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

Good morning, everyone. Thank you for joining us. Really pleased to have with us the Arcus team. We have Terry Rosen, CEO; and Jennifer Jarrett, COO. I guess to start here, it'd be helpful to get a high-level outlook. You've shared a number of encouraging data sets over the past few quarters, which we'll dive into shortly. But to start, can you give us an overview of how the company is positioned for the second half of this year? And where you're most focused on in the near term and when we might start seeing registrational data from your pipeline portfolio?

Sure. So thanks, Salveen, thanks anybody who's listening in. Obviously -- and now I'm going to go quickly so we can leave a lot of time for Q&A. We just had the two oral presentations at ASCO just within the last week. And I think those are really important for us and the field. I'll tick through those quickly at a high level. We had the EDGE-Gastric data, I'll remind you that's upper GI cancers. It's frontline. And what we showed there, the new data was median PFS on the order of 13 months. That's a median PFS that hasn't been seen anything like it in that field. There's been 3 registrational studies done. The most notable is CheckMate 649, that's a standard of care nivo chemo, but there's been KEYTRUDA chemo, Tisle chemo, they all come in about 7 to 8 months. So that 13-month PFS that we showed is actually in line with what's known to be OS in that field. The other thing I think that was important data there was that the molecule performed similarly in both the high PD-L1 and low PD-L1 population. I'll remind you, in our study, we actually had a bit of an arm tied behind our back compared to CheckMate 649. So we were 40-60 high PD-L1 to low, CheckMate 649 was 60-40 high. So what's to your point, exciting about that study, more than just the data per se is that on Monday, we just announced that the corollary Phase III study to that STAR-221 over 1,000 patients is fully enrolled. So that really starts the clock ticking towards registration. Hopefully, launching commercialization years ahead of anybody. So we're going to have a great first mover opportunity there. The second study that we reported on, really pretty exciting data, especially everyone talks about patients, there's about patients. This was in third-line colorectal cancer one of the most dismal settings out there. This is with etruma and some standard chemo regimen in colorectal cancer. And in that third-line setting, we saw in a very rigorous study, a study arm of 70 patients, control of 35 and OS that's never been seen with any combination in this -- any regimen in this setting of around 20 months. In fact, when you break out the data, we saw that even in liver met patients, and by the way, some studies even exclude liver met patients, which is kind of ridiculous given that they're about 70% of that population, we saw that same benefit of around 20 months. So that's another program we're quite excited about the data, and we'll be moving forward there. The final catalyst that I'll talk about is one that I think there's a lot of excitement in the world. And probably that excitement is most generated by call it, technical validation proof of concept. So that's our HIF-2 inhibitor program, molecule called casdatifan. We would say casdatifan is drug waiting to happen. Merck's belzutifan has validated the target. We clearly, demonstrably and in the clinic have shown that we have a better molecule. The point now will be to show that, that actually manifests itself clinically. We have a 100-milligram cohort that's fully enrolled in clear cell RCC. We'll be sharing data from that in a medical conference later this year. We've also fully enrolled the 50-milligram cohort. We did a 150-milligram cohort in dose escalation. There's so much excitement in the investigator community, an expansion cohort with that is almost fully enrolled. That will fill out our Project Optimus data set. So we'll have a steady flow starting later this year. With those 90 patients, we've also started a cohort combining with cabo, which will ultimately support where we're going with these -- this program, which will be in combination therapies. So lots of opportunity to differentiate from Merck, and we feel like we have a molecule that is already showing those signs. So I'll stop there and jump into your questions.

Great. Let's start with the TIGIT update at ASCO in gastric cancer. Walk us through any feedback you received from the physicians at the event and also your confidence that you can repeat this in the Phase III study STAR-221.

Yes. So the feedback could not have been more positive. So at ASCO, we met with well over 100 KOLs to talk about gastric CRC as well as our RCC program, but probably at least 1/3 of those were gastric. And everyone was extremely impressed by the data and said that it's unlike data that they've ever seen in gastric. So as a reminder, Terry was saying, we saw about 13 months PFS in the overall patient population almost 14 months, in the PD-L1 high patient population that compares to 8 months for the benchmarks. So getting to your second question, we're well above anything that's been seen from a historical benchmark perspective. So even though you always assume there may be some degradation as you go from Phase II to Phase III. We obviously have a huge amount of cushion there. The other thing that you might ask is, well, OS is your endpoint for STAR-221, your Phase III study, you just reported PFS data. PFS and OS tend to be very, very highly correlated. If you look at CheckMate 649, which is the Phase III study that Bristol conducted in gastric, the hazard ratio for PFS and OS was almost identical. So those two things seem to be very, very correlated. Also, the OS for the benchmark studies was 13 months, which is what our PFS was. So obviously, we're going to see an OS that's well beyond 13 months in our Phase II study, so well beyond what you would expect with comparator drugs for the primary endpoint.

