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

Well, here we are, another year, another conference. So welcome to the leadership team for Arcus, very familiar faces. Terry Rosen, the CEO; Jen Jarrett COO; Thanks. Hi, Pia and I'm Yigal Nochomovitz, I'm biotech analyst here at Citi. We've been covering Arcus now, long time. I think it's 6 years? I know it's since before the pandemic.

So well, why don't we just start. Obviously, all the focus is on the renal cancer asset, although there's some optionality, obviously, in some of the other areas too, we can talk about. So maybe we can just start, remind everyone what is the -- what we know about cas now? What are the key pieces of evidence that are supporting the advancement into the 2 Phase IIIs, which you've announced and then, maybe you can talk about those designs and the goals of those? You want to start with that.

A little bit of overview is [indiscernible]

Yes, there you go. We're good now. It should be working.

Okay. So as Yigal mentioned, a lot of focus for Arcus and about Arcus derived from casdatifan. And I'll start by just sort of highlighting why is that? And we have a very broad portfolio, and we have 3 programs in Phase III involving multiple studies. But casdatifan is probably the biggest value driver for Arcus and that's for a number of reasons. The first is we own basically 100% of the rights to casdatifan. Secondly, the molecule and the target have been validated, both biologically, clinically and then also commercially by belzutifan, Merck's HIF2 inhibitor. So that molecule is already at a run rate of $700 million in annual sales. And that's only in very late line settings were it's used now. So that's the tip of the iceberg in terms of the opportunity. The third piece is that the Arcus molecule as Miguel had alluded to, we've shown quite demonstrably that it has a better profile than that of belzutifan. So if you couple the complete ownership with the high probability of technical success, the fact that we're Phase III. And then together, you put all of that, it gives us enormous strategic optionality for the entire company. And so having that program really becomes like the bedrock for everything else we're doing. So what are the key high-level observations that would enable the differentiation that we talk about from the Merck molecule. So we have roughly 120 patients worth of data in the same setting later line where the Merck molecule has been approved in on every endpoint, whether you look at what -- it's composed of 4 cohorts, each of 30 patients, whether you look at the individual cohorts or you combine all of those cohorts to look at the 120 patients. All of those cohorts are very pharmacologically relevant doses. So that's why we did choose to look at them into 120 patients. But we also looked at the 30 patients to be sure that anyone else that wants to look at them that you're not obfuscating anything. Because that also includes a cohort of 30 patients that's using the 100-milligram tablet formulation, that's the go-forward Phase III dose of our molecule. But whether you look at the rate of primary progression, which is roughly half of that of Merck, between 15% and 20% versus 35%. If you look at the response rate, which is about 50% better, 30-some-odd percent versus 20% or PFS, which is the important that's the approvable endpoint where we're more than double. So the belzutifan PFS is on the order of 5.6 months. We're at a PFS that's over 12 months. And so we're more than double that of Merck. So whatever endpoint you look at, whether in the individual cohorts or in the combined data casdatifan looks better. We've also presented early data from what our fast-to-market strategy that's in second line. It's in general, talk about the design but that's a study looking at second line clear cell RCC and it's basically cas plus cabo versus cabo, and we showed early data where we showed response rate data. And with only roughly 5 months of follow-up, we had a confirmed response rate of 46%. Merck did a similar study with bells plus cabo. And with 2 