, E. Moitinho, M. Sala, R. Bataller, M.P. Rodríguez-Iglesias, A. Castells et al., Prevalence and prognostic value of hepatocellular carcinoma in cirrhotic patients with spontaneous bacterial peritonitis, J Hepatol 33 (2000) 423-429
[48]
I. Fleming, AJCC/TNM cancer staging, present and future, J Clin Oncol 77 (2001) 233-236
[49]
S. Ueno, G. Tanabe, K. Sako, T. Hiwaki, H. Hokotate, Y. Fukukura et al., Discrimination value of the new western prognostic system (CLIP score) for hepatocellular carcinoma in 662 Japanese patients, Hepatology 34 (2001) 529-534
[50]
A.S. Befeler, A.M. Di Bisceglie, Hepatocellular carcinoma: diagnosis and treatment, Gastroenterology 122 (2002) 1609-1619
[51]
R. Izumi, K. Shimizu, T. Li, M. Yagi, O. Matsui, A. Nonomura et al., Prognostic factors of hepatocellular carcinoma in patients undergoing hepatic resection, Gastroenterology 106 (1994) 720-727
[52]
J.M. Llovet, J. Bruix, J. Fuster, A. Castells, J.C. García-Valdecasas, L. Grande et al., Liver transplantation for treatment of small hepatocellular carcinoma: the TNM classification does not have prognostic power, Hepatology 27 (1998) 1572-1577
[53]
J. Vauthey, G. Lauwers, N. Esnaola, K. Do, J. Belghiti, N. Mirza et al., Simplified staging for hepatocellular carcinoma, J Clin Oncol 20 (2002) 1527-1536
[54]
The Liver Cancer Study Group of Japan, Predictive factors for long term prognosis after partial hepatectomy for patients with hepatocellular carcinoma in Japan, Cancer 74 (1994) 2272-2280
[55]
J. Bruix, A. Castells, J. Bosch, F. Feu, J. Fuster, J.C. Garcia-Pagan et al., Surgical resection of hepatocellular carcinoma in cirrhotic patients: prognostic value of preoperative portal pressure, Gastroenterology 111 (1996) 1018-1022
[56]
B. Ringe, R. Pichlmayr, C. Wittekind, G. Tusch, Surgical treatment of hepatocellular carcinoma: experience with liver resection and transplantation in 198 patients, World J Surg 15 (1991) 270-285
[57]
S. Yamasaki, H. Hasegawa, H. Kinoshita, M. Furukawa, S. Imaoka, K. Takasaki et al., A prospective randomized trial of the preventive effect of pre-operative transcatheter arterial embolization against recurrence of hepatocellular carcinoma, Jpn J Cancer Res 87 (1996) 206-211
[58]
W.Y. Lau, T.W. Leung, S.K. Ho, M. Chan, D. Machin, J. Lau et al., Adjuvant intra-arterial iodine-131-labelled lipiodol for resectable hepatocellular carcinoma: a prospective randomised trial, Lancet 353 (1999) 797-801
[59]
T. Takayama, T. Sekine, M. Makuuchi, S. Yamasaki, T. Kosuge, J. Yamamoto et al., Adoptive immunotherapy to lower postsurgical recurrence rates of hepatocellular carcinoma: a randomised trial, Lancet 356 (2000) 802-807
[60]
P. Moreno, E. Jaurrieta, J. Figueras, C. Benasco, A. Rafecas, J. Fabregat et al., Orthotopic liver transplantation: treatment of choice in cirrhotic patients with hepatocellular carcinoma?, Transplant Proc 27 (1995) 2296-2298
[61]
H. Bismuth, P.E. Majno, R. Adam, Liver transplantation for hepatocellular carcinoma, Semin Liver Dis 19 (1999) 311-322
[62]
J. Neuberger, Liver transplantation, J Hepatol 32 (Suppl 1) (2000) 198-207
[63]
United Network for Organ Sharing. Annual report 2000. http://www.unos.org/data.
