February 3, 2014

Nature Reviews Gastroenterology and Hepatology 10, 34-42 (January 2013) |doi:10.1038/nrgastro.2012.199

Augusto Villanueva, Virginia Hernandez-Gea & Josep M. Llovet  About the authors

The management of hepatocellular carcinoma (HCC) has substantially changed in the past few decades. Improvements in patient stratification (for example, using the Barcelona Clinic Liver Cancer staging system) and the introduction of novel therapies (such as sorafenib) have improved patient survival. Nevertheless, HCC remains the third most common cause of cancer-related deaths worldwide. Decision-making largely relies on evidence-based criteria, as depicted in the US and European clinical practice guidelines, which endorse five therapeutic recommendations: resection; transplantation; radiofrequency ablation; chemoembolization; and sorafenib. However, areas still exist in which uncertainty precludes a strong recommendation, such as the role of adjuvant therapies after resection, radioembolization with yttrium-90 or second-line therapies for advanced HCC. Many clinical trials that are currently ongoing aim to answer these questions. The first reported studies, however, failed to identify novel therapeutic alternatives (that is, sunitinib, erlotinib or brivanib). Moreover, genomic profiling has enabled patient classification on the basis of molecular parameters, and has facilitated the development of new effective drugs. However, no oncogene addiction loops have been identified so far, as has been the case with other cancers such as melanoma, lung or breast cancer. Efforts that focus on the implementation of personalized medicine approaches in HCC will probably dominate research in the next decade.

Key points

  • Epidemiological data indicate that the disease burden of hepatocellular carcinoma (HCC) is increasing worldwide, both in terms of incidence and mortality
  • The Barcelona Clinic Liver Cancer staging system provides a general framework for decision-making in patients with HCC, and facilitates stage-based unified selection criteria for clinical trials
  • Evidence-based criteria dominate recommendations for HCC management, enabling stratification of evidence according to scientific standards and providing a hierarchy of medical recommendations
  • Five treatments are strongly recommended in HCC on the basis of evidence-based data: resection; liver transplantation; radiofrequency ablation; chemoembolization; and sorafenib
  • Sorafenib, a molecular targeted agent, prolongs survival in patients with advanced HCC and is the sole systemic drug that is proved to be effective in this disease
  • No oncogenic addiction loops have so far been identified in HCC; research initiatives should aim to identify subgroups of patients with targetable dominant molecular alterations


Disease burden owing to hepatocellular carcinoma (HCC) is increasing markedly worldwide.1 In the USA, epidemiological data published in 2009 show a substantial increase in HCC mortality in the past few decades2—a trend that underscores the importance of this disease in upcoming years.3 Nonetheless, major improvements have been made in HCC management during the past 30 years. Traditionally, HCC was considered a deadly disease without curative options, with an overall survival of <6 months.4 Advances in early detection, imaging techniques and novel therapies have improved patient selection and enabled evidence-based treatment approaches.5Consequently, prognostic algorithms, such as the Barcelona Clinic Liver Cancer (BCLC) classification, have been introduced in routine clinical care.4 The BCLC approach classifies patients according to tumour burden, underlying liver dysfunction and patient symptoms (Figure 1); it additionally links each stage to specific therapeutic interventions.5 The BCLC system is widely accepted, being currently endorsed by US and European associations for the study of liver diseases and oncology.5, 6 Better patient selection has resulted in improvements in patient outcomes for almost every approved therapeutic intervention in HCC, including curative (for example, resection, transplantation, local ablation) and palliative (for example, transarterial chemoembolization [TACE]) approaches. In addition, the molecular-targeted agent sorafenib has shown antitumour activity in patients with advanced HCC, improving 3 months overall survival and delaying tumour progression.7

Figure 1 | BCLC staging system and therapeutic strategy according to EASL–EORTC guidelines.


