Arcus Biosciences, Inc. (RCUS) Earnings Call Transcript & Summary

All right. Well, thanks so much, everyone, for joining us. I'm Terence Flynn, the U.S. biopharma analyst here at Morgan Stanley, and we're very pleased to have Arcus with us this afternoon. We have Terry Rosen, the company's CEO; and Jen Jarrett, the company's COO. Thank you both for being here. Before we get started, please see Morgan Stanley's research disclosure website at www.morganstanley.com/researchdisclosures. If you have any questions, please reach out to your Morgan Stanley sales representative.

Well, again, thank you both for being here. I thought one place we'd start would be just high-level, obviously, the Gilead collaboration has been very important for the company as we think about what it's allowed you to do in terms of kind of parallel process opportunities, strength of balance sheet but then this collaboration was also expanded a couple of months ago into immunology. So maybe just give us the latest perspective on kind of that collaboration and kind of what the forward looks like from here.

Great. So Gilead has been an awesome collaborator from day 1 and it's extremely enabling. So when you think about what we're doing as a company in trying to build a long-term independent company, we have a large discovery organization. We have a large development organization. We started from scratch, but now we have 4 registrational trials, as you said, a number of earlier stage programs. And then we've also moved into immunology. And so what it's enabled us to do is to continue to build out the late-stage programs, to execute on those, but not have to cannibalize any of the earlier stage research. So from their standpoint, from day 1, we talked about being an R&D engine, and I think it enables that. But at the same time, it doesn't come at the cost of being able to compete. When you think about like anti-TIGIT, for example, where we're having to go against Merck, primarily Genentech, other companies that are in that late stage that we can do the trials that we should do. The other part of that is it's not only capital. So when you look at Gilead-Arcus, we're essentially colocalized. We're right across the San Mateo bridge from each other. The human dynamics are great. They've actually got hundreds of people working on the programs. You name a function, not only R&D, but things like even as, call it, esoteric like supply chain, we just worked together on everything. So it's broadly enabling and I would say, we expect to see more together.

Yes. And then where immunology come from? I mean, again, some of you might think that's kind of out of left field. So where did the immunology angle -- what generated that?

Yes. So it's actually surprisingly somewhat right down the middle of the fairways as opposed to [indiscernible] but I think that is like a high-level first glance, one might think that a lot of what Arcus does, because it's in immuno-oncology, our biology is immunology. So we've probably got more immunology preclinical work going on even though it's an oncology clinical endpoint. And so that's, first off, scientifically what makes sense. When Flavius Martin joined Gilead as the Head of Discovery, we had some past knowledge of each other. My Co-Founder, Juan Jaen, also knew him well. We got together even before he was officially on board. The dynamics were good. Although the original way the collaboration was structured, it really focused on later-stage programs. We felt it's always in our best interest to be aligned early on. We started talking about immunology together with Flavius early on. And Gilead, as you know, also had a substantial immunology development group. So it's that just very organically matured, we said, why don't we do something together there? Again, towards that end of us being well capitalized to enable the R&D engine, it enabled Gilead to do that with us. And I think one of the advantages that we often talk about the way that was structured because we decided to pick those targets together, it enables us to have an alignment very early. So the amount of investment that you have to go through, let's say, to get an opt-in that we feel like we're on a very good page to begin with [indiscernible]. So it's a very organic process that just has to do with the dynamics of the company that we're interacting all the time at an [indiscernible] level.

Okay. Great. I guess just to start going through the pipeline here. Obviously, front and center is TIGIT. And again, a lot of data that's come out over the course of this year, you had your Phase II data in first-line lung at ASCO. And I think one of the debates that emerged was just control arm performance. But then since then, we've seen more real-world data that started to come out. So maybe just walk us through kind of the key other data points that give you confidence in that control arm performance? And then we'll talk about some of the other competitor data that's come out more recently as well. But maybe just start on the control arm.

