Durvalumab Alone or with Savolitinib or Tremelimumab in Previously Treated Advanced Clear Cell Renal Cancer - CALYPSO Trial Results - Tom Powles

January 4, 2023

Thomas Powles discusses the results of the CALYPSO trial with Pedro Barata. Dr. Powles provides an overview of the trial design of this international, multicentre, open-label phase Ib/IIa study of durvalumab alone, savolitinib alone, and durvalumab in combination with savolitinib or tremelimumab in patients with metastatic renal cell cancer. He then discusses the trial results, plus his thoughts on the potential for future research into next-generation sequencing and biomarkers.


Thomas Powles, MBBS, MRCP, MD, Barts Cancer Institute, Cancer Research UK Experimental Cancer Medicine Centre, Queen Mary University of London, London, United Kingdom

Pedro C. Barata, MD, MSc, Assistant Professor of Medicine, Hematology & Medical Oncology, Tulane University, New Orleans, Louisiana

Read the Full Video Transcript

Pedro Barata: Hi, I'm Pedro Barata. I'm a GU Medical Oncologist and Assistant Professor of Medicine at Tulane Medical School in New Orleans. It's such a pleasure to be joined by Professor Dr. Powles is Director of Barts Cancer Center. He's obviously a internationally renowned expert in GU malignancies. Welcome, Dr. Powles.

Tom Powles: Pedro, thank you for inviting me.

Pedro Barata: Absolutely. It's true pleasure. And we just saw each other at ASCO in Chicago, it just happened and congratulations on the fantastic job you did delivering on the findings of your trial, a CALYPSO trial, which of course with the effort that you led. So, congratulations on that. Wondering if we could talk briefly about that study?

Tom Powles: I'm not sure there's too much to congratulate me about. I think the congratulations to the effort put in by the investigators to get the trial done. I think at the time, targeting met in the papillary cancer made a lot of sense. In fact, as it transpires the combination of savolitinib, the MET inhibitor plus durvalumab, PDL-1 inhibitor looked really good in papillary cancer in MET driven tumors. And so we sort wanted to ask the same question in clear cell renal cancer because clear cell renal cancer is commoner. We wanted to do it in a randomized trial using durvalumab, essentially as a control arm, because durvalumab is a PDL-1 inhibitor, and we know immune checkpoint inhibitors have some activity in kidney cancer. We wanted to look at the component parts. We also put a savolitinib alone arm in. And then as a bit of a wildcard, we put in Durvalumab and tremelimumab together, PDL-1 CTLA-4 to see what that combination might do because many people feel that may be active.

So, we ended up with a four arm randomized trial in VEGF resistant clear cell renal cancer. And though that was the trial design, we probably overcomplicated it. And you might say what about doing a two arm trial? And with hindsight, that may be better because in the end we only had 40 patients in each arm. We had ambitious endpoints looking for 50% response rates, which I think was very ambitious. But remember, we were also going to look at subsets of patients including MET driven tumors and actually personalized therapy. I think we are going to need to get up to those sort of 50% response rates if we're going to use them as a standard of care. What we showed was that none of the four arms had response rates above 30%. The highest was durva-treme with 28% and the lowest was savolitinib alone with 10%. So, I think we didn't hit our primary endpoint in any of the four trials start our trial arms.

We then looked at PFS and OS, and we didn't show that either the combination of durva-treme or savo-durva outperformed durvalumab alone for PFS or indeed for OS. And so with poor responses or lowish responses and no DFS or OS advantage over single agent PDL-1 therapy, I think it's reasonable to say that there was no rationale for taking these arms forward. Now there may be some activity. And so we look under the bonnets at those MET driven tumors for savolitinib arms and PDL-1 biomarkers for the immune arms. And what we showed actually is even the MET driven tumors, we didn't see a higher response rates or better outcomes in savolitinib treated patients. It's likely therefore that there are a number of other pathways beyond MET in clear cell renal cancer, which are more important. Doesn't mean that MET doesn't have a role to play, but MET alone savolitinib quite a specific MET inhibitor is probably not driving terrific outcomes.

Pedro Barata: So, no. So, I agree completely. And thank you so much. That's a very beautiful, elegant summary of the findings. We usually say that we learn a lot from positive studies, but I think we also learn a lot from negative studies. I guess, trying to be provocative when you go at these four, as you said, four arm study, I find it interestingly in the setting that we'll assume refer you to be [inaudible 00:04:34] you said, so minimum let's just say second line, assuming most patients got to via [inaudible 00:04:38] in the front line setting. It's interesting that when you look at the durva monotherapy PDL-1 as you said, response is 10%. When you look at the combination of durva with treme, right, you look at 28%, and then when you look at durva-savolitinib you look at response rate again, 13%.

