Exelixis, Inc. (EXEL) Earnings Call Transcript & Summary

Thank you for joining us, ladies and gentlemen. Please welcome Susan Hubbard.

Hello, everyone, and welcome to the Exelixis Investor Briefing at the ASCO GU 2021 Virtual Conference. I'm Susan Hubbard, Executive Vice President of Public Affairs and Investor Relations at Exelixis. We are pleased to have you join us following the close of a very busy ASCO GU 2021 and week in general for Exelixis. We have a very prestigious panel of physicians joining us today for a review of the array of cabozantinib data presented at the conference and to discuss the evolving renal cell carcinoma landscape. Following our panel discussion, we'll have a question-and-answer session where we'll field questions from the audience. Before I turn the session over to Mark -- to Mike, sorry, Mike, I will remind you that during the course of this presentation, we will be making forward-looking statements regarding future events or the future performance of the company, including statements related to the clinical, therapeutic and commercial potential of cabozantinib, regulatory development and submission strategies, the development path for cabozantinib and commercial planning for CABOMETYX. Actual events or results, of course, could differ materially. We refer you to Slide #2 of this presentation for applicable cautionary language. And now with that, I'd like to turn it over to Mike Morrissey, our President and CEO. On to you, Mike.

All right. Thank you, Susan. Welcome, everybody. Thanks for joining us today. Obviously, a very busy day, busy weekend at ASCO GU. We are thrilled to have 21 abstracts at the meeting. Super excited to have 4 top KOLs joining us today. So I'm really excited to be able to sit back and listen and learn. And I hope you'll enjoy it and get your questions asked as well. Top 5 cabo presentations are shown on this slide. You'll hear more about those today. I thought the timing of the ASCO GU meeting along with the -- both the approval and the launch of the cabo/nivo doublet in first-line RCC was fantastic for us. We had a lot of good discussions with prescribers over the last few days and really trying to maximize the overlap in the meeting with the approval. So we're super excited about that and looking forward to a great session. So with that, what I will do next is introduce our moderator today, and that will be Dr. Will Berg, who is the Senior Vice President of Medical Affairs. He's a card-carrying GU oncologist as well, so perfectly set up for a great discussion with our panelists today. So Will, over to you.

Okay. Thank you, Mike. And I'm pleased to introduce you to the panel for tonight. So joining us for the discussion today is Dr. Toni Choueiri, the Director of the Lank Center for GU Oncology at the Dana-Farber. I'd also like to introduce Dr. Dan George, Professor of Medicine and Surgery at the Duke Cancer Institute. We also have Dr. Rana McKay, Associate Professor of Medicine at the University of California at San Diego. And finally, we're delighted to have Dr. Sumanta Pal, Co-Director of City of Hope's Kidney Cancer Program. So as Mike mentioned, there were 21 presentations featuring cabo at the conference. We'd like to start with a quick overview of some of the key presentations for cabozantinib at this year's meeting, and I'd like to ask our panelists to briefly summarize some of the key results and takeaways on -- from the abstracts that are on the slide next to me. We'll start with Dr. Choueiri. So Dr. Choueiri, please give us a quick overview of 3 of the abstracts that you were a co-author on this year. So those are the CheckMate 9ER update, the CheckMate 9ER quality of life and the use of cabo in patients with brain metastases.