[64]
R. Freeman, R. Wiesner, A. Harper, S. McDiarmid, J. Lake, E. Edwards et al., The new liver allocation system: moving toward evidence-based transplantation policy, Liver Transpl 8 (2002) 851-858
[65]
J.M. Llovet, X. Mas, J. Aponte, J. Fuster, M. Navasa, E. Christensen et al., Cost-effectiveness of adjuvant therapy for hepatocellular carcinoma before liver transplantation, Gut 50 (2002) 123-128
[66]
J.F. Trotter, M. Wachs, G.T. Everson, I. Kam, Adult-to-adult transplantation of the right hepatic lobe from a living donor, N Engl J Med 346 (2002) 1074-1082
[67]
C.M. Miller, G.E. Gondolesi, S. Florman, C. Matsumoto, L. Munoz, T. Yoshizumi et al., One hundred nine living donor liver transplants in adults and children: a single-center experience, Ann Surg 234 (2001) 301-311
[68]
F. Sarasin, P. Majno, J.M. Llovet, J. Mentha, J. Bruix, A. Hadengue, Liver donor liver transplantation for early hepatocellular carcinoma: a cost-effectiveness perspective, Hepatology 33 (2001) 1073-1079
[69]
F.Y. Yao, L. Ferrell, N.M. Bass, J.J. Watson, P. Bacchetti, A. Venook et al., Liver transplantation for hepatocellular carcinoma: expansion of tumor size limits does not adversely impact survival, Hepatology 33 (2001) 1394-1403
[70]
S. Roayaie, J.S. Frischer, S.H. Emre, T.M. Fishbein, P.A. Sheiner, M. Sung et al., Long-term results with multimodal adjuvant therapy and liver transplantation for the treatment of hepatocellular carcinomas larger than 5 centimeters, Ann Surg 235 (2002) 533-539
[71]
T. Livraghi, A. Giorgio, G. Marin, A. Salmi, I. de Sio, L. Bolondi et al., Hepatocellular carcinoma and cirrhosis in 746 patients: long-term results of percutaneous ethanol injection, Radiology 197 (1995) 101-108
[72]
S. Okada, T. Okusaka, H. Ueno, Percutaneous interstitial therapy for hepatocellular carcinoma, In: M. Colombo and E. Schiff, eds., Update in hepatology: hepatocellular carcinoma, (2002) pages 59-67 (UTET Periodici, Milan)
[73]
L. Buscarini, E. Buscarini, M. Di Stasi, D. Vallisa, P. Quaretti, A. Rocca, Percutaneous radiofrequency ablation of small hepatocellular carcinoma: long-term results, Eur Radiol 11 (2001) 914-921
[74]
R. Vilana, J. Bruix, C. Bru, C. Ayuso, M. Sole, J. Rodes, Tumor size determines the efficacy of percutaneous ethanol injection for the treatment of small hepatocellular carcinoma, Hepatology 16 (1992) 353-357
[75]
A.B. Miller, B. Hoogstraten, M. Staquet, A. Winkler, Reporting results of cancer treatment, Cancer 47 (1981) 207-214
[76]
P. Therasse, S.G. Arbuck, E.A. Eisenhauer, J. Wanders, R.S. Kaplan, L. Rubinstein et al., New guidelines to evaluate the response to treatment in solid tumors, J Natl Cancer Inst 92 (2000) 205-216
[77]
J.M. Llovet, J. Bruix, Systematic review of randomized controlled trials for unresectable hepatocellular carcinoma. Chemoembolization improves survival, Hepatology 37 (2003)
[78]
J. Lau, J.P. Loannidis, C.H. Schmid, Quantitative synthesis in systematic reviews, Ann Intern Med 127 (1997) 820-826
[79]
C.L. Lai, J.Y. Lau, P.C. Wu, H. Ngan, H.T. Chung, S.J. Mitchell et al., Recombinant interferon-alpha in inoperable hepatocellular carcinoma: a randomized controlled trial, Hepatology 17 (1993) 389-394
[80]
E. Kouroumalis, P. Skordilis, K. Thermos, A. Vasilaki, J. Moschandrea, O.N. Manousos, Treatment of hepatocellular carcinoma with octreotide: a randomised controlled study, Gut 42 (1998) 442-447
[81]
J.M. Llovet, M. Sala, L. Castells, Y. Suarez, R. Vilana, L. Bianchi et al., Randomized controlled trial of interferon treatment for advanced hepatocellular carcinoma, Hepatology 31 (2000) 1-5
[82]
M. Yuen, R. Poon, C.h. Lai, S.h. Fan, C.h. Lo, K. Wong et al., A randomized placebo-controlled study of long-acting octeotride for the treatment of advanced hepatocellular carcinoma, Hepatology 36 (2002) 687-691
