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

So first and foremost, it's good to see everybody back in person again after 2-year live in person. Thanks for all your support and enthusiasm for attending Barclays Biotech Conference in Miami for both investors and management team from Arcus. So up on stage, I've got Terry Rosen, CEO; and Jennifer Jarrett, COO.

Today. So today [indiscernible].

Today, you're the COO. Perfect. So thanks so much. And we're trying to keep this informal as possible.

I think the first question for me for kind of Arcu's upcoming sort of massive data set in the second half. Before that, we make it Roche data as well in small cell lung cancer and 3 other indications. How do you look at those data sets as potential read-throughs to your own data in the second half in TIGIT and potential expansion sets as well?

Jennifer, why don't you talk about that a little.

Yes. So that is definitely the question of the hour for some reason that has become a popular question. So there's a lot coming for us as you pointed out. First of all, yes, we had read a poster last week, and we'll make sure that we touched on that a bit, which was on our HIF2-alpha inhibitor. So that was a nice event to get out there. But in terms of what's coming up, Gilead has their oncology day on April 14. We expect them to talk a lot about it, particularly about our joint development plan, which we've been working diligently on, and it's pretty much wrapped up kind of with the [ go ] on. And so we're excited to be talking more about that. As far as the data presentations that you're referencing, we expect to be presenting both the ARC-7 and ARC-8 data sets in the second half of the year. So both going to be large data sets. ARC-7 will be at least 100 patients, probably more like 110 patients. Medium follow-up will be about 8.5 months. So not quite as long as CITYSCAPE, but we're certainly getting close to where that dataset was from a maturity standpoint. That's also quite a bit longer than the last time we took a look at the data a few months ago, so about 3 or 4 months longer than that from a follow-up perspective. And then on ARC-8, which is our study for quemli CD73 inhibitor in pancreatic cancer that will also be a large data set, that will be about 90 patients. It will include obviously ORR data, but it will also include landmark's 6-month PFS data. So the percent of patients that have not progressed, and that's the 6-month mark, which is a very important mark for pancreatic cancer studies. The average is about 50%, a little bit less than 50%. So obviously, we want to be north of that. We think we can be north of that. You mentioned the competitor data sets, which are coming up. One of them is probably imminent, and that's a small cell lung cancer data set or Sky 2. Roche said that, that is a Q1 event. So we expect that data that's any day. In our mind, it is definitely more of the wildcard study. So we'll see what we see there. It's not a setting that we were planning on going into. We've looked at probably 8 or 9 different settings. It's definitely low on the list from our perspective. We just thought that there were like places with higher probability success. But they've alluded to seeing something in their Phase I basket study in that setting in small cell lung. Merck also just started the Phase III study that's almost identical to the Roche study. When they started that setting, Merck also alluded to the fact that they were seeing something interesting, interesting signal in their TIGIT basket study in that 10-patient population. So it felt like they saw something certainly interesting enough for them to move forward with the Phase III. So the small cell could definitely be a big surprise everywhere, and that's just pure upside. I don't think that's something people have in their models. Certainly, not something that we had been counting on. So like I said, that just represents pure upside for us. Then the non-small cell lung or Sky 1 readout, Roche said will be in the second quarter of the year. We think the presentation would be around ESMO, but they would set up some top line release. That's something we're obviously much more optimistic about just based on what they've presented so far with CITYSCAPE, based on their body language, based on the data we've seen so far in our ARC-7 study in the same setting. So that one, we're certainly very optimistic about.

Got you.

I think the one thing about the small cell trial that might be the most interesting read through, if it's positive, is if you think about anti-TIGIT is enhancing anti-PDX, that would be a great example that would make one feel good about the breadth of the opportunity because it's one of the less PDX responsive things that is PDX responsive. So to Jen's point, we see it as upside. But from a strategic standpoint, when we think about the opportunities where we're going with our anti-TIGIT, we're very unlikely to be going into small cell in any event.

So even if there's positive data, you wouldn't...

Our plan -- our strategy is not to be third to market. And Roche is obviously way ahead of everyone. Merck just started their study. But as we think about like where we want to go with our own TIGIT, our strategy is not to be third or fourth to market. We're really focusing on settings and opportunities where we can be first to market or second or where we have some differentiation. So once we start rolling out our TIGIT plan, that's what people will see based on the settings that we're going after.