years of follow-up, they were a 31% response rate. Next year, we'll be sharing data as we get to PFS, so we'll be able to compare there. So I think all of the data to date clinically point to a black-and-white differentiation between the 2 molecules. What I would emphasize, for those like ourselves, we like to see some scientific evidence that ties together with the clinical data, to make those clinical data more compelling in terms of your conviction that they represent truth, where you can see dramatic differences between casdatifan and belzutifan is when you look at what's considered the gold standard biomarker HIF2 inhibition that's the suppression of erythropoietin production. So about 80% of your erythropoietin is under the control of HIF2. So it's a very easy biomarker to measure and what we've shown is not only do we have a much deeper or a deeper inhibition than that of belzutifan. But belzutifan loses that PD effect. These are data published by Merck after about 9 weeks on therapy. And we've shown that we still have robust inhibition of that erythropoietin production out past the year. So that's something that's very quantitative, it's very reproducible and that dramatic differentiation is what underlies the clinical data differentiation. Another important aspect of those data, and we'll be sharing more on this early next year, but we've also been able now with about 120 patients worth of data to show that the ability to suppress the production of erythropoietin, while erythropoietin doesn't drive the cancer, it's correlative. So the deeper you can suppress an individual's production of erythropoietin, the more favorable their outcome is, whether you look at response rate or whether you look at PFS. So we feel that's pretty compelling. Just touch quickly on our other programs is introduction. The other program that's in Phase III is our anti-TIGIT program that actually involves 3 different Phase III studies. The reason we're so excited about that is the first one of those is going to read out in 2026. That's a study called START 221. It's an upper GI cancers. Not only is it going to read out. So that study enrolled about 1,050 patients. It was fully enrolled in June of 2024. But we just presented data from the Phase II correlative of that same dosing regimen, same patient population, where we show OS on the order greater than 2 years, whereas the standard of care in that setting, which is nivo chemo, but you also see similar data for Tisle chemo or for KEYTRUDA chemo is about 12 to 13 months, so almost 2x the OS. So that will be reading out later this year. Also, the second Phase III study there, STAR 121, which is in the PD-L1 all-comer non-small cell lung cancer, probably the biggest part of KEYTRUDA's market 1,000-patient study that will be fully enrolled at the end of this year, and we're running a Stage III lung trial also in that where we're combining with durva and actually AstraZeneca is executing that third Phase III program is with our CD73 inhibitor. We just, as of September, fully enrolled almost a year ahead of schedule in 9 months, a study called PRISM-1. That's quemli, our CD73 inhibitor on top of gem-Abraxane, the standard of care there. So that was versus gem-Abraxane that was fully enrolled in September. The readout there is OS. For gem-Abraxane that's probably in the order of 9 to 10 months. So that's not that far off. And finally, because we are an R&D centric organization. We have a very substantial discovery organization. In our investor event that we just did in October, we highlighted a group of immunology programs that we've been quietly working on. And the first one of those, which is an X2 inhibitor, we expect to enter the clinic next year. The final piece is we're well capitalized. We sit currently with over $1 billion in cash. So everything that we're talking about, we're able to execute. And our -- that cash covers well readouts for all of those programs, the initial Phase III readouts, including that PEAK-1 study that's in clear-cell RCC. Jen, do you want to comment on the design of the...

