KEYNOTE-426: Pembrolizumab plus Axitinib versus Sunitinib as First-Line Therapy for Advanced Renal Cell Carcinoma (RCC) - Brian Rini

May 28, 2020

Brian Rini joins Charles Ryan to discuss frontline therapy in kidney cancer, specifically focusing on the recent updates on the combination of axitinib and pembrolizumab in the Keynote-426 data. KEYNOTE-426 study, which is at the two-year mark of followup, is a randomized, open-label, phase 3 study that demonstrated that pembrolizumab (pembro) + axitinib (axi) significantly improved overall survival (OS), progression-free survival (PFS), and objective response rate (ORR) vs sunitinib as first-line therapy for advanced RCC (aRCC) at the first pre-planned interim analysis. Dr. Rini highlights takeaways of the updated analyses presented during the ASCO 2020 virtual scientific meeting.

Biographies:

Brian Rini, MD, Professor of Medicine, Chief of Clinical Trials, Vanderbilt University Medical Center

Charles J. Ryan, MD, The B.J. Kennedy Chair in Clinical Medical Oncology at the University of Minnesota and Director of the Division of Hematology, Oncology, and Transplantation.


Read the Full Video Transcript

Charles Ryan: Hello, welcome. I'm delighted to be talking to colleagues and an old friend, Brian Rini, who is a Professor of Medicine and Chief of Clinical Trials at the Vanderbilt-Ingram Cancer Center in Nashville. We're going to talk about kidney cancer and recent updates on the combination of axitinib and pembrolizumab and just in general about frontline therapy for kidney cancer. Brian, thank you for joining us. Happy to have you here.

Brian Rini: KEYNOTE-426, as you know, was presented and published a little while ago now. It's one of the main frontline treatment options in kidney cancer. So basically in kidney cancer, patients are getting what I would call IO based doublets, so either it ipi/nivo or commonly this regimen of axi/pembro, so either two immune agents combined, or an immune agent and a VEGF inhibitor. What's being updated at ASCO is basically an additional follow-up for the KEYNOTE-426 study.

Charles Ryan: Great. The original paper was published and demonstrated a very clear radiographic progression-free survival benefit. It's already, I think, already established as one of the standard frontline options. What's the new data about?

Brian Rini: The original data, it was actually meant to be a PFS analysis. OS was looked at. The median follow-up in that first report was only about 12 months, 12.8, I think was the number. Because the OS was so strikingly positive, it was statistically significant and obviously sort of met all the endpoints and was reported.

This follow-up is about an additional 17 months. It's about .3 minimum months and 27 I think, median. The first report was admittedly very early, although strikingly positive. And now we're getting to about that two-year mark of follow-up, so a little better. These IO studies, as you know, they really need long followup because that's the benefit of IO for some of the benefits.

Charles Ryan: Right and for listeners, I think one of the key aspects of thinking about Phase III studies is as they mature, we want to know a couple of key things. We want to know the percent of patients on the control arm who crossed over, received various therapies. That's always going to impact long-term outcomes. We also want to know how we can use the data to counsel patients. How many had complete responses, how durable were the responses that occurred, and what are the factors that come into play at the time of resistance?

Let's kind of walk through those steps to the extent that we know them. Tell us about the control arm and how they fared and what you know about subsequent therapies, for example, in that setting.

Brian Rini: That's one of the key things here, as you know, the arguments against combos are people would say, "Well, why can't you just get one drug and then get the other drug?" I think clearly there are striking benefits to the combo. The other thing is, unfortunately not all patients are alive and well enough to receive second-line therapy. In this trial, of the patients who discontinued therapy, so about 70% of the sunitinib patients received subsequent therapy, which is actually pretty high, relative to numbers, you'd see, like in a community setting, and the majority of that therapy was PD1-based therapy, presumably nivolumab monotherapy because that's one of its approved indications. Most patients got therapy and most of those patients got PD-1 based therapy.

In the pembro/axi arm, about half the patients received subsequent therapy. I think it's a lower number simply because more patients are still on that original therapy. And as you might imagine, most of those patients about half got a VEGF inhibitor.

Charles Ryan: There weren't patients though, formally or informally crossing over?

Brian Rini: Correct.

Charles Ryan: Sunitinib to axi/pembro?

Brian Rini: No. There was no crossover on trial, this sort of crossover, if you will, occurred just in the normal course of treatment after patients came off study.

Charles Ryan: Very good. What about the other points, PR rates, CR rates, that type of thing? How can clinicians use these data to counsel their patients and think of this as a frontline therapy choice based on the new data?