Got you. Is there any read-through from small cells such as PD-1 low or...

That we're planning ongoing in the near term. But certainly, there would be a read-through like that there is an opportunity in PD-L1 low. But I'd say as far as our near-term plan, it does not change our view one way the other, whether that study successes or doesn't succeed. And one other point I want to make is going to the small cell, what we might do there. We would not do exactly what Roche is doing and what Merck is doing where PD-1 plus -- I mean PD-1 plus TIGIT plus standard of care chemo. What we could consider though is TIGIT plus PD-1 plus Trodelvy or TIGIT plus PD-1 plus chemo plus etruma or quemli. So something that is differentiated based on novel molecules that are either in our portfolio or Gilead portfolio.

So you have differentiation molecule strength...

I think if you look at our -- the way our TIGIT portfolio program is going to be big and broad in the way that you should think of it playing out in a macro level as there'll be some doublets that we run where we have opportunities, as Jen said, Peter, be first or second where -- we're going in places where, for example, Genentech hasn't even started a study yet or things that where we have a triplet, where we have something that like the third arm of the ARC-7 study, where we have something that Roche, Genentech or Merck doesn't have. So as things broaden out, where are you going to be focusing on opportunities where we can be amongst the first or where we can cannibalize the whole setting because we have something that no one else has.

Got you. And then just thinking about non-small cell, what do you kind of want as a response rate? Is it -- is 50% good enough? Do you need to be higher than that? And then what do you think you're going to see an add on to it?

Yes. I mean it's really the totality of the data. I mean 50% ORR feels like an appropriate benchmark, but we want to see everything. And first of all, everyone is very focused in on this ORR for now, but ORR for micro is not approvable endpoint, PFS, really OS that's going to matter. And so we're obviously -- now that we're going to have PFS data at this next look probably even more focused on PFS than ORR at this point. But obviously, I think the benchmark that you referenced ORR is very, very reasonable. And our discussions with investigators the benchmarks that they think we should hit a 50% plus from the doublet. So I think what you're suggesting makes sense. But we really will be looking at the totality of the data. We don't want to focus just on one metric. We want to look at our ORR. We want to look at PFS. We want to look at the spider plots and see deepening of responses, longer responses. So we'll be looking at all of those things at this next quarter.

It sounded like from the interim reads, the spider plots, we're looking early trends great, looking great or something like that?

Absolutely. And the reason the spider plots tell a lot. So I often talk about the difference between reading data and text and -- versus seeing what those data look like. So you see the shape. You see the depth. You see how within the arm do the individual patients look a lot like the other patients in the arm because oftentimes you read a number and the number has a big standard deviation. So -- especially in these early-stage studies where basically, they're giving you the conviction to go long and hard that basically you learn so much. So you get the depth, you do see the durability and, of course, ORR. And you also -- you see the lack of early progression.

We're going to see them in...

Yes, absolutely. So as Jen said, it's going to be a really robust data set. And in fact, that was one of the things when Gilead opted in, we really want to make a big splash with that it's going to be not only, I think, a big data set for us, but it's going to be a big data set for the field.

Got you. Okay. And then as regards to PFS, what's the bar? Is it kind of like 16 months, 18 months? How should we think about that?

We haven't given out where we want to be. But if you look at what PD-1 does on its own, it's about 7 to 8 months in that setting. So obviously, we want to be materially about that.

Got you. Especially the first TIGIT, was it I believe, like...

16, yes.

Do you want to beat that? Or is that the -- do you have to beat that?

I mean it would be great if we hit 16 months, but what people feel about CITYSCAPE is it was -- first of all, was a subset analysis, was in the ITT patient population. It was only 29 patients per arm in that subset analysis. And they've disclosed nothing as far as like underlying patient population. So it's the only benchmark that we have today on TIGIT. Obviously, we look at it for that reason. But we almost think like Sky 1 in a lot of ways is going to be like a better benchmark because it's going to look more like ours, and ours is it going to be a larger dataset. It's a little bit pure because it's ITT. It's not a subset analysis. We blended ourselves to the data. So 16 months would be great, but we certainly don't think we need to be quite at 16 months to feel like we have a drug and to get the clinical community excited about it because we just know nothing about the underlying growth data.