Staging classification comprises five stages that select the best candidates for the best therapies currently available. Patients with asymptomatic early tumours (stage 0–A) are candidates for radical therapies (resection, transplantation or local ablation). Asymptomatic patients with multinodular HCC (stage B) are suitable for chemoembolization (TACE), whereas patients with advanced symptomatic tumours and/or an invasive tumoural pattern (stage C) are candidates to receive sorafenib. End-stage disease (stage D) includes patients with grim prognosis that should be treated by best supportive care. Abbreviations: BCLC, Barcelona Clinic Liver Cancer; DDLT, deceased donor liver transplantation; EASL, European Association for the Study of Liver Disease; EORTC, European Organisation for Research and Treatment of Cancer; GRADE, grading of recommendations assessment, development and evaluation; HCC, hepatocellular carcinoma; LDLT, living donor liver transplantation; PEI, percutaneous ethanol injection; RF, radiofrequency ablation; TACE, transcatheter arterial chemoembolization; OS, overall survival; PST, performance status. Permission obtained from Elsevier © European Association for the Study of the Liver; European Organisation for Research and Treatment of Cancer. J. Hepatol. 56, 908–943 (2012).

HCC frequently occurs in a damaged organ; liver cirrhosis owing to viral hepatitis (HBV or HCV infection) and alcohol abuse are the main risk factors.8 However, the rapid development of potent antiviral agents and the increasing incidence of cirrhosis owing to NASH and overweight will modify the aetiological landscape of chronic liver disease in the next decades.9 The unique coexistence of two diseases—cirrhosis and HCC—in the same patient complicates the prognostic prediction and therapeutic strategies. This coexistence was observed in a phase III trial assessing sunitinib versus sorafenib in patients with advanced HCC, which was prematurely halted owing to toxicity and futility in the sunitinib arm.10

This finding emphasizes the fact that patients with cirrhosis, even at early stages of liver failure, are more susceptible to toxicities associated with drugs that are otherwise harmless in individuals without liver disease (sunitinib is approved for renal cell carcinoma). This issue, in addition to the fact that conventional chemotherapy was proven ineffective for this malignancy,11 makes HCC management particularly challenging. This Review will dissect the evidence behind medical interventions in HCC, and discuss some relevant areas of controversy in disease management.

Milestones in research and management

Evidence-based approaches are progressively governing decision-making in medicine. This development is particularly relevant in oncology, in which the disease is frequently lethal and medical interventions can have major adverse effects. Hence, changes in the standard of care need to be supported by robust data. In fact, clinical practice guidelines tend to follow evidence-based criteria to select and grade clinical recommendations, as is the case for HCC.5 This approach enables the establishment of a hierarchy of recommendations based on levels of evidence, and identifies those 'grey areas' in which more research is required to provide clear recommendations. When considering HCC management, only five interventions reach the highest level of evidence, being worldwide accepted recommendations for the treatment of HCC:5 resection for patients with solitary tumours and well-preserved liver function; liver transplantation for patients with tumours within Milan criteria (single nodules <5 cm or three nodules <3 cm);12 percutaneous ablation with radiofrequency in early tumours not suitable for surgical treatment; TACE for patients with multinodular asymptomatic tumours without vascular invasion or extrahepatic spread (BCLC B); and sorafenib for patients at advanced stage (Figure 1). Beyond this framework, all other interventions require additional research to prove survival advantages when confronted with the above-mentioned standards of care. Such is the case for internal radioembolization with ytrrium-90 or the role of adjuvant therapies after resection.5 Herein, we discuss some of the areas of HCC management that require improvement and summarize the evidence currently available.