And remind us as to the size of the addressable population here.

So it's north of 40,000, it's a huge patient population. Some people think of gastric as being very Asia-centric, there is a large Asia patient population. But for the adenocarcinoma market, so gastric, GEJ and esophageal adenocarcinoma, there's also a lot of patients in the U.S. and western world. It's actually, I think, the fastest-growing tumor type in the U.S. and western world right now related to diet and other things. So it is a very large market. We think it's $3 billion plus overall. It's probably $2 billion plus just in that PD-L1 high segment. And we'll see what we see with our PFS and how long patients stay on treatment. But obviously, if we're keeping patients on treatment for 12 to 13 months, it's going to be a much bigger market opportunity than $3 billion.

And when could we see the Phase III data set here?

So we haven't said -- if you look at our investor materials and Gilead that we've had out there for a while, very consistently, we've shown 2025 plus, which means we can have the data as early as 2025. We did just announce completion of enrollment. We know OS at least for nivo chemo is 13 months. We obviously want to do a bit better than that with our regimen. But you can kind of do the math and start to get a sense for when we could see our data. But obviously, as Terry was saying, it's now around the corner.

Is there a mechanistic reason that the benefit that we've seen here might be more meaningful than what we've seen in lung? Maybe help put that in context for us?

Yes. So I would just say, actually, you might expect by the time these trials done similar types of benefit. And I like the way that you asked the question because we tend to look at whatever we're doing is you're treating biology, not an organ. And the biology that you're treating here is the same, and these are both high CD155 tumors, particularly when you bring on the chemo. So that brings in the PD-1 component, but they are high CD155. And so we think -- we're extraordinarily excited about our PD-L1 all-comer non-small cell lung trial which really is going at the largest patient population, 100,000 patients. And that study is going to be fully enrolled this year as well. So we think mechanistically, the two go hand-in-hand, both in terms of the role of PD-1, the role of chemo in driving that to lower levels and that their high CD155 tumors, which is -- what drives the TIGIT activity.

And at ASCO, you and Gilead also presented interesting data from etruma in colorectal cancer. Could you walk us through the highlights from this data set and what next steps are for that program?

Yes. So the headlines of the data set was median OS of 20 months, which is unlike anything that's ever been reported. Probably in second-line colorectal but certainly in third-line colorectal and that's something going back to what we were hearing from investigators and KOLs that they very consistently said that 20 months is well, well above anything you would expect in the third-line setting. The other thing that was very interesting about the data is that we saw just under 20 months in patients with liver mets, usually that's the patient population with a very poor prognosis that tends to progress much more quickly. Actually, the trend has been from some other studies that are looking at third-line colorectal to actually exclude those patients even though it's about 70% of the patient population. So I think that was something else that KOL sounds very striking about the data.

And you noted the potential for a second signal-seeking Phase II trial where you could determine which drugs within the combo are driving most of the activity. How necessary is this and then speak to which drug you would potentially drop from the regimen and when you'd expect to move into a registrational trial?

Yes. So we're working through those exact questions, and we're going to do it expeditiously with both Gilead and the FDA. But to give you some insight, the molecule that we have question marks about is whether we really need the anti-PD-1. So I'll remind you what you've got is sort of the standard there is the chemo bev, so the FOLFOX bev, and then we've added etruma anti-PD-1. There are very rational designs that could either be registrational Phase II or where you might do some component. We have an ongoing platform study that was the ARC-9 nomenclature itself. We'll know very shortly, again, through conversations with Gilead as well as the FDA. I'll note that, again, coming back to the concept that we're treating in biology, not an organ. This is not unlike the pancreatic data set where you're going on top. By the way, we think this is a very favorable aspect of just the general strategy that you're enhancing a standard of care that physicians are used to as opposed to displacing something in the field, so the physicians are comfortable with the other regimen. But you see the same sort of thing where we actually did in the Gem/Abraxane pancreatic setting, we did a randomized trial that looked at the effect of anti-PD-1. And in fact, we saw that the anti-PD-1 brought nothing. The other thing that is known from this field is that atezo on top of the FOLFOX bev does not bring any advantage. So we haven't made a final decision on that or whether we might generate some additional data before we went forward. But the biggest question mark for us is whether we need the anti-PD-1 or not based upon all the totality of the data that we now have that we didn't have when we started the study.

