I think we're still in the morning. Good morning, everyone. My name is Daina Graybosch. I'm a Senior Equity Research Analyst here at Leerink, and I'm happy to be joined by Arcus for 30 minutes. We were just commenting, there's no way any of us on the stage, if you know us, do anything in 30 minutes. We're gonna do the best we can. I'm joined here by Terry Rosen, the CEO, and then Jen Jarrett, the COO of Arcus. I think we'll spend the bulk of our time talking about casdatifan, which is your HIF-2α inhibitor. Really exciting program and an exciting space, HIF-2α inhibition, in my opinion. I'm on that. You can go read my notes.
If we have time, maybe we'll get to the rest of the pipeline as you're building some autoimmune programs as well. I think your clinical and pharmacodynamic outcomes do point to Cas being best in class, but there's still some debate there. Can you talk about what you've shown and why you think it's best in class over the only other HIF-2α inhibitor with meaningful clinical program, which is Merck's belzutifan?
Sure. I think that there's that debate, if there was any, is pretty much gone as of ASCO GU. To just reiterate what are the key differentiators. The first thing about casdatifan versus belzutifan is that the biggest driver of the differentiation starts in a very simple way. It's PK/PD profile. It has linear dose proportional pharmacokinetics. Belzutifan has absorption-limited kinetics, so it allows us to really hit the target harder. How does that manifest itself? I'll move through the different clinical efficacy parameters, and we'll show those numbers, and then I'll refer to some biomarker data to reinforce all that. The first thing that you see differentiation is if you look in the late-line setting, and what's nice about that is you're comparing single-agent casdatifan versus single-agent belzutifan.
It's a very clear comparison. belzutifan has a rate of primary progression that's approaching 35%. casdatifan is in the high teens. How does that manifest itself then in the next set of data? Which let's talk about response rate. What we're looking at with Arcus is we have 120 patients in that late line monotherapy setting, and they're broken into four cohorts. Those four cohorts are similar in dose, so we tend to think of them similar from an efficacy standpoint. One of those cohorts includes our phase III go-forward dose and formulation, 100 milligrams. What we tend to do is we speak to the data for the 120 patients, and we speak to the data for the 30 patients that are in that cohort.
If you look at the confirmed response rate for the 120 patients, as we sit here today, it's now about 35%. In the 100 milligram go forward dose and formulation, it's 45%. The confirmed response rate for belzutifan in that setting is just over 20%, so a little more than twice. If you look at the median PFS for the 120 patients, it's stabilized. It's at about 12.2 months. If you look at the 100 milligram go forward cohort, it's up above 15 months. Belzutifan is consistently shown about 5.6 months. We're somewhere between two-three -fold the PFS, and that's the approval endpoint.
The thing that really ties it all together and just amplifies those differences because you can tie scientifically those data, the efficacy data, is we have biomarker data. What we've shown, the gold standard for HIF-2 inhibition in a human is looking at suppression of erythropoietin production, which is under the control of this transcription factor, HIF-2α. What you can see is that not only do we get a deeper suppression, but the durability, the Merck molecule belzutifan loses that PD effect after about nine weeks. We've shown that we robustly inhibit it out past one year. Finally, on the biomarker front, we've actually now shown while erythropoietin is simply a biomarker, doesn't drive the cancer, we've been shown a very clear correlation between the ability to suppress erythropoietin production and efficacy, so a clinical outcome.
Connecting all the way from PK/PD to ORR to PFS to biomarker data, it's very consistent and very different, profiles between the two molecules.
One of the interesting things we see with HIF-2α inhibition is you get really prolonged disease control. I think you see that in all the Bel studies, you see it in the Cas studies. I wonder, why, what is it about HIF-2α that leads to such a prolonged response when a patient does respond? How much longer do you expect duration of treatment because of that on Cas to be than even what we've seen with Bel's?
What Daina Graybosch is referring to is that, You see patients, for example, in the late line, we'll see patients that are out past two years on the therapy even patients that have stable disease. Once a patient. That's why that primary progression is so devastating because if a patient can get past that first scan they have a very good chance of doing very well and you see patients with tails that look very much like immunotherapy. What we believe, there's two factors probably driving that. First of all, in clear cell RCC about 85%-90%+ of the patients have HIF-2 as a driver. It's an important driver across that patient population. That's why when you look at the waterfall plots, they're pretty amazing.
You'll see like 90% of the patients are getting some level tumor reduction. The other thing that points to where the world has been pushing, where you want cancer therapy to go is that the mechanism is extraordinarily non-toxic. In the only AEs you see, and this is really with both belzutifan and casdatifan is on target AEs. In their anemia, which is very well managed, that comes from that suppression of erythropoietin production. A much more rare, you see hypoxia, which is also under the control of HIF-2.
