All right. Hello, everyone. I'm Stephen Willey, one of the senior biotech analysts here at Stifel, and glad to have with us here for the next presentation. Compugen representing Compugen, we have Anat Cohen-Dayag, who is the Chief Executive Officer, and Eran Ophir, who is the Chief Scientific Officer. We're going to kind of have an informal discussion. I think most of you know that there's a Q&A function that you can populate should you have any questions. We'll try to get those asked and answered as appropriate. It's good to see you both. Always appreciate the time. Always good to chat. Anat, not sure if you just want to kind of make any opening statements here before we get into Q&A. Maybe just give us a quick overview of the Compugen story. Sure. Thank you, Steve and Stifel, for inviting us. Yeah, I'll give an overview. Compugen is a clinical stage cancer immunotherapy company. We discover new drug targets by a computational discovery platform that we developed over the years. That's basically a mix of artificial intelligence and machine learning with human expertise. We develop first-in-class drugs to address these drug targets. We have a rich pipeline, validating partnerships. We're eligible for more than $1 billion in milestone payments, plus royalties. We have a solid balance sheet with cash runway into 2027. We're well positioned for growth. We're looking forward to this discussion with you. All right. Maybe we can start in on COM701. This is an anti-PVRIG antibody. You're one of, I think, two companies that are currently in the clinic with a version of this. You've talked about looking at this as single-agent maintenance therapy in recurrent platinum-sensitive ovarian cancer. I wanted to maybe just kind of get into the design specifics of the trial a little bit. Maybe first, we can just kind of talk about the rationale for this, right? I know you first generated a lot of COM701-based triplet data in the platinum-resistant setting, both in combination with PD-1 and with TIGIT. You're now pursuing this single-agent COM701 trial as maintenance therapy in platinum-sensitive patients. I guess we know the history of IO-based therapy in ovarian isn't a great one. I guess my questions are, what's the data that you have to support monotherapy activity within this tumor type specifically? What's kind of the underlying biological rationale to support this pivot that you're now making into the platinum-sensitive setting? Sure. First of all, I would say, yes, checkpoints did poorly and miserably in ovarian cancer. The reason is because ovarian cancer has a very strong suppressive microenvironment. It's difficult for checkpoints to induce sufficient anti-cancer immunity, right? Why PVRIG will do what others have failed. I think everything we do, first of all, is the biological rationale. PVRIG has a very, very different biology from all the other checkpoints. We published last year a very elaborated paper in a high-profile peer-reviewed journal how the biology is different. That biology enables PVRIG to drive additional waves of T- cells in indication which are typically not responsive to other checkpoints. This is the biology. This is the signals we have seen in few trials that we did in difficult-to-treat patient population, in patients which are PD-1 negative, in patients in indication which are not responsive. We have seen that PVRIG blocker, COM701, is active. This is, in general, about the biology. Why should it be active or other checkpoints have failed? We had our own data in ovarian. In ovarian cancer, it was a target indication for us from the beginning. The PVRIG pathway is very highly expressed in that indication. We started actually by seeing a monotherapy signal. We didn't treat many patients in monotherapy because in these platinum-resistant ovarian cancer patients, which failed everything, we thought that the best chance for the patient would be eventually to combine more drugs. TIGIT and PD-1, we have the hypothesis why we thought it will be more active. We started by having a small cohort in monotherapy. We have seen clinical response in monotherapy. We also saw that in most of the patients we treated, we saw translationally that we induced T- cells exactly with the MOA of COM701. We moved to the triplet mostly. What we wanted to see there was, do we have a signal? The historical data was PD-1 or PD-1, even plus TIGIT, giving something like 8% of response. We had, across 40 patients in triplet, a 20% response or close to that. Therefore, in addition to the monotherapy data, we said, okay, we have a signal that COM701 is active in ovarian cancer. Now, looking at the data of the trial, we saw that many patients relapsed so fast. They probably never even had the chance to respond to chemotherapy-free regime like the triplet IO that we used. The patient who did respond had very long-term responses, excellent safety profile. The rationale is we take this drug that showed activity in the last-line patients, which is safe and durable in responses, into the maintenance settings. We are going into patients who are second or third-line platinum-sensitive. They are still platinum-sensitive. They responded to the platinum. They have low tumor burden, more active immune system. We are starting this adaptive trial design. We are going to discuss the trial in a sec. We are starting with monotherapy maintenance settings to see if the signals in the last line and the biology in translating indeed to an efficacy in PFS in the maintenance settings in patients who responded to