|Home | About | Journals | Submit | Contact Us | Français|
Medical therapies have entered center stage in the treatment of hepatocellular carcinoma (HCC) little more than a year after the positive results of a large phase III trial of sorafenib showed a clear survival benefit with sorafenib, a targeted agent, in this setting. Even though this marks a breakthrough in the treatment of HCC, the narrow patient profile necessary for the study to be successful has generated a number of questions regarding the efficacy of this approach in other clinical settings. New studies aiming to define the role of sorafenib from in the adjuvant setting, through patients with more advanced liver disease, all the way to combination treatments of HCC have been initiated. The success of one targeted drug has stimulated enormously the efforts of competitors to develop additional and better drugs, either in a first-line or a second-line setting. These are exciting times for the treatment of HCC, both for physicians and patients. In the years to come we will see an extension of treatment options in different clinical situations in patients with HCC, and survival will be improved in many stages of the disease, except for the most advanced. The rapid increase in knowledge about the molecular mechanisms underlying the development and progression of HCC will lead to a more tailored approach to treatment depending on the molecular characteristics of the tumor and the disease stage.
Hepatocellular carcinoma (HCC) is the sixth most common cause of cancer worldwide with an annual incidence of over 711,000 new cases, and is the third most common cause of cancer death with an annual mortality of 679,000 patients [Garcia et al. 2007]. Most commonly, HCC develops in cirrhosis, irrespective of the etiology.
In the Western world, chronic alcohol abuse and nonalcoholic steatohepatitis are about as important etiologic factors for cirrhosis as chronic hepatitis C. In chronic hepatitis C it is estimated that about 20% of patients will eventually develop cirrhosis after 20–30 years of infection. Once cirrhosis is established, the annual risk of developing HCC is estimated to be between 3 and 4% [Llovet et al. 2008], largely irrespective of the etiology of cirrhosis.
Chronic viral hepatitis by itself is less commonly associated with the development of HCC. The annual risk of developing HCC in patients with chronic hepatitis B without cirrhosis is reported to be around 0.5% in an Asian series with no corresponding data available for the Western world. Whether chronic hepatitis C without cirrhosis can lead to HCC is not clear at present.
HCC is mostly asymptomatic in early stage disease. Without proper surveillance programs of cirrhotic patients, diagnosis is only established in advanced stage disease.
The efficacy of surveillance by ultrasound (and to a lesser extent by alpha-fetoprotein measurement) has been established in prospective trials in the West [Sangiovanni et al. 2004] as well as in the East [Zhang et al. 2004]. Surveillance by experienced sonographers makes curative treatment possible in up to 75% of patients [Sangiovanni et al. 2004], while there is no curative treatment without proper surveillance [Zhang et al. 2004]. Once a lesion is detected by ultrasound, the diagnosis can be established radiologically in lesions with a typical appearance above 1cm in diameter. Biopsy is mandated only in cases with atypical presentation on imaging [Bruix and Sherman, 2005].
Staging of HCC can be done using several systems. Currently, the most widely used staging system is the Barcelona Clinic liver cancer (BCLC) staging system, which takes the underlying liver disease, tumor characteristics, as well as the general performance status into account [Bruix and Llovet, 2009]. This staging system is popular as it is directly linked to treatment, making treatment decisions easy (Figure 1).
In Western countries, about 30% of patients are identified with an HCC in BCLC stage 0 or A, either through surveillance or by chance. For those patients curative options can often be applied, which currently involve only surgical or interventional treatments (Figure 1). However, curatively treated patients, except for those patients who underwent transplantation, will have a tumor recurrence in 70 to 80% of cases within 5 years of therapy and will eventually progress to BCLC B or BCLC C stage disease [Livraghi et al. 2008; Llovet et al. 2005]. Another 20% of patients are diagnosed at a very advanced stage BCLC D, being either symptomatic from the decompensated cirrhosis (Child–Pugh C) or having an advanced tumor. Those patients have a very short survival, which cannot be influenced by any therapeutic intervention and are only eligible to receive best supportive care.
Currently, the domain of medical therapies for HCC is in the setting of advanced stage BCLC C.
Conventional chemotherapy of any kind has never shown any meaningful therapeutic benefit, particularly in overall survival in randomized controlled trials in adult patients [Thomas et al. 2008; Mathurin et al. 1998] and cannot be recommended for the treatment of HCC today. Conventional cisplatin-based chemotherapy (with or without doxorubicin) only has a place in the treatment of childhood hepatoblastoma, which is a distinctly different disease entity. In hepatoblastoma cisplatin-based chemotherapy does improve survival and can even provide a cure in over 80% of patients when combined with resection [Perilongo et al. 2009].
