Liver cancer: From molecular pathogenesis to new therapies

Liver cancer: From molecular pathogenesis to new therapies: Summary of the EASL single topic conference

Jnl of Hepatology Jan 2010

"The major advances described above were presented at this EASL single topic conference on HCC. Much work remains to improve the outlook of patients with HCC.....Successful treatment of underlying liver disease prevents the development of cirrhosis and decreases the risk of HCC.....Surveillance improves management of patients who subsequently develop HCC.....Ultrasound is the recommended screening test"

Jean-François Dufour12, Philip Johnson3


Received 29 June 2009; received in revised form 17 September 2009; accepted 22 September 2009. published online 19 November 2009.

Corrected Proof

Abbreviations: DEB, drug-eluting bead, HBV, hepatitis B virus, HCC, hepatocellular carcinoma, HCV, hepatitis C virus, hTERT, human telomerase reverse transcriptase, PDGF, platelet derived growth factor, FAK, focal adhesion kinase, Pyk2, proline-rich tyrosine kinase 2, MAPK, mitogen activated protein, c-FLIP, cellular-FLICE inhibitory protein, cccDNA, closed circle cDNA, NS5A, non-structural protein 5A, PI3K, phosphoinositide 3-kinases, TCF1, transcription factor 1, mTOR, mammalian target of rapamycin, TTP, time to radiologic progression, VEGFR, vascular endothelial growth factor receptor, VEGF, vascular endothelial growth factor, p-ERK, phosphorylated-extracellular signal-regulated kinases, PEI, percutaneous ethanol injection, RFA, radio-frequency ablation, TACE, transarterial chemoembolisation, TAE, transarterial embolisation, OLT, orthotopic liver transplantation, MELD, model of end-stage liver disease


This report summarizes topics which were covered at the EASL 2008 single topic conference on hepatocellular carcinoma (HCC), following on from the EASL 2000 HCC conference [1]. As such, no attempt has been made to grade the strength of the evidence presented.

Epidemiology


The incidence of hepatocellular carcinoma (HCC) (620,000 cases/year worldwide) nearly matches its mortality (595,000 deaths) demonstrating the aggressive nature of the tumor and the limited therapeutic options [2]. Eighty percent of the cases of HCC occur in China, Africa and South East Asia. In Europe, the geographical distribution of HCC varies. It is less frequent in Northern Europe with mortality rates under 1/100,000 [3]. High HCC mortality rates are observed for men in France (6.8/100,000), Italy (5.7) and Switzerland (5.9) and slightly lower for women, in Italy (1.9), Switzerland (1.8) and Spain (1.5) [4]. In European populations, the factors associated with an increased risk of liver cancer are gender specific. In a male population, an HBV or HCV prevalence higher than 2% and high alcohol consumption (greater than 11L of pure alcohol/(person-year) figure prominently whereas in women, HCV predominates [3]. After adjustment for these three risk factors, a higher incidence of liver cancer is found in both genders in South-Eastern Europe [3]. These disparities in HCC mortality rates and trends across European populations represent important medical challenges.


Despite the identification of well defined epidemiological risk factors and intense research activity, an immense knowledge gap persists in elucidating the molecular pathways involved in hepatic carcinogenesis. This is particularly evident for HBV. Smith and Blumberg proposed, in 1969, a link between Australia antigen (now recognised as HBs Ag) and HCC [5]. This prediction has since been confirmed; a large epidemiological study found that a positive serology for HBs antigen is associated with a relative risk for HCC of nearly 100 [6], and universal vaccination against HBV at birth has since been shown to reduce the incidence of HCC [7]. Refining this association, other HBV-related factors have been identified as predictors of HCC development and these include: HBe antigen seropositivity [8], high viral load [9] and genotype C [10] as well as an environmental factor, exposure to aflatoxin. Aflatoxin, derived from a mould that grows on grains, is probably the most potent natural carcinogen known to man. Nearly half of African or Asian patients with HCC are infected with HBV and have an aflatoxin characteristic mutation at codon 249 of TP53 gene. Social intervention programs such as improved storage of grain can significantly reduce exposure [11]. HBV contributes to liver carcinogenesis by multiple mechanisms: chronic hepatitis B sustains chronic inflammation, hepatitis B proteins impinge on oncogenic pathways, HBV favours genomic instability and HBV DNA can insert into the human genome. Among the many different insertion sites described, five gene families are recurrently targeted: human telomerase reverse transcriptase (hTERT), PDGF receptor, calcium signaling related genes, mixed lineage leukaemia and 60s ribosomal protein genes [12]. The protein HBx still attracts most attention because it modulates a wide variety of cellular processes, including transcription, in particular CREB-mediated transcription [13]. HBx elicits calcium signals [14] which activate focal adhesion kinase (FAK) and proline-rich tyrosine kinase 2 (Pyk2), and subsequently Src tyrosine kinases and downstream signaling pathways such as Ras–Raf–MAPK pathway [15]. HBx has been found to interact with c-FLIP and to inhibit its recruitment to the death-inducing signaling complex, resulting in hyperactivation of caspase-8 and caspase-3 by death signals [16]. HBx was reported to induce angiogenesis by stabilising hypoxia-inducible factor-1alpha [17]. The HBV DNA can also form a closed circle cDNA (cccDNA) structure, which exists as a minichromosome in the nucleus associated with histone and non-histone proteins, regulated by epigenetic changes [18]. HBc may also act as a gene expression regulatory protein and inhibit the expression of p53 [19] and Death Receptor 5 [20].