the second or third- line of platinum. Okay. I guess that's kind of an overview of the biological rationale and the monotherapy approach here. As you think about the trial now, I think it's been posted to clin-trials.gov. I don't think it's actively recruiting yet. Maybe you can just kind of talk about some of the key design features in terms of patient numbers, eligibility criteria, endpoint assessments. Maybe you can also just kind of talk about how quickly you think you can complete patient enrollment into this study and how big of a trial site footprint you need in order to accomplish that. The trial is an adaptive trial design, which gives us a lot of flexibility. We start, as mentioned, by COM701 monotherapy. We have the flexibility to add more combinations. I will discuss it. We are talking about first, second, or third-line platinum-sensitive patients. They are still platinum-sensitive, meaning they respond to the platinum. They already received Bevoparp and/or PARP, or they are not eligible to receive Bevoparp. These patients have no standard of care. They receive the platinum, six cycles. Normally, they sit and wait, or they maybe receive some off-label. Eventually, they will relapse after a median of 5.5 months and eventually will become last-line or platinum-resistant. We are talking about this line of patients. We are excluding patients with liver metastasis. We are stratifying by prior treatment with PARP because we thought that this is also, like a key thing that could homogenize properly the population, the previous PARP treatment. We are randomizing. It's a randomized blinded study in which we're going to enroll in the first part 40 patients, COM701 monotherapy, and 20 patients, so it's a two-to-one, that we're going to see a placebo. This is going to read in half to 26. We have the flexibility to move forward, either to continue in the monotherapy, maybe even some fast recognition. We can then add combination. As mentioned, it's an adaptive trial design that after the first part of the monotherapy we have the flexibility to make different choices. Yeah, maybe to get to the. Yeah, go ahead. Yeah, maybe to get to the enrollment part, you were asking about this and the sites. Look, this specific indication, the platinum-sensitive, is less competitive than the platinum-resistant. Specifically, the patient population that we're targeting is a patient population that is not a candidate for maintenance treatment. Since most of the patients are going to be third-line platinum-treated patients, it's about 30% of the second line, which is second line is about 34,000-35,000 patients. On one hand, it's less competitive. On the other hand, it may still be challenging due to the size of the population. We feel that we know the investigators from prior studies. These are ovarian cancer investigators. We know them. They are engaged. We have the relationship. We will open enough sites in order to address it. We'll keep flexibility in how many sites we need to open in order to keep the pace of the enrollment. Okay. You talked about progression-free survival as a primary endpoint. Can you maybe just kind of speak to the data that you have that informs that assumption around progression-free survival, both in the control arm? I guess, what is the magnitude of difference that you would deem to be clinically meaningful that you would want to see COM701 do in terms of magnitude of benefit above and beyond a control arm? In this case, in this patient population, actually, there is quite solid historical control. There are quite few phase III trials in a similar population that showed quite consistently that the PFS is around 5.5 months. This is what we expect to see in the placebo. We think, and after the discussion we had with KOL, the three month improvement. 5.5 of the placebo and then three months improvement in PFS in the COM701 monotherapy will be a win. Again, we can have the flexibility to choose how we continue depending on how big is the effect. It could be more than three months of effect. You can think also on an accelerated approval and other options we can think about after seeing first that you have these three months of improvement. Okay. It sounds like you do have some confidence around the control arm in terms of the assumption that you're making here. Yeah, very confident. Which is probably very important in winning in any randomized control trial. I know the randomization scheme that you're using, obviously, you want to get as many COM701 data points as possible. Your control arm consists of 20 patients, right? There's obviously some potential for inherent variability to pop up. How are you controlling for variability in this trial such that it's not heterogeneity that then overrides this control arm assumption that you have on the progression side? Yeah, it's a very good point. Yes, we made the decision to have 40 to 20, to have, as you mentioned, more COM701-treated data points. The reason we did that, because of that solid historical control. I mean, we know and we have clinical development in different trials. In this case, we really have solid data in few phase III studies to show what we should expect in the placebo. We think that's why we could allow ourselves to go a bit lower in that to 20 patients. In addition, we also have a very strict inclusion criteria to make sure we have monogenized the population going to the trial. As mentioned before, these are patients who respond to the platinum. These are patients who are excluded patients with liver