[83]
M.V. Madden, J.E. Krige, S. Bailey, S.J. Beningfield, C. Geddes, I.D. Werner et al., Randomised trial of targeted chemotherapy with lipiodol and 5-epidoxorubicin compared with symptomatic treatment for hepatoma, Gut 34 (1993) 1598-1600
[84]
Y.H. Chung, I.H. Song, B.C. Song, G.C. Lee, M.S. Koh, H.K. Yoon et al., Combined therapy consisting of intraarterial cisplatin infusion and systemic interferon-alpha for hepatocellular carcinoma patients with major portal vein thrombosis or distant metastasis, Cancer 88 (2000) 1986-1991
[85]
C.L. Lai, P.C. Wu, G.C. Chan, A.S. Lok, H.J. Lin, Doxorubicin versus no antitumor therapy in inoperable hepatocellular carcinoma. A prospective randomized trial, Cancer 62 (1988) 479-483
[86]
T. Ishikawa, T. Ichida, S. Sugitani, Y. Tsuboi, T. Genda, S. Sugahara et al., Improved survival with oral administration of enteric-coated tegafur/uracil for advanced stage IV-A hepatocellular carcinoma, J Gastroenterol Hepatol 16 (2001) 452-459
[87]
P. Mathurin, O. Rixe, N. Carbonell, B. Bernard, P. Cluzel, M.F. Bellin et al., Review article: overview of medical treatments in unresectable hepatocellular carcinoma - an impossible meta-analysis?, Aliment Pharmacol Ther 12 (1998) 111-126
[88]
D.Y. Lin, Y.F. Liaw, T.Y. Lee, C.M. Lai, Hepatic arterial embolization in patients with unresectable hepatocellular carcinoma - a randomized controlled trial, Gastroenterology 94 (1988) 453-456
[89]
G. Pelletier, A. Roche, O. Ink, M.L. Anciaux, S. Derhy, P. Rougier et al., A randomized trial of hepatic arterial chemoembolization in patients with unresectable hepatocellular carcinoma, J Hepatol 11 (1990) 181-184
[90]
Group d'Etude et de Traitment du Carcinome Hépatocellulaire, A comparison of lipiodol chemoembolization and conservative treatment for unresectable hepatocellular carcinoma, N Engl J Med 332 (1995) 1256-1261
[91]
J. Bruix, J.M. Llovet, A. Castells, X. Montana, C. Bru, M.C. Ayuso et al., Transarterial embolization versus symptomatic treatment in patients with advanced hepatocellular carcinoma: results of a randomized, controlled trial in a single institution, Hepatology 127 (1998) 1578-1583
[92]
G. Pelletier, M. Ducreux, F. Gay, M. Luboinski, H. Hagege, T. Dao et al., Treatment of unresectable hepatocellular carcinoma with lipiodol chemoembolization: a multicenter randomized trial, J Hepatol 29 (1998) 129-134
[93]
C.M. Lo, H. Ngan, W.K. Tso, C.L. Liu, C.M. Lam, R.T. Poon et al., Randomized controlled trial of transarterial lipiodol chemoembolization for unresectable hepatocellular carcinoma, Hepatology 35 (2002) 1164-1171
[94]
J.M. Llovet, M.I. Real, X. Montana, R. Planas, S. Coll, J. Aponte et al., Arterial embolisation or chemoembolisation versus symptomatic treatment in patients with unresectable hepatocellular carcinoma: a randomised controlled trial, Lancet 359 (2002) 1734-1739
[95]
S. Elba, V. Giannuzzi, G. Misciagna, O.G. Manghisi, Randomized controlled trial of tamoxifen versus placebo in inoperable hepatocellular carcinoma, Ital J Gastroenterol 26 (1994) 66-68
[96]
F.J. Martinez Cerezo, A. Tomas, L. Donoso, J. Enriquez, C. Guarner, J. Balanzo et al., Controlled trial of tamoxifen in patients with advanced hepatocellular carcinoma, J Hepatol 20 (1994) 702-706
[97]
A. Castells, J. Bruix, C. Bru, C. Ayuso, M. Roca, L. Boix et al., Treatment of hepatocellular carcinoma with tamoxifen: a double-blind placebo-controlled trial in 120 patients, Gastroenterology 109 (1995) 917-922
[98]
C. Bordignon, M. Roncarolo, Therapeutic applications for hematopoietic stem cell gene transfer, Nat Immunol 3 (2002) 318-321