The most relevant advances in HCC management have resulted from randomized studies, meta-analysis and cohort studies providing different levels of evidence and strength of recommendations in guidelines of clinical practice (Figure 2). Some of these pivotal studies have been widely cited, and represent scientific milestones in liver cancer research (Table 1); even though clinical decision-making should rely on recommendations based on levels of evidence, showing how frequently these recommendations reflect the importance of published research when evaluating the total number of citations is interesting. Table 1 summarizes the most relevant achievements and milestones in HCC research during the past 30 years, represented in hierarchical order according to number of citations and topic. These milestone studies in HCC address different aspects of clinical management such as epidemiology (association between viral hepatitis and HCC development), surgery (Milan criteria for liver transplantation, intention-to-treat analysis for resection), locoregional therapies (percutaneous ablation in small tumours, randomized trials and meta-analyses of TACE), chemoprevention (HBV vaccination or interferon therapy and decreased HCC risk) and systemic therapies (such as sorafenib). Besides clinically oriented studies, few other studies represent true breakthroughs in the understanding of the pathogenesis of HCC, such as TP53 mutations and HCC in aflatoxin endemic areas, or the oncogenic role of HBV and core proteins in transgenic mice.13 With respect to HCC management, to date, the highest cited paper (Table 1; n = 2,331 citations) is the landmark study published in 1996 by Mazzaferro et al.12 that introduced the Milan criteria for selection of optimal HCC candidates for liver transplantation. When analysing the manuscripts according to number of citations per year, the sorafenib study ranks first with an average of 325 citations per year. Not surprisingly, both studies provide the rationale for strong clinical practice recommendations in HCC management.5, 6 In fact, substantial overlap exists between findings from the top cited papers and the clinical recommendations made in guidelines.5

Figure 2 | Therapeutic interventions in HCC according to level of evidence and grade of recommendation.


Level of evidence (based on NCI classification and grade of recommendation based on GRADE criteria. Abbreviations: GRADE, grading of recommendations assessment, development and evaluation; HCC, hepatocellular carcinoma; LDLT, living donor liver transplantation; OLT, orthotopic liver transplantation; NCI, National Cancer Institute; PEI, percutaneous ethanol injection; RF, radiofrequency ablation. Permission obtained from Elsevier © European Association for the Study of the Liver; European Organisation for Research and Treatment of Cancer. J. Hepatol. 56, 908–943 (2012).

Table 1 | Milestones in HCC research reported during the past 30 years*

Citations Reference Title Thematic area Direct influence on CPG Citations per year
Data accessed on Web of Science® (Thomson Reuters, New York, USA), using the search terms “hepatocellular carcinoma” or “liver cancer” on 20 September 2012.
*Sorted by number of citations.
Abbreviations: CPG, clinical practice guidelines; HCC, hepatocellular carcinoma.
Molecular pathogenesis
1,640 Knowles et al. (1980)83 Human HCC cell lines secrete the major plasma proteins and hepatitis B surface antigen Basic science No 49
1,285 Hsu et al. (1991)84 Mutational hotspot in the p53 gene in human HCCs Basic science No 56
1,085 Bressac et al. (1991)85 Selective G-mutation to T-mutation of p53 gene in HCC from Southern Africa Basic science No 48
788 Kim et al. (1991)86 HBx gene of HBV induces liver cancer in transgenic mice Basic science No 34
736 Moriya et al. (1998)87 The core protein of HCV induces HCC in transgenic mice Basic science No 46
Epidemiology and natural history
1,867 Beasley et al. (1981)88 HCC and HBV: a prospective study of 22,707 men in Taiwan Epidemiology No 57
1,661 El-Serag et al. (1999)89 Rising incidence of HCC in the US Epidemiology No 116
1,215 Okuda et al. (1985)90 Natural history of HCC and prognosis in relation to treatment: study of 850 patients Prognosis No 42
978 Kiyosawa et al. (1990)91 Interrelationship of blood transfusion, non-A, non-B hepatitis and HCC: analysis by detection of antibody to HCV Epidemiology No 42
943 Saito et al. (1990)92 HCV infection is associated with the development of HCC Epidemiology No 40
881 Chang et al. (1997)23 Universal HBV vaccination in Taiwan and the incidence of HCC in children Chemoprevention Yes 54
899 Chen et al. (2006)25 Risk of HCC across a biological gradient of serum HBV DNA level Epidemiology No 118
720 Tsukuma et al. (1993)93 Risk factors for HCC among patients with chronic liver disease Epidemiology No 35
683 Nishiguchi et al. (1995)94 Randomized trial of effects of IFN-α on incidence of HCC in chronic active hepatitis C with cirrhosis Chemoprevention No 37
668 Bruix et al. (1989)95 Prevalence of antibodies to HCV in Spanish patients with HCC and hepatic cirrhosis Epidemiology No 28
2,331 Mazzaferro et al. (1996)12 Liver transplantation for the treatment of small HCC in patients with cirrhosis Surgery Yes 129
1,777 Llovet et al. (2008)7 Sorafenib in advanced HCC Systemic therapy Yes 325
984 Llovet et al. (2002)96 Arterial embolization or chemoembolization vs symptomatic treatment in patients with unresectable HCC: a randomized controlled trial Locoregional Yes 84
831 Llovet et al. (2003)11 Systematic review of randomized trials for unresectable HCC: chemoembolization improves survival Locoregional Yes 78
791 Livraghi et al. (1999)97 Small HCC: treatment with radiofrequency ablation vs ethanol injection Locoregional Yes 56
791 Lo et al. (2002)98 Randomized controlled trial of transarterial lipiodol chemoembolization for unresectable HCC Locoregional Yes 59
722 Llovet et al. (1999)99 Intention-to-treat analysis of surgical treatment for early HCC: resection vs transplantation Surgery Yes 50
730 Curley et al. (1999)100 Radiofrequency ablation of unresectable primary and metastatic hepatic malignancies: results in 123 patients Locoregional No 49
663 Livraghi et al. (1995)101 HCC and cirrhosis in 146 patients: long-term results of percutaneous ethanol injection Locoregional Yes 36