What we believe enhances that durability is because it's such a safe mechanism that essentially what you're seeing is instead of a typical therapy like if you use on the extreme something like chemo, where basically you hit the patient with a nuclear bomb, you get some tumor reduction, but then they get sick, and then they progress. In the case of HIF-2, you're giving the patient a therapy that's very non-toxic. As they do better, as their tumor burden goes down, they get healthier, they get stronger, their own immune system kicks in. That's one of the best prognostics of doing well for a cancer patient, is that tumor reduction and that they do better. They start to contribute. That's a key aspect of the durability.
In fact, we've seen patients that have been stable out past a year and then actually, went on to have a partial response even. Where you can see not only a flat lining, but even tumor reduction past a year on therapy.
The other important point is there's no real known mechanism for resistance. It's very different than a kinase as an example where you tend to develop mutations very quickly. Here, I think there's only been one mutation that's been discovered, and only two patients have been known to have that mutation. The fact that you're not developing mutations quickly and leading to resistance is something that also helps with this mechanism.
You know, the other important thing about that when you start to think about the utility of HIF-2 inhibition, it's likely that a patient's gonna be able to get HIF-2 inhibitor in every line of therapy, similar to like what you see with Bev in colorectal cancer.
There's a question I get a lot, and it's a confusing question, so I'm gonna see if you can answer it.
I'll give you a confusing answer. Yeah.
Bel's, Merck's drug, has these really durable responses too. Just a lot of patients aren't getting to that response. How do you square that with their loss of ability to keep the pharmacodynamic marker EPO? That comes back up after nine weeks, but the patients stay in nice response. What's going on?
I love the question because it to some people, that might be intuitively make sense. It's actually, I think, quite easy to explain because in general, when you think about anticancer agents, they're usually like it's sort of a monogenic target. What you've got is HIF-2 is a transcription factor that regulates hundreds of genes. That erythropoietin that you're measuring in the periphery is simply a marker. What you have to recognize is that each gene has its own dose response, and it has its own environment that might reflect that response.
That could sound like an esoteric answer, but I think there's a very tangible example that I can share that can help you see why erythropoietin blockade, while it's a surrogate, doesn't linearly correspond with what's going on in the tumor. The best way to exemplify that is the second on target AE that you see, which is hypoxia. Anemia, you see about 80%+ of patients will get some level of anemia, and it's anemia, and it's very well managed. Hypoxia, on the other hand, is much rarer, and it's involves different genes, and you only see it in maybe 10% of patients. You can imagine the genes involved in the tumor is another set that's different.
They all have their own, different dose response and environment that can affect the activity. It's not surprising that they're not linearly correlated, but they're just qualitatively correlated.
Speaking about toxicity, do you think as we get to larger studies, we're gonna see that CAS's potency comes with some detriments in terms of on-target toxicity, either anemia or hypoxia?
Well, Jen can describe the data per se, I'll just make one comment on the anemia that as you said, that's one of those things intuitively you might expect them to be similar. I'll let Jen comment on what we actually know about the data. The reason you don't see more anemia is because only about 80% of your erythropoietin production is under the control of HIF-2. You can whack HIF-2 as hard as you want, but you're still gonna have that 20% that bleeds through. You can't get that down to zero. That's why you don't see more anemia. Jen, why don't you just comment on what we've seen to date across our studies?
Yeah. We've shown the comparison from the phase I, ARC-20 study, where we now have data from 121 patients, which is, you know, almost phase III like in terms of size. In that 121 patients, we've shown that anemia and hypoxia rates, both all grades as well as Grade three, four, are almost identical to Merck's and certainly no worse. If anything, we've heard this anecdotally as well, you know, the hypoxia, especially the Grade three hypoxia seems to be a little bit better than what Merck has shown.
Mm-hmm.
Right now, like, we have zero expectation that this would change in the phase III study. The other thing that I'd say, and this is like for any therapeutic, is, clinicians are always gonna get better at managing these toxicities over time. We're hearing that, you know, people are getting better at, like, identifying the patients at highest risk of hypoxia, getting better identifying it really early on, you know, before a patient's O2 sat is super low, et cetera. We think over time, these rates probably come down a little bit just as clinicians do get better at managing them. Punchline is, like, we expect no difference in the AE rates relative to what Merck has shown. No one is thinking that as like a potential liability for casdatifan in the clinician community.
With ARC-20, that's your large umbrella study. Can you just remind us what kind of clinical updates we should expect this year from that?