Adenosine, does -- what does that add? Is it depth or...

Is thus far what we've seen is adding to everything. So to your point about the whole PFS discussion, when we present these data, they're going to be much closer to like what CITYSCAPE looked like first time they presented. So you're going to have about 8 months of follow-up, maybe a little bit more. So we don't know an absolute number yet on what the long-term PFS will be. We'll have PFS. We already even saw it the first -- it the second interim analysis, while we don't overemphasize it, and we shouldn't overemphasize it because it was early, and it's unstable, but we're already seeing separation between the arms. So we think the PFS is going to be very, very meaningful. But at this point, we're still early to be throwing around numbers and cost as much as cross-trial comparisons are hard to do. It's even harder to do when you're thinking for the reason Jen sets 29 patient per arm study. But the -- I think that Skyscraper 1 will start to at least put something out there that it's more meaningful data set that you can look at and say, okay, it starts to look more like truth. 29 patients could be up, could be down, who knows.

And the separation you're seeing is that in all 3 arms, is that still there?

So the thing you -- when we -- we often get questions about what's new. We have not looked at the data since the last interim analysis. So we're -- we -- we're treating this as a blinded study. So we looked at the interim analysis, closed the open back up and now we'll obviously look again before. We make the presentation, but we haven't looked at.

Is it too hard to read into, I don't know, investigator excitement about the...

Sometimes we get asked about that. So in contrast, so like in ARC-8, you have a very concentrated number of investigators and calling up, they want to put more -- the study is fully enrolled. They want to put more patients on. And so you get investigator enthusiasm very palpable. This is a 3-arm study. It's a very large number of sites. So you're not having that same dynamic of really being able to sense from one person who's looking at enough across 3 arms to be telling you what different is this.

Kind of as a side for adenosine, do you think you ever see single-agent activity with adenosine?

So we think you might do an esoteric experiment and be able to show some patients that had a single agent activity. But mechanistically, you simply would not expect. And I think it's going to end up being like anti-TIGIT, where it enhances and there's less likely to be a single agent. But what you want to keep in mind about what adenosine does. These things are not like -- when you think about historical anticancer agents where you're killing a cell and you're killing it a little better, the cancer cells than the normal cell, in this case, you're modulating very well understood human biology. And so when you remove the effects of adenosine, that doesn't drive a T-cell response. It enhances -- if there's a T-cell response, it enhances it. So the type of situation that you can conceptualize and probably does and will exist is a patient who is like right on the inflation like some other -- something else was driving a T-cell response, but that in adenosine was just inhibiting enough so that it didn't happen. So that particular patient could gain benefit. But you really want to think about CD73 inhibition A2 receptor antagonism is combining that with something else whether it's anti-PD-1, whether it's immunogenic chemotherapy that's known to induce a T-cell response. So oftentimes, when you talk about the metaphor where the adenosine is like water, and basically, you're trying to get a fire going, got wet newspaper. Is the wet newspaper sitting in a pool of water? Or is the wet newspaper just a damp newspaper, but you want to first drive that response and then the adenosine modulator pulls away the water, so it enables the fire to better burn.

Perfect. The second half data so we get PFS data systems for the CD73. Is it -- what should we be looking at for like -- what kind of number we want to see or additional metrics that we should be focused on around that data?

Yes. So if you look at how gem-Abraxane has done on its own, even like a more contemporary studies that you look at the halos on Phase III or the ibrutinib Phase III in [ PEG ] and how the gem-Abraxane arms performed. Confirmed ORR and some people confuse unconfirmed and confirmed. The confirmed ORRs are still mostly in that 20s range, kind of getting the upper 20s. Unconfirmed ORR going to be like low 40s, but yes, first confirmed ORR, they still tend to be kind of mid- to upper 20s. From a PFS perspective, those have ranged from 5.5 to 6 months. So we think getting north of that 6 months is important when we've spoken to investigators very recently. They feel like anything worth of 6.5 months in this disease, where there's been no improvement since 2013 would be a win and something that we get them excited. So we obviously want to be north of that 6.5 month mark.