Breakthroughs in the management of human cancers have been dominated by the discovery and selective blockade of so-called oncogenic addiction loops.14, 15 The concept of oncogene addiction describes the selective dependence of cancer cell proliferation to certain molecular aberrations. Tumoural cells become dependent on and/or addicted to a specific molecular alteration responsible for its own controlled proliferation.16 Many of these loops have been therapeutically exploited and some have substantially improved patient survival.14 Examples include BRAF-mutated metastatic melanoma and response to vemurafenib,17 or ALK rearrangements in lung cancer and response to crizotinib.18 Strikingly, ALK oncogenic addiction in lung tumours was discovered after a fairly short timeframe: 3 years passed between the identification of ALK rearrangement19 and publication of the phase II trial.18 Unfortunately, not a single oncogenic addiction loop has thus far been identified in HCC.15 However, several efforts aimed at improving patient selection for novel therapies in HCC, such as deep sequencing and genomic profiling,20 are ongoing and could elucidate oncogenic addiction loops in the setting of HCC.

Evidence-based management of HCC


Identification of patients at risk of HCC development has become a public health priority. Assuming that close to 1% of the global population has cirrhosis21 and one-third of patients develop HCC in their lifespan,5 the expected number of lives saved following successful prevention of HCC is substantial. Cirrhosis of any aetiology is, by itself, a risk factor for the development of HCC.22 In terms of primary prevention, studies from Taiwan clearly demonstrated that prevention of HBV-related liver disease with universal vaccination is highly effective at diminishing HCC rates.23 Consequently, since 1991, the WHO has recommended vaccination of all newborn babies and high-risk individuals.24 In the setting of chronic hepatitis B or C, the main risk factors for the development of HCC are the presence of advanced hepatic fibrosis or cirrhosis, and high viral load.25 In fact, current guidelines recommend antiviral treatment to suppress virus replication and prevent the progression of fibrosis.26 Few randomized controlled trials (RCTs) are available that evaluate the role of HBV treatment for HCC prevention as a primary end point. The only two RCTs designed to measure the importance of treatment on HCC development showed that both interferon27 and lamivudine28 are effective in increasing HBV clearance, decreasing cirrhosis progression and subsequently reducing HCC risk. Studies with PEG-IFN-α are limited and restricted to the demonstration of improvement of surrogate markers (for example, viral DNA suppression, hepatitis B e antigen seroconversion and hepatitis B surface antigen loss),29 although the clinical benefits are expected to be at least similar to those with conventional IFN-α. Regarding chronic HCV infection, achievement of sustained virologic response has been shown to decrease the risk of HCC in patients with chronic hepatitis C.30 However, once cirrhosis is established, the benefit of antiviral treatment for HCV infection in terms of HCC prevention remains unclear.31 Moreover, maintenance treatment in nonresponders with IFN-α was neither beneficial for fibrosis progression nor for HCC development.32