Sure. So the first Phase III study that we have ongoing for cash is combining cas with cabo and comparing that combination to cabo and IO experience clear cell RCC. So these are patients that got IO, either in the adjuvant setting, or in the frontline setting. So this is a very big patient population, a market opportunity. We're going right on top of the most widely used TKI in that setting cabo. It's used about 2.5 to 1x more than lenva in that setting. So the reason that we liked using cabo is that clinicians are so comfortable using that. When you're dealing with TKIs, TKIs do have a fair amount of toxicity. And so choosing a TKI, the clinicians are very comfortable using very comfortable managing toxicity and therefore, keeping patients on drug is super, super important. And then cabo is also what we're using in the control arm. So using the same TKI and experimental arm and the control arm also just makes us a very clean study. And that is a key difference from the Merck study, which is similar LITESPARK-011, where they use lenva and the experimental arms, so they combine [indiscernible] with lenva, and they're using cabo in the control arm. So I don't think we'll have any contribution component issues, but there's a little bit of like if they show a benefit, which they seem to have shown with the combination with some of that coming from lenva. Is it from belz? So we feel like we have a much bigger study design. We expect to...

How are the investigators liking that they're enrolling that fast because they like it? They like that design, it's easy.

Well, yes, that's exactly where I was going. And so investors love the design, our expectation is that this study will enroll very, very quickly. If you look at the typical RCC study, either frontline or second-line study, they tend to enroll 15 to 18 months. There's no directly competing studies right now, which is really nice. RCC is a much less competitive space than a lot of other oncology market like [indiscernible] We're using a 2:1 randomization. So that means patients are twice as likely to get an experimental arm versus the control arm. So both patients and clinicians love that. And then the third thing that's important is that belz is still not paid for in a lot of regions, including in Europe. So Europe is particularly enthusiastic about the study. I mean there's a lot of enthusiasm everywhere. But because there's so much awareness around belz in Europe and it's used for PHL, but it's not paid for yet in RCC, the only way to get access to HIF2 alpha inhibitor RCC is to put them in a clinical study. So that's driving a lot of the enthusiasm for in Europe. So we do think that study is going very quickly. It's a PFS endpoint. So PFS for the control arm is probably around 9 months or so. So that should also enable us to get to a readout very, very quickly. So we're calling this our fast-to-market opportunity.

Did you give a time line for the...

We haven't. I'd say our goal is to get the study completed, enrollment by the end of next year. So that would be in that kind of 15-month window that is pretty consistent for other RCC studies. And then it's just going to come down to like the event rate. What we do have in ct.gov right now, is the primary completion date of April 2028. We definitely want to beat that. Our goal is to beat that.

Sounds like are some months are there.

Yes, we have all that. So I think as we get closer to the data, that start to get a sense for what the event rate looks like, then I think more specific guidance. Because obviously, the goal is to get to data.

So PFS is the critical. Is there an OS in there too?

We're looking at OS as the key secondary endpoint.

Is there a crossover or you want to do that?

No, no crossover. So OS is the key secondary endpoint. We had a discussion with the FDA, very, very comfortable with PFS as the sole primary, which is what, 80% to 90% of the RCC studies have had.

And you're just going to go to the prespecified number of PFS events. You don't have -- it's fast. They don't want to put interims in.

Right. So this is very [indiscernible] One thing we have said is there's no interim. So we're putting all of the alpha on that on that PFS endpoint. So not taking any risk there. So very excited about that study. The next thing is frontline. So we have a collaboration with AstraZeneca to combine cas with volru, which is their anti-PD1.

One other thing, sorry. So on the IO experience versus the -- so was it, you can just -- do you have a ratio? Or is it just you take you take in that population or...

Right. So our sense is, is that just based on what we see with that cas cabo cohort. About 20% of patients enrolled in cas cabo cohort had gotten adjuvant pembro. And that's pretty consistent with what we've seen in some other contemporary studies. We'll see what they got for LITESPARK-011. That will be an interesting data point to see what percent of patients have previously been treated in the adjuvant setting. But our expectation based on everything we know is that about 20% of our patient population is adjuvant pembro. And these patients -- because most of these patients either progressed while they're getting adjuvant treatment or right after their adjuvant treatment completed, they're considered to be resistant to pembro. And so that's why they're considered to look just like a patient that failed frontline anti-PD-1 therapy in the metastatic setting.

Okay.

And when we look at -- we talked about this, too. So when we look at those different patient populations in the cas cabo cohorts of patients that got treated with pembro in the adjuvant setting versus a patient that got treated in frontline metastatic patients, they performed very similarly.

Okay. And one other thing before we get to EVOLVE, just maybe comment a little bit on the data that's supported that they play well together, cas and cabo, just say if there's no issues there with this.

Yes, safety profile. I mean it's something else that we highlighted, the investor -- or I'm sorry, the ASCO event where this was presented. Safety profile actually look at the safety profile and think that it looks better than what you would expect with cabo mono. So I don't know if there's anything to that, but the data looks really good. One of the things that we're keeping in close eye on which is super important is just the dose intensity because obviously, the goal is to keep patients on the highest dose possible of both drugs as long as possible. And the cas intensity is like 95%. I think cabo was 90%. So that's something that clinicians have been very impressed by because a lot of times when you combine 2 active anticancer drugs, you will see added tox and so you'll have to dose reduce or take patients off one of those therapies. So it's a very, very positive sign that we've been able to keep patients on either full doses or drugs after treatment.

Okay. So, and then EVOLVE?