And could you also speak to mechanistic reasons as to why we are seeing a benefit targeting the adenosine pathway here in colorectal cancer versus the prior setting, I think in prostate cancer where it did not add as much therapeutic value?

Yes. Absolutely. So it comes back to that same concept where our original hypothesis as to where we thought the adenosine pathway would have its greatest utility was in the context of immunogenic chemotherapy. So by definition, briefly, immunogenic chemotherapy produces a ton of adenosine and that adenosine inhibits the action of -- if you can generate a T cell response, those T cells are inhibited by the adenosine. So to get the benefit of that, the hypothesis was remove that adenosine fog and let that immunogenic chemotherapy play out. Now if you think about we've actually generated 3 data sets in like the last 6 months that have shown that. Our ARC-8 pancreatic data set, again, profound OS. The Genentech MORPHEUS pancreatic data set also with etruma profound OS, 16-plus months OS versus around 10 months for Gem/Abraxane and now this molecule. If you actually go back and look at our prostate data was interesting as we did -- we had multiple arms in there. They were relatively small. It was the early days. And the ones that did not include chemo we basically saw no signal. Interestingly enough, we -- there was 1 arm in later line with docetaxel. And if you look at those data, there actually is somewhat of a signal there. It would be something that you may or may not convince yourself as 17 patients on -- whether you might go forward. But if you were to look, that was -- we saw PSA decreases. We didn't have bone measurements on all of them. We didn't have tumor reduction on all of them. So I would say the key thing to keep in mind in terms of mechanistic differences. And if you look not only at our studies, but across the field, where you will see some [indiscernible] that we think is where you have that immunogenic chemotherapy. I'll remind you of a data set that, again, it's under the radar for most people, it's not under the radar for AstraZeneca is their cost data set. And they just presented updated data at ASCO continues to look quite good. And I'll remind you what that is, is that's in Stage 3 lung where you're treating with CRT and then you're treating in their case with their CD73 antibody and their anti-PD-1 durva. So they're actually running a registrational trial there. It's the same sort of thing. You've got the CRT, the chemo. The other thing I'll point out is that these are also in colorectal and pancreatic are poster trials for being high CD73 settings as well. So you are definitely getting those high production of adenosine in the combination with the immunogenic chemotherapy.

Shifting back to TIGIT before we move to the rest of the pipeline. Now we've seen a number of data sets in addition to yours from Roche and Merck. Can you share your thoughts on the potential for targeting this pathway and where you're most confident and where you see the most risk?

Yes. So we feel like lung and gastric. We've probably seen the most supportive data. So we have a lot of confidence in our lung and gastric studies. And just as a reminder, SKY-1 data was obviously inadvertently leaked. The takeaway from that data set was that we'll probably end up with a hazard ratio from that study somewhere between 0.76 and 0.81. So depending on their stock plan, it will be a successful study. If they don't have their stocks exactly right, they may just miss. But that, in addition to all the data we generated internally, both from ARC-7. And then as a reminder, we had the ARC-10 study, which was a Phase III study of dom plus zim versus chemo. We shut down the study for a whole bunch of strategic reasons, but we obviously have the data from that, that is also supporting what we're doing in lung. It gives us more confidence around our Phase III studies in line of which there's now 2. On the gastric side, Roche has also reported some really supportive data sets, both at esophageal adenocarcinoma as well as squamous cell most recently at the end of last year, I think it was at ESMO-IO or ESMO, but they've reported data sets in both those esophageal tumor types that were both very, very encouraging. And then we have our own data set, EDGE-Gastric -- on PFS, that also gives us a lot of confidence in what we're doing in gastric. So we think we've picked the right tumor types, the right settings, the right combinations. The other thing that I think is really unique about our program versus others as we have the only Fc-silent TIGIT antibody. So if you look at some of the data sets from the Fc-enabled TIGIT antibodies, there's definitely signs of more toxicity, particularly with the BeiGene and the Merck data sets. And I think most interestingly, Merck just reported that they did not meet an interim analysis for the study they were running in melanoma where they were looking at their co-form of their TIGIT antibody plus PD-1 versus KEYTRUDA and they said very explicitly in the release, but the reason they were unlikely to hit the endpoint was because so many patients had to discontinue due to immune-related AEs. So we think the fact that we have a more mild or could have a more mild immune AE profile and we'll be able to keep drug on patients longer will be a huge advantage for us and should result in better data sets.