The epidemic in chronic viral hepatitis C has contributed to the rise in HCC incidence observed over the last few decades. HCC is the most common cause of death in patients with hepatitis C-induced cirrhosis. In many countries more than 30% of the HCC patients have a positive serology for HCV, when less than 1% of the general population is HCV positive. The risk of developing an HCC is slightly higher with genotype 1b [21]. Increasing experimental evidence raises the possibility that HCV contributes directly to malignant transformation of hepatocytes. Three different HCV core transgenic mice develop steatosis and HCC, which supports the hypothesis that HCV core protein itself is oncogenic. The mechanism may involve oxidative stress, but remains to be defined [22]. In contrast, mice transgenic for the non-structural protein NS5A do not develop HCC. However, an oncogenic mechanism has been proposed for this protein. NS5a can activate PI3K resulting in activation of survival pathways [23] and stabilisation of beta-catenin [24].


Worldwide, more than 80% of the HCC are associated either with HBV or with HCV [25]. Besides alcohol, obesity has recently been recognised to be another important risk factor. In men with a BMI>35, the relative risk of death from HCC is greater than from any other cancer [26]. The HCC that develops within the metabolic context of obesity displays specific features that we are just beginning to unravel, such as its tendency to occur in non-cirrhotic liver [27]. Thus the major addictions in Western societies – drugs (associated with viral hepatitis), alcohol and food (associated with obesity and hence steatohepatonecrosis) are all strongly associated with liver cirrhosis and HCC. HCC will remain a significant public health problem for many years.

Molecular classification


At first glance, the broad diversity of HCCs at the genetic and epigenetic levels would seem to preclude their molecular classification but various molecular criteria are now being used to successfully classify HCCs. Methylation profiling revealed activation of Ras and downstream Ras effectors due to epigenetic silencing of inhibitors of the Ras pathway [28]. Some promoters are frequently methylated such as CDKN2A (p16) [29] and CDH1 (E-cadherin) [30]. The recognition of genetic alterations leads to HCCs being divided into two main groups according to chromosome stability status. HCCs without chromosome instability have frequent beta-catenin mutations as the single genetic alteration and tend to be of large size and HBV negative, whereas HCCs with chromosome instability have frequent axin1 and p53 mutations and tend to be HBV positive [31]. Applying cDNA microarray-based gene expression profiling, Iizuka et al. identified a set a 12 genes whose expression in surgically resected tumor correctly predicted early intrahepatic recurrence [32] and Ye et al. found that intrahepatic metastatic lesions were indistinguishable from their primary tumors, whereas primary metastasis-free HCC was distinct from primary HCC with metastasis [33].


Thorgeirsson’s group have classified HCCs into two categories with different prognoses; tumors from the low survival subclass have strong cell proliferation, antiapoptosis gene expression signatures and display higher expression of genes involved in ubiquitination and histone modification, suggesting involvement of these processes in the progression of HCC [34]. By integrating gene expression of rat fetal hepatoblasts and adult hepatocytes, a human HCC subgroup which may arise from hepatic progenitor cells was identified and associated with a particularly poor prognosis [35]. This classification was refined by the group of Zucman-Rossi, who identified six groups of tumors: tumors of the first group are associated with low copy number of HBV and overexpression of genes expressed in fetal liver and controlled by parental imprinting; tumors of the second group are associated with a high copy number of HBV and have mutations in PI3K and p53; tumors of the third group are characterised by overexpression of genes controlling the cell cycle; HCCs of the fourth group are more heterogeneous with TCF1 mutations. The last two groups are strongly related to beta-catenin mutations leading to Wnt pathway activation. Importantly, this classification carries therapeutic implications: the first two groups, characterised by Akt activation could theoretically be treated with receptor tyrosine kinase inhibitors, Akt or mTOR inhibitors, the third group by proteasome inhibitors and the last two by Wnt inhibitors (Fig. 1).