Yes. So I think one of the stories that was probably just starting to emerge, as you're saying, was that KEYNOTE-024 and KEYNOTE-042 were probably not real-world studies. And so actually, at ASCO Plenary when our data was presented, Solange Peters, who is one of the SKY-1 she's actually very close to us as well, actually watched her 2 recent real-world data sets that have been presented, I think as ESMO IO, one of the big conferences, which showed that in the real world, pembro performed very differently than it did in KEYNOTE-024 and 042. And we actually have a slide that's now in our corporate deck online, where we looked at other studies aside from some form of the more contemporary studies relating to KEYNOTE-024 and 042, where you see the same trend, whether it's how nivo performed in the study that supported the approval of dual-LAG, but there's like multiple examples of when PD-1 is a control arm and not the experimental arm, it performs differently. So I think that's now very well accepted. I think the other thing that's very well accepted is that at the time, KEYNOTE-024 and 042 were conducted, it was like a very different time in the space. There was a lot less competition. PD-L1 testing then took about 3 weeks. So you tended to lose all of the rapidly progressing patients. So you ended up with your healthier patients. Now PD-L1 testing turnarounds are much faster. So for a whole bunch of reasons, we think that you would never be able to replicate the results that Merck then got with KEYNOTE-024 and 042. So we feel very good about how zim has performed. There's nobody at Arcus, right, Gilead has any doubt that zim looks just like any other PD-1 antibody. I think the clinical world is very, very comfortable with zim as PD-1. And then also as a reminder, like even if you don't believe all of that, we're always combining zim with somebody else, dom or whatever it is. And so at the end of the day, if dom plus zim beats pembro, it won't matter. So we think it's just as good or better than pembro but like I said because we're always using pembro as a standard of care -- sorry, as a control arm, that's the standard of care, it doesn't really matter what zim would do on its own.

Okay. And maybe to segue into the Roche SKY-1 interim OS data that came out. Just reflect on that data in light of what you're just talking about, Jen, in terms of control arm performance, but then also what this means in terms of confidence for the TIGIT hypothesis.

Yes. So first of all, I think our confidence in the TIGIT hypothesis has not changed. We have been consistently very confident based on our own data set. The external world is probably now more confident like the debate has shifted like, is TIGIT active? Is TIGIT going to work? To like, do they hit? Is it statistically significant or do they just miss? And so, it kind of feels like at this point, the worst-case outcome for SKY-1 is that they end up with a 0.8 hazard ratio and they need to be at 0.79 or 0.79 versus 0.78. So it's a just miss and we think either way, that's a win-win for us. Either they hit, the whole world is very confident into the success or they just miss and Roche loses what was probably going to be -- its only real advantage, which was having a first-mover advantage on the market. So we're excited with the SKY-1 data and the fact that like the external world is not even more confident. We actually just had a team that was at World Lung and meeting with a bunch of clinicians there. And I think a lot of clinicians once they saw the SKY-1 data are also more confident in TIGIT, that's obviously what we care about even more than investors. So we feel really good about the setup and we continue to be very confident in TIGIT. And I forgot the other part of your question was…

No. I mean the real world -- I mean the control arm…

Yes. So just as another example of the control arm performed a little bit differently than when TIGIT with experimental arm and IMpower110. The issue that they have is that they did not use what's considered the standard of care, which is pembro. So even before this cycle, we were having [indiscernible] 6 months or 9 months ago, we were hearing them from clinicians that it was going to be always hard to interpret their hazard ratio because they weren't using pembro as a standard of care, they're using atezo, there's a view that atezo may be inferior to pembro. So I think the fact that they did not use pembro is the standard -- comparator arm is going to end up hurting them from a commercial perspective.

Okay. Understood. And then maybe just give -- remind us on your Phase III lung program in terms of kind of status, enrollment and timing of data?