Dr. Berg, colleagues, thank you very much for this opportunity. It was really an overwhelming GU '21. There was a lot of new data presented. In relation to what I am going to mention, first is the abstract 308 presented by my colleague, Dr. Bob Motzer, the first author. Here, there was an update about CheckMate 9ER with more follow-ups. And with more follow-up, the combination of nivolumab and cabozantinib continued to maintain the trifecta of progression-free survival, overall survival and response rate benefit relative to sunitinib. The PFS hazard ratio was 0.52. The OS has a ratio of 0.66, and the response rate were -- remain doubled, 55% versus 28%. In addition, there was a look at patient with sarcomatoid histology, which are anywhere between 5% to 15% of all patients with metastatic RCC. Here, the hazard ratio is even lower and the benefit higher in terms of progression-free survival and overall survival, as you can see. Second abstract is also from CheckMate 9ER I was an author on. It was presented by Dr. David Cella out of Northwestern. This abstract addresses quality of life and patient-reported outcome. This is not just one metric was quality of life; many, many metrics. This is one of the most extensive description of patient-reported outcome metrics across all combination of VEGF-IO. And it clearly shows along multiple health-related quality of life the benefit from the combination of cabozantinib and nivolumab are consistent over sunitinib and clinically meaningful. The third abstract comes from one of my mentees, Dr. Hirsch from France. Well, we looked in a retrospective fashion a patient with uncontrolled as well as controlled brain metastases. There was 2 cohorts here. And we looked back at the scans on an investigator assessment. And cabozantinib, actually, in these patients with brain metastases, an unmet need, so the 61% response rate, an intracranial response rate, justifying a very, very important Phase II study, prospective study, going with our French colleague by the name CABRAMET in France. Thank you, Dr. Berg.

Thank you, Dr. Choueiri. So Dr. Pal, perhaps you can now review the data from the PAPMET trial, which you presented earlier today and which you simultaneously published in the Lancet as well.

Yes. Thanks a lot, Dr. Berg. This study is very near and dear to my heart. Basically, I wrote the study over a decade ago, in fact, and it was initially designed ironically as a trial comparing sunitinib to cabozantinib. The study subsequently morphed to a 4-arm design comparing sunitinib, cabozantinib, savolitinib and crizotinib, the latter 2 being perhaps more specific MET inhibitors lacking VEGF activity. And ultimately, those latter 2 drugs failed at an early futility analysis. Cabozantinib was shown to -- a significantly prolonged progression-free survival relative to sunitinib. You see the numbers there, 9 months versus 5.6 months. There was also an impressive discrepancy in the response rate, 23% response rate associated with cabozantinib versus just 4% with sunitinib. And fortunately, or unfortunately, the poor efficacy with sunitinib mirrors what we've seen in previous studies. I think the study automatically catapults cabozantinib into a standard for patients with metastatic papillary kidney cancer. And truthfully, this is really the first time that we have randomized data to support the systemic therapy specifically in this disease. So it's really quite exciting. Thank you, Dr. Berg.

Yes. That was a fantastic report, Dr. Pal. So Dr. McKay, can you walk us through now the updated results from Dr. Apolo's presentation on cabo in combination with nivo and/or ipi?

Absolutely. We heard the results from Dr. Andrea Apolo, the Phase I study looking at the combination of cabozantinib plus nivolumab and also the combination of cabo, nivo plus ipi. And really, this was a landmark Phase I study that has set the stage for a lot of the Phase III studies that are looking at this combination in RCC and other tumors. And so it was exciting to see the final analysis get presented. The study was a 2-part study. In part 1, they looked at the combination of nivo and cabo at various doses. And really, this study established the dose of cabozantinib with IO therapy. For part 2, they looked at the combination of the triplicate of cabo, nivo plus ipi. There was a total of 120 patients that were enrolled on the study. Over 50% of patients had rare histologies, and they demonstrated an overall response rate of 38% with a complete response rate of around 11% from this study. And again, this really set the stage for a lot of the Phase IIIs that are currently being designed of combination therapy with cabozantinib and IO.

Thank you, Dr. McKay. And now Dr. George, before we go into further discussion, I'd like to have you just give us your perspective on the Phase III CLEAR data, which we saw earlier today that Dr. Motzer presented. So if you could just give us a high level on that, that would be great.

