Hello, everyone. Welcome to our Fireside Chat with iTeos. I'm Li Watsek, a biotech analyst at Cantor. I'm thrilled to have Michel Detheux with us from the management team. So I guess we can, you know, before we get into your exciting ESMO data, we can start maybe with an overview, and then what's coming up. Yes, then in terms of overview and where is iTeos today, we have used our twelve years of expertise in tumor immunology to develop first-in-class or best-in-class assets. We have today three assets in clinics: our TIGIT antibody, partner with Pfizer, with GSK, sorry, and two adenosine pathway inhibitor, which are currently in early-stage development. Okay, great. So let's jump right into your ESMO data, right? I think you guys showed some pretty exciting data in lung. So I wonder if you can just, you know, walk us through the data and what are some of the key takeaways there. It's an early cut for the moment. We have disclosed data on 122 patients in first-line non-small cell lung cancer, high PD-L1, where we have compared our PD-1 dostarlimab from GSK with the doublet of dostarlimab and three different doses of our TIGIT antibody, belrestotug, 100, 400, and 1000, and we have seen all what we are looking to see from this study. We have a very impressive response rate around 60% for the three doses and a significant delta versus dostarlimab, 30% delta versus dostarlimab. Some additional data with ctDNA, which is an emerging biomarker for non-small cell lung cancer, and what we consider to be a manageable safety profile. Most important, we have very consistent data through the three doses. We have data on almost 100 patients. That makes the most robust data set for TIGIT antibody in this indication published so far. I guess, when would you guys be in a position to sort of pick a dose and... I guess you're already in phase 3, so? Indeed, we started the phase 3 in June. We got a $35 million milestone payment by GSK. One of the rationale for this phase 2 was to satisfy FDA Optimus. We have selected 400- Mm-hmm. -which is, I would say, the best balance between risk and benefit. And we are moving forward with 400. What I want to clarify is that we have started outside of U.S. for, I would say, timing reasons, and we have a final discussion with FDA, with a full package, not only with these 122 patients, but also with a total amount of more than 200 patients, in order to go to FDA and confirm the recommended phase 3 doses. Okay. I guess, how are you guys thinking about, you know, differentiation of your doublet versus some of the competitors out there? Obviously, they're a little bit ahead of you guys. Yes, indeed, this is something which is super important. We started the year being in position number fifth compared to the competition. Today, non-small cell lung cancer, we could be number one with the quality of the data that we've released. You have, I would say, three major component: the quality of your assets- Mm-hmm. which is super important. I'm going to come back to that. This is the first one. The second one is really, indications. Mm-hmm. Where do you go, and what is the rationale to go there? The third one is the design of your phase 3. Mm. What is the robustness of the phase three? Which population are you targeting? Coming back to the first component with preclinical data set, the clinical data set, including the one we released this weekend and translational medicine, we have very strong argument that we have a best-in-class TIGIT, and we can come back to that later. For PD-1, GSK has been the only company which has performed a head-to-head comparison with pembro in non-small cell lung cancer, and they showed that they were as good as, if not better than, pembrolizumab. Then we have really an excellent doublet here and probably a best-in-class doublet. In terms of indication, there are multiple indication which is being explored with TIGIT so far. Mm-hmm. There have been more than 15 phase 3. The most robust data set, and we want to be data-driven, has been in high PD-L1 non-small cell lung cancer in combination with a PD-1. This is what we are doing. Mm-hmm. And finally, being behind the other one has helped us to learn a lot from the failure of the other companies and the design of the studies. We are confident that today we have the optimal design for our phase three in terms of number of patient, statistical design, and the specific population to be successful in our phase three. I want to get your thoughts around some of the immune-related side effects, and I think it goes to the debate whether you should use Fc-active versus inactive. And we've seen some, you know, sort of side effects from Merck as well. So I guess if you can share some thoughts in terms of what you see from your own study, is that potentially a concern for you guys? And it's not a concern for us. Mm-hmm. And I want to make clear that both David Spigel, who presented our data, the discussant will analyze our data after at ESMO, and at ESMO, we've been one of the top story of all the meeting. Then it's something which is super important, and both David Spigel and the discussant were really confident that it's a manageable safety profile. Analyst report and the press report were on the same page. Obviously, the market did not consider that, and on Monday we have seen what was the reaction. I want to make very clear the following: When we consider the mechanism of action of a TIGIT antibody- Mm-hmm. an Fc-active isotype, we have a multifaceted mechanism to trigger an antitumor response. Interestingly, even if we take the average response, discontinuation rate that we see with our two initial dose, 100, 400, it's 19%. Mm. ARC-7, which is a similar study from Arcus with the Fc-silent isotype, reported a discontinuation rate of 18%- Right. which is really identical. Mm. On top of that, we have a significant number, one-third of our patients, which have been discontinued for grade one and grade two, and we have worked very hard over the last few weeks with our clinical site to be sure that we have a better management of this profile. Then in conclusion, we are very confident with the safety profile that we have