A profiling of HCV-induced HCC also identified categories characterised by beta-catenin mutations and cell proliferation markers, but revealed two further classes one related to interferon pathway and one typified by polysomy of chromosome 7 [36]. Profiling at the level of the microRNAs revealed that miR-21, 10b and 222 are upregulated. More specifically, miR-375 is downregulated in HCCs associated with beta-catenin mutation, miR-96 is upregulated in HCCs associated with HBV infection and miR-126 is downregulated in HCCs associated with alcohol intake [37]. These molecular classifications define groups that share overlapping oncogenic processes. Tumor suppressor and oncogene mutations are major drivers for these group assignments (e.g. Wnt/beta-catenin, receptor tyrosine kinase/Akt/ras). In order to integrate all the data of the molecular classifications and signaling pathways, as well as to provide the rationale for molecular targeted therapies, one has to recognise that the vast majority of HCCs present activation of angiogenesis, activation of telomerase reverse transcriptase providing limitless replicative potential, and checkpoint inactivation as a result of inactivation of p53, Rb and cyclin D1 [38]. Against this common background, HCCs can be classified into specific pathway subgroups. Classifying HCCs into homogeneous subgroups using gene signature is a promising tool for constructing rational protocols with targeted therapies. Interestingly, analysis of noncancerous hepatic tissue adjacent to HCC revealed gene-expression signature associated with recurrence and survival after resection [39], [40].

Systemic targeted therapies


Systemic targeted therapies have been investigated in patients with Barcelona Clinic Liver Cancer advanced stage, since specific therapies are established for earlier stages: surgery for early stage and transarterial chemoembolisation for intermediate stage [41]. The rationale for systemic targeted therapies is based on the following three premises (1) the relative inefficacy of conventional chemotherapy for HCC, (2) the pivotal role of angiogenesis in the development of HCC and (3) identification of pharmacological inhibitors/blockers of key pathways involved in cell proliferation and survival. Systemic targeted therapy for HCC was deemed successful with the results of two randomized phase 3 clinical trials, which demonstrated a prolonged survival and a delayed time to progression with sorafenib in patients with HCC too advanced for local therapy and conserved liver function (Child A). The magnitude of effect was remarkably similar between the two phase 3 studies with hazard ratios of 0.69 and 0.68. In the trial reported by Llovet et al. (a predominantly European population) the overall survival was 10.7months vs. 7.9months in the patients treated with sorafenib vs. placebo, respectively, and the time to radiologic progression (TTP) 5.5months vs. 2.8months [42]. In the Asia-Pacific trial the analogous figures were 6.5months vs. 4.2months and 2.8months vs. 1.4months [43]. This difference of a few months between the studies may have been due to the fact that the Asian trial enrolled patients who had had more extensive previous locoregional treatments and hence were at a more advanced stage.


Sunitinib is, like sorafenib, a multi-kinase inhibitor, but has different targets, predominantly PDFR-b, VEGFR-1, 2 and 3. A phase 2 study, using 50mg in cycles of 4weeks on/2weeks off, was associated with toxicity including death [44]. A second phase 2 study at the lower dose of 37.5mg was better tolerated and resulted in an overall survival of 9.8months [45]. Several other VEGFR tyrosine kinase inhibitors (such as brivanib, which is also a fibroblast growth factor receptor inhibitor) are in clinical trials. There is a strong rationale to target angiogenesis in the treatment of HCC [46]. VEGF mediated angiogenesis is an important component of HCC growth and maintenance. HCC is a hypervascularized tumor and increased circulating levels of VEGF have been associated with aggressive HCC behaviour and poor prognosis [47]. Bevacizumab, a humanized monoclonal antibody that binds to VEGF-A has been used in advanced HCC, but its benefit remains unproven. A phase 2 study combining erlotinib and bevacizumab showed indications of antitumor activity with a progression free survival of 39weeks and a median overall survival of 68weeks with a high overall response rate of 25% [48].