Yes. So we have 4 Phase III studies, 3 of them are in lung. The first study, which is targeting the largest market opportunity within line lung function data [indiscernible] our studies is STAR-121, which is our all-comer lung. TIGIT plus PD-1 plus chemo combo. So Roche has a similar study SKY-6. Merck has a similar study as well. So we're super excited about that study. It is enrolling well. I'd say that and then STAR-221, which is our first one all comers in gastric. Those are our #1 priorities, both us and Gilead from an execution standpoint. The other 2 lung studies are PACIFIC-8. So that's a study in Stage III LUNG that AstraZeneca is operationalizing. So there, we're combining TIGIT or TIGIT -- dom with durva -- versus durva. And then the last lung-style study is ARC-10. So that's the equivalent to SKY-1. So that is in the PD-L1 high patient population. The study that actually could read out versus STAR-221 in gastric. So actually outside -- right now, that an STAR-121 or neck and neck. Those are all comer studies. So easier to enroll. STAR-121 has a little bit more competition, obviously, because it's in lung versus gastric. But like I said, both the studies are getting a lot of interest and are enrolling well.

The one thing I would add to that in the context of anti-TIGIT is Jen was talking about 221, our next data set in the TIGIT field is going to be probably later this year at the latest, very early next year, and that's in exactly the same patient population as STAR-221. So it's a study called EDGE-Gastric or ARC-21. It was really designed. That study is enrolled -- the Phase II study is already enrolling very well. But this platform study included the arm that I'm about to talk about. And the idea there was in certain regions of the world, they want to see exact combination, exact setting as what you're going to do in your Phase III. So this study was designed to do that. And so we're going to have 40 patients in this exact setting. We'll break them out by PD-L1 status, you'll see ORR, and you'll see 6-month PFS. I think what's also important is that not only will it give a feel for that particular setting. But if you just think about our doublet, there's actually a very contemporary study that BeiGene just presented in the same patient population with their anti-PD-1 their anti-TIGIT. And what you've got there is a very good surrogate for the field like whether you think about Merck with their anti-PD-1 and the anti-TIGIT, the combination of an anti-PD-1 with an anti-TIGIT that's Fc-enabled and data sets to date have consistently shown that when people report on it, the Fc-enabled anti-TIGIT deplete Tregs in the periphery, and you see immune-associated AEs. In the BeiGene data set, you'll be able to look at the safety data from that set and you'll have a very contemporary data set with the only primary difference being in our case, we have the Fc-silent anti-TIGIT. You'll be able to look at the efficacy as much as, whatever you feel like you like to do when you look across trial comparison, but you'll also see data in the context of the AEs associated with that. And I think it's going to really reflect not only on that setting but when you start to think about the competition and we view our primary competition in the broader field is Merck and when you start to think about Vivo, KEYTRUDA versus dom, zim and the big difference being the Fc-silent anti-TIGIT, I think you'll start to see perhaps a foreshadowing of what those differences might be and which might really be the better doublet.

Yes. Okay. Great. Maybe just to go back because I just want to ask one question on the stats. Are you planning to make any changes to your lung program based on what you saw or you were already kind of planning for that based…

Yes. I mean I don't think we were super surprised by what we saw. We've had our own data set with ARC-7, and so that certainly informed how we think about stat. So we're always pressure-testing our statistical analysis plans and yes, making sure we continue to feel good about our stats as is Gilead. But I think right now, we feel very good about our study designs and our statistical analysis plans.

Okay. Great. And the other question, again, it relates probably to both indications is just this question of competing commercially with not just Roche and Merck but also this issue of coformulation. And so would love your perspective on that because that's another question we get pretty frequently is like, all right, it seems like the coformulation is going to be a lot easier for centers, physicians, et cetera. Again, I asked Gilead the same question earlier today, but I would welcome your [indiscernible]

Yes. Not sure what Andy said, but yes, should be convincing because I think we're all aligned.