Medical treatment of HCC changed dramatically in 2007, when the first data from the successful use of targeted agents, in particular tyrosine kinase inhibitors, in advanced stage HCC were presented.
Sorafenib is a multikinase inhibitor with antiproliferative (through the inhibition of Raf-kinase), pro-apoptotic (through Mcl-1 inhibition), and anti-angiogenic (through the inhibition of vascular endothelial growth factor receptor-2 – VEGFR-2 – platelet-derived growth factor receptor – PDGFR – and Raf-kinase) properties [Wilhelm et al. 2004]. Positive data on 137 patients with advanced stage HCC treated in a single arm, phase II trial were published in 2006, showing a median survival of 9.2 months and a time to progression of 4.2 months in a mixed population of Child–Pugh A (72%) and Child–Pugh B (28%) patients [Abou-Alfa et al. 2006]. These data led to the initiation of a large multicenter randomized controlled phase III trial (SHARP trial) testing sorafenib against placebo in 2005. The SHARP trial for the first time after many trials on the medical treatment of HCC was able to show a significant survival benefit (10.7 versus 7.9 months, p=0.001; hazard ratio 0.69) (Figure 2) for sorafenib over placebo [Llovet et al. 2008]. This positive effect was maintained in almost all subgroups analyzed and there was also a significant improvement in the time to progression with sorafenib over placebo (5.5 versus 2.8 months). Sorafenib seemed to be of particular benefit in patients with chronic hepatitis C, where survival was prolonged from 7.9 to 14.0 months (hazard ratio 0.58) [Bolondi et al. 2008]. This could be due to the antiviral action of sorafenib that has been described recently in the hepatitis C virus-replicon system [Himmelsbach et al. 2009], but this observation is still awaiting confirmation in patients. At the time of publication of the SHARP trial, sorafenib was already the accepted standard of care for advanced HCC in Europe, the USA, and other parts of the world.
A parallel study to the SHARP trial was conducted in the Asia-Pacific region, which had the same inclusion and exclusion criteria and was designed as a smaller confirmatory trial with 226 Asian patients [Cheng et al. 2008]. Again, sorafenib was able to show a highly significant improvement in survival versus control (6.2 versus 4.1 months, p=0.014, hazard ratio=0.68) and in time to progression (TTP) (2.8 versus 1.4 months). The main difference between the two trials is in the patient characteristics. In the Asia-Pacific trial, over 80% of the patients had chronic viral hepatitis (over 70% chronic hepatitis B), more patients had multilocular disease, and 50% of patients had lung metastasis. Most likely due to the different patient and disease characteristics, the absolute survival and TTP figures were much lower in the Asia-Pacific trial than in the European trial, highlighting the effects of the underlying liver disease on outcome in HCC patients, as repeatedly mentioned [Llovet et al. 2008].
Of key importance for the success of the SHARP trial was patient selection, allowing only Child–Pugh A patients, BCLC B and C, and performance status 0–2 to be recruited into the trial. In that way the large variability in survival caused by the natural course of the underlying cirrhosis was minimized and allowing to investigate the drug effect on the survival of HCC- patients.
Even though sorafenib is the standard of care for advanced stage HCC and is registered for the treatment of HCC without restrictions by the European Medicines Agency (EMEA) and the Food and Drug Administration (FDA), the narrow inclusion criteria of the SHARP trial leave many patients without a proven efficacious treatment with regard to their disease stage. One point of discussion is the applicability of the treatment benefit to patients with higher Child–Pugh stages. In the phase II trial [Abou-Alfa et al. 2006], 38 patients (28%) had Child–Pugh B status. Their overall survival was significantly shorter compared with Child–Pugh A patients, but their antitumor response was identical, with about 50% of tumors showing an objective reduction in size (unpublished data). The uncontrolled experience from the Austrian extended access program, which included Child–Pugh A as well as Child–Pugh B and C patients, indicates quite clearly that there is no role for sorafenib in the most advanced patients. Child–Pugh C patients in this series had no clinically meaningful benefit with a median survival of 1.6 months with treatment [Pinter et al. 2009]. In the same series, 23 patients with Child–Pugh B cirrhosis had a median survival of 6.5 months with treatment. The effectiveness of this approach has to be regarded as unproven to date, owing to the lack of a control group without treatment, but the survival of Child–Pugh B patients was the same as the survival of the patients receiving sorafenib in the Asia-Pacific trial. The current approach to patients with Child–Pugh B cirrhosis and advanced HCC in many centers is to evaluate each patient individually and decide on a patient by patient basis [Peck-Radosavljevic et al. 2010], taking into account variables like decompensation, Child–Pugh score, performance status, nutritional status, and desire to be treated.