EVOLVE as a clinical collaboration that we're doing with AstraZeneca in the frontline setting where we're combining cas with their anti-PD-1 CTLA-4-bispecific. Anti-PD-1, CTLA-4 is a very widely used regimen in the frontline setting. So that was what we loved about this collaboration with an opportunity to create, to kind of build on IO-IO in the frontline setting and add additional efficacy without the added toxicity and to avoid TKIs in the frontline setting, which is what people really like about using Anti-PD-1, CTLA-4 versus anti-PD-1 plus TKI, which is the alternative. So we started this study a few months ago. It enrolled very, very quickly. We decided to pause enrollment and see how the patients already on treatment were doing because we saw some immune-mediated AEs, not unexpected at AEs just based on the fact that we're combining anti-PD1/CTLA-4. So we're tracking those patients now, and then we'll make some decisions. probably late this year as to whether or not we want to modify the dosing regimen for or and then likely keep going there. So that's

Where did you get with the enrollment? Or did you... With the enrollment or deny...

We didn't say -- what I would say is enrolled quicker than we thought.

Whatever you saw on this AE , this was what you would have expected?

It definitely was not something that we would not have expected. It is a if you look at late It definitely was not something that we would not have expected. It's classic. If you look at like...

What was the -- I mean, then it wasn't than it was in the...

Yes. So I think it just -- just based on the incidents, we just wanted to be careful conservative, given these are like novel metal drugs. And so we just decided like we have a decent number of patients on therapy. Let's pause not enroll additional patients, see how these patients do. And then probably modify how volru is being given. And if you look across their Phase III studies, the Phase III studies are all using like slightly different dosing regimens of volru. So this wasn't totally unexpected. So we'll see how these patients do, probably make a decision on twealing the dosing of volru and then we can keep going here. In the meantime, we are adding some other cohorts to ARC-20, to look at cas with other regimens. So ipi/nivo will be one of those. And second half of next year, we'll make a decision on which of these regimens.

Wait, you said ipi/nivo separately? I didn't hear that.

So with ipi/nivo.

So it's triple?

Yes. So ipi/nivo plus cas, and this will be in ARC-20. One of the nice things about ARC-20 when we set that up is it's very easy to stick additional cohorts in there and evaluate other combinations. And so we're taking advantage of that and just decided like to be safe, let's just see what like cas, ipi/nivo looks like. And we'll probably look at some other things in the study as well, which we haven't talked about yet. And then second half of next year, we'll be in a position to start a Phase III study in the frontline setting and what we think is the best combination.

We should think of that as -- and this was like an intent all along is somewhat of a bake-off. So Jen did mentioned about like one cohort that we already have that's almost fully enrolled, is we have anti-PD-1 plus cas. And the idea there was cas was looking as good or better than TKI alone and then top of it had better AE. So a big concept Jen alluded to, but I want to highlight is the whole TKI strategy, TKI free strategy, that's something that's feeding in what the investigators are quite enthusiastic about. And it's an opportunity. It's rare that you get call it dropped in your lap, an anticancer mechanism that essentially has AEs but no toxicity. So HIF2 inhibition, I've had investigators tell us that patients who experienced TKI and they go on HIF-2 inhibitor, they basically never want to go on TKI again. I think it's like a vacation. So you can imagine -- and we're talking about the CTLA-4, PD-1 plus cas, PD-1 plus cas, you could imagine that you could get 2, 3, 4, 5, 6 years of therapy that's robust and then they'll get their TKI in the next regimen as opposed to giving them the TKI.

Are you giving serious consideration to just do PD-1 cas?

It's certainly in...

Why you see CTLA-4 is a tough target?

Because well, the interesting thing with -- and this is why we feel we should export and it's why we not only we -- it was our intent all along to not only investigate with volru, but ipi, anti-PD-1 as well, because it does turn out. The reason that's used in about 35%. So it's the most used regimen in frontline, and it's the highest probability to get a cure. So it has the most risk as you said, CTLA-4 is a tough therapy. So it will be part...