This one is like a [indiscernible]

We have a lot of discussions about this with them actually but actually probably 2 years ago, we did some market research because we were trying to figure out what we want to do if we want to coform, if we wanted to do just coadmin and that research shows that clinicians actually prefer having individual agents versus a coformulation. So they like idea if you are seeing some AE, you don't necessarily have to withhold 2 drugs. You can potentially just withhold one of the drugs. Interestingly, also like payers, particularly don't like coform. I mean, they're smart. Obviously, they see coforms for what they are and the way to get around patent expiries. So for both those reasons, we decided to go down the path of coadmin. So what we're going to be doing is 2 vials that will come in the same package, so it'll probably be marketed under one brand name. And then the natural solutions will be mixed into one bag. So you get all the same advantages, the [indiscernible] advantages that you get from the coform and what really matters at the end of the day is how long the patient is in the infusion chair. So our goal is to meet whatever Merck is doing from that perspective. But we have decided to go to the coadmin copackaging path. We'll continue to keep an eye on coform. And if all of a sudden, we feel like coform is a way to go, we'll certainly pivot a bit. But right now, we feel very strongly that coadmin is the right way for us.

Okay.

Now I'll make one comment that gets to -- we talk about the safety advantages, but the safety advantages can actually translate to efficacy advantages if you're forced to either take patients off or back down in the dose that you're using of your anti-TIGIT because of Treg depletion. So remember that Treg depletion is an on-target activity. So those are Fc-enabled anti-TIGIT, if they're depleting Treg, that's through direct interaction with TIGIT. So if you back down to avoid that, then you're also risking whether it's in certain patient populations, certain individual patients. Obviously, you're reducing target engagement, and you could imagine manifestations and so far, as efficacy as well.

Okay. Maybe just one more on the lung cancer side before we go back to gastric is just obviously, there are multiple mechanisms going forward now in frontline. So you've got TIGIT as we just talked about, you have TROP-2, B6As and another one. And so kind of our view is that there's going to be a fragmented market more as opposed to now, it's just KEYTRUDA 80%, 90% of the market. And so do you think about like evolution of the frontline lung market? What's kind of the take on how that plays out here?

Yes. We actually think we have the easiest path because we are building upon the standard of care, which is actually much easier to do than trying to displace part of the standard of care. And so that's, I think, the primary difference between the TROP-2s and what we're doing. I mean I think TROP-2s at the end of the day is going to try to displace chemo. I mean AZ is obviously, I think, still toying around with TROP-2 plus chemo plus PD-1, but that, I think, looks too toxic. So I think at the end of the day, like Gilead, they're going to be combining TROP-2 with just PD-1 and they're trying to displace chemo, which I think is a high bar. I mean I think the TROP-2s are really, really exciting class, including in lung. But I think so far, we probably haven't quite seen data that shows that it's better or less toxic than chemo. I do think very strongly that it's going to have a very significant place in second line. I think that's how you might see the lung cancer market really evolve as in the past, you kind of had chemo and then chemo plus PD-1 and then you adopt the [ paclitaxel ] on the second line, that was it, if you were a patient that didn't have a mutation. So I think what is going to change is now you start to have the KRAS inhibitors. So I think that will become more established for KRAS patients in second line. I think you'll start to see the TROP-2s, other ADCs used in the second line. So I think what you'll start to just see is like revelation about sequential therapy in lung which is something we just really haven't had in the past.

Okay. Great. So maybe just going back to the gastric data and the EDGE study you're going to have. So have you talked about a venue yet where we're going to see that data?

So for EDGE-Lung, we have not, right?

Sorry, EDGE-Gastric.

EDGE-Gastric, sorry, we haven't said exactly when we're going to present that data. So our goal is to present the data either by the end of this year or very early next year at the latest. So we have a plan in mind, but we haven't disclosed exactly what that…

Because we have -- basically, we know what we want to do, but until we have an accepted abstract.

Okay. Got it. Understood. Since you mentioned 40 patients are stratified by PD-1 status, we get ORR, 6-month PFS. What -- just remind us the relevant benchmark here for Opdivo chemo as we think about what you guys are trying to look to compare against?

Yes. So I mean there's actually like 3 studies out there that are all pretty much in line with one another. So the study that you're referencing for Opdivo plus chemo CheckMate-649, there's also a KEYNOTE study, KEYNOTE-859 and then actually a BeiGene study with Rationale 305. And if you look at all of those studies, the highest ORR that you see is 60% in the CPS-high patients and then it sort of goes down to 50% for patients that are CPS-low. BeiGene was a little bit lower for some reason. So they were, I think, at the high -- sorry, low 50s in the CPS high patients. And then the PFS ranges kind of anywhere from 7 months to 8 months. So 8 months at the very highest. So what that tells you, I think, like anything above 60% ORR would certainly seem like you're adding something. And then I think once you get beyond sort of 7.5 months PFS, that's when it seems like you're adding something to PFS.