Sure, Will. This has been a long-anticipated study, really the latest in a series of Phase III studies that have looked at the combination of TKIs and IO versus a historical control of sunitinib. And this was actually a 3-arm study, 1,000-patient trial looking at pembrolizumab and lenvatinib versus sunitinib versus a third arm of lenvatinib and everolimus, an FDA approval in refractory disease setting. Really, the top-level data really focused on the lenvatinib/pembrolizumab combination versus sunitinib. And there, we actually saw a 24-month investigator-assessed progression-free survival rate, median progression-free survival rate for the pembro/lenvatinib combination versus just 9-month median progression-free survival rate for sunitinib with a hazard ratio of 0.39, a very impressive end point in and of itself. To go along with that, an overall survival benefit as well was seen with a hazard ratio of 0.66. And this study had a median follow-up of 24 months. There was also an impressive overall response rate, objective response rate for the pembro/lenvatinib combination of 71%. It did come with some toxicity, and we did see a high degree of grade 3 or greater toxicities in 72% of patients in the pembro/lenvatinib arm. So there's a balance there. But overall, very impressive results.

So thank you all for reviewing the presentations. Now we have a series of questions for the panel that will delve into the evolving landscape in RCC. We'll start with a discussion of first-line RCC. And then towards the end, we'll touch upon ongoing trials that are relevant to the life cycle of cabozantinib and XL092 and which we hope will further define the future treatment landscape for RCC. To kick things off, let's discuss how you all look at first-line IO/TKI trials. And to start us, we'll have Dr. Choueiri again. So Dr. Choueiri, what is the potential impact of differences in patient populations when considering key results of Phase III trials in first-line RCC?

Thank you, Dr. Berg, again. I think this is a very, very important point. We have combinations that have proven superiority over sunitinib. And not just any superiority, we have combination that have an overall survival benefit against sunitinib. And these combinations have not been compared head to head. So why are the results different in term of progression-free survival, overall survival response rate or in terms of the performance of the control arm, sunitinib? And we can have a glance into that by looking at the baseline characteristic. The trials are different. The baseline characteristics are not the same. So let's take a look here contrasting CheckMate 9ER, the cabozantinib and nivolumab study. Let's take the arm from Checkmate 9ER that deals with cabo/nivo; KEYNOTE-426, the axi/pembro study; and the CLEAR, which was just presented, the len and pembro study. And we can see clearly here between 300 and 400 patients there enrolled. So let's focus first on the most, I would say, recognized and famous prognostication here in this disease, the IMDC risk factor. As you can see, the KEYNOTE-426 as well as the CLEAR study have more patients with favorable risk. And overall, they have less patient with poor risk by IMDC. On the other hand, if we look at the sarcomatoid feature here, the KEYNOTE-426 have more patient. And we know patients with sarcomatoid feature tend to respond better despite their aggressive biology to immune checkpoint inhibitor. Radiation therapy could be another adverse aspect when patient receive radiation therapy often to control brain or bone metastases. And on CheckMate 9ER, the proportion, at least numerically, is the highest. Prior nephrectomy. And while Will, you and I, remember, during our training, where almost everyone can get cytoreductive nephrectomy, these studies happen around the same time that CARMINA was reading or read. So we saw a drop in those patient that did not have cytoreductive nephrectomy. And to that end, CheckMate 9ER has the lowest rate of cytoreductive nephrectomy or prior nephrectomy, actually, you could have prior nephrectomy -- at 69%. And that could influence the responses, as we know responses in the kidney tend to be hard. If you look at patients with 2 or more organ metastases site, that also has been described independently as an adverse prognostic factor. And numerically, these seem to be higher with 9ER. Bone metastases, Dr. McKay has a large body of evidence and publication about bone metastases. Interestingly, it's the same across all trials. However, liver metastases, Dr. McKay in the past had papers describing the impact, the negative impact of liver metastases alone or in combination with bone metastases. These also seem to be numerically higher with liver metastases. Not to mention, of course, and I'm sure my colleague going to talk to, sometimes the reporting of the trial and the median follow-up is not the same. Thank you, Dr. Berg.

Thank you. Dr. McKay, since Dr. Choueiri brought your name up, you want to comment on baseline characteristics and maybe add to what Dr. Choueiri said?