Brian Rini: The overall survival update, not surprisingly, is still significant. There's no median reached in pembro/axi. It's about 36 months for sunitinib, so about three years, which is pretty good for the sunitinib. It's about the same as the sunitinib treated patients did in the CheckMate-214 study. PFS still had a significant median a little over 15 months. And then the response rate is a little over 60% in the pembro/axi arm and that's, I would say, a distinguishing feature of the VEGF containing regimens. The IO plus VEGF regiments are that their response rates are in that 50-60% range. Sunitinib was 40%, almost 40%, so actually sunitinib did pretty darn well in this trial, as it did in CheckMate-214.

One of the important updates from this trial is the complete response rate, which was originally reported at 5.8%, is now 8.8, almost 9%. Some of these evolving responses, which we somewhat come to expect with immune therapy, happened in this trial and this, I think, puts the numbers certainly in the range of the CheckMate-214, which I believe was 11% in the ITT at the last update.

I do want to talk a little bit about a depth of response analysis we did because I think to me, one of the important points is that it's not just the recess CRs who do well on immune therapy. Clearly more than 9% of patients are doing well long-term with this regimen or with the ipi/nivo regimen. But just in terms of counseling patients, 60% response rate, and 9% CR in this update.

Charles Ryan: So that by the depth of response, you're referring to long-term PRs? Tell us about that analysis.

Brian Rini: We did, it's admittedly post-talk exploratory, all those caveats. We did a six-month landmark analysis where we took patients, maximal tumor shrinkage, so of their measurable disease, what was their percent shrinkage, 22%, 68%, 90%, whatever it was as measured the independent review. Whatever that status was at six months and then look at their survival after the landmark obviously from six months onward.

Really this was to try and look at what I just said, are there patients... My sense was that in my clinic, a lot of my patients doing well long-term on this and similar regimens with worse CRs, but they were 70, 80, 90% shrinkage. They were those DPRs, we might call them.

What was found is that patients in the 80 to 100% shrinkage and did about the same as RECIST CRs in terms of survival. Their survival was similar. It was only true on the pembro/axi arm, not on the sunitinib arm.

There was a similar analysis presented at ESMO from CheckMate-214. They looked at different cutoffs. Nobody knows what the right cutoff is. They showed actually that anything above 50%, 50 to 75 and 75 to 100, did about the same from a survival standpoint.

There are bits of data emerging. There are efforts to look at data across all these trials, but it's a data emerging that it's not just about a RECIST category. I mean, RECIST categories are just buckets and they're broad and they're very imperfect, but a more granular and nuanced sort of depth of response that probably is going to identify patients who do well long-term.

Charles Ryan: It's a good example of RECIST existing for the purpose of clinical trial reporting so that we're all sort of speaking the same language to some extent there for regulatory approvals, but it doesn't always impact how we think through an individual patient in the clinic on an average day.

Brian Rini: What I want to know, and I think what the patient wants to know when we're sitting on day one, is what's my chances of doing really well long-term, "I have small abnormalities on the skin, but I feel great and I'm maybe off treatment and doing well, then I don't really care what my skin looks like." Patients want to be alive and feeling well long-term obviously.

Charles Ryan: And ultimately we don't really know, I don't think, maybe you do, what's happening in those small residual tumors. They're radiographically present but are they biologically in character et cetera.

Brian Rini: FDG PET isn't useful in kidney cancer. We don't have a PSMA PET in prostate cancer. We don't have that in RCC to assess viability. We don't commonly resect these lesions because patients are doing well. So we're not going to go in and start whacking things out if they're doing well clinically.

My sense to your point is that a lot of these DPR patients just had dead tumor left or scar left, and we're just picking it up radiographically.

Charles Ryan: Where do you think the line is? Is it a 90% PR, a 60, or do you want to speculate?

Brian Rini: We didn't do that receiver operating curve analysis that you need to do to say what is the inflection point above which those patients do better? It's something that I think we need to do, but it wasn't done for this analysis. We pick these landmarks, frankly, arbitrarily. Eighty seemed to be the cut point here. I have a feeling it probably could be a little lower, but I think it's somewhere in that 70 to 80 range.

Charles Ryan: You brought up a point that I want to come back to, and I think this is useful for treating clinicians. There are conversations going on in the clinical trial community about looking at treatment discontinuation formally. Now you're bringing up this issue that a really good PR may be functionally close to a CR.

What do you recommend we do for patients who maybe are having one of these DPRs? Let's divide it between obviously the patient who is tolerating treatment poorly versus a patient who is tolerating treatment well. How do you integrate that response and the tolerability into subsequent treatment decision making?

Brian Rini: Good question. I will say in this study we haven't yet looked at patients after discontinuation and how did they do and could they stay off treatment? It's hugely important. I totally agree. There's just not a follow up in this trial to do it yet.