So, first, I guess my one question is we know as you pointed out in your presentation from check May 25, that PD-1 nivolumab responses will be closer to 30% or 25 or plus percent. So, it seems like this PDL-1 either mono or in combination didn't perform as well. And so I guess one provocative question is, to some analysis out there is whether you see a difference biologically with offering a PD-1 versus a PDL-1 in the refractory setting? Is that something there or not? What are your thoughts on that?

Tom Powles: Look, I mean, I think it's a very provocative question. I think it's fair to say in lung cancer, most doctors feel that the PD-1 therapies are more potent than the PDL-1 therapies. That's what my lung colleagues tell me. And pembrolizumab has a big role to play there as does nivolumab but atezolizumab is also licensed. It's striking in kidney cancer, that the Emotion 150 trial and the Javelin 101 trial, both PDL-1 therapy trials, the different types of [inaudible 00:06:00] therapies didn't hit overall survival. Whereas on the other hand, we know clear, we know 46 and obviously nine [inaudible 00:06:09] with the PD-1 therapies all hit overall survival. So, I think that if you speak to my... By the way, your bladder cancer, I can't see any difference between the two. I sometimes think we are looking at the stars and seeing patterns that don't really exist. I sometimes feel that way.

And I think that's the case in bladder cancer, but I think in kidney cancer, it's reasonable to say at the moment, the body of evidence for PD-1 therapy does support PD-1 over PDL-1. Having said that, I don't think the CALYPSO trial is powerful enough to support that a great deal, because it's only 39 patients in the control arm. I know that nivolumab alone got 25% response rate and durva alone is lower than that. But with only 39 patients, you only need three or four responses to push that up. And the confidence intervals are quite wide. I think that there's going to be a study called Emotion 010 which is the [inaudible 00:07:04] trial. We recently had the pembrolizumab study. They're quite similar populations. And it'd be interesting to see if the 010 trial is negative. And if that's the case that I think you'd say, well, it looks like it does seem to be this trend. So, I think there's an increasing body of evidence in this direction, but I do also think that sometimes we're looking at the stars and seeing patterns that don't exist.

Pedro Barata: Right. And there's a follow up question on that. Just going at the combo now for a second. Of course, you are very well aware of all the data we got with the [inaudible 00:07:33] in the refractory space, right? We have a number of studies and we're all very excited. Then we saw that it became for selected patients and going back to CALYPSO, right? We look at your combo of durva-treme there. And actually didn't seem to force with small numbers again, as you said, 39 or 40 patients in that cohort, but it doesn't suggest to be a synergy compared with, let's just say we cannot compare studies. We all do. We need a monotherapy. So, do you think that in the context of, IOIO with the CTLA-4 and the PD-1 or PDL-1, do you think the story is over or complete right now to say in general, we do not consider a CTLA-4 based approach in the refractory RCC space?

Tom Powles: Pedro, I think we need to think very carefully about what evidence we have CTLA-4 adds a great deal. So, single agent pembrolizumab in the frontline setting has response rates of 39%. ipi-nivo has response rates at 40%. So, don't... There may be, I don't think anyone feels that, I mean, maybe pembro is better than nivo, and maybe we shouldn't be doing these incomplete comparisons, but both those are quite 427, and 204 quite big trials. Of course, we never did the monotherapy arm in 214. I know the EMA asked for it. And then the EMA on to ask for a 8Y8 which is nivo versus ipi-nivo. I think that this durva versus durva-treme arm in CALYPSO made the first indicator that from a PFS and indeed from a response perspective, that there isn't a big difference.

And if you said to me, what's the most potentially provocative thing that comes out this trial, it might be that actually from a response to the PFS perspective, durva performs the same as durva-treme and maybe nivo will perform same as ipi-nivo. We don't know that yet. I mean, when one looks at the... And this isn't being talked about as much as perhaps it should, but when you look at the censoring of patients with these long term progression free patients five years in the ipi-nivo trial, the number is 30%, but the number of risk is only 10% at the whole population. And so there's heavy censoring that's going on before the patients are dropping out for sorts of reasons. And we just need to be maybe a little bit careful about saying that we know ipi he adds a great deal and you didn't show that. And therefore that mean there's a problem with durva-treme.