metastasis. We are stratifying by previous exposure to PARPs. We think that being very strict on the population, in addition to the historical data, should make this trial that we could read into the conclusion if COM701 is active in these settings. Okay. You mentioned data would be available in the second half of 2026. Is that a final analysis? Have you structured in any kind of interim or futility analyses into this trial? I guess, if so, what triggers any kind of interim or futility analysis? The planned interim analysis that will happen anyway is half to 26. In addition, there was a data monitoring committee that reviews regularly the data. If they will think either the trial is futile or that the reserves are actually to unblind the data and move to the next step because we see activity, it could also happen earlier. Okay. Okay. Second half, 2026 for the interim, but a DNC decision could potentially occur before that. All right. Let's assume that COM701 achieves the benefit that you're hoping in this trial. What do you think are the next steps? I mean, do you just run this study back as a single-agent registrational study in a randomized control manner? Do you start to look at combination-based approaches here, maybe in earlier lines of maintenance therapy? Obviously, it's all data-dependent, but what are the options here that you think you may potentially have? Yeah, I think, as you said, first, it would be data-dependent. As always, we're data-driven. That's the way we make the decisions. I think that positive data opens the door for us for multiple paths forward. One would be to go and discuss with the FDA, with the regulatory authorities, and to understand what is it that we should do in order to take this COM701 as a single agent for an approval in this specific patient population. What would it require us to do? Build a path to registration. That's one. The second would be to expand the study, the platform study, and test COM701 as a backbone for different combinations. Either go to earlier settings, as you indicated, in combination with Bev and PARP, other combinations. Obviously, we're still believing the triplet combination, and we may see how this would fit. It will open a door for us to use it as a backbone for drug combinations. I think that with this, we're not looking forward also to take we're not looking to take this program forward alone. Maybe part of it will be done internally, and some of it with partners. We'll see. We'll be data-driven. Okay. We know Glaxo also has a PVRIG- targeting antibody in the clinic. I think this was procured through the surface transaction, if I remember correctly. Right. What kind of competitive intelligence do you have on this asset at all? Can you maybe talk about what GSK is currently doing with it in the clinic? GSK, as far as we know, they did not present any data, anything in the public. We do not have specifically any competitive intelligence, but they progressed the program to phase II, and they tested as a triplet combination. That is one. They are following basically the DNA hypothesis that we came up with, and we will wait to see their data. Obviously, any data from GSK, and by the way, also from others that are testing it in phase I studies, there are others as well. For example, BioEos is testing it as a bispecific. Any data from any other party, any clinical data may serve as a validation. I want to remind that IP-wise, we have solid IP. I think that maybe if you follow in the public domain, you will see that the IP that we have is triggering an interest. There are attempts to challenge our IP. As of now, these were not successful. We see it also as a validation to what we're doing. Okay. Maybe we can talk a little bit about the partner programs, right? You have the partnership with AstraZeneca. This is a PD-1 triggered bispecific, rilvegostomig, which I usually have a very difficult time saying. Astra's put a ton of resources behind this, right? Astra's putting a lot of muscle behind this program. I think, what, seven phase III trials now, another 5-10 phase I, II trials. I guess when you look at the suite of phase III studies that are now enrolling, which are you the most excited about? Where do you think the biology makes the most sense? I think before going more to the specific studies, I think that I'll say first, by now, it's eight pivotal studies. We're still counting. Eight, okay. Yes, it's eight. They are spread across lung and GI, gastrointestinal. They use a very broad and comprehensive strategy, which is really to replace the PD-1, the PD-1, PD-L1 inhibitors. They are testing it in multiple combinations, chemo and ADC. They are addressing really different patient populations. In lung, they have so many studies where they are testing different subpopulations of non-small cell cancer in a very methodical way, comparing to standard of care. They are really allocating a value of a peak year revenue target of more than $5 billion for this. I think that they already presented compelling, promising data in gastric and in lung in 2024. It seems comparable to what's been there with the standard of care or with the PD-1s. They are pursuing indications where PD-1 works. When you're talking about the biology, I think that AstraZeneca is taking the strategy to address the tumor types that are the more inflamed ones, those where PD-1 works. They're investing their resources there. They plan to present data in 2025 for the combination of rilvegostomig with ADCs. By the way, they're also going to present trial in