Thank you so much. And I think that as we -- Toni, so I think we went through each of these line items. But in general, I think when we step back and look at a bird's eye view, the CheckMate 9ER trial just enrolled a sicker patient population. There was more patients that had porous disease. There was more patients that has received prior radiotherapy. So probably, those patients were symptomatic if they had -- before they even started systemic therapy, 15% of them were getting radiation to something to control a disease -- an area that couldn't actually be put into check. They had their primaries intact. There was more sites of metastases. So I think when we just step back and take a bird's eye view, compared to CLEAR and KEYNOTE-426, this patient population was a little bit sicker. And then I think as we interpret the efficacy data, understanding that is going to be important when we look at numbers and trying to look at why is SUTENT performing in a certain way or why is the PFS at a certain level for each of these trials. So the fact that this was a sicker patient population, I think, is probably going to speak to some of that. But I think it also speaks to, gosh, the efficacy as well despite the patient population being a sicker patient population.

That's great. Thank you, and we will get into those discussions a little bit later. So Dr. Pal, anything to add to the discussion of baseline characteristics?

I will say that I think that Dr. Choueiri and Dr. McKay did a phenomenal job giving an overview of baseline characteristics. Just to sort of reiterate some of the key points here, I think very important to, again, acknowledge that in CheckMate 9ER, lower proportion of favorable risk, higher proportion of core risk versus the CLEAR study. I think that's the other study perhaps that we want to sort of try to compare our data against here. I would also propose that we do a little exercise with the prior nephrectomy groups, and Dr. Choueiri had already sort of walked you through this. But keep in mind, 69% prior nephrectomy implies that 31% of patients in CheckMate 9ER still had their kidney intact. That's a population of patients where it's going to be harder to demonstrate a complete response. It's probably a more poor prognostic category as the IMDC criteria imply. And that's versus 26% in the CLEAR trial if we look at the inverse of 74% there. Also, the next row down, greater than or equal to 2 organs with metastasis, take the inverse of that. This implies that just 20% of patients in the 9ER study had one organ with metastasis versus 28% in the context of the CLEAR study. Again, we know when you're thinking about response rates, when you're thinking about PFS, these all sort of play into those absolute numbers. So again, I would just reiterate a focus on hazard ratios and the relative benefit that we're seeing across these studies. Dr. Berg?

Yes. Before we move on, Dr. Pal, one more quick question. In terms of looking at -- when you're talking about the distribution of numbers of patients that are favorable, intermediate and poor, which of those 3 trials do you think is closest to sort of typical clinical practice out in the community?

Great question. I think that the breakdown that I always sort of suggest to my fellows is that 20-60-20 is what we're seeing in clinical practice. And I think that many of the real-world data sets really reflect that. So from that standpoint, the 9ER data set is really spot on.

Great. Thanks a lot, Dr. Pal. So with that context, Dr. George, maybe you can help us understand the key results of the various ICI-TKI combinations. Please comment first on efficacy and safety, and then we'll toss that back out for the panel as well.