mTOR is a central regulator of cell growth and angiogenesis and this pathway was found to be activated in the majority of HCCs [49], [50], [51], [52]. Functionally, it operates downstream of growth factor receptors and upstream of the protein synthesis machinery. Experimental work that demonstrated an effect of mTOR inhibition on tumor growth and survival, has paved the way for clinical trials [53], [54]. Clinical evidence is coming from retrospective analyses of patients transplanted for HCC and subsequently immunosuppressed with an mTOR inhibitor [55]. A case control study of patients immunosuppressed either with calcineurin inhibitor or with calcineurin inhibitor in combination with sirolimus reported better disease free survival in the group receiving an inhibitor of mTOR: 93% vs. 79% at 1year and 75% vs. 54% at 5years. The overall survival improved but not significantly [56]. The ongoing multinational SiLVER study testing sirolimus after liver transplantation for HCC should provide a definitive answer. Due to the complexity of this pathway and presence of feedback loops, mTOR inhibitors are likely to exert most of their potential benefits in combination therapy [57].


The multitude of trial possibilities for targeted therapies to treat HCC raises the following challenges: (1) to identify the right drug for the right patient, (2) to assess correctly drug efficacy and (3) to pay attention to new side effects. The discovery and validation of new markers are required to give the right targeted therapy to the right patients and spare those patients the side effects of inadequate treatments. The phase 2 study with sorafenib identified baseline immunohistochemistry for p-ERK as a useful marker for better response [58]. It is likely that quantitative radiological changes detectable with contrast agents will provide early information on the biological effect of a drug. This could be combined with changes in circulating levels of growth factors [45]. We need to have the adequate technology and knowledge to correctly design trials and assess activity. Otherwise, effective drugs will be discarded because of inadequate assessment, and resources will be wasted in futile assessment of inadequate drugs. An important progress has been the publication by a group of experts of recommendations on design and endpoints of clinical trials in HCC [41]. Proper staging is key for the selection and evaluation of HCC patients in clinical trials, and it is recommended to use the Barcelona Clinic Liver Cancer staging system. Regarding endpoints, the expert panel recommended time-to-event as primary endpoint of randomized phase 2 trials and survival as main endpoint to measure effectiveness in phase 3 trials [41].


Phase 2 trial design typically involves around 30 patients enrolled in a single arm and takes tumor response measured by RECIST or other volume related measurement as an endpoint, assuming that change in tumor size reflects changes in viable tumor and therefore is a good index of response to therapy. According to such endpoints, sorafenib would have been dismissed. The percentages response and the percentages of stable disease are nearly the same between the sorafenib-treated group and the placebo group in the phase 3 trials [42], Modified RECIST criteria may be used to take into account tumor necrosis and assess viable tumor volume [41]. To capture the benefits of targeted therapy, the times to events, such as time to progression and death should be taken as endpoints already in phase 2 studies [38] Their interpretation requires adequate knowledge of the natural history of the disease. The solution to avoiding uncertainty in the historical response rates is to design randomized phase 2 studies; the control group provides a benchmark for the treatment arm [41], [59], [60].

Local therapies


Rates of local and late relapse after percutaneous ethanol injection (PEI) are, in the long term, higher than initially estimated with a local relapse rate of 30% at 3years, even if the tumor is smaller than 3cm [61], and the rate of treatment failure is 40% at 2years for lesions larger than 3cm [62]. HCC as small as 2.5cm may have microsatellite lesions, but they are only a few mm away and can be included in the volume of thermal ablation, thereby providing a safety margin that mimics surgery (Fig. 2) [63]. A randomized trial recruiting Child A patients with early stage HCC revealed a significant advantage of RFA compared to PEI in terms of local recurrence-free survival and event free survival compared, and a multivariate analysis found RFA to be an independent prognostic factor for local recurrence-free survival [64]. This was confirmed by a recent systematic review of randomized trials testing percutaneous ablation therapies which found a 3-year survival benefit for RFA [65]. In comparison to surgical resection, RFA is as effective for ensuring local control of HCC smaller than 2cm, and it is associated with similar survival rates [66], [67]. Histological analysis of HCC treated by RFA before liver transplantation shows that radiological assessment often overestimates the efficacy of the procedure. Perivascular localisation of the tumor results in incomplete eradication in about 50% of the cases due to the heat sink effect of blood flowing in nearby large vessels [68]. Combining RFA with doxorubicin-eluting beads augmented the ablated volume leading to a complete response at the final assessment in 12 of 20 patients who were recruited after ablation failure [69]. The 5-year survival of Child A patients with HCC within the Milan criteria treated by RFA is about 50% [70]; this survival figure increases to 68% in tumors smaller than 2cm [67].