What's nice about ARC-20 is the way it was set up is very easy for us to add additional cohorts. It's a very efficient way for us to look at either dosing regimens or different combinations or different settings for casdatifan. As of today, we now have 10 different cohorts that have been added to ARC-20. The first four were monotherapy. We added a CAS plus cabo cohort. We added three early line cohorts, so looking at CAS without a TKI in early line settings at RCC. We're just now adding an open enrollment for two cohorts looking at CAS plus anti-PD-1 and CTLA-4. What's nice about all this is there'll be a very steady stream of data flow
Mm-hmm
You know, over the next two to three years from this study. You know, it's also likely we add additional cohorts over time as we wanna look at different combinations. As far as the next datasets that we expect to come out of this study, first will be that Cas plus cabo cohort, so the fifth cohort. This is evaluating the same combination, same setting as we're evaluating our first phase III study, PEAK-1. We've said that dataset can come any time from around July to October. When it comes will probably depend mostly on whether we decide to just present it at Investor Event.
Mm-hmm
O r wait for a medical meeting. The next dataset will be Cas plus anti-PD-1 in frontline RCC. This would actually be the first dataset ever presented for a HIF-2α inhibitor in frontline without a TKI. We've already disclosed the primary progressive disease rate from that setting, which showed 9% versus what you'd expect for PD-1 alone is probably more like 25%-30%. That obviously bodes well in terms of what we think we're gonna see from that cohort. The other combination that we're also likely to now see by the end of the year is the one that we just added to ARC-20. That's the one looking at Cas plus anti-PD-1 plus CTLA-4. That we're exploring as a likely regimen for our first phase III study.
At a minimum, you know, we should have some safety data on that by the end of the year. As we get into next year, you know, obviously, you just have efficacy data and, you know, more mature data from all of those cohorts.
Will we have more of the early line mono Cas cohort data this year?
We do this too, yeah. I do wanna, like, confuse too here. Yeah, given there was three early line cohorts, yeah,
Mm-hmm
D ata from at least one of the additional cohorts as well.
Probably the most important is casdatifan plus anti-PD-1 because that's likely to be the backbone. As Jen mentioned, we've got PEAK-1, our second line study's up and going and rolling, but our first phase III is likely to be casdatifan, anti-PD-1, anti-CTLA-4. We think that seeing the data from the anti-PD-1 cohort, which is fully enrolled with Cas, will be very reassuring insofar as how that the benefit that you're getting from IO plus Cas. In fact, if you look at what is known to be the rate of primary progression for ipi/nivo, it's about 20%-25%, and we're seeing less than 10% with Cas plus anti-PD-1 alone.
We expect to be in, you know, starting that phase III study by the end of this year, and we think that that dataset will be particularly important when people see the efficacy there and knowing that we're also looking with anti-CTLA-4 on top of that.
Let's talk about your first registration study, PEAK-1, which you've mentioned many times. We've seen the first data from ARC-20 with Cas-cabo, this year will be the more long-term follow-up data. You're, you decided to test of cabozantinib in the second line setting, and the trial excludes patients of prior cabo exposure. For those who don't know, Merck's study, which we just saw the outcomes from at ASCO GU, LITESPARK-011, similar study, they tested with their TKI lenvatinib and compared that to cabo. Why PEAK-1? Why was this the first registration trial for Cas? You didn't go single agent, Merck did. Do you believe that there's synergistic benefit, or additive, either in efficacy or safety?
I think additive plus synergistic are almost the same thing. We definitely think that adding a TKI to cas mono as great as cas mono looks, is going to add even greater activity. The reason we like cabo as the TKI partner is, first of all, we had an ad board and, like, asked them, like, "Which TKI would you use?" They were, like, unanimous that you should be using cabo. They said, like, for RCC clinicians, cabozantinib is almost like water. We're just so used to using it. We're so comfortable with it. That was what really led us to the choice of using cabo, and we also felt like cabo is a better-tolerated TKI. Cas is so well-tolerated, the last thing we wanted to do is, like, bring a bunch more toxicity onto Cas.
We were definitely solving to have a better-tolerated TKI. The other thing that's nice about cabo, it's just a lot easier to dose, which is critically important once you're adding it to something else. cabo, you know, typically start with 60 mgs. If you see some tox, you just go down to 40, and then more tox, go down to 20. Very simple.
For lenva, it's available in a lot more different dosages. If you look at the different dosages approved for RCC for them, they've got three different RCC approvals. It's a different dose of lenva for each of those, so a lot of clinicians say we're not even sure what the right dose of lenva to use is. Lenva's used all the way from 20 mgs down to 4 mgs, so just a lot more complexity in dosing lenva.