Absolutely, Will. Thank you. And let me just preface this by our statisticians will always tell us that you really have to be careful about cross-trial comparisons. And there's a reason for that. And the reason for that is because there's going to be variability in each of these studies, in both the patients that were accrued, the sites that were involved, the timing of when those studies were done. And I think also in the methodology, even things like which independent review panel reviewed which study, and they have different nuances and measurements to each. So with that caveat in mind, let's look -- let's do just what I said we shouldn't do. And let's look across these studies. And the reason I say we need to do that is because as clinicians in the field, this is exactly what we're left with. We don't have head-to-head studies of cabo/nivo versus lenvatinib/pembrolizumab. We may never have studies comparing those 2. So this is probably the closest we can do in terms of understanding how do we choose one frontline regimen over another. And as you can see, they have different median follow-ups here. We see different responses and on down the line. There's a few key factors, as a clinician, when I look at studies, that I would say are really the notable end points. And the first is probably the primary end point. If you look at the primary end points for each of these studies, investigator-assessed median progression-free survival shown here. And really, there's the median, but then there's the hazard ratio. And I really look at the hazard ratios as really driving for us the results here. And what you can see is hazard ratios here that are really strong, particularly for the CheckMate 9ER and for the lenvatinib/pembrolizumab arms. And if you look at these, they're fairly comparable. When you look at the blinded independent review, which were about the primary end points, you can see maybe 0.4 versus 0.5, a little bit more in favor of lenvatinib. But be careful of that because, again, how each blinded independent review committee reviews that may be different. So for instance, in Checkmate 9ER, you see the investigator-assessed PFS is 19 months, it actually goes down on blinded independent review. It's kind of the opposite trend in the CLEAR study with len/pembro. It's 22.1 months in the investigator-assessed. It actually goes up to almost 24 months in the blinded independent review. So you do see some differences in these study results, and it probably has to do with who's on those independent review committees and how they judge these results. The investigator-assessed in some ways is a little bit more organic. That's a little bit more of how people practice and their view of things on these patients. And it may be a little bit more representative of the real world. As Dr. Pal said, this is, at the end of the day, a disease that has a lot of heterogeneity in the real world, the 20-60-20 populations of favorable, intermediate and poor risk. And we're left to assess these responses in that context. And so there is some subjectivity that goes into this. And as well it should, is a patient tolerating a therapy well -- are they responding? Are they getting a clinical benefit from the therapy? Those do factor into our assessments of response and of progression as well. So recognize that each one of these end points is not perfect. Really, what we like to look at is the totality of the data. So I would say PFS is one of the end points that's part of that totality. Overall survival is another. And if you really come down to it, what patients are really most scared about is dying of this disease, not necessarily the progression, the symptoms, but of death and particularly early death. And one of the really exciting and encouraging things about all of these studies is we do see a separation in the rates of early death associated with all 3 of these trials, which is really, I think, validating across the field for this combination of IO/TKI. But what we do see is that, overall, these median overall survival differences in hazard ratios -- we don't have the medians for any of these trials, but we see these hazard ratios somewhere in the 0.53 to 0.66 range. And I would say those are relatively tight numbers. So for me, when I'm looking at differences in hazard ratios, something around 0.1 difference is probably not going to sway things. So I look at this and say, these are pretty comparable like the PFS data. There's overall response. And I think overall response of the 3 end points is probably least important. I know it's important for 2 reasons, one, because patients like to see their tumors regress. And two, individual physicians like to see on an individual patient basis that they're benefiting. And it's hard to see that with a PFS end point or an OS end point on an individual basis. So it is meaningful from that, a practice perspective. But again, if we look across the board here, you have to look at it as an intra-study control. So although you might say 71% response rate really looks impressive from lenvatinib/pembrolizumab, our sunitinib arm in that study was 36% response rate. It's about half. I would say, similarly, for the CheckMate 9ER, we see about a 56% response rate with cabo/nivo. And we have about half of that with a sunitinib response rate of 27%. And with KEYNOTE-426 in pembro/axi, it's similar, although I think it's maybe a little bit less than half in that group. And then the last piece of this is what Dr. Berg mentioned, safety. And balancing out all of this efficacy has to be how patients are tolerating therapy. And one of our biggest concerns with safety is grade 3 or greater adverse events. And the reason for that is because those represent serious adverse events, many of which require hospitalization, discontinuation of therapy, complication, symptoms and risk of other problems. So when we look at these rates, we see it pretty high across the board for all 3 of these, 75%, 82%. These are common events. What we really look for in putting those into context is how much of that is really disrupting permanently the treatment. And that's where this last line comes in, this discontinuation due to AE. And you'll see here, it's got a lot of colors, which is nice. But you'll see the either, so stopping either drug, is in purple. I think that's purple. Then you see, call it, a bluish green, that's the IO therapy. That's your nivolumab or your pembrolizumab. The TKI is in the darker green blue, and that's your nivolumab or your axi or your lenvatinib. And then stopping both drugs is in gold. And what you see across the board here is that CheckMate 9ER is a pretty well-tolerated regimen in that patients are able to stay on drug and not have the discontinuation most of the time. Discontinuing both drugs was only 6%. Either drug was only 20%, and individually, each drug was 7% or 8%. If you look on the other hand, for this latest data with lenvatinib/pembrolizumab, you can see almost double that rate for discontinuation either drug, 37%. You can see for the immunotherapy, the discontinuation was 29%. Somehow that was much worse than what we saw in the other studies. Likewise, lenvatinib, which is really kind of the biggest variable in this regimen, 26% had to discontinue this drug. And overall, for discontinuing both drugs was 13%, again, about double of what we see with CheckMate 9ER. So when I look at that data and I think about my clinical trial population versus my real-world population, it does cause me to pause a bit to say how well will this regimen translate to a broader population of patients who might not have otherwise qualified.