Yes. Okay. Got it. Okay. Great. And then maybe just remind us that how to think about the commercial opportunity there? I mean, again, that is one of the kind of more below-the-radar cancers where again...

Yes, for sure. It's a huge cancer. I mean the other thing people get a little bit confused is by there's another histology for esophageal cancer, esophageal squamous cell carcinoma. So that's what Roche is pursuing with their SKY, I forgot the number, I think SKY-8.

Okay.

That study is much more prevalent in China and much less prevalent in U.S. and Western Europe. So the histology that we're looking at, which is the adeno histology is much more prevalent in the U.S. and Western Europe. So there's about 25,000 patients in the U.S. alone. There's 100,000 if you look at the U.S. plus other G7 countries. You just the other day figured out when the G7 countries are but if anyone wants to know, they can ask us afterwards. But anyway, if you look just at those countries, that translates into a market opportunity north of $3 billion. And then obviously, you tack on other countries and it grows from there. So it's a big opportunity. So we'd like to say there's obviously people that want to believe our work is just going to take the whole market in lung, which we don't think that's going to be the case, but let's say, you want to believe that. We are well ahead of anybody in gastric. So we're the only company with a TIGIT antibody that's running a Phase III study today in gastric, there's also just very little competition from other mechanisms. The only thing that's really out there is that we're keeping an eye, I don't know, say cloud in 18.9%. That's only like prefilled said set of gastric cancer. So it's a huge market opportunity, and that's a market that we could really own for a few years.

Okay. Great. Maybe I just want to move on HIF-2 alpha here. Again, I think TIGIT and all the IO work is taking up a lot of the focus from investors but HIF-2 alpha. We've been talking about it for a year now, and you've been excited about it and I saw some Phase III data from Merck in RCC setting. So maybe just remind us kind of the strategy, differentiation and then kind of next milestone here as we think about this asset?

Sure. So this is one. And fortunately, as this SKY-1 overhang disappears, we get -- we're getting tons of interest now from investors because this is AB521, which is our HIF-2 inhibitor, is the drug waiting to happen. And so the differentiator is very simple. The belzutifan is given clinically the dose of 120 milligrams. The reason of that 120-milligram dose has nothing to do with an MTD. It has nothing to do with modeling. It's simply that it has absorption-limited pharmacokinetics. So if you go above 120 milligrams, you simply don't get substantially greater drug levels. HIF-2 is a very challenging target to get a small molecule. It's a transcription factor. That's why you don't see 1 million programs despite the fact that you have an approved drug. So you've got a validated mechanism, and improved drug. And what we've shown first in healthy volunteers and now our next data set, which is imminent, will be from our dose escalation study in patients and what you can expect to see is similar to what we showed in those healthy volunteers that basically, we achieved 3 to fourfold, so it's not just incremental, 3 to fourfold the exposure that Merck gets with our 120 milligram dose at the go-forward dose that we selected which is 100 milligrams. There's a very good pharmacodynamic marker that has nothing to do with the anticancer activity, but it's easy to measure that if you inhibit HIF-2 in the healthy kidney, you see reductions in EPO levels. So you'll be able to see pharmacodynamically that we hit and reduce that same as belzutifan. What they do at their clinically used dose, we can demonstrate in a dose that's roughly 3 to fourfold lower than that will similarly show a very similar safety profile. And the next data set that we share will be from that dose escalation study. And while the numbers are small, there are a good number percentage-wise of RCC patients, and I think you'll be able to pick out the RCC patients, which starts to bode well for the efficacy piece as well. And so far as what else is happening, we started a dose expansion study in clear cell RCC patients, that's enrolling incredibly well, a start of a combination study with a TKI is imminent. We haven't shared exactly that partner, but we feel it's going to offer an advantage of the doublet with a better TKI than lapatinib and our aim there is to be in the Phase III study by the end of next year.