So it's interesting because like the KOLs are saying no TKI, like let's keep it like a nice, very well tolerated. But then you kind of -- you would think that to be intellectually consistent, you're like, well, we don't want to do CTLA-4.

I had that same intellectual argument with myself. And I think the thing comes -- I mean it's interesting, Jen's pointed this a couple. The last ASCO, they were literally -- BMS gave a presentation. It was like the 9-year follow-up on ipi/nivo. So you can get some dramatic cure. So you do have that -- I would love to see if you gave me the choice, I think I'd rather see anti-PD-1, HIF2 push the TKI out further. But if you could pull something off where HIF2 basically lower the rate of primary progression and enhance the efficacy of CTLA-4 PD-1, that would be that we want to.

So when will you have resolution?

Second half of next year. So the idea -- we're choosing this all is a bake-off and everything has been like working, if you're sort of think of it as a retrosynthesis, we're working backwards to have the data to inform that.

And not necessarily paying for, obviously...

Obviously, yes, there is a big cost income.

And then you do -- if it's just a PD-1, you're going to use it nivo? What are you...

Yes, In Phase I, but we'll say a lot more, I think, early next year exactly what we're planning on doing.

But our data set I would foreshadow today, so the study is up during clinicaltrials.gov is ZIM plus cas. It's almost fully enrolled. The early data looks good.

Okay. So for the purposes of this take off, you're using yours?

Yes, right.

But that wouldn't...

Right. It may be different. And so we are -- we will have an ipi/anti-PD-1, and we haven't decided if it will be nivo, zim, or an arm of each.

Okay. So that's important. We got that.

Exactly.

And then that will coincide pretty much with potentially almost having the... Sorry. No, we're my year, a year off, sorry -- it's the year.

So like one is the fast to market, the others capture that frontline, but their frontline durability we think that really could be something that you've got patients 4, 5 years. I mean, keep in mind, so this is a data point we like to point out, in our late-line monotherapy, cohorts. The one that's the furthest along, 40% or so of the patients were still on study after 2 years starting treatment. So you're getting that durability even in the later. That's what drives the concept of why for someone to take the TKI early in their therapy push that into a later line of therapy and let them have the quality of life.

What is the KOL you had at the lady from -- what was her view on which way to go for the frontline? Did you have a or see what -- did she like [indiscernible] ?

Yes, TKI is I perineal -- and like going back to your...

Right. So I meant like whether to do the triple like whether to do the bispecific or the...

She's involved with the bispecifics study. I don't think she has a strong view either. But we're excited about the combination. So we'll see where things go and start these other options and important thing is having options to make an educated decision.

The individual antibodies gives you a different flexibility with when you turn off the CTLA-4. So we'll see how that plays out.

Okay. Anything else you want to highlight on cas? I mean that's -- we covered a lot of territory there. Any other -- I mean any other updates from what you've already shown, I mean, we saw the update recently and you've shown the combo data, you've shown the doses, the different doses. We know the dose going forward. You've done the dose exploration.

Right. I think the update would be what we expect for next year very succinctly. Early in the year, we'll update on those monotherapy cohorts. We see a further refine durability PFS. We'll talk more about the biomarker correlation between erythropoietin suppression and clinical outcomes. Roughly in the middle of the year, we'll update on the durability of the cas cabo cohort. And then in the second half, you will share the data from that those -- the bake off. You'll see what drives our selection of at least one Phase III study with the intent to start that by the end of next year.

Okay. Now the one that's sort of -- obviously, everyone has written off, but I think... it's very interesting. So like what's your view here? I mean STAR-221, I mean, you have basically a very good setup, right? I mean...

So I'll describe it. So STAR-221 is our first anti-TIGIT trial that we'll read out next year. We and AstraZeneca are the only ones who have utilized an Fc silent configuration for the anti-TIGIT We, in a molecule domvamalinab, they have a bispecific that has Fc silent. Between the 2 of us, we've had 5 readouts.

Which is their Fc silent?

The bispecific.

It's that one.