There is a second explanation, right? And that second explanation is that actually ipi from a response from a PFS perspective may not be adding that much. It may be adding something in long term survival, but we don't know that yet. And I think that some of the studies where we add ipi in after progression in some of the trials, Mike Akins on the grim did one he's showing some responses. So, I think this question hasn't been answered, and I'm slightly concerned that there may not be as much of a bounce, a response, and a PFS for ipi when you add it to [inaudible 00:10:55] perhaps people expect.

Pedro Barata: Very important points. Thank you for that. And I guess just switching gears slightly and going back to the maths, right. Which in essence was strong of the rational for CALYPSO, if you will. Of course, we talked about NGS and genomic testing, right? In our clinical practice. And we do it all the time and some would argue that in kidney cancer, it's not ready for prime time. In other words, a way to say that a lot of folks, and colleagues that I talked to tell me, well, for the most part, I'm not using that information to make decisions. And so, that's on one side on the other side, as you elegantly showed as well, the role of MET in papillary, but even in that scenario, some folks come and say but well with cabozantinib seems to perform really well including the MET population.

So, when we put all that together, I guess the larger question and one could ask is, how do you see the value of genomic data in clinical practice today? In other words, are you using NGS testing results to derive your decisions in clinical practice? And on the other side, I'm going to ask you, is anything you're going to wait and you'd like to see more, whether it's gene expression signatures, or any other biomarkers you would like to highlight that you think that might be kind of the next steps for RCC biomarkers?

Tom Powles: Well, I think that it's fair to say that perhaps one of the reasons why next generation sequencing is not, or any other biomarkers aren't ready for prime time in kidney cancer is we've never given it a prime time slot. We've always put it at the back of the agenda, and we've always gone for a one size fits all approach. Now that's been quite successful actually in kidney cancer, survival has at least doubled, and we've got two terrific classes of drugs, but I don't see that third class of drug around the corner or is it currently stands. And I think we have reached a point now where we do need to start trying to select patients for therapy, particularly in the adjuvant that's setting where we know we're over treating large groups of patients.

Next generation sequencing identifies a series of biomarkers that do seem to be prognostic, at least back on PBRM1 seemed relevant in that respect and we showed a bit of that in CALYPSO, not to a robust degree, but we showed some early signals there. Of course, there's the second gene RNA, seven gene RNA signature that Brian and Bob developed out the Emotion program, which looks at immune and antigenic signatures, as well as some cell cycle signatures, which I think are really important too. We need to explore those, but dare I say it, even the PDL-1 biomarker in isolation it looked actually pretty good in the first trial that we did with ipi-nivo from a PFS perspective. But we didn't pursue it any further. I know it's not worked well with VEGF TKI therapy with immune therapy, but you wouldn't expect it to work well in that respect.

So, I think when I summarize the picture of the biomarkers, I think that the next piece we are doing will be involving these biomarkers quite carefully. I also saw some really nice data on CTD in ASCO, and I thought that was really cool. We made a short video on that, and lots of people looked at it. And about 90% of made metastatic patients, you can find circulating mutations, which is higher than we expected. And people always say, oh, there are no circulating biomarkers and kidney cancer into quite doesn't appear to be the case. So, I think we need to dive into that chapter. And I'm going to talk a lot about that over the next two or three years, to try and encourage people to pursue that because doing triplets and quadruplets and using the same drugs and sequencing inside different ways A plus B plus C is going to equal ABC in the end.

Pedro Barata: Yeah. And by the way, congrats on your work. You've been doing a lot of their work with liquid biopsy and of course, CT DNA. And of course you did very elegant work in the adjuvant [inaudible 00:15:02] and it's nice to see it, that being transported into the RCC arena. We also conducted kind of a small study using CT DNA with the same [inaudible 00:15:12] and in RCC. And we found the same thing that it might be a role for the use of that kind of technology in RCC as well. So, I couldn't agree more, but thank you so much Dr. Powles. I think you've made such great points. This was a fantastic discussion. Again, congratulations, not easy to conduct such a a large study sense of multiple cohorts. You got collaborations from colleagues in [inaudible 00:15:35] and thank you for taking the time to check with us today. So, thank you.

Tom Powles: Yeah. Well, thank you. It was a group from Spain and France, the UK working together [inaudible 00:15:45]. Well, we want to do more of that in Europe. So, I would really like to take my hat off to those investigators who really drove this, it was a terrific team.

Pedro Barata: Yeah. Great point. You had five star investigators working with you and of course the result is very elegant data that you present.

Tom Powles: Thank you.

Pedro Barata: Thank you so much.

Tom Powles: Pedro, see you soon.