Is there an opportunity for personalised medicine in HCC?

Sorafenib became the standard treatment for advanced HCC five years ago after findings from the SHARP trial showed that treatment with this multi-targeted tyrosine kinase inhibitor (TKI) resulted in a 2.8 month improvement in overall survival (OS) compared with placebo. However, since then, results from Phase 3 trials evaluating targeted agents in this setting have been disappointing.

In April 2010, the SUN trial was discontinued after early data showed that sunitinib was inferior to sorafenib in terms of OS and was more toxic, despite the fact that this agent has a very similar molecular target profile to sorafenib. Then in December 2011, Bristol-Myers Squibb announced that the VEGFR- and FGFR-targeted agent, brivanib, did not improve OS as 2nd-line therapy compared with placebo in the BRISK-PS trial. This news was followed by a similar announcement in July 2012 for the BRISK-FL trial of brivanib as 1st- line therapy, which also failed to meet its primary OS endpoint. Continuing this pattern here at ESMO, findings from the SEARCH trial, reported during Sunday’s Presidential Symposium, showed that the addition of erlotinib to sorafenib does not improve OS compared with sorafenib alone (median OS: 9.5 months [erlotinib + sorafenib] versus 8.5 months [sorafenib]; HR 0.929; CI: 0.781–1.106, p=0.204).

Commenting on this increasing tally of negative Phase 3 trials in advanced HCC, Professor Roberto Labianca, an expert in the treatment of gastrintestinal cancers from Riuniti Hospital, Bergamo, Italy, said “we’ve been conducting the same trials for 15 years – it’s time we started selecting our patient populations based on the molecular target profile of the agents being studied, as has been done so successfully in other areas, for example, with trastuzumab in HER2-positive breast cancer”.

Suggestions that this may indeed be the way forward in HCC were strengthened at ESMO 2012 when Dr Bruno Daniele from Rummo Hospital, Benevento, Italy, presented final results from a randomised Phase 2 trial of the MET inhibitor, tivantinib (formerly known as ARQ 197), as 2nd-line therapy in unresectable HCC. In this trial, 107 patients with Child-Pugh (CP) A liver function and a Performance Status (PS) of <2 were randomised 2:1 to receive tivantinib or placebo. Although the primary endpoint of time to progression (TTP) showed a small benefit in favour of tivantinib in the overall intent-to-treat (ITT) population (median TTP: 6.9 versus 6.0 weeks; HR 0.64; 95% CI: 0.43–0.94, p=0.04), efficacy results among MET-positive patients treated with tivantinib were much more promising: median TTP was 11.7 weeks (versus 6.1 weeks for placebo; HR 0.43; 95% CI: 0.19–0.97, p=0.03), disease control rate (DCR) was 50% (versus 20% for placebo), and median OS was an impressive 7.2 months (versus 3.8 months for placebo; HR 0.38; 95% CI: 0.18–0.81, p=0.01). Dr Daniele commented that “the pronounced activity of tivantinib in MET-positive patients seen in this trial warrants further evaluation, and a Phase 3 trial in MET-positive HCC patients is being planned.” Indeed, if these findings are confirmed in a Phase 3 trial, tivantinib would represent the first step towards personalised medicine in HCC.

Given this glimmer of hope, Professor Labianca suggested that it may be appropriate to “dig a little deeper” into the results obtained from recent negative Phase 3 trials. For example, as preclinical data suggest that the FGF pathway is implicated in the development of resistance to anti-VEGF therapy, could brivanib provide clinical benefit for relapsed patients with high FGF expression? Similarly, as the benefits of anti-EGFR therapy have been shown to be greater for patients with EGFR activating mutations in NSCLC, could erlotinib + sorafenib be an option for EGFR mutation-positive patients with advanced HCC? Or, put another way, would the BRISK-PS and SEARCH trials have been positive if patients had been selected based on expression of key molecular targets?

Negative data from the SEARCH trial reported here at ESMO also raise an important question regarding the viability of combination therapy in advanced HCC. “We mustn’t forget that the vast majority of patients with HCC also have underlying liver disease, most commonly caused by alcohol abuse, hepatitis B or hepatitis C infections.” Professor Labianca warned. “Many of these patients are very sick and it may be that combination therapy is just too toxic” he added.

Interestingly, the effect of liver disease aetiology on sorafenib therapy was the subject of a presentation yesterday by Professor Jean-Pierre Bronowicki from the University Hospital of Nancy, Vandœuvre-lès- Nancy, France. Based on findings from the second interim analysis of GIDEON, a large observational study that is gathering information on the use of sorafenib in everyday clinical practice, Professor Bronowicki presented data to suggest that toxicity (drug-related adverse events and serious adverse events) was greater among patients with hepatitis C infection, whereas median OS appeared to be longer in this group compared with patients with hepatitis B infection or alcohol abuse as the underlying cause of liver disease. However, despite these apparent differences, Professor Bronowicki warned that these variations may reflect differences in prognosis associated with the natural history of the underlying liver disease rather than a difference in response to therapy. Nevertheless, data from studies such as GIDEON highlight the importance of considering both the cancer and the liver disease when treating patients with HCC, and serve as a strong reminder of the need for a multidisciplinary approach to patient care.

So, how can we move towards personalised medicine in HCC?

Answering this question, Professor Stefano Fagiuoli, from Riuniti Hospital, Bergamo, Italy, stresses that "We need to refine our selection criteria for patient eligibility for treatment. These criteria need to be based on a comprehensive evaluation of both the aetiology of the underlying liver disease and a more reproducible assessment of portal hypertension, the key factor for the risk of decompensation." Adding to this, Professor Labianca comments that “We need to learn from our mistakes and ensure that future clinical trials of targeted therapies are conducted in selected patient populations based on the expression of key molecular targets”.