Listen, thank you guys for joining us again. Excited to have Arcus Management. I know you guys had important HIF-2 data recently. There were some tidbit updates in the broader space as well. Maybe let's kick it off on the HIF-2 side, if I may. Could you remind us, perhaps, what's the window for opt-in that Gilead has, just so we know, because that could be an important catalyst as well?
Yeah, so we've been saying late this year, early next year.
Late this year, early next year.
Yeah. It's based on having a certain amount of data with a certain amount of follow-up, and we deliver a package with that information, and that's what triggers their opt-in window.
Is there a cutoff date for Gilead to opt-in by?
Have you delivered the package yet?
No.
We have not, no.
Oh, I see. So you deliver the package, and then they start.
And then it's disclosed.
30-day clock or 60-day clock?
Yeah, we haven't disclosed exactly what it is, but it's in that range, 30-60 days or so. And then you had another question?
One thing I would just say, just a little clarity. We haven't delivered a package because that's a formal part. Because of the dynamics of our relationship, they're basically totally aware of all of our data.
Oh, I see. So it's not like they need to start.
No, they've been. They know all of our data.
Can they give you a complete response? I'm joking.
No. And John's question, or your question, was they have to make a decision by a certain date. If they decide not to opt-in for whatever reason, the asset is then ours forever. So they get one bite at the apple, essentially.
Excellent. And was there any preliminary feedback shared by them on whether they think this data is looking a little differentiated over Merck or not? Any?
Yeah. Do you want to go?
I mean, they're excited about the data.
Got it. OK.
Let's talk about level of differentiation versus Merck. Obviously, one of the things that you've been highlighting for a while since early in the year is progressive disease rates. I guess let's start there. How significant is the PD delta that you've seen so far? Realistically, when we think about late-line monotherapy, how much benefit to PFS does the amount of PD delta that you've shown indicate?
I think it's going to be dramatically different, so I think we have a rate of primary progression that is roughly half that, maybe a little bit more, half to slightly higher than that. I think what you're going to also see is by the time we have four cohorts of data, they're all going to look that way. And what that leads to is not only more patients that obviously make it through the funnel and are going to contribute to that PFS, but I think also because you are hitting the target harder, you are getting deeper responses and more durability. And one of the things that we'll be sharing in our next data release will be PFS from both the 50 milligram and 100 milligram cohort. And I think that'll bear out. And that's, by the way, the provable endpoint.
You'll see a very substantial difference from Merck.
Remind us the comp again for Merck.
Yeah, they're 5.6 months PFS.
We don't need to adjust for the prior treatment, et cetera, when we do that?
Yeah, it's a good question because in that study, like LITESPARK-005, patients could only have had one, two, three prior treatment regimens. So these were patients that were sort of on the earlier lines from a lines of therapy perspective. We've said that in the 100 mg and 50 mg cohort of ARC-20, that about a quarter, a little over a quarter of patients had four or more prior lines. So those patients would not have even been eligible in LITESPARK-005. So it's the best comp out there, which is why we use it. But it's still tough to, I think, make an apples-to-apples comparison because we're enrolling patients that are more advanced.
Do you guys have visibility on your PFS right now?
Yeah, we have. Yeah.
OK. And remind me, I remember John and I had a long debate on this at one point. This PFS you're about to report is not a mean of all the PFS values. This will be a Kaplan-Meier PFS. Is that right?
This is a Kaplan-Meier PFS, yeah. So looking at the curves. And as a reminder, what we said when we presented the data at the Triple Meeting was that the median had not been reached. So what that means is your curve is still above that 50% PFS event line.
This is median calculated separately for 50 and the 100 milligrams group?
Yes. Yes. So for each cohort.
OK, got it.
So, I think by the time we're going to have four, we'll have two that you've already had a look at. We're going to have two more. One of them is the 100 milligram tablet that we're actually taking forward, and then one will be the 150 milligram. And by the time you see all four of those cohorts, I think what you're going to see is improved ORR, improved rate of primary progression, and improved PFS. So you're going to have four 30. Basically.
Sorry, 50, 100, and what's the other two cohorts?
The two 100s.
There's two different 100s. One was capsules, one is a tablet.
Now, when we talk about PD, especially in a combo setting where PD rate in general is much lower, thanks to the TKI effect, obviously, how much differentiation can you expect to see on PD in that setting, which is where you're developing?
I mean, it's going to be less, to your point. But you still do see some PDs with TKI mono. So it's probably 5%-10%. So there is an opportunity to improve a little bit on that. But to your point, where we think we're going to see most of the benefit is in PFS. So if you look at LITESPARK-003, so that was a study where they looked at belzutifan plus cabozantinib, they showed over a 13.5-month PFS in that study. So it was small sample size, but we think it's a.