As previously mentioned, the vast majority of HCC occurs in cirrhotic livers and increasing evidence suggests that fibrosis per se confers carcinogenic risk.33 Molecular data also point towards a major pathogenic role for diseased liver microenvironment ('field effect') in HCC development.34 Hence, reversal of fibrosis regardless of primary aetiology could theoretically prevent HCC development. However, studies targeting fibrosis as a chemopreventive strategy are scarce. Only two large, prospective, placebo-controlled studies have tested antifibrotic agents and showed no effects on fibrosis after 1 year of therapy.35, 36 Overall, besides primary prevention efforts aimed at avoiding known environmental risks for liver disease (including, HBV vaccination and withdrawal of alcohol intake), or antiviral therapy to suppress viral load in chronic hepatitis, no clear measures to decrease HCC incidence in patients with cirrhosis are available. In addition, little is known about the predictors that confer higher risk among individuals with cirrhosis, which justifies the recommendation of surveillance with abdominal ultrasonography every 6 months in all patients with cirrhosis.5 Preliminary evidence at the molecular level in experimental studies has identified different signalling pathways as potential targets for chemoprevention (for example, EGFR and PDGFR37, 38). However, there are two major drawbacks to developing effective drugs in the chemoprevention setting: first, the expected long duration of these trials requires the inclusion of enrichment strategies to increase feasibility by selecting those patients at very high risk of HCC development; second, a pressing need exists to identify accurate and noninvasive biomarkers for diagnosis and monitoring of fibrosis progression.


Stage-oriented HCC management based on the BCLC algorithm (Figure 1) classifies medical interventions as potentially curative (for example, surgical resection, liver transplantation or percutaneous ablation) or palliative (for example, TACE and sorafenib).5 Treatment allocation is based on levels of evidence as defined by the National Cancer Institute, which rely on strengths of study design and end points. Herein, we summarized the different therapeutic interventions for HCC according to these evidence-based parameters (Figure 2).

Surgical treatments

Surgical resection and liver transplantation are first-line options for patients with early stage HCC (BCLC)0–A), as they confer 5-year survival rates of 70%.39 Improvements in surgical approaches and refined selection of candidates for surgery (single nodule without liver dysfunction of portal hypertension40) have contributed to increased patient survival, minimization of complications and reduced recurrence.41 Overall, recurrence after resection reaches 70% at 5 years, either because of true metastases or de novo HCC,42and no adjuvant therapies able to reduce recurrence are currently approved. Local ablation has been suggested as a competitive alternative to resection in patients with a single tumour <2 cm.43 As no RCT has been designed ad hoc to address this issue, the role of local ablation as a first-line option in this setting remains controversial. Patients within Milan criteria (that is, single tumour <5 cm or three nodules <3 cm), without vascular invasion (BCLC A) should be evaluated for liver transplantation.5 These criteria have been independently validated by several groups and are widely adopted in transplant centres in Europe and the USA.39 Despite their utility, Milan criteria might be too restrictive and preliminary evidence indicates that some patients with tumours exceeding Milan criteria are potentially curable by liver transplantation.44, 45 Studies using clinical or pathological variables to expand Milan criteria have a number of methodological limitations, such as small sample size46 and their retrospective nature,47which prevents providing any recommendation owing to lack of robust data. A retrospective study suggested that microvascular invasion was a limiting factor to discriminate good and poor outcome among patients with HCC who underwent liver transplantation within and beyond Milan criteria (under the so-called up-to-seven rule).45These data emphasize the need to unravel molecular readouts of tumour biology to enable proper decision-making.48