I think that -- I mean just to call it out, I think that Merck press release read exactly what we've been saying for some time, including the fact that with the co-form. So not only do you see the AEs, but then you have to take the patient off of both drugs. So it's not only you're having an AE difference, but it affects efficacy. The other tagline, I think this is implicit in what Jen said the settings that really you want to go, and this is -- I won't go through the biology, but anti-PD-1 and anti-TIGIT are not orthogonal mechanisms they're related to each other. So a lot of things in anticancer combination therapy, you're combining two things and almost "hoping". In this case, the anti-TIGIT activity is facilitating the intrinsic activity that you can get out of anti-PD-1. So the settings that will make the most sense are settings where anti-PD-1 actually works. So a lot of the studies where people might have just done things, they say, hey, there's high medical need, it's strategic. It makes sense in a PD-1 refractory tumor are pretty much destined for failure. So we think we've picked strategically not only the right sort of market opportunities and need opportunities but also scientifically and feasibility doesn't work. It doesn't matter how big of an opportunity is.

And one other data point is AstraZeneca had an investor presentation or meeting, I think it was either right before ASCO or at ASCO, one of the programs that they highlighted was their PD-1/TIGIT bi-specific and they called out lung and gastric is the two areas that they were likely to focus on. So we still, again, that reinforces our strategy with our [ own TIGIT ].

You clearly have a lot of enthusiasm for the HIF-2 alpha program. Maybe help us understand the confidence here as you think about not just having that -- hitting that clinical benefit, but also in the context of competitive dynamics. And frame the upcoming dose expansion data that's set to be released in the second half, I believe.

Yes. So we could not be more excited about the casdatifan program, especially because it's progressing so quickly. And the data that we're seeing is everything that we'd hope to be seeing. So as a reminder, our 100 mg dose, which right now is our going forward dose, we're getting 5x the PD equivalent that you see with the belzutifan 120 mg dose, which is the dose that they're using belzutifan as Merck's drug. They're currently on a $500 million runway, just about 3 months or 4 months after they got their drug approved in RCC. So I think that really highlights the market opportunity and the unmet need in RCC. From a competitive landscape point, we really think it's just us and Merck. We are probably a few years behind, but we're doing everything that we can to catch up. And at our Q2 earnings call, we'll shed a lot more light on what our Phase III plans are, and we'll get to the specifics of our first Phase III study where we'll be combining our HIF-2 inhibitor with a TKI. So we'll disclose what that TKI is and what exactly we're planning to do. And then probably talk about another Phase III study later in the year, early next year. As far as the data sets that are coming in the fall, we will have our first data set from our first expansion cohort, so that's a 30-patient expansion cohort from the 100 mg dose, so that will be at a medical conference in the fall. That will be mature ORR you'll see waterfall, spider plot, et cetera. And we're right on the cusp of having that 50 mg expansion cohort data set as well. So we may be able to talk a little bit about that. We'll certainly know what the progressive disease rate from that toward, as we know what that is today, it looks very, very good. And -- but we may have a mature ORR number for that as well. And then the 150 meg extension cohort. So this is another 30 patients on top of the 60 from the 100 mg and 50 mg cohorts. We'll have that data set probably early next year at this point. So we're just about to complete enrollment of that cohort. Originally, we thought we had data probably mid next year that could actually move up a little bit, just given how quickly that cohort enroll. So it's going to be nice study from the data. It will be about 90 patients of expansion cohort data, all patients getting treatment in second-line plus settings. So it's going to be exciting. Very exciting time for the program.

You're clearly not very busy.

No. No. And I'll tell you the other thing about that is from the competition standpoint, we clearly have a better molecule. We think that is demonstrating itself and manifesting itself clinically, not just from the pharmacokinetic dynamic and pharmacokinetic advantages but we do feel also that our combination that will not involve lenvatinib and that's where this field is really going to be expanding and getting into the full population that's certainly going to be another advantage that we're going to have relative to Merck.

And should we expect Gilead to opt-in post this? And what would those financial implications be?

You can ask them later today, but I think they're excited about the program, and I think they would say that they do have a GU presence because of Trodelvy, so RCC that might -- it's very adjacent to GU. So we think they'll be interested in the program. We're seeing strong -- very strong ORR as a single agent, so like everything you want to see now today with an oncology drug. If they do opt-in and then it would be $150 million upfront and then cost sharing, so it would be nice to get that capital. And I think if cas, just talked about some of the investors, cas was a stand-alone company today, and we're just about to report 29 -- I'm sorry, 90 patients worth of data in a single-agent line settings and you'll have, I think, a pretty interesting ORR. I think that's the company that would work through the $3 billion. So we think the $150 million opt-in will be a pretty good deal for them.

