My name is Peter Lawson. I'm one of the biotech analysts at Barclays. I've had the pleasure for the last, I guess, five years or so covering Arcus Biosciences. Up on stage with me, I've got Terry Rosen, CEO, and Jennifer Jarrett, the Chief Operating Officer. I'll be moderating the Q&A. I've been asking our companies, kind of just on the macro level, any impact you were worried about, anticipating from tariffs on the supply chain. I guess, drawing back to COVID, I guess that helped insulate the company as well.
Yeah, no. So far, keeping a close eye on everything. By the day, it's sort of a moving target. Right now, we're not anticipating anything major that would be an impact from the tariffs.
Gotcha. FDA interactions, have there been any changes?
Not at all, zero. Maybe they're letting people go in other areas or withholding funding in other areas. We haven't seen, we're having a lot of ongoing interactions with the FDA as we speak and have seen zero impact.
No kind of incoming resumes from the FDA or NIH?
Yeah, I mean, that's our regulatory team. No, not that I've seen, yeah.
OK. NIH spend, I know that's kind of essentially a very large negative if you think about it longer term. Anything near term that could trickle through to clinical trials?
No, I've never seen. Yeah, I mean, we're not working with NIH or whatever. Everything that we're doing is with institutional sites. We expect zero impact from anything that might be happening in Washington.
Honestly, our trial enrollments across the board are going quite well. Nothing we anticipate.
OK, perfect. I guess, if we think about your pipeline, HIF-2α, I guess, combination data in the middle of this year, what should we expect from that? How much follow-up data should we see?
Yeah, so what's nice about how the ARC-20 study has been designed is there's a lot of different cohorts. We're looking at a lot of different things within that study. As a result of that, there should be a very steady flow of data really over the next two years. The next data set that we've been talking about presenting will be around mid-year. That'll be the first look at the cohort that's evaluating casdatifan plus cabozantinib. It'll probably include efficacy data on around 25 patients. The safety database will probably be a little bit bigger just because I think on the efficacy side, we'll probably focus on patients that have had a certain number of scans. We want to give people as broad a look at the safety as possible because we know that that's important to people.
So far, the safety is looking really good. Later in the year, we expect to present more mature data from the monotherapy cohorts that were the original cohorts that were enrolled as part of ARC-20. As we go into next year, we should have more mature data from the Cast+Cabo cohort. Maybe even later this year, we'd have some more mature data from that cohort. We have added some new cohorts to ARC-20. Next year, we would actually start to generate some data from those other cohorts as well. Back to this year, as I said, the two data sets to focus on will be the one around mid-year for Cast+Cabo and then later in the year for Cas-Mono, where we'll have more mature data. Maybe we'll have PFS from the 50 mg QD cohort at that time.
Good. That is the key thing you think to focus on for the combo data—is we'll have PFS or?
More mature ORR. As we've seen, ORR can increase over time. Responses can happen late with this mechanism. It'll be more mature ORR. PFS for the cohorts that we do not have mature PFS for at the ASCO GU presentation.
Gotcha. OK, thank you. And then.
To the extent we've hit a median, we still have not hit a median for the 50 mg QD or for the 100 mg QD.
Right, perfect. That kind of brings me on to the next question just about the, I guess, the current casdatifan level of, like, what, 100 milligram dose in the combination. Is there any need or desire to push it higher or bring it down? Or kind of how should we think about that dose going forward?
Yeah, we're very, very comfortable with the 100 mg QD dose. We've talked about that with the FDA. We had our meeting with them last year. They were comfortable with that dose. We're starting to see some data from the 150 mg monotherapy cohort. There's nothing in that data set that would tell us that we should be going with a higher dose. As we've been saying, we think 50 mg QD is getting towards the top of the dose response curve and the flat part of the dose response curve. We probably are squeezing a bit more efficacy out of the 100 mg QD dose. We don't think we're going to get a whole lot going to 150 mg.
Gotcha. OK. The combo versus the mono, what does that do? What's the benefit? Is it response rates? Is it duration?
Yeah, it's going to be all of those things. I mean, duration of response is something, the duration of response for this mechanism is so long that that's something I don't think we're going to know for a while. We do expect to see an improvement in duration of response. As we said, just going back to the monotherapy expansion cohorts, when we presented that data, we had 26 responders across the three cohorts. Only two of those responders had progressed. Duration of response, both for Cas-Mono and Cast+Cabo, is going to be incredibly long. We do expect to show improvement on the other measurements. You mentioned ORR and PFS. ORR would obviously be the most near-term one that we'd be able to demonstrate. PFS will take a while to mature. We do expect to show improvement on both those metrics and then ultimately overall survival.
Gotcha. OK. Do you have a goal in mind for the improvement around ORR or DCR or prevention of?
