The Role of TKIs in Hepatocellular Carcinoma

MedPage Today brought together three leaders in the field of liver cancer to discuss the latest research and clinical advances. In this third of four exclusive episodes, moderator Ghassan K. Abou-Alfa, MD, of Memorial Sloan Kettering Cancer Center in New York City, is joined by Amit Singal, MD, of UT Southwestern Medical Center in Dallas, and Peter R. Galle, MD, PhD, of Mainz University Medical Center in Germany, for a roundtable discussion on the role of tyrosine kinase inhibitors (TKIs) in the treatment of hepatocellular carcinoma (HCC).

Following is a transcript of their remarks:

Abou-Alfa: Hello everybody, and thanks for joining us on this roundtable on liver cancer. It’s a great pleasure to have you on board and my name is Ghassan Abou-Alfa from Memorial Sloan Kettering Cancer Center, New York. I’m joined today by dear colleagues Dr. Peter Galle from University Mainz, Germany, as well as Dr. Singal from University of Texas Southwest in Dallas.

We were all at ESMO [European Society for Medical Oncology] meeting and we heard about the LEAP-002 study, a study of another doublet combination of a checkpoint being pembrolizumab [Keytruda] plus lenvatinib [Lenvima] as a tyrosine inhibitor. And shockingly it was negative for its primary endpoint.

But at the same time it did showed some other intriguing things, which really makes us revisit the role of the tyrosine inhibitors, specifically single-agent lenvatinib — 20-plus month median survival for the combination of pembrolizumab plus lenvatinib and 19.2 months for lenvatinib alone.

Dr. Singal, your thoughts on that?

Singal: Yeah, Ghassan, I think a lot of us were quite interested in this study, given the fact that the combination of lenvatinib and pembrolizumab had very exciting data in the phase II study, as you know. And then it looks like that sort of promising data actually bore out in the phase III study. But I have to say myself, and I think many others, were surprised by this 19-month survival for lenvatinib. And I think there’s a couple of things that probably contribute to this.

There’s some comfort with TKIs that happens over time. We saw the same thing with sorafenib [Nexavar], where you saw a maturation of the survival between SHARP [Sorafenib Hepatocellular Carcinoma Assessment Randomized Protocol] and more recent studies that showed a gradual improvement in survival over time as we got better in terms of managing TKI-related AEs [adverse events], keeping people on drug.

When you take a look at the duration on treatment for lenvatinib in the LEAP-002 study, it’s actually quite long — some of the longest duration on treatment that we’ve seen in a study for TKIs.

I think the other thing that we can see is that we clearly have become better … of getting people on systemic therapy earlier. When they’re doing better they have more preserved liver function, not waiting for a very large disease where that patient has a worse prognosis. So I think you’re starting to see us be better at treatment selection and patient selection over time as well. And I think those things likely contribute to sort of the improved survival that we see with lenvatinib.

And then the third thing is post-regression treatment. Clearly we have more and more therapies on market. When somebody fails first-line therapy, we have effective options in the second-line setting.

And so I think these three factors together likely contributed to what we can see and achieve if we do a good job with patients in the systemic therapy landscape, even with some of these older TKI agents.

Abou-Alfa: No, no, by all means, I totally agree with you. It was a little bit of an awakening for all of us. And I go to Dr. Galle here, and Peter, so does it mean that we have to revisit or at least re-study the idea of doublet versus sequential therapy — i.e., maybe a TKI followed by checkpoint [inhibitor]? Your thoughts on that? Where are we going?

Galle: It has proven to be particularly difficult to perform clinical trials in HCC, and that has in a way gotten worse in the time of immunotherapy. Think about the RATIONALE trial versus CheckMate — nivolumab [Opdivo] investigation in the first-line. These trials were basically investigating the same sort of situation — IO [immunotherapy] versus sorafenib.