That's actually a great segue because when I look at the Merck combo studies, because they have LITESPARK-003 with Cabo, but they also have KEYMAKER with Lenva, which is obviously where they're developing. When I look at the Belz Cabo versus Belz Lenva setup, it doesn't look to me like Cabo is necessarily that much stronger. The ORR is lower in that combo cohort. Even DCR is about the same in the Cabo combo as it is versus the Lenva combo. When you chose Cabo for your combo, can you talk a little bit about how the potency of the TKI matters and whether you're going to be able to see a benefit in that?
Yeah. So we actually have a slide. It was definitely in our ARC-20 deck. I think it's in our corporate deck as well that lays out those two data sets next to one another. So the Belz plus Len data set, as well as the Belz plus Cabo. You're right on ORRs, but PFS tells you a totally different story. So the LITESPARK study, where they combined Belz plus Cabo, showed over a two-month improvement in PFS relative to what was shown with Belz Len. The Belz Len study was also very small. It was part of a platform study. It was only 24 patients. The ORR was higher, but we also think that they were really maxing out on the dose in that study and probably seeing a lot of toxicity, which was one of the reasons why we think their PFS ended up being shorter.
To your point on why we chose Cabo, I was actually just having this conversation with one of our investigators yesterday who's in the Cass plus Cabo cohort, and she thinks that they're going to see a lot of toxicity with Len plus Belz, and so we think that is a huge opportunity for us if we can see less toxicity and then be able to manage whatever toxicity we're seeing better, and that's going to result in even longer PFSs.
Makes sense. Now, Terry, you've said a couple of times you think ORR could rise over time, the late responses. It's certainly something we saw in the Merck trials. A pretty material number of their responses happen after 12 months. But at the same time, and certainly you've had some delayed responses yourself, but at the same time, if you're hitting the target harder, shouldn't we be expecting a more rapid time to response? Can you square the circle for me there a little bit?
Yeah, so no, I think it's both. So I think what it's like everything else where you bring something where you're hitting the target harder is that you're going to see some patients respond earlier. But it's almost like if you think of a conceptual shift of a dose response curve. So some patients who might otherwise not even have been amenable to the therapy now are going to be, and then that's how I rationalize that you end up still seeing these patients that do well later on. I think it's a product of the type of therapy almost analogous to what you have when you have immunotherapy, where because it's so non-toxic, it's not like one of these things where you have slightly different toxicity between the cancer cell and the normal cell that you can get these patients in check.
Then with time, some of them hit an inflection where they actually get a deepening of response and you get a formal response. I think the corollary to that is, as you know, the definition of a response is somewhat arbitrary. So a lot of those stable disease patients that might be sitting at 25% are going to contribute to PFS and benefit equally well to those who might have gotten to 35% tumor degradation.
Makes sense. I guess what's the, just so we understand the path forward now from a clinical development perspective, would it be a registrational cohort in a refractory kidney setting, or would you guys consider some sort of PD-1 combination trials as well now?
So the first phase III that we're going to be doing is a study called PEAK-1, where we'll be combining Cass with Cabo and comparing that to Cabo alone. And that is in patients that are IO experienced. So they receive PD-1 either in the adjuvant setting or in the first-line metastatic setting. It'll be a 700-patient study, PFS endpoint, and we will be starting that study by the end of mid-2025, so first half of next year.
Got it. I guess why not almost right away, considering all the data is coming in right now?
For a.
For the first phase III start? Or you need the Gilead opt-in?
We need the start dates. We met with the FDA.
We're going as fast as we can.
Yeah, in September, and then it's just protocol writing.
Gilead needs to be activated.
No, this is independent. Whether Gilead was involved or not, the day will be the same.
Got it.
And you want to make sure you have enough safety data as well as just experience with casdatifan Cabometyx.
Why not first line? Because presumably that's where you have the cardio benefit.
Yeah, so we haven't said exactly what we're doing in the study with AZ. Also, we said that that's an IO naive patient. So you can maybe extrapolate into what we might be doing. But that study is looking at Cass in combination with AZ's PD-1 CTLA-4 bispecific.
I guess why not do it with a Keytruda collaboration? Or even if you need to buy Keytruda, why not that?
So it's funny. I was just having this conversation again with the investigator last night. So when we've talked to investigators, they were much more optimistic and excited about combining with PD-1 CTLA-4. They just think that Cass is going to bring a lot more to that combination. One of the biggest issues with PD-1 CTLA-4 is that it has a high rate of PD. So if a patient responds, you get a really, really durable response. And these patients can live a very long period of time. But 25% of patients progress out of the gate. There's no way to screen for those patients. It's not like another tumor type where you can screen for PD-L1, and that gives you a good sense of.
Isn't the current market share of Keytruda, I think, like two-thirds and Bristol's one-third in kidney?