For all these reasons, we felt like cabo was the better partner. As Terry Rosen always says, you can just, like, look at the numbers. If you look at cabo use versus lenva use for RCC in the U.S., it's about 2.5x more use of cabo. In some European countries, it's 10x. Literally, there are some countries where it's, like, 40 times more cabo use relative to len use. We feel very, very strongly that cabo was the best TKI partner, and we do feel like it's gonna bring additional efficacy to Cas.
I think, the other thing when you think about, it's a, it's a far cleaner study where you've got cabo in both arms. As you know, Merck For those who aren't aware, Merck also has commercial rights to lenva, there certainly could have been a commercial element to that decision. When you think about the hazard ratio, they've got cabo in the control arm, lenva in the study arm. Interestingly, to your point, when we think about this study, Cas cabo versus cabo, we think when you're sure gonna show a really favorable hazard ratio particularly because of the AE profile, you should basically expect to take 100% of what was cabo. It becomes a no-brainer.
What's interesting in thinking about the real differences between lenva and cabo is that, in fact, after the data came out and you had in that LITESPARK-011 in this latest analysis, they had a you know, it had a 0.7 hazard ratio. Despite that, with all the investigator discussions that followed and the, you know, they were very public, there was some you know, whether debate is the right word, but there was a lot of discussion about whether they would use it, whether they might still prefer cabo. Certainly, it's going to be adopted and used, but not necessarily unanimously. I think that points to the differences in the clinical practice world of lenva versus cabo.
Let's talk more about, like, PEAK-1. Maybe you could give us an update on when we could potentially see that readout, if you have interim analysis or you're guiding more towards the final analysis. And then also, you know, you have PFS as your sole primary endpoint, which is different than Merck and their similar trial that had a PFS OS co-primary with most of the alpha on OS. Maybe you could talk about that decision and whether you still have a lot of confidence in that after seeing Merck's LITESPARK-011 data.
There was like 10 questions.
I know.
All really good ones.
I, Yeah.
Well, I'll start with the last question 'cause I think that's a really important one. Yeah, we absolutely feel very, very comfortable with PFS as the sole primary endpoint. OS is a key secondary endpoint, so we will absolutely be testing for OS, and we think it's important to for sure not show a detriment for U.S. The update was very clear that for second-line RCC, PFS is the gold standard regulatory endpoint. We feel very, very good about our selection of PFS as the sole primary. If you look at most RCC studies, both front line and second line studies, they had used PFS as the sole primary just like us, and then OS as the key secondary. We're not, you know, being a renegade here.
The second thing that I would say is, we don't like to share details of our stats plan, but, you know, we have a saying. It's a very simple study. There is one primary, there's no interim analyses. You know, the advantage of that is that you're not having to split your alpha, so we're really maximizing the probability of success of this study, which is easy to do when you know that, like, you're using the endpoint that the FDA is very comfortable with. We feel great about our choice of endpoint. We feel great about the study design. It's, you know, as Terry was saying earlier, it's as simple as you can get. Same TKI in both arms. You know, there's not gonna be any confusion, you know, contribution of components, et cetera.
We feel great about it. What did I? I probably left something out.
I think you. Did you?
Merck should have done that since-
Yes.
They never presented the interim analyses anyways.
I know.
What was the point?
Exactly, yeah. They tend to get overcomplicated with their study designs and. Yeah. I think we looked at their study. That was also nice when you're like the number twos, you can look at their stats plan, how did things work out from them, take a look at our stats plan. Do we still feel good about everything that we've done? We've already done that, we're like, we feel even better honestly about our stats plan than we probably did, even though we already felt good about it. What we saw from LITESPARK-011 was just like very confirmatory.
The other thing that reflects, y ou know, when we started and we talked about the differentiation, and the discussant noticed this, noted this. I'm sure you noticed it as well. If you looked at the curves from LITESPARK-011, you really didn't see any separation between the study and the control arm for six months. The discussant, that caused her to remind people about that 35% rate of primary progression as a monotherapy that the molecule isn't hitting the target as hard, as fast, but it's bringing some durability. We fully expect that, with casdatifan, all of our data, every single cohort we've shown, we see low rate of primary progression. We expect those curves to start separating sooner.
While OS is not a required endpoint for approval, we actually they showed a trend, but statistically it's probably a pretty low probability that they hit on OS, and we think that if we hit there could be another differentiator from the a long-term perspective of differentiating the molecules.
You know, it brings up a point always the second in class has a challenge and that the first in class is approved, and so your studies have to anticipate use of Bel's post-study. How do you think that could impact your OS, your ability to show OS?