The liver is an ideal organ for chemoembolisation given that its dual blood supply allows embolisation without infarction since hepatic tumors such as HCCs are almost exclusively supplied by the hepatic artery. There is a discrepancy between the histological analysis, which finds a complete necrosis in about 44% of cases, and the radiological assessment, which finds a complete response according to WHO criteria in only 6% of the cases [71]. When assessing the effects of TACE and TAE radiologically, it is important to use the amended RECIST criteria which take into account the viable tumor [41]. A meta-analysis of transarterial chemoembolisation (TACE) vs. transarterial embolisation (TAE) alone could not detect a survival difference between the two procedures [71] confirming two previous meta-analyses [72], [73]. Such analyses are limited by the great variability in transarterial techniques. With the introduction of drug-eluting beads (DEBs), the same device induces toxic and ischemic necrosis and the procedure will be better standardized. DEBs adsorb the chemotherapeutic agent (doxorubicin or irinotecan) and once injected and entrapped in tumor vasculature, release the drug in a sustained manner. Peripheral blood concentration of the chemotherapeutic drug is much lower than after conventional TACE [74]. It seems that the rates of objective response and survival are higher after DEB–TACE than after conventional TACE [74], [75], but abscess formation is more frequent [74]. Embolisation with radioactive microspheres loaded with 90-Yttrium has been reported to have significant antitumor effect in highly selected patients [76], [77]. This method needs to be evaluated in adequately designed trials.


Tumor embolisation leads to a central necrosis with a massive hypoxia on its border. This stimulates the cells at the edge of the lesion to produce survival and growth factors such as VEGF. These factors may act locally and systemically to promote the growth of adjacent satellites as well as more distant tumor lesions. It is thus clear that the combination of embolisation with systemic targeted therapy would make sense. The timing of the combination is relevant [78]. The targeted therapy should be introduced before the embolisation without interruption in order to take advantage of the antiangiogenic effect (Fig. 3); whether this approach leads to specific complications needs to be investigated in phase 1 trials.

Liver transplantation

The 5-year survival after liver transplantation for HCC has improved over the years [79]. The Milan criteria (single tumor ⩽5cm in size or ⩽3 tumors each ⩽3cm in size, and no macrovascular invasion) are associated with an excellent outcome and have become the benchmark for liver transplantation. To explore the survival of patients with tumors that exceed the Milan criteria, Mazzaferro’s group collected information on 1112 patients with HCC exceeding Milan criteria and 444 patients within Milan criteria [80]. Based on this unique database, the authors observed that microvascular invasion has a major effect and that HCC recurrence and risk of dying correlates linearly with HCC size, whereas the relationship with the number of nodules flattened beyond 3, an important observation since the number of nodules is generally more difficult to assess than their size (Fig. 4). They derived a prognostic model based on pathological objective tumor characteristics

(http://www.hcc-olt-metroticket.org/calculator/index.php). Patients without microvascular invasion, but who fell within the new ‘up-to-seven’ criteria (HCCs with seven as the sum of the size of the largest tumor [in cm] and the number of tumors) achieved a 5-year overall survival of 71.2% [80].

Once placed on the waiting list for transplantation, HCC patients accumulate arbitrarily MELD points over time but this prioritization system does not take into account the biology of the tumor. There is a need to replace waiting time by tumor characteristics and by combining HCC-specific parameters with the MELD score. The MELD score at listing for HCC patients remains predictive of de-listing [81]. Such a system is already in use in Bologna where MELD score and tumor characteristics were combined in a continuous scale. This decreased the time on the list for HCC patients, and allowed a more equitable prioritization of HCC patients without affecting non-HCC patients since the drop-out risk was not different between HCC and non-HCC patients [82].