As you may know, across the pembro program, patients got 35 doses or about two years. So in this study with a median of two years of follow-up, patients are really just getting to that point or were just starting some data, and probably next ASCO or something, we'll have an update on how those patients did. There are data with ipi/nivo because that's two years ahead in terms of maturity that a substantial fraction of patients can stay off treatment and do well. There's no data yet from this trial.

What do I do in clinic? The tolerance is important. So it's [inaudible 00:10:31] tolerance, patient preference, et cetera. If a patient's not tolerating treatment well it makes the decision easier. If they have a nice DPR and they're really not tolerating treatment well, then I can reasonably say to the patient, "Listen, let's give you a break. Let's bring you back in two-three months with scans. And let's kind of see what we see and if disease grows, then we'll figure out how to restart treatment. If it doesn't, then maybe we've, at the very least, bought a break from therapy, of short duration or long duration."

If the patient's doing wonderfully on treatment and sort of swimming along, again in this study and in practice, I would stop the pembro at two years. But let's say they're just on axi monotherapy, but doing well, the motivation to stop is low. Obviously there can be long-term toxicities. I think they're relatively rare, but they exist. I think then it's a conversation with the patient.

I'm always comfortable stopping therapy because I've done it a lot and, a very general statement, patients tend to do well, but obviously I'm not the one with the disease. The patient has to have either trust in me and-or comfort level that we [inaudible 00:11:32] comfortable stopping, more than you might think.

Charles Ryan: It's a really important point in terms of setting out your initial treatment goals and managing the patient long-term. Finally, let's go back to the start of therapy, really impressive data emerging. You've mentioned ipi/nivo as another frontline therapy. Where should clinicians sort of consider in the algorithm, the choice of axi/pembro versus ipi/nivo for their frontline metastatic kidney cancer patients?

Brian Rini: That's a complicated question, but a good one. I think those are the two dominant regimens being used. They're the ones that have the most impressive clinical data, including survival benefit. Cabozantinib plus nivo just had a press release that also has similar survival benefits. So that'll be coming.

But it's really kind of that choice of categories, a dual immune regimen versus an immune plus VEGF. I could probably argue both sides of this. I usually argued the IO VEGF side. I think axi/pembro is getting more use, I think simply because it's easier to give. So in a community setting, et cetera, it's just easier to give. There's less upfront toxicity, although probably more chronic toxicity. With ipi/nivo, you deal with this upfront inflammatory toxicity. But if you get through that unscathed or minimally scathed, the nivolumab maintenance is pretty darn easy and they have quality of life data to show it.

It's a little bit of mindset. My colleagues who are used to giving high dose interleukin-2, tend to be ipi/nivo givers, the upfront maximal toxicity, but the potential for long-term benefit. But I think the use is shifting towards axi/pembro and maybe similar regimens. But I think at the end of the day where we're headed is triplets because clearly all these pathways are important, VEGF, CTLA-4, and PD-1 are all important pathways. I don't think anybody would debate that.

And so you can imagine naturally that we're going to move towards triplets. There is an ipi/nivo plus cabo study called COSMIC-021 that's ongoing. That'll be the first of those triplets to report out, but I think there are others coming.

I evaded your question a little bit. I tend to give axi/pembro. It's easy to give. Obviously I've been involved with both regimens, but more so this one. I think again, there are clear progression-free survival and response rate advantages. Those aren't necessarily important for every patient, but they are important for some, so complicated question and I think this will continue to evolve.

Charles Ryan: I think it's not an evasion of the question. It's a recognition of some of the ambiguity between the clinician's choice. And it really, I think hearing you say it validates the choice of either regimen and as you point out, clinicians in different settings and different environments and patients with different goals and different backgrounds require different considerations. I think it's good.

It's really encouraging to hear about the triplet studies. And I'll look forward to talking to you about those maybe in future years and I want to thank you for this update and all your hard work on behalf of kidney cancer patients and in the development of these new drugs. It continues to be an exciting time in kidney cancer. I remember saying it was an exciting time in kidney cancer, but we started saying that 18 years ago and it's continuing. It has been a really nice progression.

Brian Rini: I think, put it in perspective maybe as a final note, when I was a fellow at the University of Chicago, median survival was 12 months. The initial sunitinib data was about two years. Now the sunitinib arm on trials is living three years. The mediums have not been reached in any of these double trials, but are in the four- to five-year range. So that's a fivefold increase in survival, again over 20 years, so over a long time, but quite a difference. Yeah.

Charles Ryan: All right, Brian, Dr. Rini, thank you so much for joining us and we will look forward to future conversations.

Brian Rini: Thanks, Chuck. It was fun.