That's great, Dr. George. Let me bring in Dr. Choueiri. If you want to just comment now, Dr. Choueiri. This is updated -- what you're seeing on this slide, this is updated follow-up from CheckMate 9ER, KEYNOTE and CLEAR trying to normalize the median follow-up. So it's a little more similar in that regard. So let me have you jump into this slide, Dr. Choueiri.

Yes. Thank you, Dr. Berg. And I think one of the interesting thing where we try to over-accommodate is by bringing the trial together in terms of median follow-up. I don't know how much that works. I know Dr. George and I always get yelled at by our statistician for a good thing, and we like it. But we cannot imagine today how drug development happening in the U.S. and outside the U.S. that to be sure that there'll be another trial, let's say, a Phase III trial comparing your favorite 2 combination. It may be a very important question, but it doesn't mean it will happen, how drug development usually goes in general. So trying to bring the 3 trial that showed an overall survival benefit here with VEGF-IO, not counting CheckMate 214, to a range between 2 years and 30 months, 23.5 and 30 months, we can see the CR rate overall. We can see the PD rate. The PD rate, really with all the combinations, remains quite low. And the lowest is with nivo/cabo as well as pembro/len, less than 10%. Blinded independent center review for progression-free survival, while it's longer for pembro/len like we've seen, again, how much these populations are different. One way to look at it actually is to look at the control arm, sunitinib. And the performance of sunitinib ranges between 8.3 and 11.1 months. 3 months difference, with the lowest being for CheckMate 9ER likely, in my opinion, because the baseline characteristic suggests a bit of a less favorable population. Now let's focus on, overall, the safety. One thing I really like, let's go back, is the treatment discontinuation because of adverse event of both the drugs so that you have nothing left on board because sometimes if you stop one drug, the other drug may take over, you have something on board. And if you recall this, in orange, overall, it's 6.6% with CheckMate 9ER, 7% with KEYNOTE-426, and it's a bit higher overall with CLEAR. Next. Now a couple of comments perhaps before giving it to another colleagues here. This is the updated actually adverse event in first line, Will, not the previous slide. And if we look at the treatment discontinuation of both the TKI and the PD-1, contrasting the 3 VEGF/IO combination with survival benefit, it's anywhere between 6% and 13%, the lowest being for CheckMate 9ER. There was no -- even if you look at the 3 combination, nothing in terms of -- new side effects that got amplified by -- folks have to be very familiar with the side effects of a TKI and with immune checkpoint inhibitor now. Now what about hepatotoxicity, and we know how this is -- can be important overall. If you look and compare the hepatotoxicity, whether you look at all cause grade 3, 4 at AST specifically and ALT, you can see clearly that there are differences here. And the rate of the utilization of steroid, we still don't have that information for CLEAR, but it's less than 20% with 9ER. It is overall the lowest at least reported. So that's in a nutshell what I see when I look at the AEs. The good thing also is that grade 5, so that is treatment-related, in general, seem to be also very, very low despite these are handful of cases. We know these are very important and part of every safety data, and these are the lowest in CheckMate 9ER. Nivolumab and cabozantinib have been a drug for a long time in renal cell carcinoma, both as single agent or in combination, and I'm not surprised here with the safety profile overall. Dr. Berg, back to you.