So first of all, it's a very fragmented market. But PD-1 plus CTLA-4 or Nivo plus Ipi is about a third of the market today, and it is growing. It is also very dominant in the academic setting. They really love that combination because they don't like using TKIs if they can get away with it in the front line. That's another reason why people are so excited about combining PD-1 CTLA-4 with the HIF-2alpha. It's another TKI-free regimen. So that was just another reason why we liked the idea of not starting off with a PD-1 TKI HIF-2alpha combination. We just felt like it was going to work.
I guess my last one on this would be presumably Opdivo could have biosimilars by later in the decade, perhaps even the CTLA-4. So if you guys go down the path of AstraZeneca bispecific, aren't you limiting yourself to a non-biosimilar approach, whereas?
We can always do that later. What we liked about this is we do think that they have a best-in-class molecule, and it gives us the opportunity to work with a collaboration partner. I think people get confused with collaboration agreements sometimes because there's like a clinical supply agreement where someone is just giving you free drug, which is nice. But in an ideal world, you want them to also help fund the study with AZ. We haven't disclosed the economic terms, but they are providing funding to the study. So it's not just free drug that we're getting, and they are operationalizing the study. So it's great to be working with what we think of as the best.
And this trial would not be OK, but this trial, the comparator arm would be what, just PD-1 CTLA-4? Or would it be versus the TKI?
We haven't said, but you can extrapolate. But the thing is, involvement may end up. That is one example. In general, I'll take two separate entities versus the polypharmacology of a bispecific. But in this case, you do have the fact that because the anti-PD-1, the anti-CTLA-4 are in the same molecule, there does seem to be some predilection for the immune cells that are tumor-experienced as opposed to CTLA-4, where you get all of the peripheral toxicity.
Maybe we should definitely save some time to talk a little bit about TIGIT too. But just last question on HIF-2. The near-term data we're going to get here is mono, late line. Obviously, we've just been talking about combos. That's what the development path is. So what should we be really looking at in the incremental mono data that comes next year that you think reads most across to combo setting? What can we actually take from mono data to combo?
So I think, first of all, with the mono data, what I think is a bit unique about what we're doing is we're going to have four different monotherapy cohorts. And to the point Terry was making earlier in differentiation, we think we're going to show a very consistent effect across all those studies, including better confirmed ORR, lower rate of primary progressive disease, better PFS than the LITESPARK studies in each of those different cohorts. And so we think that is going to be very compelling. The first data set that we're going to get out there is for the 50 mg BID and then the 50 mg QD. So those are the data sets that we presented earlier this year. It'll be more mature versions of that, so more mature ORR, and I think, importantly, median PFS.
Two other cohorts that we're excited to get out there that will answer your question as well. One is the 100 mg QD cohort. So as Terry mentioned earlier, the 50 mg BID cohort that we presented earlier used sort of our first formulation, which was capsule. We had to give 10 capsules a day. So this next cohort, the 100 mg QD cohort, is our tablet formulation, which is what we're taking forward. It's a 25 mg tablet. So we think that'll be a really interesting cohort to show. And then we'll also be able to start showing data from the casdatifan Cabometyx cohort. So to your point, how do we extrapolate to.
But you've said that next year we might not see a lot of efficacy data from that combo cohort.
Yeah. We're being careful to what we're promising, but at a minimum, we will definitely see safety. We're going to try to get that out early in the year. That cohort is starting to enroll very, very well, and we'll try to get efficacy data out as soon as we can. It'll be ORR initially, and then eventually.
No early PD looks?
Oh, for sure, but it's going to be very low rate of PD overall. It'll be very low.
But in the last three minutes, maybe if we touch upon the TIGIT side as well. I guess maybe just a level set, everyone. Could you remind us, when is the first randomized readout we could get from the Arcus side? And we can talk about your conviction on the program, but let's start with the timelines.
So we've been guiding to the earliest could be the second half of next year.
That would be not long?
That would be gastric. So that one was the first to complete enrollment. So it actually completed enrollment now almost six months ago. And as a reminder, in the previous studies and first-line gastric with PD-1 plus chemo, so CheckMate 649, the OS was about 13 months. So you can start.
Earlier to enroll and shorter OS for the standard.
So this is not an interim look. We're talking about a mature OS, you think, second half of next year.
Yes, all we've said is readout. We're not really.
Do you guys have passive utility in this trial?
There was no futility analysis yet in the study.
Well, I guess why not? Wouldn't you need one?
A lot of times it's not that advantageous, particularly given you're not saving a lot of costs. We think futility, just based on everything that we've seen, is a really unlikely outcome too, just based on the amount of phase II data that we have.
Don't forget the data we've already generated.
Remind me, in this trial, it's your TIGIT plus what's the PD-1?