Bel's is still very hard to access outside of the U.S. It's still not paid for in most of the European countries. It is obviously like in late lines, you know, it's getting to be pretty widely used here. I think our geographic footprint is likely to look very similar to Merck's, which had about less than 20% the U.S., about 60 Europe, and then the rest was rest of world. I actually don't think we're gonna get that much Bel's use as a subsequent therapy. To your point, what I do think helps, I know you and I have traded emails on this, but because they use cabo in one arm and lenva in the other, patients as a subsequent therapy were likely getting what they didn't get in that prior line.
If you were in that cabo mono arm, you probably got len next. If you were in the Bel's plus len arm, you probably got cabo next. Your subsequent therapies were different depending on what arm you were in. For us, given that patients of both arms are getting cabo, it's probably gonna be more consistent in terms of what those different arms are getting for subsequent therapy, which I think will help us.
The other, the other thing I'll just say, like fast-forward to a world where everybody's had their approvals. The way I would see it is first off, it's likely that HIF-2 and a HIF-2 inhibitor is gonna be used in every line of therapy. The way we given our holistic development strategy, you could imagine that we start with a TKI-sparing regimen, then in second line they get Cas cabo. If you picture a world where ultimately Merck's front line is approved, let's say that's lenva, Bel's, pembro. If a patient progresses on that, by definition they're gonna get Cas cabo.
Mm-hmm.
We don't really see that affecting things. As you also know implicit in that question, you'll see Bel's belzutifan use in adjuvant setting. We don't think that will affect things at all. Interestingly enough, we take away from the positive data there, just a reinforcement of the importance of HIF-2, even in early lines, even though we had no doubt and we actually have a small study that's ongoing in the period, a neoadjuvant setting with casdatifan. We don't really see the Bel belzutifan usage affecting where we go, largely because of that point that Jen was talking about, that you're not gonna have resistance development to HIF-2 inhibition either.
It's a really interesting point because if you can be used in every line, this gets to be an even larger market.
Oh.
Given it's so important, do you actually have a study testing that? When are we going to have real data testing sequencing?
We will. We don't have any patients at this point that are post, Bel's, but there's a number of things we are thinking about and by thinking, we would probably do. I think the most interesting one to do that would just be most illustrative because it, there's a lot of good science behind it as well and it'll, you know, confirm something important, is that since the rate of primary progression is so different in the late line, you would basically, if you took the primary progressors from belzutifan or in fact took all progressors from that third line plus patient population that progressed on belzutifan then put them on casdatifan, I think you'll see some differentiation there.
'Cause particularly the primary progressors, as you know, where there's no sign that you're really having resistance development, a patient that's progressing before the first scan was probably one of those that if you have one that's a high teen and one that's 35, the delta you would've thought would've been susceptible to casdatifan and they certainly didn't become resistant during that period. We would expect to see some meaningful tangible response in patients that failed on belzutifan.
That proves Cas is potent, but it doesn't necessarily prove you can keep it online and switch to combo partners. You almost need two studies.
The reason is I mean, the HIF-2α inhibitor could still be bringing some benefit.
Yeah.
Maybe it's like that TKI that they've developed resistance to. If you keep the HIF-2α on board, so keep Cas on board, but switch out, you know, cabo for whatever, or, you know, they're getting Belslen and it was the len that they got resistance to and then you put them on Cas cabo, they could, you know, both between having a more potent HIF-2 inhibitor and a TKI than, you know, the patient could-
inevitably you are gonna see patients that progressed on Belslenva pembro go on to a Cas cabo at some point.
Yes.
I mean, what is very different about RCC and shirpa about the market opportunity here, it's just very different than other tumor types. Like, you look at lung and pancreatic, it's just, you know, there's frontline treatments if you fail, there's second-line treatments if you fail that, there's probably nothing on it. It's like RCC, these patients can stay on treatment for five to 10 years and clinicians just like sequence different therapies and take them off things, put them on something else. Again, I think it speaks to the point that you could definitely see a scenario where HIF-2α inhibition is used over and over again.
'Cause as you know, even if you look at the current world, building on Jen's point, you sort of, they sequence through. At some point they're gonna get every TKI. That's one of the nice things about getting the, you know, pushing the TKI to later where you're gonna get, you know, several years without a TKI that only we can do because of the properties of the molecule, and Merck can't do that 'cause of the high rate of primary progression. It, the reason that works is because none of those TKIs are really the same. They all. That's what's gonna be your, drive your progression into the next line, not the HIF-2 inhibitor.
Awesome.
Cool.
Okay.
Thank you.
Thank you.
Thanks
for all for your attention.
Appreciate it.