Thank you, Dr. Choueiri. So Dr. McKay, how do you consider all of this data when you're in the clinic making decisions?

That's a very good question. And maybe we can go back a slide to the updated data and just take a look at that as we interpret things. And I think that some of these things, when we think about complete response rates, while complete response rate matters, I think, really, at the end of the day, why we look at that is because we want to assess durability of that response. And you can see actually in the updated data, we see some conversion of the partial responses and the CR rate actually go up for CheckMate 9ER over time as some of these patients convert into complete responders. And the PD rate is still incredibly low with nivo/cabo, which is an incredibly important thing when you have a patient in front of you. Now I know we don't necessarily capture it here on this slide, but actually in the presentation by Dr. Motzer earlier today regarding the CLEAR data, we actually see the curves cross at the end, and we actually see that the sunitinib curve crosses over the len/pem curve towards the tail end of the curve. And so that makes me really question the durability of these responses. The reason we care about complete responses is that we -- there could be some association of somebody with a complete response actually having durable disease control. But I actually do worry because it's actually the first study where we're actually seeing the OS curves cross like that. And I think time is going to tell what happens. But I think as we think about end points that matter for our patients, it's long-term disease control. It's disease durability. It's how do I feel when I'm on therapy. I'm going to be on therapy for a long time. The responses to these drugs are long and the duration of therapy is long, and am I going to feel good, am I going to have a lot of toxicity? I think these are all things that clinically matter when we have a patient before us and we're coming to a treatment decision about what to do. I think also when we look at the median OS, here at the updated follow-up, the hazard ratios are pretty similar across the 3 studies. But again, I think with the len/pem combo, the curves are crossing towards the tail end, and maybe that's a reflection of the toxicity and patients needing to stop therapy, not for a progressive event but because it's actually too toxic and then that could be a reason why the curves are crossing there. And again, I think when we're looking here at the landmark OS, both 12 months, 18 months, 24 months, because of the differences in the arms of these -- and the baseline characteristics of these trials, it's really tough to just look at the raw numbers and look at -- and compare raw numbers. And really, I think if we're going to compare, we should look at, well, what is -- how is that combination doing better over the control for that specific trial. And you can see that the delta there for the landmark OS are about 10% across the studies for what's been reported thus far. So I think those are the things that really matter to me when I have a patient before me, comparing the data. It's how good does the regimen work, how durable is that regimen and how safe and toxic that regimen is. I think these are critically important items.

So that's great. A lot of data to discuss here on these few slides and appreciate all that discussion. Maybe, Dr. Pal, if you want to try to wrap it up and maybe give a bird's eye view of everything we've been talking about for the last 10 minutes.

Yes. No, absolutely. Thanks, Dr. Berg. I'll try not to be redundant with really excellent points that Dr. Choueiri and Dr. McKay have already brought up around this issue of toxicity as we focus here on this plot. I think as you look to the 4 of us on the panel, I think all of us really had extensive experiences with both cabozantinib -- Toni and I were part of a Stage 1 clinical trial evaluating cabozantinib in kidney cancer together. And we've also had extensive experiences with lenvatinib and everolimus. I just chaired a 300-patient study comparing 2 dose levels of lenvatinib and everolimus at 18 milligrams and 14 milligrams. And I think that the word in the investigative community and the word on the street is that lenvatinib is a difficult drug to tolerate. And so going into this era of reporting for CheckMate 9ER and CLEAR, I actually had some predictions that really seem to bear out on this slide. I had a thought that when you increase the dose of lenvatinib to 20 milligrams as we do in the frontline setting with pembrolizumab versus the 18 milligrams that we use in the salvage setting, second line and onwards, you're going to run into issues of toxicity. And that's precisely what we see reflected here. I also thought that as you go from 60 milligrams of cabozantinib down to 40 milligrams and apply that in the frontline setting as we do in CheckMate 9ER, you're going to see a lesser degree of toxicity. And lo and behold, you see the data that's on this slide here. It literally doubled the rate of discontinuation due to adverse events with lenvatinib and pembrolizumab versus cabozantinib and nivolumab. I think this is something that really is in line with what I anticipated going into these studies. How does that play out into the real-world utilization of these drugs? I think that increasingly, we as oncologists realize that we're playing the long game with our patients, particularly with frontline therapy. When we get into the second and third line setting, we're willing to perhaps take on a more aggressive toxicity profile. But in the frontline setting, we know our patients are going to be on an extended duration of treatment. We know these patients are going to have to live with the day-to-day side effects associated with therapy. And I think that, that makes the cabo/nivo regimen all the more palatable, 40 milligrams of cabo, again, versus the 20 milligrams of lenvatinib that's used in the CLEAR study. I'm not going to steal Dr. Choueiri's thunder. I think he's going to highlight the quality of life data shortly here. But obviously, that significantly plays into my choice of frontline therapy as well. Dr. Berg?