This is Zim versus Nivo.
This one could fall on the basis of Zim approval as well, this study?
Yes.
The lung data is 26?
We have not provided guidance on the lung data yet. So there's a lot of different variables. It's event rate driven, et cetera.
But enrollment there complete by the end of this year, you've suggested, and with OS and first-line lung tests.
20 months. Yeah, so it's going to read out after gastric for sure.
So, to start, maybe just remind us of the design again.
Yeah, so it's Dom, Zim, chemo versus Nivo chemo in first line gastric and esophageal adenocarcinomas. So it's a huge patient population. We think the market opportunity just in the CPS greater than one, so this is excluding patients that are PDL1 negative, we think is $3 billion, is probably $5 billion in total. So we think this is a huge first opportunity.
I guess the only question would be, let's say this trial hits and let's say a 0.75 hazard ratio, but the comparator is Nivo chemo and not Keytruda chemo, could that create pushback? How do you think about that?
Nivo chemo is the standard of care. So Keytruda plus chemo was approved later than Nivo chemo. So unlike lung where Keytruda.
Oh, I'm sorry. We're talking about the last one.
Oh, sorry. I was talking about gastric. Yeah, sorry about that.
Oh. OK, sorry. The lung study, what's the design?
Oh, it's TIGIT versus Keytruda chemo.
OK. That's what I was.
Every one of our studies has the global standard of care as the comparator.
OK, got it. Excellent. And maybe then, I guess in the last 30 seconds, obviously we all saw SKYSCRAPER-01 fail. So we saw second interim data. It looked like it had no OS benefit. 0.8 at second interim looked like it was developing reasonably well. But it obviously failed. So was the failure a stats issue? Did OS degrade somehow from the second interim? What's the ideal outcome?
Since we're well, I'll just speculate. Less ideal, just obviously they said nothing. If there was a degradation, so first of all, we don't even know their stat plan. As you know, they lost patients along the way because they switched PD-L1 assays, and they had to look for concordance, and the patients that didn't concord were tossed out. So they obviously had less power than they thought when they designed the study. If there was a degradation, though, keep in mind that Treg depletion in the AEs is real. And so that is something that's an area under the curve and cumulative. So you could imagine that as you go along, you might have lost more patients in the study arm. And I think that's borne out a bit by other data sets, their Sky Six data and then the two studies from Merck.
If you're having a negative hazard ratio in those studies, the only way you're going to get that negative hazard ratio is either you're depleting Tregs, which has been demonstrated, or you're having an effect on T effector cells as well.
Terry, sorry. Maybe just I know we're at time as well. Maybe one last quick one from us. I know you guys unblinded an ongoing phase III, which you guys terminated early. The hazard ratio looks very intriguing. I was just curious what feedback you guys got, but also it was so much of the performance from the ex-US sites. I guess how do you guys think about that?
Yeah, so we think.
This is a randomized TIGIT.
Yeah, so we felt at the time that STAR-121, which is the all-comer lung study, was the better bet. We were hearing that more and more clinicians were using PD-1 plus chemo in PD-L1 high. And so we felt like we're capturing a very big segment of that, as well as the PD-L1 low, PD-L1 negative patients from STAR-121. So that was the decision to close ARC-10 at the time. We think that the readout was meaningful. What was awesome about that is it gave us another randomized readout. It looked very, very similar to ARC-7. The sites were XUS, but in many countries. I know there was something out there about it, Korea. There were no patients actually enrolled in Korea in that study, but multiple other countries. So this wasn't like a typical U.S. academic institution only type study.
It was many countries, very heterogeneous patient population. It was randomized. So we think that there should be very.
Remind us the hazard ratio was what again?
0.64.
For OS. Yeah.
The median?
Zim, he had like a 24-plus month OS. And the OS hadn't actually been reached for the anti-TIGIT. Keep in mind, the reason that was designed that way is when it started, you couldn't enroll that in the U.S. anymore. And that's another one of the reasons where we had redesigned that along the way before we decided that it was just smarter to go off there to the PD-L1. The other thing I would just point out that's interesting in the anti-TIGIT field, we've had three positive studies: the ARC-7, ARC-10, EDGE-Gastric, all looking very consistent. AstraZeneca has had two positive studies, and they're running six phase III. And now they're running with a bispecific anti-TIGIT, anti-PD-1, and they have the Fc-silent anti-TIGIT.
I think one thing that is going to become very apparent is that in this particular field, where normally you might not get that Fc- enabled translates into effector functions and you deplete cells, there's a profound depletion that's been demonstrated. I think you can end up thinking that Fc- enabled versus Fc- silent is almost like two different classes of molecules. It's not a subtle difference.
Excellent. I know we're well past time. We should just wrap it up here. Thank you guys so much for joining.
Thank you so much. Thank you.