progress for TROPION- Lung 12 as an oral poster in the AST conference in May. That's very soon. I think the fact that they are targeting the biology with the specific tumor types and really focusing on the various patient populations in these indications is a smart way to move. Eran can give maybe examples, maybe better examples, diving into the details of the studies on the biology. Maybe, Eran, maybe in speaking to this, you can maybe offer up your opinion on their strategy to pursue ADC-based combos, right? That appears to be kind of a core element of what they're doing. We'd just be curious on your thoughts around that as well. They have rilve as a potential new IO backbone, which is safe because it is not binding FC receptors, and it is potentially more active than PD-1 because it is a PD-1 triggered bispecific. They have shown it, at least in single-arm studies, to be superior to historical data of PD-1s. You have the smart chemos, the ADCs, which are maybe a bit like chemotherapies, inducing immunogenic cell death, reducing tumor burden, and making the environment more favorable to the triggered PD-1 rilve to be active. I think this is really exciting. What is interesting also in some of their studies, even though we think, and it was shown that triggered is active, because they have a bispecific molecule, they do not really have to prove it. If they are doing a combination of rilve and Imfinzi and compare it to Pembro plus chemo, they are combining the forces of two strong drugs versus Pembro plus chemo, and they could get approval, potentially even without showing directly that rilve is beating Pembro, even though I think it will. This is, I think, what's interesting about their strategy. Okay. We had a discussion this morning with Tom Marin from Sinai. He gave a shout-out to this drug. He finds it interesting. He is obviously very well-schooled in all things lung cancer. I mean, it does not seem like AstraZeneca gets a ton of credit for this asset. Not that I understand the intricacies of AstraZeneca valuation all that well. I am guessing it is probably a byproduct of the triggered component here. Look, it has been a space littered with attrition. I think BeiGene dropped out of this space last week. You have COM902. You have kind of paused development there. Is there anything that you are kind of monitoring now on the competitive landscape? Anything that you are kind of anxiously awaiting here to maybe give you some kind of confidence that COM902 might be an asset either worth dedicating resources to internally or maybe trying to pursue opportunities externally? Yeah, sure. Obviously, we're monitoring the field. I'll explain our approach to it and what we are looking for. I think in fairness, there were many failures, and there is a lot of skepticism. We were always saying that not all triggers are the same. We're focusing not only on the format, which is FC silent or inactive, but in general, we believe that our antibody, the COM902, is a very good antibody. We reached the conclusion that there is no study that we can do that is a phase I or II study with our triggered antibody that will make a difference. We really believe that the field needs to see a very good phase III study, positive phase III study, in order to make a difference. AstraZeneca and Gilead are having the right format, and they're pushing forward. They have the studies. We'll wait to see how their studies are panning out. We believe that this will also reflect on us, not only by what we're eligible to from AstraZeneca in terms of financial reward, but also in terms of how it reflects on our COM902. At the end of the day, we're not familiar with that many companies that have an FC silent that is ready to be tested if we're correct with what we're saying in this triggered space and the antibody format. Okay. Maybe last question because I know we're pressed up against time here. I mean, look, the company has a very good history, very validated history of leveraging this computational platform that you have to discover novel targets. You guys were first with triggered, first with PVRIG, first with IL-18 binding protein, which is now in phase I, partnered with Gilead. We didn't get a chance to talk about that. How big of an effort do you still have ongoing there with respect to kind of target discovery and validation? Should we expect those efforts to yield any kind of pipeline expansion over the course of the next one, two, three years? Definitely, I mean, we use this AI-based platform, which is focusing on target discovery. Not many companies are taking the challenge of identifying new drug targets to develop first-in-class. We're doing this when combining the very strong computational capabilities and AI, which is something that people sometimes tend to forget, and non-AI. We also use human intelligence, and we have the know-how to take this kind of novel drug assets all the way from clinical prediction to validation and clinical testing. The platform is proven, as you mentioned yourself. Definitely, we use all the resources we got from our collaborations to continue and build our internal computational platform and continue to build our internal pipeline, which is derived from this computational platform. Definitely, we do expect to bring more assets with our first-in-class potential into the field. All right. I guess we'll stop there. Anat Eran, always a pleasure to speak with you guys. Best of luck. I'm sure we'll be catching up here over the near term. Thank you. Thanks, everyone. Thank you.