So yes, that's a great segue, Dr. Pal. So let's now do exactly what you were referencing. Let's turn it over to Dr. Choueiri. So Dr. Choueiri, we've seen a detailed presentation today, which you briefly discussed earlier, by Dr. Cella for the CheckMate 9ER quality of life. And results of KEYNOTE-426 have been reported previously in terms of quality of life. You didn't see any from CLEAR today, but on this slide, we have what we can show you from the 2 trials that have reported. So let me turn it over to you to have you give some opinions about the quality of life and how it relates to looking at these trials overall.

Dr. Berg, thank you very much. I think quality of life more recently, I would say, the past 5 years and the patient-reported outcome is the patient voice -- are very important, at least to me, in picking first-line or second-line therapy or any therapy. And here, the most extensive reported quality of life data -- after I would say CheckMate 214, which we're not going to bring the nivo/ipi IO-IO combination, we're focusing on VEGF/IO, is CheckMate 9ER. We know how important this is because I grew up in the era of COMPARZ, where we did not have immune checkpoint inhibitor. And the better that we could come up with is a non-inferiority trial of pazopanib versus sunitinib with a significant focus on 14 quality of life questionnaire that clearly showed at that time that in 11 out of 14, pazopanib was somewhat superior. Now remember, different quality of life, the timing, where you take the questionnaire, on and on and on. But at least in these studies, what do we have? With CheckMate 9ER, David Cella today presented a very interesting data that built on the presentation from ESMO 2020, which we presented 2 quality of life metrics, the FKSI 19, 19 symptoms, or a subset of FKSI-DRS, which showed a superiority, statistically significant and clinically relevant superiority of cabozantinib plus nivolumab in terms of quality of life over sunitinib. During this GU '21, Dr. Cella extended that and presented way more metrics, including the time to true deterioration from baseline, looking at FKSI-DRS, the EQ-5D-3L, and all of them showed consistent superiority with cabozantinib and nivolumab combination over sunitinib. And the timing of the questionnaire was usually around the end of the cycle of sunitinib. So this is not around 4 weeks when the maximum side effects of sunitinib happen. On the other hand, if we look -- and to my knowledge, there hasn't been any dedicated publication, I would say, full publication for any of these quality of life. But if we look at KEYNOTE-426 and you compare here sunitinib to pembrolizumab and axitinib, the quality of life here captured by the time to true deterioration from baseline seems to be inferior to sunitinib. There was an updated presentation, I believe, by Dr. Jens Bedke looking at sunitinib versus pembro/axi in the KEYNOTE-426 that did not show any difference, did not show superiority of the pembro/axi regimen over sunitinib. Why is that? Could it be the dose of cabozantinib? Could it be just that interaction between cabo and nivo better tolerated than pembro/axi? Or could it be that axi is full dose? We really do not know. What we know is the data. So that's in a nutshell what we can have, really a very extensive report of many, many metrics, quality of metrics -- quality of life metrics with the combo of nivo plus cabo. So a consistent superiority over sunitinib.

Thank you, Dr. Choueiri. So let me toss this one out to Dr. McKay. If you can let us know how quality of life, how you use it in your clinic, what you think of the data presented here as well.