[ Quemli ] targeting CD73, you're advancing that next year into pancreatic cancer in a Phase III study. What gave you the confidence to move this program forward and help us understand why Gilead will not be joining you?

Sure. So first off, the confidence primarily come from the data themselves. So this is like just under 16 months OS. When you have another study that's similar but different, the MORPHEUS pancreatic data set, that's also confidence enhancing. They showed over 16 months OS against the Gem/Abraxane is that -- and coming back to what I said these 3 studies reinforce that, their biology makes sense. I think the misconception about Gilead is they still have the option to that. The -- when we did the $300 million equity purchase, that was -- keep in mind, to fund as part of what that did, it was to partially to fund this study. So Gilead is involved. There's no interaction we have with the FDA where they're not involved. There's nothing about the program. So I think the -- and they are opted into the program. So I think there may be a misconception as to we're executing it. The way it was funded came through the funding, but they definitely have retained their interest in this.

While you've recently had several positive early-stage readouts, investors continue to view the profile that it's high risk, high reward. What do you view as the most misunderstood aspects of your portfolio and the key debates here?

Sure. So we think the portfolio is really very highly derisked, and I'll just start with -- if you run through what we talked about today, the HIF-2 program, there's probably no program on this planet, we're being second, but having a molecule that's demonstrably better on a validated mechanism, that's a huge opportunity. So to the extent that someone doesn't see that is about as derisked as you can get in an early-stage company, and we're -- we have the resources to run a Phase III program there. I tick that off first. The anti-TIGIT field for the reason Jen said, I think is extraordinarily derisked. There's multiple data sets, the data sets make sense. As we all know, there's been there's -- all sort of investor risk evaporates that somewhat has stemmed from all of the saga, if you will, of SKY-1, and what's -- how is that going to affect investor sentiment. So I view that as a bigger investor sentiment risk than where we come to work every day, which is patient-centric and investigator centric. And obviously, we care about our investors. But from a long-term standpoint, I think that's going to turn out to be great for investors. And when they look back at time, the positive data, derisking it is staring them in the face, and it's from multiple companies with multiple different molecules. And other companies talking about why they're going to be best of class, but you still have some concern otherwise. And then I think the -- on the adenosine front, we think we've generated quite compelling data that's derisking, but we recognize there's still left from what we would say we're obviously inferior molecules. And I'll show you a really quick -- experiment to do. So Novartis just shared data from their HIF-2 inhibitor, which, frankly, are not good data. And frankly, they don't have a good molecule and their properties or the molecular out there. What I would tell you is if -- those data were the first HIF-2 data ever to be out there, and you didn't have belzutifan, and you didn't have our molecule. From an investor sentiment standpoint, I think there'd be a lot of investors that would say HIF-2 is [indiscernible]. And then they have a lot of skepticism about the next HIF-2 data set. That's a thought exercise. I think it's actually very true, but it gets to a true technical derisking versus short-cycle time investor mentality, which is part of the ecosystem, and it should be there or shouldn't be there. It's just -- it is what it is. I think from a long-term perspective, our later-stage programs are really, really quite well derisked.

Great. So closing out, could you briefly walk us through the next key data sets for Arcus and any that you're monitoring as well from a competitive read-through standpoint?

Yes. So the second half of the year for us will largely be around cas, the HIF-2 alpha inhibitor with that first data set coming in the fall, the 30 patients worth of data, maybe even more if we're able to put the [indiscernible] in there. Also [indiscernible] the OS data from EDGE-Gastric probably at some point early next year. So that will hopefully be more validating data on TIGIT. We can also start to see some competitive data sets from the other anti-TIGIT antibody companies. Obviously, Sky-1 will be in the second half of the year. Iteos has a data set coming, I think maybe at ESMO for their anti-TIGIT antibody. We're also starting to get close probably to Merck's first readout in lung. So that's all-comer study where they're combining their anti-TIGIT antibody, KEYTRUDA plus chemotherapy. There's a lot coming in the TIGIT space. The other data set is a little bit more of the [indiscernible] is our ARC-10 data set, which I alluded to a [indiscernible] so that was the Phase III study that we're doing that we discontinued, but we've about 100 patients worth of data there, and we'll probably get that data set out at some point this year as well.

Great. Well, with that, Terry and Jen, thank you so much.

Thanks.

Thank you.