Correspondence Prof. Massimo Colombo, Department of Medicine, Division of Gastroenterology, Fondazione IRCCS Cà Granda Ospedale Maggiore Policlinico, Università degli Studi di Milano, Via Francesco Sforza 35, 20122 Milan, Italy Tel: ++39 025 503 5432 Fax:++39 025 032 0410 e-mail: email@example.com
In the late 1990s, the development of a standardized staging system for hepatocellular carcinoma (HCC) (Barcelona clinical liver cancer) using evidence-based data made it possible to stratify HCC patients to receive appropriate treatment with substantial clinical benefit. In the past, treatment failures were often because of faulty staging of tumour disease. This was the case for hepatic resection, as patient survival after surgery was endangered by tumour recurrence (70% at 5 years) as well as the risk of unresolved post-surgical hepatic decompensation, a complication that can be avoided by restricting resection to patients with low portal hypertension and normal serum bilirubin. It was also true for liver transplantation, whose outcome clearly improved when stringent patient selection criteria were adopted based on the volume and number of tumour nodules and venous invasion (so-called Milan criteria), resulting in <15% risk of tumour recurrence and a 75% 5-year survival rate. Switching from an empirical to an evidence-based patient selection approach improved the outcome of local ablation therapies, mainly because ideal candidates were identified, resulting in>50% 5-year survival rates, similar to those obtained with hepatic resection. While the prognosis is still poor in patients with advanced HCC, the recent availability of multikinase inhibitors has generated hope for improving survival in these patients.
The treatment of hepatocellular carcinoma (HCC) has progressed in the last few decades from a single option of surgical resection in selected patients with small tumours to an array of choices including orthotopic liver transplantation, locoregional ablation procedures, transarterial chemoembolization in patients who do not fit the indication for resection (1), and more recently sorafenib, a drug that increases survival in advanced HCC patients for whom no therapy was previously available (2). A better knowledge of the natural history of the tumour as well as the development of clinically based staging systems, such as the Barcelona clinical liver cancer (BCLC) classification, which stratifies patients according to the stage of the tumour and liver disease, has meant that life expectancies can be confidently predicted and the appropriate treatment can be chosen according to stage (1, 3) (Table 1, Fig. 1). Not surprisingly, the BCLC system has been endorsed by both the American and the European associations for the study of the liver as a backbone for both guidelines in clinical practice and study designs to validate new treatment modalities (1, 4). There is no doubt that in the past many treatment failures were the consequence of faulty staging of patients and empirical decisions, resulting in inappropriate treatment.
Table 1. The Barcelona clinic liver cancer staging classification for hepatocellular carcinoma
Tumour volume, number and invasiveness
BCLC, Barcelona clinic liver cancer.
A – early
Single<5 cm or three nodules<3 cm each
A and B
50–75% at 5 year
B – intermediate
A and B
C – advanced
Vascular invasion and/or extrahepatic spread
A and B
D – end-stage
Any of the above
Early hepatocellular carcinoma (Barcelona clinical liver cancer stages 0 and A)
The current definition of early HCC is a tumour <2 cm in size confined to the liver in a patient with compensated cirrhosis (5). This definition still includes a heterogeneous set of tumours, such as very early HCC, i.e. a vaguely nodular tumour that usually escapes diagnosis with dynamic imaging techniques such as contrast-enhanced ultrasound, computerized tomography scan or gadolinium magnetic resonance, and early HCC, i.e. a distinctly nodular tumour that is detectable by contrast imaging because of arterial neovascularization. In clinical practice, however, the definition of early HCC is broader, including patients with a single nodule, <5 cm or up to three nodules each 3 cm or less (Milan criteria) (3). Because the BCLC system also includes the clinical parameters of liver function, early-stage HCC is further stratified into four classes according to the number and size of the tumours and predictors of survival after surgical resection, such as portocaval gradient and serum bilirubin level (Fig. 2). The early application of radical treatment while preserving compensated liver function is the therapeutic paradigm for HCC according to the volume and number of nodules. This predicts vascular tumour dissemination as well as late mortality because of tumour recurrence after radical therapies (6, 7). Tumour nodules ≤2 cm, which represent 80% of all tumours detected in cirrhotic patients being monitored by ultrasound, have a low risk of extranodal spread (8), and can theoretically be treated by radical therapies, such as local ablation, resection and transplantation. The final therapeutic decision, however, is based on patient characteristics such as age, liver status, presence of comorbidities, potential for improving underlying liver disease and the length of the transplantation waiting list. Carefully selected patients with a tumour ≤2 cm and compensated cirrhosis benefit equally from all radical therapies in terms of survival (>70% at 5 years), but with different costs and risks of intervention-related mortality. The latter is very low for percutaneous ablation, while it is 1–3% for resection and 10% for liver transplantation (4).