Got you.

Peter, one thing that I think is worth pointing out again that I think if it holds up, will be very exciting data as part of this data set. And it was something we saw early, and we continue to see throughout the expansion phase, and we'll see how it plays out in the randomized phase. But with gem-Abraxane, if you're going to have a response, you generally have it in a first or second assessment. This is very typical of chemo. And then patients tend to rapidly progress. One of the things that we've seen in a fair number of cases is patients who have gone 3, 4, 5, 6 assessments. And they've either had -- you have some patients that are just very prolonged -- long -- prolonged stable disease. But patients who are hovering between, let's say, a 20% and 30% tumor reduction and then all of a sudden have a response, as I said, of 3, 4, 5, even 6 assessments. It's clearly a phenotype of an immunotherapy. So I think that's something that's going to end up being -- if it holds up, one of the most interesting things out of ARC-8 in terms of something that very much differentiates from the effects of chemotherapy. And keep in mind, so we don't know this yet, we'll find it out. But in general, those patients somewhere along the line go off of the chemo. So we're more likely than not have had some patients that may have achieved their response where they may have just been for 100-plus days on the immunotherapy. We've had patients respond out past the year, where they've had their initial response.

That's an interesting narrative around. With CD73 replace adenosine, do you think is that a better molecule or is it just...

So they're both -- we're thrilled that we have both. And I think as this all plays out there'll be a placement in the toolbox for both. And I can give you very specific examples. So -- and everyone should be very clear so there's no revision of history. The reason that we want with our CD73 inhibitor quemli into pancreatic cancer was simply that was the first population. There's a strong rationale for why you would go with CD7 -- it's one of the most high CD73 tumors out there, very KRAS driven. But we just as well could have gone with etruma, but that was our first study looking at that molecule after we did the healthy volunteer study in patients. In certain cases, there's a strong rationale actually for going with etruma. So like in prostate cancer, there's a -- what I would just say is a CD73-independent mechanism for producing adenosine. So there's an enzyme called PAP. That's so prevalent in that setting that it actually used to be used as a marker much as PSA is used today. So in a situation like that, you're going to probably want to brought the actions of adenosine as opposed to just as formation or a combination of both. So we think there'll be a place in role for both of those mechanisms.

Got you. And then I guess we had really much validating data from Astra with the CD73. Is there a benefit of having a small molecule versus antibody?

So we do think that -- so we think the AstraZeneca antibody, oleclumab, is a good molecule. We think it can anything a drug. We definitely see advantages of quemli on that. And there are 2 principal advantages and they're very tangible. So they're not -- it's not make up a story try to figure out something that's there. The first thing is simply that small molecule does a better job of penetrating tumor, does a better job of penetrating target tissue. The second piece is that biologically, the antibody was engineered. So this enzyme has found, at least initially on the cell surface. So the antibody was engineered to cause internalization of that enzyme. But it turns out there's a lot -- it's very easily clipped from the self-service. So there's actually tons of soluble CD73 floating around, which obviously the antibody can't internalize. The antibody only maximally inhibits either the bound or the free enzyme. The enzymatic activity only inhibits maybe 50% to 60%. Our small molecule inhibits fully on both the membrane bound and the soluble. AstraZeneca, they were doing the same studies with their own molecule, they would say the same thing. I'll point out that, that ability to get into the tissue and the tumor is very real because if you think about it in that ability then to inhibit the enzyme is important. Because if you think about what's going on, those tumor cells are decorated with CD73. This is like they're loaded with it. So in that micro environment, it's very likely that you're having a lot of soluble. So you do have that enzymatic activity and the ability to act at the site of action is probably going to be very critical. So we think -- like I said, I think out of the field and when I compare to A2 receptor agonist, where some of those molecules out there are 4 molecules. The AstraZeneca antibody is a good molecule, but we think the small molecule is better.

Perfect. Thank you so much. Thanks for the time, and looking forward to the read-throughs from Roche and your end date during the second half.

Appreciate it. Thanks.

Thanks.