Yeah, I think, I mean, it's always tough to quantify improvement. I think what we've been saying very consistently is that Cast+Cabo ORR will be higher than what we've seen with Cas-Mono as well as what's been seen for Cabo-Mono. There's a few different studies out there. That is our goal, to show that we can get to a higher response rate versus both agents as monotherapy. I think when you see the waterfall chart and other graphs, I think it will show clearly that there's something more happening there than you would expect with either mechanism alone.
OK. At the moment, you're not kind of guiding to a success rate for the ORR?
No. I think the way you should think about our Cast+Cabo data, it's going to be early. The most important thing is that you will look at those data and be able to look and say, like, they're going to be running a trial of Cast+Cabo versus Cabo. It's going to look like you will believe that their trial is going to be successful and it's going to be clinically meaningful difference. I think the data will already be telling that story.
OK, gotcha. Thank you. Have you seen and will we see, I guess, patients that have had prior HIF-2α?
No, that's easy. We're not looking at patients that have had prior HIF-2α inhibitor. As a matter of fact, the patient population really doesn't exist. You're not seeing patients getting the approval for belzutifan is in late-line patients. Obviously, you won't be having patients coming into this that have had HIF-2α inhibitor. It's not even sort of a feasible thing at this point.
Yeah. Would you assume they're too similar to if you're not responding to HIF-2α, it would just be?
In general, we would feel that way. However, if you think hard about it, given now I'm talking more downstream, you could imagine, since we're seeing that dramatic difference in the rate of primary progression, we see about half the rate versus belzutifan. What that tells you, if that's true, that means that those 50% that are progressing on belzutifan that might not have otherwise, were not intrinsically resistant to the HIF-2α mechanism. You could imagine at some point a study where you would look at, for example, those early progressors and that that would not have represented some sort of resistance generation. You could imagine those patients going on to casdatifan after belzutifan. There might be some benefit there. That is not in our near-term plans.
Gotcha. OK, perfect. Thank you. The dates are in the fall, the additional monotherapy. I think we touched upon it. Kind of what should we expect to be seeing from that? I know there's a couple of doses that we should be thinking about. Does that data in any way change the registration strategy?
No. You mean anything that we see in the monotherapy or casdatifan-cabozantinib?
Yeah.
No. I mean, we're obviously looking at the data on a regular basis. I mean, there's nothing that we've seen that would change our registrational strategy. As a reminder, our first registrational study is a study called PEAK-1, where we'll be looking at casdatifan plus cabozantinib versus cabo. That's why that Cas+Cabo cohort is so important. As Terry was saying, first, we obviously want to look at safety and make sure that we can combine those two mechanisms safely, which so far, based on everything we're seeing, we can. Obviously, you want to start to look at efficacy as the data set matures. We're super excited about the study. It's a very simple study, about as simple as you get. 700 patients, PFS endpoint. We really set up the study so that it will enroll well and enroll quickly.
As an example, we're using two-to-one randomization. So patients have twice the likelihood of getting in the experimental arm versus the control arm, which is something both patients and clinicians like. We don't see a lot of competition out there, direct competition for patients in the study. That's something that our investigators have been pointing to as well and why they're really excited about the study and why they think the study is going to enroll quickly. Obviously, the fact that it's a PFS endpoint should allow us versus overall survival will allow us to get to an answer pretty quickly as well. So far, I'm not seeing anything in the data that would make us change our plans. We're as excited as ever about getting the study started and expect it started by mid-year.
To your specific question about the data sets, the monotherapy data sets that we'll show later in the year, I think they will have, as Jen said, zero impact on registrational strategy. I think the only thing they'll do is reinforce confidence that casdatifan looks both better than belzutifan and looks better than TKI monotherapy. Because what you'll see is more mature data from those earlier cohorts. You'll see deepening of responses. You'll see actual, where we had PFS that was immature or we couldn't even state a median PFS, PFS that's likely to be double or more than that of what's been reported for belzutifan. It'll just be confidence enhancing, probably more to the external world than to ourselves or investigators. Nothing that will change any of our strategy.
Gotcha. Thank you. And then on the safety, what should we be looking for? What's the worry? I mean, I guess KABO has been a fantastic drug. But it's sometimes difficult to take just.
Yeah, I think there's actually, the data will end up showing what was anticipated. You need to play the game to know the outcome. There are orthogonal mechanisms, substantial AEs that you see from casdatifan, like belzutifan, are those that are on target, anemia, and more rarely hypoxia. When you combine the two mechanisms, there doesn't seem to be anything that ends up being some sort of synergistic tox. As a matter of fact, we've been looking at the clinically used dose of cabo in combination with the 100 milligram dose of casdatifan. That seems to be quite good.