Sunitinib is an oral multikinase inhibitor with strong anti-angiogenic effects through inhibition of VEGFR and PDGFR, but lacking the antiproliferative Raf-inhibitory activity. Its similar mode of action and efficacy to sorafenib in renal cell carcinoma led to two phase II trials of sunitinib in advanced stage HCC. The European trial used the standard dose of sunitinib (50mg/day) for treatment of renal cell cancer [Faivre et al. 2009] and showed a median overall survival of 8 months in HCC-patients. The USA trial used a lower dose (37.5mg/day) [Zhu et al. 2009] with a median overall survival of 9.8 months. Toxicities were considerable: grade 3/4 thrombocytopenia and leucopenia occurred in 38 and 24% of patients, respectively, and hand–foot skin reaction in 10%, even though sunitinib was given for 4 weeks with 2 weeks off in between cycles. About 10% of the deaths in the trials were judged to be treatment-related [Faivre et al. 2009]. Even though the European trial was judged a failure due to the low response rate and high toxicity, and the USA trial used an intermittent dosing regimen for sunitinib, a large multinational phase III trial of sunitinib versus sorafenib is currently being conducted using 37.5mg/day sunitinib given continuously. Results are eagerly awaited, but it appears from the available data that safety will be a major issue in this trial.
Brivanib is another oral multikinase-inhibitor with predominant action against the VEGFR-2 and fibroblast growth factor receptor (FGFR)-1 tyrosine kinases [Huynh et al. 2008]. It has been tested in a phase II trial that included patients with advanced stage HCC in two different patient cohorts [Raoul et al. 2009]. One cohort included patients naïve for an anti-angiogenic treatment and showed a disease control rate of 60% with an overall median survival of 10 months, similar to what can be expected from treatment with sorafenib. The second cohort investigated the effectiveness of brivanib in a second-line setting in patients that had failed one prior anti-angiogenic therapy (94% progression under sorafenib, 6% thalidomide failures). This cohort had a slightly lower disease control rate of 46% but a very remarkable median survival of 9.8 months. These data are the first reported for any targeted therapy in a second-line setting after sorafenib failure. Since treatment failure will happen eventually in all patients on sorafenib, there is a great medical need for a second-line treatment with proven efficacy, and the median survival of 9.8 months after progression with sorafenib (median TTP in the SHARP trial was 5.5 months) offers hope for another step forward in improving survival of patients with advanced stage HCC. Brivanib is currently being tested in three phase III trials: a first-line trial of brivanib versus sorafenib in advanced stage HCC, a second-line treatment of brivanib versus placebo in sorafenib failures, and a randomized trial of brivanib versus placebo in combination with transarterial chemoembolization (TACE). The results of these trials, in particular the second-line trial, may be available within 2 years.
Erlotinib, an orally active inhibitor of the epidermal growth-factor receptor (EGFR) signaling, has been tested as monotherapy for advanced stage HCC in two phase II single-arm trials [Thomas et al. 2007; Philip et al. 2005]. The disease control rate in the study by Philip and colleagues was 60%, the median overall survival was 13 months, and 88% of the patients had EGFR1-expression in the tumor, which did not correlate with outcome [Philip et al. 2005]. These findings were largely confirmed in the trial by Thomas and colleagues [Thomas et al. 2007]. Erlotinib is not being tested as a single agent treatment for HCC but in several combination trials (see below). Again safety will be a concern in those trials, since erlotinib has been reported to be able to induce acute liver failure in rare instances [Liu et al. 2007] and some of the side effects may be additive.
Bevacizumab is an intravenously administered monoclonal antibody blocking binding of VEGF to VEGFR1 and VEGFR2, with proven efficacy in metastatic colorectal cancer [Hurwitz et al. 2004] and several other cancers. It has been and is being tried in several trials up to phase II as monotherapy [Siegel et al. 2008] or combination therapy with either interventional or medical therapies [Thomas et al. 2009; Zhu et al. 2006]. Median overall survival was heterogeneous in these trials as was patient selection and treatment algorithms, and ranged from 9.6 months with gemcitabine and oxaliplatin in combination with bevacizumab (GEMOX-B) [Zhu et al. 2006], through to 12.4 months with bevacizumab monotherapy, to 15.6 months in the bevacizumab–erlotinib combination trial [Thomas et al. 2009]. All of these trials were single-arm trials without a control group, precluding any conclusion about efficacy [Llovet et al. 2008]. There are, however, safety concerns with bevacizumab, which caused grade 3/4 gastrointestinal-hemorrhage in 11% [Siegel et al. 2008] to 12.5% [Thomas et al. 2009] of patients treated, even though prophylactic obliteration of esophageal varices was standard in most of these trials. With other potent and less problematic candidate drugs available, it seems less likely that bevacizumab will play a major role in drug treatment of HCC.
