Ghassan K. Abou-Alfa, MD: And, Ruth, while we’re here in San Francisco for the 2018 GI ASCO, there will be data on the use of cabozantinib in regard to HCC, and the phase III trial looked at cabozantinib versus placebo in the second-line setting. Before I ask you about that study though, my understanding is that cabozantinib is a c-MET inhibitor. But my understanding is that there was an active trial on a c-MET inhibitor with tivantinib beforehand?

A. Ruth He, MD, PhD: Yes. Tivantinib is a little different from cabozantinib. It does have an offset of target effect. And it was a very promising study. We have a lot of hope for that study. It’s the first study that actually selects patients based on a biomarker, c-MET level expression, in patients with HCC tissue. But that study was negative. We believe that because the drug has 2 targets, that sort of makes the data hard to interpret. And so, cabozantinib is a different drug compared to tivantinib. And the current positive data in the randomized phase III trial seems likely that it will bring us another second-line agent into the treatment of advanced stage HCC.

Ghassan K. Abou-Alfa, MD: So, Catherine, back to you. That notion that Ruth is bringing up is in regard to the c-MET expression and the specificity of tivantinib. But as you know, and you’re the expert in that, hepatocyte growth factor is a target for this c-MET, and of course the liver is full of hepatocyte growth factor. The tivantinib study suggested that the only patients who should go on that study, if I recall, was 50% 3-to-4 plus expression. Is this really like the make it or break it point? Or, how much c-MET do you need with that liver, per se?

Catherine Frenette, MD: I don’t think we know the answer to that honestly. And I think that’s part of why the tivantinib study failed. And unfortunately, we’ve looked at a lot of different biomarkers on biopsies, testing for PD-L1 on biopsies, which didn’t pan out for nivolumab response. So, c-MET didn’t pan for tivantinib response. We’ve looked at EGFR, we’ve looked at FGF, we’ve looked at HGF, and really none of these biomarkers yet have panned out. And I think that part of the reason is become HCC is a really heterogenous cancer, even within 1 tumor, or within 1 patient. There can be really multiple different factors affecting the growth of it.

Ghassan K. Abou-Alfa, MD: I totally agree. And another thing—I’m honored now that the meeting has completed—is the cabozantinib study that I presented where the cabozantinib arm showed a median survival of 10.2 months. And in addition to that, the placebo arm had a mean survival of over 8 months. And if anything, the study was statistically positive in regard to the outcome. And, of course, it shows that benefit. If anything, I comment on what Catherine said and I totally agree because when we designed the study to begin with, we specifically avoided that selectivity of the patient with c-MET expression because, you’re right, we don’t know how much. Probably the same logical use in regard to hormonal therapy for breast cancer, because it doesn’t matter how much hormone expression, but you’ll end up with the application of hormonal therapy despite even if it was a whiff of hormone therapy, per se. So, that’s really one of the reasons why the tivantinib story was really looked at a little different from the cabozantinib story.

But to go back to even more into those multivariant choices of therapy, we spoke already about sorafenib, we spoke about regorafenib, we spoke about nivolumab. Then we brought up lenvatinib. We just brought up cabozantinib. But also, and Anthony you already mentioned, there are data also that were reported at that meeting in regard to the KEYNOTE-224 of pembrolizumab. Can you tell us more about that?

Anthony El-Khoueiry, MD: Right. So, KEYNOTE-224 brings us some phase II data about the activity of pembrolizumab in the patients who were previously treated for advanced hepatocellular carcinoma. I think the study had slightly over 100 patients. And the punchline really is that we see a response rate of 16%, which is certainly consistent with what was seen with nivolumab. And the toxicity profile also seems to be consistent with the profile of pembrolizumab, or checkpoint inhibitors in general in other tumor types. So, it provides more confidence in the whole activity of checkpoint inhibitors in this disease.

Ghassan K. Abou-Alfa, MD: Yes, I know, it’s fascinating. If anything, what we just heard about is there are multiple therapies and no question, the key question that we all have is, how can we ultimately decide which therapy to use? And I think we had already some suggestions but just to kind of relay them, we have the number 1 question: the TKIs versus the checkpoint inhibitors, and which role is each one going to play, the first- or second-line? We heard very clearly from Catherine that not necessarily every patient will be eligible for the checkpoint inhibitors. They could very much not be the right patients for the checkpoint inhibitors. And then, of course, if we are to talk into the TKI component of the discussion, is it sorafenib followed by regorafenib? Is it lenvatinib maybe followed by cabozantinib? No question that many variables will come along here. Some of them, as Ruth suggested, might be directly related to the side effects. But maybe a different thinking that can come along as we discussed can come, for example, from the etiology, or the risk factor for developing the HCC, the ethnicity of the patient—Asia versus New Asia—the unfulfilled 14 level, and, of course, also any indirectly related or correlated markers—including, for example, the c-MET, even though we just agreed that the cabozantinib worked in everybody regardless of the c-MET expression.

Transcript Edited for Clarity