Gotcha. OK, perfect. Thank you. Just onto PEAK-1, you've touched upon it. Just how should we think about interim data and kind of if it's built in, what triggers it, and what should we expect to be seeing for that interim?
Yeah, we haven't really talked about the statistical analysis plan and whether we'll look. Yeah, for competitive reasons and all sorts of reasons, we probably won't. Like I said, I think the important point is it is a PFS primary endpoint. All of the alpha is on PFS. If you look at what Cabo alone has done on PFS, it's been about 10 months. You've done a good job, I think, of pulling together that data. Actually, the range has probably been about 7 and 1/2 months to 10 months. Meteor was about 7 and 1/2 months. CONTACT-03 was 10 months. That's the range that we'd expect for the control arm. Not a huge amount of time. That, again, is why we think we can get to data relatively quickly.
Gotcha. OK, perfect. What else? Oh, I guess the anti-CTLA -4 PD-1 combination as well. Kind of when do we see that data? I guess what should we be looking for? It's an interesting combination.
Yeah, so the study that you're referring to is part of the clinical collaboration between Arcus and AstraZeneca, where we're combining quemliclustat with their PD-1 CTLA-4 bispecific. They presented data on their molecule about a year and a half ago now at ESMO and RCC, which looked very, very interesting. We work with their same investigators. Investigators are very excited about that molecule. We are combining volrustomig with quemliclustat as part of a study that they will be operationalizing. We haven't talked about the economics of this study. Obviously, there's some cost sharing. It's a really efficient way for us to be exploring that combination. That combination will be targeting patients that are IO naive, which are essentially going to be first-line patients. That study will start right around the same time as PEAK-1.
The focus for this study will really be on safety, just again, similar to what we're doing with Cas+Cabo right now, just to make sure that those two mechanisms can be safely combined. We don't expect there to be any overlapping toxicities, again, like Cas+Cabo. We expect very much that those mechanisms will be safely combined. I think we'll talk more in the future about what we want to do next. I think around mid-year, we'll be able to say more about the study for competitive reasons. We've been pretty quiet, as has AZ, about what we're planning. Like I said, when we're getting the study off the ground, I think we'll be able to say a lot more exactly what our strategy is with that program and in that setting.
Gotcha. OK. Being several years behind Merck, what's the go-to market strategy? Is it like having the right combinations? Is it just having a better molecule?
It's all of that. I think early on, we would tell people, including you, even if our molecule looks just like belzutifan's, we think we have a differentiated development strategy and differentiated combination partners. We still very much believe that. I think the thing that has changed as we generated more and more data from ARC-20 is we are very, very confident now, I think, the investigator community, that we have a better HIF-2α inhibitor. I think Terry touched upon the efficacy data that we've seen. We're seeing half the rate of primary progressive disease. They've seen ORRs kind of ranging from 18%-21%. We're now clearly north of 30%. Our PFS is much longer than what's been seen with belzutifan. We have a better backbone HIF-2α inhibitor molecule.
We are also evaluating combinations that we think are going to be better than what Merck is doing. As an example, they are combining belzutifan with lenvatinib, which is their TKI. The challenge with lenvatinib is that it's a more toxic TKI. I think it's going to be more difficult to combine with relative to cabozantinib, which we're combining with. The other challenge with lenvatinib is that it's a lot more difficult, complicated to dose. It's available in a lot of different dosing regimens. When you talk to investigators about their experience with lenvatinib, they would just tell you that just the titration up and down is a lot more complicated than with cabozantinib, where cabozantinib is available in three dosage forms, 20, 40, 60. That's it. We think we have a better TKI combo study.
The other challenge with their study is they are combining with lenvatinib, but they're using cabozantinib in the control arm. If lenvatinib underperforms cabozantinib, that could create some issues for them in that study. Just the last point that I'd mention, our first-line strategies are just completely different. They are combining with PD-1 and TKI. We're combining with PD-1 and CTLA-4. We're just going after different segments of the market. We're competing, but with different combinations in the first-line setting. Again, because ours is TKI-free, we think we have a better combination.
I think the way you should think about that, because we get asked that question a lot, and I think it's not a situation based in actual reality. Jen talked about the difference in strategy. The reality is Merck is ahead on one study, Lenvo. That study may or may not even be successful. It's obviously the readout's been pushed out since it was initially reported. We get asked sometimes, oh, but they're combining with Zanza. How is that going to be? That'll actually end up being behind us. What's clearly emerging is a better molecule. Not only are our strategies differentiated, Jen said, they're not even going right now with the TKI sparing regimen, which we see is the direction that the field is going to go.