And what -- and so the data eminent, I think you used to talk about maybe this year, more likely next year. So it seems like that's pulled forward a little bit. Is that fair?

That's probably pulled forward potentially a little bit like we're still thinking for exactly when we would present the data. But yes, it could be pulled forward a little bit.

Okay. Got it. And the Phase III program, maybe just again, what's limiting to starting that because you said by end of next year?

Yes, this time, just Phase III will take about 12-month process from beginning to end. So some of its time and some of it is you need to generate like a sufficient amount of safety and efficacy data to support your filings in different countries, so just supporting that data. But we're being as aggressive as we can, and we're excited to continue on this path to get a Phase III study going by the end of next year.

As a reminder, this Gilead has an option on all these programs. And so maybe just remind us like what's the minimum data you have to deliver to them and how does all that stuff work?

Yes. So for every molecule, it's a little bit different. The idea is that it's supposed to be whatever would be considered clinical proof of concept for that molecule. We actually just went through the process to align with them on what the opt-in trigger would be. So we do know when it's going to be, it's not tomorrow, but I'd say it's not 2 years out either. I think it's a very reasonable period of time for now. And we're seeing a very close content, let's say, about the program. So I mean they still see all the data. Anything we would do clinical collaboration-wise, they know about it, they can have some influence just so that they want to provide advice on study designs or whatever else. So we'll say we know they like the mechanism, and we hope they're there. But if they decide not to opt-in for some reason, it's also an asset that we'd love to own on our own or we'll find another partner or whatever else so.

So that -- so sorry, the opt-in is not standard for every program. Do you have like some discussion about it in that…

Yes. It depends on a little bit on like what you consider to be clinical proof of concept for that molecule. And so obviously, like HIF-2 alpha, is just be a little bit different than like an across like IO molecule. So I'd say because of that also, the opt-in triggers relatively soon. So we'll see.

Okay. All right. Sounds promising. Okay. And then the Phase III, so would the Phase III program differ if Gilead opt-in versus if you did it alone or you think it's pretty boilerplate?

Yes. No. I think…

Pretty standard.

Pretty standard and yes.

And would you have to do like how would you handle the control arm given Merck's already approved like, how would that work?

It wouldn't be approved by the time we started the Phase III and it's in -- sort of not do opt-in a lot of countries.

Because you'd only do [indiscernible]

Yes. Exactly. Yes. And I think we feel pretty good about like what our strategy would be. So we haven't said a whole lot about it, obviously, but I think we have a good strategy in place.

Okay. Got it. The other -- I mean we only have 20 seconds left. I was just going to say quemli, with 20 seconds.

Yes. Quemli is very exciting. So it's our CD73 inhibitor. We're looking in frontline pancreatic cancer. This is a study that's been ongoing for some time. The last patient in, keep in mind was like end of 2021, we're going to share very shortly OS data. So we think we have the opportunity to be the first meaningful change potentially in pancreatic cancer therapy in the better part of a decade. The -- since we only have 20 seconds, I would just say the gem/Abraxane data are out there, there's a number of studies. You can tend to think of OS for gem/Abraxane in the probably 8- to 10-month range. You can think of FOLFIRINOX slightly better than that. It tends to be a healthier patient population because it's more toxic regimen. We'll have 120 patient population, 90 of which were randomized to determine whether anti-PD-1 brought any benefit. Anti-PD-1 does not bring any benefit. And what we'll be sharing is the OS data, and we think they're exciting. They look meaningfully different no matter how you cut these historical data and compare gem/Abraxane, when you look at patients with liver mets, with a liver met, with treatment, without pretreatment, you combine them all together, we look favorable compared to historical gem/Abraxane on all those fronts, and we think it will be a very exciting data set, both for that pancreatic cancer and for the molecule itself and the role it might have in other settings where you're using immunogenic chemotherapy.

Great. Well, thank you so much, Terry, Jen, really appreciate the time.

Thank you. Appreciate it.

Thanks Terence.

Great.