Resection and local ablation
Certain results have shown that resection is not cost-effective in patients with a portocaval gradient above 10 mmHg and/or >1 mg serum bilirubin, since this includes a high risk of irreversible post-operative clinical decompensation and reduced survival (<50% at 5 years) (9). Percutaneous local ablation with radiofrequency (RFA) or ethanol injection (PEI) is less effective in tumours above 3 cm, because of a substantial risk of tumour recurrence on the periphery of the nodule (10, 11). Both resection and percutaneous ablation can be proposed to patients with up to three nodules of 3 cm or less, but with less encouraging survival rates than in patients with a single ≤2 cm nodule, once again because of the increased risk of post-operative recurrence (12) (Fig. 3).
Patients within the Milan criteria (a single nodule, <5 cm or up to three nodules each 3 cm or less) treated with orthotopic liver transplantation have an overall predicted survival of 75% at 5 years (13). There may be an increased risk of tumour recurrence in patients with >2 cm nodules and those not responding to tumour ablation by RFA or transarterial chemoembolization (TACE) while on the waiting list (bridge therapy), but this assumption is not evidence based. Although perprotocol survival rates in transplanted patients are better than in similar stage patients treated by resection, the survival rates are less encouraging when calculated by intention to treat because tumour progression occurs in patients while on the waiting list, leading to them being taken off the list or increased rates of tumour recurrence after transplantation (14). Although RFA and TACE are extensively prescribed to halt the progression of HCC in patients listed for liver transplantation, their cost-effectiveness is the subject of intense debate (15, 16). This is also true for interferon therapy to prevent tumour recurrence in patients treated by hepatic resection or local ablation. In a recent meta-analysis, adjuvant therapy with interferon was shown to be associated with increased patient survival rather than the prevention of tumour recurrence, suggesting that patients benefitted more from the control of hepatitis than from the prevention of new primary tumours (17). While liver transplantation is felt to be the best therapeutic option for patients with early HCC while also preventing the late-onset complications of cirrhosis, this option is limited by the few number of organ donations worldwide. As survival figures improve in transplanted patients and indications for this treatment are extended, living donor liver transplantation may seem promising, although it is currently associated with high complication rates (>30%) in the recipients and a small but not negligible, risk of donor death (0.5%) (18). In most countries, including Italy, the limited number of donor livers does not favour strategies to extend the oncological criteria (Milan criteria) for transplantation in patients with HCC, since extending criteria is associated with an increased risk of tumour recurrence and shortened survival (19) (Table 2). At present, both hepatitis B- and hepatitis C-related HCC patients are considered for transplantation whenever they correspond to Milan criteria and have no general contraindications to transplantation. However, unlike hepatitis B, since the survival rates of patients transplanted for end-stage hepatitis C are reduced because of severe recurrent HCV, the aetiology of underlying liver disease should be taken into consideration when placing HCC patients on the list for liver transplantation (20).
Table 2. Survival rates of early hepatocellular carcinoma treated by liver transplantation within Milan criteria
To improve the outcome of RFA in patients with tumours >2 cm, sequential treatment with RFA and TACE/TAE has been suggested; however, there is no evidence that the survival rates are better with this approach than with RFA alone. Multimodal treatment of HCC patients is a sound approach in patients with more than one nodule who do not reach the end-point of a radical cure with one type of treatment alone. Since results of liver transplantation are better than hepatic resection for tumour recurrence, liver transplantation may be used as rescue therapy in these patients who are found to have microscopic vascular invasion during tumour resection (21).