As I was mentioning before, we see that casdatifan is looking even better than TKI monotherapy. We see a paradigm shift where HIF-2α will be going earlier into the treatment paradigm, with potentially the TKI being put off till later so patients can avoid the TKI. We do not even really see ourselves as substantially behind Merck in general, other than in that single study.
Gotcha. OK, perfect. Thank you. And then TIGIT, so the gastric data and kind of expectations, or can you help us set expectations around the OS for the fall of this year?
Sure. Let me remind everybody, we're looking at our registrational studies called STAR-221. It's in upper GI cancers. It's a very straightforward trial. It's domvanalimab chemo versus nivolumab chemo. N ivolumab chemo is the standard of care. We reported a median PFS some time ago of around 12-13 months. That's actually the equivalent of the OS for the standard of care. We'll have data later in fall where we will share the OS. We have said publicly that with 18 months of follow-up, we had well over 50% of the patients still on study. We think that's going to look quite positive for the standard of care there. In addition to nivolumab chemo, there's also Keytruda chemo is approved and tislelizumab chemo approved, very similar 13-ish month OS. We think we have opportunity to very substantially beat that in a very clinically meaningful way.
More importantly, the phase III study that's correlative to that, STAR-221, was fully enrolled as of June of last year with 1,047 patients. It is very well-populated. You can do the math. By the end of this year, it will be 18 months since the last patient was in. With a standard of care of 12 or so months, you can get a sense as to that readout. We've been guiding to a 2026 readout. We'll probably give tighter guidance as the year goes along.
Gotcha. That trial over-enrolled. Does that kind of help with the stats?
Yeah, it did over-enroll. It was one of those that was enrolling, I think, like 100 patients, maybe more than that.
18 months.
A month towards the end. It was almost impossible to not over-enroll because you can't just all of a sudden stop and all the slots go away. Yeah, it did over-enroll, which is obviously a good thing because it gives us a bit more power as well in the study. The over-enrollment was a function of just how quickly it was enrolling at the end.
How are you thinking about kind of the probability of success in intent to treat versus the PD-L1 high population?
Yeah, again, without going to the stats plan, it's probably a little bit difficult to give you a specific answer. I think we're very optimistic on both. I think when we look at the EDGE gastric data, which obviously includes the ITT patient population, and we've been looking at different cuts of the data, I think we're still optimistic, obviously very optimistic on PD-1 high, but also pretty optimistic on ITT. If it were to hit just in PD-1 high, and to be honest, I think we're thinking about it more about PD-1 positive versus PD-1 high. We actually changed the stat plan a bit where we'll initially, just like we were, look at PD-1 greater than five and then ITT as dual primary endpoints. If you hit on one of those endpoints, there's been a hierarchical analysis.
We will look at PD-1 greater than one. If we do have a situation where we did not hit an ITT for some reason, but we do hit on PD-1 greater than five, you then have a really good opportunity to expand the label to PD-1 greater than one, which is 80% of the patient population. As a reminder, the allocation across the PD-1 positivity is different than what you see in lung. It is very weighted towards PD-1 greater than one. All that said, we are still very optimistic about ITT. We will see the data actually pretty soon.
Gotcha. OK. And would you have to do, so if you did not hit, would you have to do PD-1 testing or?
Regardless, there's going to be PD-1 testing. I mean, just like today. We've been doing actually a lot of work on this. Almost everybody is tested for PD-1, just like lung and gastric today. Irrespective of what happens, patients will likely be tested before they receive something.
The complexities around the testing is sometimes we see with lung or?
No, no. It's like a very routine testing that everybody is doing now.
Gotcha. OK. What are the expectations around the control arm for 221?
Really sort of straightforward because it's been so consistent. The nivolumab chemo comes in about 13 months OS. Similarly, it comes in around seven months PFS, even though PFS isn't the readout. As I mentioned before, tislelizumab chemo and Keytruda chemo also are in that 13-month range. It's very clear, very consistent standard there.
Gotcha. Maybe in the last few seconds, just on CD73, kind of what should we expect? The translation of that phase I into phase III, any differences in the trial and how we see the risk?
We think that's a great trial. That's a little under the radar. It's a phase III trial in frontline pancreatic cancer. It's off to an incredibly quick start. Enrollment is great. I think it's very realistic that it could be enrolled even within a year and certainly a readout in 2026. Simple design going on top of standard of care, GEM-Abraxane versus GEM-Abraxane Quemliclustat, very similar to what we did previously. We think supported very well by a synthetic control arm that we had done in parallel to the initial trial as well as biomarker work that actually suggested we were doing better in patients that from a prognostic standpoint actually should have done worse, but those patients that have high adenosine. We feel, A, very optimistic about the reproducibility of those earlier results, and B, that we'll have those data very near term.
Perfect. Thank you so much.
Thank you.
Thanks, Peter.
It was a pleasure, Jen.