The risk of HCC recurrence after transplantation can be evaluated on the explanted tumor taking into account pathological features such as degree of histological differentiation, presence of a capsule, presence of satellite nodules, percentage of necrosis and most importantly, vascular invasion. Molecular signatures based on microarrays have been developed to predict recurrence [83]. In patients with an underlying chronic hepatitis B, HCC recurrence was associated with HBV recurrence [84]. The choice of the immunosuppression may have an influence on HCC recurrence. The strategy to decrease calcineurin inhibitors in favour of mTOR inhibitors is so far empiric and is awaiting confirmation. There has been one randomized trial testing neoadjuvant treatment with systemic low-dose doxorubicin starting from acceptance onto the waiting list. This strategy improved neither survival nor freedom from recurrence in patients with HCC undergoing transplantation [85]. The data are more convincing for TACE. Patients undergoing TACE while on the waiting list showed a reduced drop-out rate [86], [87] and this was not at the expense of more recurrences after transplantation [88]. The advantage of TACE while waiting for a liver transplantation is dependent on the waiting time [89]. If the waiting time is short, TACE is not necessary to keep patients on the list. If the waiting time is too long, TACE is not able to keep patients on the waiting list. It has been proposed that TACE is useful when the waitlist time is between 4 and 9months [89], but randomized controlled trials are lacking and this approach should not be considered a standard of care.

Prevention, screening


There is only one randomized controlled trial addressing the outcome of HCC screening. This study was conducted in China and found that mortality rate was 83/100,000 in the screened group and 132 in the non-screened group giving an odds ratio of 0.63 [90]. This study, which is unlikely to be repeated, has several limitations; it is not stratified for cirrhosis and it had suboptimal compliance (58%). Ultrasound is the recommended screening test [91]. A randomized study found a better survival with 6months interval than 12months [92]. Currently used serological markers are associated with portal vein invasion [93], large tumor size [94] and microvascular invasion [95] and are predictors of late stage disease and poor outcome [96]. There is a need for better markers. Serum N-glycan profile [97] serum glycipan 3 levels [98] and GP73 [99] are candidates that deserve further consideration.


Surveillance improves management of patients who subsequently develop HCC. Tumors diagnosed within a screening program are detected at earlier stages and provides access to a potentially curative treatment to two third of these patients [100], [101] resulting in improved median survival [102], [103]. In a prospective study of 447 patients with compensated cirrhosis enrolled in a surveillance program, mortality in treated patients decreased significantly over time due to refinement in the selection criteria for therapy [104].


Twenty-five percent of the nodules found in a surveillance program fell in the 1–2cm category [103]. A diagnosis of HCC can be established in this size category without a positive biopsy if two dynamic imaging studies show coincidental typical vascular pattern [91], but the sensitivity of these noninvasive criteria is only 33% [105]. About 70% of such lesions are located in segments easily accessible to biopsy [106]. However, a ‘wait and watch’ approach with radiological follow-up delays the diagnosis by only a few months without jeopardizing treatment options [107]. Pre-transplantation fine-needle aspiration biopsies may increase the risk of extra-hepatic recurrences post transplantation [108], and tumor grade assessed on needle core biopsy correlates poorly with tumor grade or presence of microscopic vascular invasion on final pathology [108].


Successful treatment of underlying liver disease prevents the development of cirrhosis and decreases the risk of HCC. A meta-analysis of three randomized controlled trials and 15 nonrandomized controlled trials showed a slight reduction in the occurrence of HCC in hepatitis C patients who experienced a sustained virological response to interferon [109]. A retrospective analysis of patients with HCV-related cirrhosis who achieved sustained virological response after interferon-alpha therapy found a reduction of liver-related mortality lowering both the risk of complications and HCC development [110]. However, another retrospective analysis failed to show a statistical significant difference in the development of HCC in patients who had responded to interferon vs. those who had not [111]. Irrespective of sustained virological response achievement, all cirrhotic patients should continue surveillance because the risk of occurrence of HCC is not completely avoided. Long-term pegylated interferon administered to cirrhotic patients who did not have a virological response, did not prevent the occurrence of HCC [112]. In terms of prevention after surgical resection, a randomized controlled trial showed a benefit regarding late recurrence in caucasian patients infected with HCV and treated with interferon monotherapy, but this is derived from a post-hoc analysis based on a small subset of patients [113].


After 15years of follow-up, HBe Ag seropositive patients treated with interferon-alpha therapy had a decreased cumulative incidence of cirrhosis and HCC but the reduction of HCC development was significant only in patients with pre-existing cirrhosis [114]. A retrospective survey of patients with chronic hepatitis B found that lamuvidine decreased the annual incidence rate for HCC from 2.5% to 0.4%/(patient-year) [115]. In a randomized control trial, maintenance therapy with lamivudine significantly decreased HCC occurrence in patients with advanced fibrosis [116]. Development of resistance, with virological breakthrough was associated with a loss of the beneficial effect on HCC incidence [117].


The major advances described above were presented at this EASL single topic conference on HCC. Much work remains to improve the outlook of patients with HCC.