Intermediate hepatocellular carcinoma (Barcelona clinical liver cancer stage B)
Patients with compensated cirrhosis, a tumour stage beyond the Milan criteria and neither macroscopic vascular invasion by tumour cells nor symptoms of neoplastic disease, are rarely considered for radical treatment. A meta-analysis of seven studies (22) showed that repeat treatment with TACE was associated with improved survival of an average of 16–20 months in patients with intermediate HCC (Fig. 4). As a result, TACE has became the standard of care in BCLC-B patients, with the recommendations to restrict treatment to patients with Child–Pugh A or B cirrhosis, with or without small oesophageal varices to minimize the risk of hepatic decompensation that may follow procedure-induced liver ischaemia. Nevertheless, it should be noted that the two studies in the meta-analysis that showed a benefit in patients treated with TACE enrolled patients with a 3-year predicted spontaneous survival of 8 and 23%, respectively, which is significantly worse than the 50% predicted spontaneous survival of patients with intermediate HCC originally described by Llovet et al. (23). Interpretation of the results of this meta-analysis is also difficult because previous studies with TACE were not standardized for embolization procedures while recent developments with TACE, including new, more effective embolizing agents (16), make standardization better, thus requiring a redefinition of selection criteria and endpoints in TACE studies. Indeed, TACE with doxorubicin-loaded (DC) beads can cause complete necrosis in 2–5 cm-sized HCC nodules, suggesting that TACE may be a radical treatment modality in selected patients with adjuvant pharmacological therapies like patients treated by RFA (24). The multinational SPACE study is investigating the effectiveness of TACE with DC beads in patients with intermediate HCC, who are randomized to receive adjuvant therapy with the molecular-targeted agent sorafenib vs placebo.
Advanced hepatocellular carcinoma (Barcelona clinical liver cancer stage C)
In these patients, the presence of tumours is complicated by radiological evidence of portal vein thrombosis or extrahepatic tumour invasion as well as symptoms of neoplasia (performance status 1 or 2). While until recently no effective treatment modality was available to these patients, whose spontaneous survival does not exceed an average of 6 months (1), in 2007, the multikinase inhibitor sorafenib was approved by the FDA and EMEA as a standard of care for patients with advanced HCC and compensated cirrhosis (25, 26) (Table 3). In Italy, sorafenib therapy is restricted to BCLC-C patients with compensated liver disease (Child–Pugh A) who have either a clinical or a radiological response during the first 2 months of therapy. As a result, the recently amended RECIST radiological criteria have been replaced by the EASL criteria that include early modifications of arterial vascularization of the tumour caused by sorafenib and better identify patients who respond to sorafenib (2, 26). Several trials are now ongoing (27) with new agents inhibiting either the same pathways as sorafenib or different pathways such as mTOR and growth factors involved in liver cell carcinogenesis and angiogenesis.
Table 3. Comparison of randomized-controlled trials of sorafenib in Western (SHARP) and Asian patients with advanced hepatocellular carcinoma
BCLC, Barcelona clinic liver cancer; HBV, hepatitis B virus; NA, not applicable; TTP, Time to Progression.
HBV aetiology of cirrhosis
5.5 months (2.8 months)
2.8 months (1.4 months)
Median survival (control)
10.7 months (7.9 months)
6.5 months (4.2 months)
Grade 3/4 toxicity
End-stage hepatocellular carcinoma (Barcelona clinical liver cancer stage D)
These patients have tumours of any size with symptoms of neoplasia and damaged liver function (Child–Pugh C), with an average predicted survival of 3 months. There is no specific treatment available in these cases, except the best palliative care.
In the last few decades, the management of HCC has changed significantly, thanks to breakthroughs in the treatment algorithm. A better understanding of the natural history of the tumour has led to the development of evidence-based staging systems, thus improving patient selection and treatment. At present, treatment of HCC patients is based on individual evaluation of the patient to guide the decision-making process. Half of the HCC patients presenting to a tertiary referral centre will have an early tumour corresponding to criteria for radical therapy with resection, local ablation or transplantation. One-third of the patients will have intermediate HCC, which can benefit from treatment with TACE if these patients have compensated cirrhosis (Child–Pugh status A or B) and have no portal vein complications such as thrombosis or large oesophageal varices. For the remaining patients presenting with advanced HCC, sorafenib is an option as long as cirrhosis is well compensated (Child–Pugh A) and a radiological or a clinical response is obtained in the first 2 months of treatment. Because of the extended survival in patients when tumours are detected early, monitoring of patients with chronic liver disease and early diagnosis is the only practical approach to improve the treatment of HCC. However, there is also a hope for cure in the many patients with more advanced tumour disease especially by taking a multidisciplinary approach and combining different treatment modalities.
Conflicts of interest
Massimo Colombo receives a grant and research support from Schering-Plough, Roche, Gilead, Bayer. He sits on advisory committees and is a speaker and teacher for Bristol-Meyers-Squibb, Schering-Plough, Roche, Novartis, Gilead, Vertex, Bayer. Angelo Sangiouanni is a speaker and teacher for Bayer.