Locoregional therapies

Local ablation is the standard of care for patients with early stage tumours who are not suitable for surgery; survival rates are 50–70% at 5 years with this approach.5Radiofrequency is the first option, and percutaneous ethanol injection is only reserved as a complementary treatment in cases of difficult tumour location.5 Patients with multinodular disease, preserved liver function, without tumour-related symptoms and absence of vascular invasion or extrahepatic spread (BCLC B, intermediate stage) are candidates for TACE, with the goal of delaying tumour progression and extension of patient survival.5, 6 Evidence to support this recommendation comes from a meta-analysis of pooled data.11 Improvements in embolization devices (for example, drug-eluting beads) provide median survival rates beyond 30–40 months, at least in referral centres.49 Findings from a Cochrane Review have challenged the role of TACE as the standard of care for patients with intermediate stage HCC,50 but the inclusion of studies using inadequate control arms and suboptimal patient selection might bias this conclusion.51 Regarding other embolization approaches (such as radioembolization with microspheres loaded with yttrium-90), data suggest it has a favourable safety profile,52but only well-designed, properly powered RCTs will determine the therapeutic niche, if any, of this intervention.

Systemic therapies

Until 2007, and despite substantial efforts, no single systemic agent had shown survival benefits in patients with HCC. This finding can be partially explained by the frequent coexistence of HCC and cirrhosis, its high molecular heterogeneity and a relative resistance to conventional chemotherapy.53 The SHARP trial demonstrated that sorafenib, a tyrosine kinase inhibitor with a broad inhibitory profile (for example, BRAF, PDGFR and VEGFR), was able to substantially increase survival in patients with advanced HCC (stage C of the BCLC classification: patients with well-preserved liver function and extrahepatic spread or vascular invasion) from 7.9 months to 10.7 months (HR 0.69).5, 6The effects of sorafenib in delaying tumour progression and improving survival were further validated in Asian HBV-infected patients.54 Adverse events such as diarrhoea or hand–foot skin reaction associated with sorafenib were manageable. On the basis of these data, sorafenib (400 mg, twice daily) is the standard of care for patients at advanced stages of HCC, and it should be maintained at least until radiological progression with periodic monitoring of adverse effects (mostly cardiovascular, diarrhoea and skin reactions).15 Thus far, no strong biomarkers are available to accurately predict a patient's response to sorafenib.50

Numerous targeted therapies are currently under evaluation in different developmental phases for the systemic treatment of HCC, both as first and second-line therapies (thoroughly reviewed elsewhere15). Regarding potential registration trials, initial results have been negative: the sunitinib trial was prematurely halted because of adverse events and futility in the sunitinib arm as a first-line therapy. Brivanib showed an antitumoural effect that did not markedly improve survival compared with placebo in second-line therapies,55 and the phase III trial testing the multikinase inhibitor linifanib has therefore been halted.56 Furthermore, the combination of sorafenib and EGFR inhibitors (erlotinib) failed in comparison to sorafenib alone as a first-line therapy.57 The remaining ongoing phase III trials are testing monoclonal antibodies against VEGFR2 (ramucirumab) and mTOR (mammalian target of rapamycin) inhibitor (everolimus). In addition, >250 early phase clinical trials of HCC therapies are ongoing, testing 56 molecular therapies.15 If more trial results are negative, alternative approaches will be needed to improve on the results already achieved with sorafenib. Potential alternatives are, first, the identification of drugs that can be safely used in combination with sorafenib. Second, the discovery of oncogene addiction loops that restrict survival benefits in a subpopulation of patients with HCC (for example, MET-positive patients), and third, to develop drugs targeting novel signalling cascades (for example, WNT–β-catenin and Notch) or molecular mechanisms (for example, epigenetic aberrations). In any case, it will be necessary to implement changes to the current paradigm of trial design (Box 1). Progressively, enrolment criteria should include molecular information of the mechanisms responsible for tumour progression in each patient, following previous examples in other solid tumours.58