for our doublet. Okay, great. So I wanted to talk a little bit about your phase three. So if we go back to Roche's, you know, phase two CITYSCAPE, obviously they show very nice data. But once they move into, you know, phase three SKYSCRAPER-01, obviously we still have to, you know, wait for the data later this year. So the assumption is, well, maybe the signal is gonna dilute. So how are you guys thinking about, you know, when you move from phase two? Obviously, you're seeing very nice delta here, but going to a much larger phase three, how confident that, you know, the signal will not dilute to a point where you will have. You know, you will still have a pretty good chance of success there. I think, Michel, you touched on, you know, trial design, so I wonder if you can expand on that a little bit and the learnings from the field? So, you know, Michel already used the term once, and throughout the presentation, we'll probably use it again, which is being data-driven. Mm-hmm. We look at both what's happening externally and the data that we're generally generating internally in our decision-making process. Of course, we look at SKYSCRAPER-01 and CITYSCAPE, and what gets lost sometimes are two things from CITYSCAPE. One, that was 29 patients. Yes, it was a high signal, perhaps artificially high, due to the small sample size. It did translate into a six-month survival benefit. There's an OS delta of six months in SKYSCRAPER-01, so it's clear TIGIT is driving a significant benefit in a much larger study. We look at our data again, 90 patients with the doublet compared to 30 patients of dostarlimab. We're beating every benchmark that's been set by pembrolizumab. Michel already went through the reasons why we think we have a better TIGIT, why we think we have a better PD-1 than what Roche has, and you mentioned already the learnings of those studies in the design of our own phase three is gonna be incredibly helpful, both in terms of statistical design, so when we look at the data, and it was a good to see or important to see in SKYSCRAPER-01, that the curves didn't separate until six months. We've incorporated that aspect into our plan, as well as a much larger study, about 60%-70% larger than SKYSCRAPER-01. We have 1000 patients in GALAXIES-301, and we think we have an optimized study, with two optimized components in our doublet. And so, you know, we learn from our data. We learn from data in the field in our go-forward decision making. How are you thinking about, you know, enrollment in frontline PD-L1? 'Cause, we heard from some companies, maybe they have some challenges there, but is that going to be a concern for you? We've seen no challenges in 201. It's early in 301, so I say, look, we're on track there, but it's early days. One of the reasons we wanted to save the data, remember, we press released that we'd seen a strong signal back in the second quarter, but we're preserving the presentation until ESMO, is to make sure we have the clinicians, investigators in the seats. We're meeting with them. GSK is meeting with them to ensure that, you know, our positive strong data is not lost in the general headline of, you know, TIGIT has been unsuccessful. And, you know, it was a great meeting for us. Strong engagement, great feedback. I don't expect any challenges in 301, and certainly 201 continues to enroll very nicely. How are you guys thinking about the market opportunity in frontline PD-L1 high non-small cell lung cancer, given that, you know, assuming Roche is gonna, you know, show positive data, and they're, you know, market there? There is a major limitation in the SKYSCRAPER-01 study. Mm-hmm. They are comparing to atezolizumab as the control arm and not pembro. Yeah. We are comparing to pembrolizumab in our 1,000-patient phase 3. We believe that the best case scenario for us would be a positive readout from SKYSCRAPER-01 that will confirm that the small subset of phase 2 data translate into a positive phase 3, at least for OS, but it will not be clinically meaningful in terms of market practice because they will not be able to compare to pembrolizumab within the study. Okay. So, there are other players in the field sort of, you know, testing the triplet in front line. Obviously, we've seen the results from SKYSCRAPER-06, there was a disappointment. So, I guess, how are you thinking about, you know, combining with chemo in perhaps, you know, PD-L1 low population? Is that something that you guys are thinking about? You think, like, TIGIT can play a role there? Yes, this is part of our strategy. But first, I want to stress that we've already evaluated, PD-1, TIGIT, and chemo. We have data on 42 patients, and for example, for the- Yeah ... safety profile, we are exactly, within what has been reported for, PD-1 plus chemo. Then we are very comfortable that we have a manageable safety profile. But we still believe that to go for the low PD-L1, a biomarker strategy will be very useful, and we are currently, starting a biomarker, trial in order to refine our strategy there. Okay. Let's switch to head and neck cancer. Obviously, you guys have a very nice, perhaps, first-to-market opportunity here. So maybe just talk a little bit about, you know, your strategy here, what data that you've shown that gives you confidence that TIGIT can actually work. Yes. Again, we want to be data-driven. Then, when we started the first study called TIG-006, it was a single-arm study where we have evaluated the high PD-L1 CPS above 20% and the low PD-L1 for our doublet of dostarlimab plus belrestotug. The reason why we did this single-arm study was due to the fact that Roche was supposed to release almost one year and half ago- Mm-hmm ... a randomized phase two on head and neck, and we were using our study as an add-on to support our investment into a major phase three. With time, it appeared that this SKYSCRAPER-09 study was lost in translation- Mm-hmm ... and we don't even know if it will be published one day. In order to continue our strategy to be data-driven, we have decided to start our own randomized phase two, which is GALAXIES H&N-202, and we will get our first data next year. We are going to disclose