Cetuximab, a monoclonal antibody against the EGFR, was tried as a monotherapy as well as in combination with chemotherapy for advanced stage HCC. In the monotherapy trial, cetuximab showed no objective response with a progression-free survival of 1.6 months and an overall survival of 9.6 months [Zhu et al. 2007]. In combination with gemcitabine combined with oxaliplatin (GEMOX), cetuximab showed a disease control rate of 60% with an overall survival of 9.5 months and grade 3 toxicities in 49% of patients [Asnacios et al. 2008], indicating that cetuximab is not a likely candidate for further development.
Imatinib showed very little effect in advanced stage HCC in a small trial [Lin et al. 2008], and is also not a likely candidate for treating HCC in this setting.
Several new molecules are currently being tested. Among those, ABT-869, a VEGFR- and PDGFR-inhibitor has shown promising results in phase II studies [Toh et al. 2009], just like BIBF-1120, a VEGFR-/PDGFR-/FGFR-inhibitor blocking the MAPK- as well as the mTOR-pathways [Hilberg et al. 2008].
One very obvious strategy to improve the outcome for patients with advanced stage HCC would to be to combine targeted agents with single agent activity against HCC and complementary modes of actions. So it seems a logical choice to combine the only agent with proven positive effect on patient survival, sorafenib, with an agent that has different targets other than VEGFR2 or PDGFR. Erlotinib with EGFR-blocking activity could be such an agent. Phase III testing of this combination strategy versus sorafenib monotherapy started in May 2009, recruitment is ongoing, and results of this first combination therapy trial (SEARCH trial) are eagerly awaited. Potential downsides could be additive side effects like diarrhea, as seen in a pilot trial of this combination in up to 88% of patients [Duran et al. 2007], and the peculiarities of erlotinib metabolism, which is influenced not only by smoking but also by a number of drug interactions.
Before the results of the SHARP trial became available, a phase II combination treatment trial of doxorubicin with or without sorafenib had already been initiated in the USA. This was carried out even though the rationale of using a chemotherapeutic drug without any proven effect on patient survival in such a trial seems at least debatable, and the use of doxorubicin in the control group seemed even at that time ethically questionable. Survival in the control arm of that trial was actually worse than survival in the control arm of the SHARP trial, indicating that there should no longer be a place for doxorubicin in the systemic treatment of HCC. Median survival in the treatment group was 13.7 months and significantly better than in the control arm [Abou-Alfa et al. 2007]. If this combination strategy is to be pursued any further, this would mandate a trial comparing the combination treatment with sorafenib monotherapy.
Another similar approach was taken in a combination therapy trial exploring GEMOX-B [Zhu et al. 2006]. Median survival in this single-arm trial was 9.6 months, which was shorter than in a bevacizumab monotherapy trial and in the bevacizumab-erlotinib combination therapy trial. Since it is virtually impossible to compare different single-arm trials of HCC due to the potential heterogeneity of the patients included, no conclusions can be drawn from these data. It appears, however, that GEMOX, which was never shown to be beneficial in the treatment of HCC, should not be used in the treatment of HCC, neither as monotherapy nor in combination with targeted agents.