Yes. It's that one, same. And they're running 11 Phase III studies with it including in the SKY-1 population, they started it after the failure of Genetec. But our data, we showed OS in the exact population, that's the STAR-221 population, an OS of just over 2 years where the standard of care, which is nivo chemo, but you all have the same data for Tisle chemo, the same data for KEYTRUDA chemo is about 13 months. So almost 2x the OS. We've also had a very positive readout in a randomized study in PD-L1 all-comer non-small cell lung drop. So this study, which is event-driven, it enrolled 1,050 patients. It was fully enrolled as of June of 2024. It's going to read out sometime in 2026, and we think this will be the first time that you'll see a Phase III readout for Fc silent anti-TIGIT. All of our data to date they've only pointed to positive. So from an investor standpoint, I think we have like literally 0 built into our value from that.

Just remind everyone that just the trial design... Yes, just go through that.

And the hierarchical readout.

Yes. So it's pretty simple, which is like another simple study design, which is also like what we tend to do is like going on top of standard and using standard of care in the control arm. So we're looking at dom, zim, plus chemo versus nivo plus chemo. So nivo plus chemo is by far and away the most widely used standard of care in the frontline gastric setting. We're enrolling all-comer frontline gastric with the exception of HER2 -- sorry, HER2-positive patients. I know [indiscernible] just had a positive readout. That patient population is not included in the study. For the primary endpoints, we're using test OS, which is really the gold standard from the FDA perspective. And we're using dual primary endpoint. So one primary endpoint is the entire patient population, so ITT. And then the other is looking at just the TAP 5, which is essentially your PD-L1 high patients. And that's about 48% of the STAR-221 patient population. So it's still a big percentage. It's not like we're looking at a sliver in that other endpoint. And so we can win on either of those to have a successful study. If you win on just one of those endpoints, let's say, you win on the PD-L1 high or TAP greater than 5 patient population, you can then recycle the alpha into the ITT. So it's a very simple...

That one is first? That's like hierarchical?

No, they're tested at the same time. And so that's why...

So ITT and this TAP 5..

In the TAP 5, yes, those are tested at the same time. If you hit on one but not the other, you can then recycle alpha from one of the endpoints to the other endpoint

If you hit on the TAP 5, wouldn't it be weird not to hit on the ITT. I don't know.

Not necessarily just because we think most of the benefit is probably going to be in the TAP positive. So either TAP greater than 1 or TAP greater than 5 patient population. So I think if you were to predict today, we would say that's the patient population that's most likely to benefit. So that's how we did it like we did, but we think there's some especially since the greater than 1 is like 80% of the patient population, there's still a decent chance we get...

Okay.

So also, as Terry said, no one is giving us value for it, which is fine.

Which people are -- do you know?

No. The reason we don't know is because it's the event rate. And the thing that for those -- like we're not being cute about it, but to give a sense of how it looks, we've actually shared Kaplan-Meier curves from a randomized data set in the high PD-L1 lung patient population. And what you have there is anti-PD-1 with an OS of about 2-plus years, just what you would expect in the anti-TIGIT arm of that anti-TIGIT plus anti-PD-1, the hazard ratio was 0.64. But if you look at the Kaplan-Meier curve, you'll get -- you'll understand when you see how flat it is that the event rate that was the Arc-10 study. We stopped at or are we in favor of...

You're saying you need a lot -- even if the hazard -- I mean, you can have a good hazard ratio, but still have not a lot of events if you have good data.

Exactly. So the thing is that until you hit the event...

You're tracking the pool.

We know the pool...event rate.

And you never said what, how many are...

No. And we don't ever comment on the stat plan and -- but -- we can't make a call as to when we would want to give guidance that's going to read out in what part of 2026. Maybe as we get into early 2026 as soon as we can, we will.

Okay. And then can you go through the argument about the Fc silent and the fact that, that -- what you said, but just to resay it. I think there's a lot about the immunity.