Box 1 | Main issues for phase II/III trial design in patients with HCC102, 103

To select the target population

  • BCLC stage: include patients at specific BCLC stage (A–C)
  • Child–Pugh classification: include Child–Pugh A

To choose the appropriate end points

  • Overall survival (main end point in cancer research)
  • Time to progression*
  • Objective response rate*

To decide the adequate control arm

  • TACE for intermediate stage
  • Sorafenib for advanced stage (first-line therapy)
  • Placebo plus best supportive care for advanced stages (second-line therapy)

To stratify factors before randomization

  • ECOG performance status
  • Tumour burden (extrahepatic spread and/or vascular invasion)
  • AFP levels >200 ng/ml

*Time to progression and overall response rate should be assessed according to modified RECIST criteria. Abbreviations: AFP, α-fetoprotein; BCLC, Barcelona Clinic Liver Cancer; ECOG, Eastern Cooperative Oncology Group; HCC, hepatocellular carcinoma; OS, overall survival; RECIST, response evaluation criteria in solid tumours; TACE, transarterial chemoembolization.

Adjuvant therapies as a first unmet need

Tumour recurrence is a major complication after resection, occurring in 70% of patients at 5 years after surgery.39 Unlike most malignancies, two clear patterns of recurrence exist in HCC. Early recurrences (<2 years) are related to early dissemination of the primary tumour (that is, true metastasis) whereas late recurrences (>2 years) are a result of the carcinogenic cirrhotic milieu present in the remaining liver that promotes the progression of new malignant clones (that is, de novo tumours).42, 59 Feasibly, each type might have a different molecular background and therefore would require different therapeutic approaches for its prevention.60 Unfortunately, despite the fact that several RCTs have been conducted to evaluate adjuvant therapy, no agent has shown robust efficacy in preventing or delaying tumour recurrence.5

The role of IFN-α as adjuvant therapy after resection has been frequently evaluated in different studies, including RCTs.61, 62 One of the largest trials analysed the use of IFN-αin 150 patients with HCC, finding a negative trend for the primary end point (recurrence-free survival), but a positive trend in terms of prevention of late recurrence of HCC in the patients tested.63 Additionally, different meta-analyses have tried to provide a more comprehensive answer for the role of IFN-α in preventing HCC recurrence.64, 65, 66Owing to conflicting data and issues related to study design (such as trial selection for meta-analysis and mixing end points), IFN-α is currently not recommended as an adjuvant therapy after resection.5 In HBV-related HCC, a systematic review including 230 patients with HCC treated with ablation or resection found no evidence to support the use lamivudine with or without adefovir as adjuvant therapy.67

Besides IFN-α, other agents such as vitamin analogues have been evaluated in the adjuvant setting. Vitamin K was believed to have a beneficial effect in prolonging disease-free survival on the basis of findings from small studies.68, 69, 70 However, when tested in a large RCT including >500 patients, vitamin K was unable to prevent either HCC recurrence or death.71 Vitamin A analogues also have controversial results. Although the acyclic retinoid prevented the development of secondary HCC after resection and improved survival in one RCT,72 this result was not validated in an, as yet unpublished, large, multicentre RCT.73 Regarding adoptive immunotherapy with IL-2-activated lymphocytes, despite initial encouraging results,74 this strategy has not been subsequently validated and it is not recommended in HCC clinical management.5 Also, no survival benefits have been established with other immunotherapy regimens including cytokine-induced killer cells.75 Currently, two large phase III RCTs are ongoing that are evaluating the role of sorafenib and rapamycin in preventing tumour recurrence after resection and/or ablation (STORM trial)76 and after liver transplantation (SiLVER trial)77in patients with HCC.