The current standard of care in treatment of intermediate stage HCC, which makes up between 30 and 40% of HCC patients, is TACE [Bruix and Llovet, 2009]. TACE has been performed with a mixture of a chemotherapeutic agents (in Western countries mostly doxorubicin, in Asia often cisplatin) and lipiodol until recently, often followed by the application of gelfoam particles to induce complete stasis of blood flow to the tumor tissue [Bruix et al. 2004]. Aside from some theoretical inconsistencies in the concept of TACE, such as using chemotherapeutic agents without proven survival benefit and no consensus on how to prepare the embolizing agent before application [Marelli et al. 2007], it took at least seven prospective randomized trials and a meta-analysis to demonstrate any survival benefit of this intervention [Llovet and Bruix, 2003]. TACE is an imperfect standard treatment with an objective response rate between 16 and 61% and a median survival of 11–20 months. If only the best candidates with Child–Pugh A status, performance status 0, and without segmental portal vein invasion are selected (less than 15% of patients with HCC), response rates between 30 and 50% can be observed and survival was linked to response in several studies [Bruix et al. 2004]. Eventually, however, 70–80% of BCLC B patients will die of tumor progression. This is often due to neo-angiogenesis originating at the rim of treated lesions, which is caused by the up-regulation of pro-angiogenic and anti-apoptotic mechanisms and factors through the induction and stabilization of the transcription factor HIF-1α by the hypoxia inflicted by the disruption of blood flow to the tumor tissue [Sergio et al. 2008; Yang et al. 2004]. Since VEGFR levels increase immediately after TACE [Li et al. 2004], are more prominent if TACE is incomplete, and correlate with survival after TACE [Sergio et al. 2008], it seems very reasonable to apply anti-angiogenic drugs at the time of TACE or, maybe even better, before TACE in order to prevent an increase in the production of pro-angiogenic mediators and signaling through their receptors. To date, no data on this strategy have been reported in full but a phase III trial of sorafenib after TACE has been completed in Japan and several trials with drugs such as bevacizumab, sorafenib, or brivanib are currently ongoing. The most advanced is a large phase II trial of the combination or sorafenib starting before TACE with doxorubicin-coated beads (DC-beads), which is planned to complete patient inclusion in 2010 (SPACE trial). Toxicity will be an important issue in these combination therapy trials and strategies with low systemic exposure to cytotoxic chemotherapeutic agents like the use of DC-beads will be of advantage.
Medical therapies for HCC have not been evaluated to any extent in the adjuvant setting but several trials are currently ongoing. Since tumor recurrence 5 years after curative treatment with resection or radiofrequency ablation (RFA) amounts to 70–80%, adjuvant treatment approaches constitute a great medical need. The largest current trial in this setting is the STORM (Sorafenib as adjuvant Treatment in the prevention Of Recurrence of hepatocellular carcinoMa) trial. In this randomized placebo-controlled phase III trial, sorafenib or placebo are given after curative resection or RFA to 1100 patients for up to 4 years with a primary endpoint of recurrence-free survival. A different approach is taken by researchers trying to evaluate the effect of sorafenib in a neo-adjuvant setting for patients awaiting liver transplantation for HCC. Even though the concept is intriguing, it should only be employed in clinical practice after rigorous clinical testing with proven survival benefit. It seems clear that a potential benefit in survival in the adjuvant and neadjuvant setting will come at a cost: this will not only be monetary, but will be the considerable side effects in a population that otherwise would live a normal life with minimal health problems and symptoms. And it should not be forgotten that even though sorafenib leads to better local tumor control, its effect on metastasis formation and proliferation has not been assessed sufficiently. There are experimental data with angiogenesis inhibitors that give us reason to worry about the use of targeted agents, in particular those that block VEGFR, in the adjuvant setting after curative treatment [Paez-Ribes et al. 2009].
The current role of medical therapies for HCC is in patients with advanced stage HCC (BCLC C) and good liver function, where they have a proven positive impact on patient survival. Here, a sizable number of new drugs are being tested in phase II and III clinical trials and will extend treatment options. Most importantly, we will also have second-line and maybe even third-line options in this setting, and studies exploring combination therapy with targeted agents may give an answer to the question of whether sequential or combined treatments will be the preferred first-line option. Extension of medical therapies to intermediate stage patients (BCLC B) is being tested. Here, combination strategies with interventional treatments will most likely show the greatest benefit but to date we do not have the data to endorse such strategies. Medical therapies for HCC in the adjuvant setting after curative treatment are being tested rigorously and should not be introduced into clinical practice prior to a positive outcome of those clinical trials. Proof of efficacy of medical therapies for patients with advanced stage HCC and less favorable disease characteristics, such as portal invasion or reduced liver function, is not available to date and studies should be conducted in order to answer these important questions. No effective treatments are available for patients with end-stage disease (BCLC D), and it seems unlikely that current treatment approaches will offer any promise in the near future for these patients.
Dr Markus Peck-Radosavljevic has served as a consultant for Bayer Schering Healthcare, Boehringer-Ingelheim. He has received honoraria for speaking from Bayer Schering Healthcare, AstraZeneca. He acts as investigator for Bayer Schering Healthcare, Roche, Novartis, Abbott, Boehringer-Ingelheim, and Britol-Meyer-Squibb. He has also received grant support from Bayer Schering Healthcare and Roche.