I'm going to explain this like clearly because it's going to finally matter. So in Fc-silent antibody, cannot have effector function. And Fc enabled may, but not necessarily have it. It turns out, in the case of anti-TIGIT that you actually see very strong effector function with an Fc enabled antibody. What is the effector function? Effector function means that the antibody not only binds that cell, but potentially it kills the cell. And so what's driving that most likely is your peripheral T regulatory cells which are basically keeping your immune system in check. They're your peace keepers of your immune system are loaded with TIGIT. And about 80% of them are getting whacked by the Fc-enabled anti-TIGIT. Why is that important? Well, it causes immune AEs go look left and right with either the Merck molecule, [indiscernible] the Genentech molecule. And what you'll see left and right is that the immune AEs are dramatically higher in the study arm. Now the thing is it doesn't just cause immune AEs, but then that causes the patients in the study arm sometimes they have to go off therapy. So you lose efficacy as well. With the Fc-silent anti-TIGIT antibody -- and by the way, this effect is even more profound in the presence in chemotherapy. With the Fc-silent anti-TIGIT, the mechanism, it's very elegant, it was worked out by Genentech, the blocking mechanism meds to none of that. So your safety profile, when you do anti-TIGIT, anti-PD-1 chemo versus anti-PD-1 chemo, the AEs look almost identical as if the anti-TIGIT wasn't there. So you're getting this blocking effect, which is central to the mechanism of basically activating the immune system selectively in the tumor to a very safe mechanism and that's the big difference. So sometimes I'm going to try to frame this, give people a metaphor in their mind. The Fc-enabled anti-TIGITs are almost like you should think of them as an ADC that targets peripheral Tregs and blows them up. So you're getting this collateral activity that is nonbeneficial and, in fact, detrimental to your anticancer activity because I'm giving you these AEs and that's the big difference. Both we and AstraZeneca have seen similar data sets where the safety profile enables the study arm to not lose out because of AEs.

I wonder if an enterprising academic like ever took the data or someone at AstraZeneca or somewhere or one of the -- and took the data and impute out all the stoppages for all the AEs and try to see if they could recover like an actual therapeutic because that would be a good proof point for what you are saying.

One thing that was interesting, I mean this wasn't quite making point, but AZ did present some data, I think maybe at ASCO. And so they have a bispecific, they went TIGIT bispecific. And so what they did is they looked at the bispecific with their Fc silent TIGIT and then they created another bispecific with an Fc active TIGIT antibody. And they showed exactly what was referring to of Treg depletion, most importantly, maybe T cell depletion. So that was kind of an interesting experiment that they showed that it made that point, I think, very specifically like why you want to block TIGIT and not target TIGIT containing...

Anyway, so this is a silent catalyst here which is like a one of the...

By the way, the synergy on the catalyst is that if that hits the PD-L1 all-comer non-small cell lung trial, based on those good data, now we're going to start that. So that's probably the biggest component of KEYTRUDA's market. The 1,000-patient study with dom will be fully enrolled by the end of this year. So you're going to go from something that the world looks at is probability of technical success is 0 from an investor standpoint, not from our standpoint to like immediately overnight, I think particularly since we have the data and the PD-L1 highs that look quite compelling. So you combine that with a chemo regimen from the upper GI cancer. Don't forget, you're treating biology, you're not treating organs. And so when you combine the lung data we have with a potentially positive STAR-221 readout, I think there'll be the probability that people start to ascribe to that lung trial being positive, which is a huge market will be dramatic. But we're at this point, we don't want to twist anyone's arm or we're comfortable with...

I mean that's still running. And we haven't heard much about it.

Because just we'll be fully enrolling okay.

So if STAR-221 hits, that will become very topical. And then, of course, it doesn't whatever. But if it hits, then when would you have that? When would you be able to say something like an endpoint on the lung trial? That's like 20 later, much later?

Keep in mind, standard of care is probably 2 years ago.

Why did you do -- the reason you did nivo and you did your.

Well, that's my standard of care.

No, I know that. But why didn't -- you didn't want to do nivo in the active arm. You wanted to use your...

Because it's a lot cheaper, cheaper and then we're pricing not on top of whatever nivo is priced like want to price...

Okay. Cool. Well, that was fun. All right.

And we didn't even get to quemli.

You talked about it at the beginning. Thanks, everybody, for checking in. Appreciate it.