The CheckMate 459 trial was negative because it was going for superiority and the TESLA trial was positive because it was going for non-inferiority. Basically the same results — one trial positive, the other negative. This is arbitrary. So we do have a clinical trial-design problem here.

And the other issue is if you look at all these curves, the Kaplan-Meier typically with the immunotherapies, they separate in a way quite nicely, and if you go for HIMALAYA, they separate late, and this is very difficult. I mean definitely you don’t capture with PFS [progression-free survival] what’s happening late. The lack of biomarkers is coming into play here. We would need to know who is in the end at the tail of the curve where the plateau occurs and is benefiting.

So I would say we have here issues which are going beyond the simple question of whether drug is active or not. I think most of us believe that the combo [pembrolizumab plus lenvatinib] is actually an active one, and that this trial failed is actually very disappointing. But in the end there are aspects — and what Amit said about post-trial medication — that’s probably most relevant. I mean now we have many lines of therapy and OS [overall survival] is no longer a reliable endpoint for a single agent.

Abou-Alfa: Thanks so much, Peter. Both Dr. Singal and Dr. Galle, you brought up very important points. Let me try to help summarize, because this is really … it’s getting complex by the second.

Number one, we should not consider that any combination is just another combination. Combinations do differ because it all depends on where you are affecting the cancer immunity cycle. Because as we all know, anti-PD-1/anti-PD-L1 activity is to be really inhibited with the anti PD-1. In other words inhibiting that activity of PD-1/PD-L1.

But now as Dr. Galle just told us, the 459 study of nivolumab by itself did not really show superiority. Because clearly now we understand that there’s a need for kind of like a phone call from up in the chain of command of something enhancing that anti-PD-1/anti-PD-L1 activity happening. That call could be from anti-VEGF.

Again, as Dr. Galle mentioned to us about the atezolizumab [Tecentriq] plus bevacizumab [Avastin], it could be the benefit, in other words, causing that enhancement. Or it could be all the way up on the top of the chain of command in the lymph node with the anti-CTLA-4.

Interestingly, automatically, to just throw in anything onto that cancer cycle is not necessarily always the key element or key players. We can’t really assess, we can’t really critique, the pembrolizumab. I think anti-FGF [fibroblast growth factor] does have a role in regard to the cancer immunity cycle, but for example if you take anti-c-MET with cabozantinib [Cabometyx] plus atezolizumab was again a negative study because probably is a bit far off in regard to that sequence or that combination per se. That’s one component.

Another component is, which a little bit more mathematical but very critical, which I think Dr. Singal brought in also, is what are we exactly trying to look into in regard to the outcomes. No doubt, as also Dr. Galle mentioned, these studies probably are all positive. It depends what we’re reading onto. Because as we know, this documented perspective that we are used to with chemotherapeutic interventions is a little bit old now in regard to how checkpoint inhibitors work. Because there’s a certain priming of the T cells that’s needed to prime the effect as nicely, Dr. Galle just brought into us how the curves might separate and they can have what we are now used to here, the long tails on the curve.

And interestingly, in other words, do the medians matter when it comes to checkpoint inhibitors or shall we look into what we call landmark analysis at a certain point in time that’s showing the separation per se? This is an evolving science, and no doubt that version 2.0 how clinical trials are to be run, especially with checkpoint inhibitors, is coming.

Now where will the role of the tyrosine kinase inhibitors as single agent play? No doubt that the LEAP-002 study with the negative outcome — ironically, where some of us are calling it like the most positive negative study ever, because it showed 19.2 months for the lenvatinib single agent.

I think Dr. Singal mentioned to us about maybe we’re seeing patients earlier anyway, getting more therapeutic intervention. And remember also the role of the therapies that happen afterward, and as such does not efface with all of this, the role of TKI as a single agent.

Watch episode one in this series: Recent Advances in Systemic Therapy for Hepatocellular Carcinoma

Watch episode two in this series: Combination Treatment Approaches in Hepatocellular Carcinoma