at the same time- Mm-hmm ... the interim analysis of GALAXIES H&N-202 and TIG-006, and we'll get the data that will allow us to decide if we have the rationale to move into phase 3 or not. So later this year, are you gonna share the data from the small trial? We've decided to wait for having the first interim analysis of head and neck 202. Okay, I see. So I guess for the phase 2, study 2-02 that you just mentioned, I believe that this is a randomized control study. Maybe just walk us through the trial design. How do you power the study? What do you hope to show? What is sort of the bar? Then, this is a study where we compare dostarlimab in monotherapy with two different doublets: either dostarlimab and belrestotug, our anti-TIGIT antibody, or dostarlimab and the anti-CD96 antibody developed by GSK. Mm-hmm .. which is really unique and an additional component of the TIGIT CD226 pathway. And finally, there's also an arm with the triplet PD-1, TIGIT, CD96. Well, KEYNOTE-048 is the reference, but obviously we need to take into account also the recent data from Merus and Bicara- Yes ... in order to support our decision to move forward. Okay, so I guess, what do you need to see before you move into phase three? What is sort of the benchmark here? Here, as much as for non-small cell lung cancer, there is a very good correlation between response rate, PFS, and OS. In head and neck, the important parameter will be really PFS. So how are you thinking about sort of the registrational, you know, path in neck, head and neck? Obviously, there's some companies, for instance, as you mentioned, you know, Merus and Bicara, think maybe they can get a accelerated approval sort of based on, you know, response rate rather than, you know, OS. Is that something you guys, can do here, or do you think you may have to show sort of PFS and OS? It's still under discussion, but as I said, PFS will be a very important criteria to define how and when we move forward. Okay. Now, how are you guys thinking about sort of the competitive landscape, you know, frontline head and neck? 'Cause, you know, the bispecifics have shown pretty nice data that we don't know the OS yet- Mm-hmm ... but, you know, response rate looks great. PFS looks, you know, decent. So just how are you thinking about- Again- Yeah ... we'll be data-driven- Mm-hmm ... and we'll integrate the data of the competitive landscape, in order to refine what would be the best population and how to move forward with head and neck. Okay, then maybe switch to your adenosine pathway, so you have maybe some data coming up in non-small cell lung. Maybe talk a little bit about that and sort of frame the expectations for us. ... Yes, and adenosine pathway is another, I would say, challenging mechanism. Several companies have failed by modulating or to modulate this pathway to get a clinical benefit. Here again, we want to be data-driven. We have a best-in-class A2A receptor antagonist, which has shown some nice clinical benefit and also, biomarker associated with, the modulation of A2A. We are going to release before the end of the year, the data of a part one of a phase two called A2A-005, where we are in second line non-small- Mm-hmm. cell lung cancer in chemo-naive patient post-PD-1. This dataset will define if we move forward in the part two of this study or not. So, yeah. I told you Michel would say data-driven again. This is another indication where PFS is really important. Okay. We're combining with chemo. Chemo is going to drive a pretty high level of response. A lot of those are short-lived responses with fairly rapid progression. Mm-hmm. So PFS is going to be critical as we make that further investment decisions for the program. So is there a PFS bar that you guys need to see internally? Yes. -to make? Yes. One of the challenge of this specific trial- Mm-hmm. is the lack of reference data. Mm-hmm. Then we work very hard with KOL and expert and define the bar as being a PFS of six months in the chemo-only patients, and this is what we have to outperform in order to be confident that we can move forward and bring a clinical benefit to the patients. Okay. Maybe just quickly on EOS-984. Talk a little bit about the target and why are you excited about this one? Yes. We've been very excited with this target, and this is a target that our team has discovered in this pathway. Mm-hmm. ENT1 is a transporter for nucleoside. In fact, when we characterize our A2A receptor antagonist, inupadenant, we have shown that this antagonist was very efficient to restore activity of T cell, but not the proliferation into high adenosine microenvironment. What we have done is that we have characterized this inhibition of proliferation, and we have found that ENT1 was playing a critical role in this proliferation of immune cells. Mm. We have developed, ENT1 inhibitor, EOS-984, that we are currently evaluating into a dose escalation. We are just starting the combination with, pembrolizumab now. Mm-hmm. That would make a lot of sense in terms of mechanism of action, and we are planning to report the data in the near future of this phase one. Okay, great. Maybe just last question on your partnership with GSK. Maybe just talk about how that is going, and then any near-term milestones that we should pay attention to. It remains a strong collaboration. Mm-hmm. You know, we've said data-driven multiple times, and GSK shares that approach, and you see it in the decisions we make in terms of what criteria is needed for further investment. Again, we think, you know, internally and externally. I think the big milestones, you know, Michel has talked about the two oh two data, the head and neck package- Mm-hmm ... with TIG-006 and 202, that'll come next year. I think likely more impactful is going to be the next cut of GALAXIES 201, and that's a study of 340 to 360 patients, so there's a lot left to do. We had 120 patients of ORR, so next year you're going to see both that response rate and get much, much larger, and you're gonna start to see some durability and PFS data as well, so. Okay, great. I want to see if there's any question from the audience. Not? Okay, great. Thank you so much, guys.