One of the main limitations of liver transplantation for HCC is donor shortage, which causes longer waiting times and increases drop-out rates as a result of tumour progression whilst waiting for a graft. Therefore, transplant groups are applying locoregional therapies upon listing to prevent drop out.78 Evidence behind this approach comes from uncontrolled studies and cost–benefit analysis,78 showing that when expected waiting time exceeds 6 months it is cost-effective to offer the patient neoadjuvant therapies. This approach has been incorporated in a consensus document on HCC and liver transplantation that recommends locoregional therapies in patients with HCC within Milan criteria or United Network for Organ Sharing stage T2 (one nodule 2–5 cm or ≤3 nodules each ≤3 cm), with an expected waiting time longer than 6 months.5, 79Medical therapies, including sorafenib, have not been able to demonstrate a beneficial effect in this particular setting.

Towards molecular medicine in HCC

The future paradigm for treating patients with HCC includes customization of medical interventions on the basis of molecular alterations that govern tumour development and progression on an individual basis. In this regard, the most straightforward approach is to identify and validate oncogenic addiction loops, following the path opened in other solid tumours (for example, mutated BRAF melanoma, ALK rearrangements in lung cancer, and so on). Deep analysis of the HCC genome (such as exome and RNA sequencing) could provide new candidates. Several failures of systemic treatment in first-line therapy (brivanib versus placebo,80 sorafenib plus erlotinib versus sorafenib,81 sorafenib versus linifanib56 and sorafenib versus sunitinib)82 and second-line therapy (brivanib versus placebo)55 alert us to several concerns: the heterogeneity of the disease; toxicity of multikinase inhibitors in patients with cirrhosis; complexity of the disease; and the need for a more stratified approach by using reliable biomarkers and drugs for specific molecular aberrations. Targeting oncogenic addiction loops is expected to add survival benefits to the existing HCC therapy that currently relies on sorafenib.

Besides trial enrichment, combination therapies can also improve the current standard of care. The wide inhibitory profile of sorafenib, in addition to an antiangiogenic effect and good safety profile confirms this drug as the benchmark for systemic management of HCC. The bottleneck of the combination approach will be drug toxicity. The trade-off between efficacy and adverse effects is pivotal in this scenario, as patients with cirrhosis are clearly more susceptible to adverse effects of drugs than patients without cirrhosis. When considering changing the standard of care for HCC management, any intervention should be evaluated in the context of well-designed, properly powered, high-level RCTs—only these types of studies can currently change accepted treatment recommendations in guidelines.


The management of HCC has changed substantially in the past few decades, and five treatment options are accepted in clinical guidelines. Decision-making relies on evidence-based criteria; however, despite the recent advance in the understanding of HCC pathophysiology and development of new therapies, several clinical areas lack strong recommendations. The approval of sorafenib changed the understanding of HCC and brought in a new era in the management of liver cancer, led by the effort in understanding the molecular regulation and the identification of new molecular targets. Efforts on the implementation of personalized medicine will probably dominate research in the next decade.

Review criteria

A search for original articles focusing on hepatocellular carcinoma was performed in MEDLINE and PubMed. The search terms used were “hepatocellular carcinoma”, “liver cancer”, “treatment”, “management”, “medical therapies”, “randomized” and “systemic review”, alone and in combination. All articles identified were English-language, full-text papers. We also searched the reference lists of identified articles for further relevant papers.

Author contributions

All authors contributed equally to all aspects of this manuscript.

Competing interests statement

The authors declare competing interests.

Author affiliations

A. Villanueva, V. Hernandez-Gea & J. M. Llovet
Hepatocellular Carcinoma Translational Research Laboratory, Barcelona Clinic Liver Cancer Group, Institut d'Investigacions Biomèdiques August Pi i Sunyer (IDIBAPS), Liver Unit, Hospital Clínic, Villarroel 170, Barcelona 08036, Catalonia, Spain (A. Villanueva, V. Hernandez-Gea, J. M. Llovet).

Correspondence to: J. M. Llovet jmllovet@clinic.ub.es

Published online 13 November 2012