All right. I am Stephen Willie, one of the Senior Biotech Analysts here at Stifel, and glad to have with us from Compugen, CEO Eran Ophir. I believe this is your first official fireside chat as CEO.
Absolutely.
Did you think that Billie Jean would be your walk-up music? No? Okay. Maybe before we jump into Q&A, do you want to just give any kind of introductory comments?
Sure. I think it would be good to give a few words about Compugen. Compugen is a pioneer in computational AI-based target discovery. We use our computational platform to identify new drug targets in the field of immuno-oncology. I think what is unique about this platform is mostly that it is validated. We identified multiple first-in-class assets, targets, and then developed the drugs. We identified, actually, TIGIT around the same time Genentech did and published a paper head-to-head with them. We then identified PVRIG, and we have in our pipeline two wholly owned assets, COM902 and Fc-reduced TIGIT blocker. We will discuss a bit later why we think it is critically important that it is Fc-reduced. We have a PVRIG blocker antibody, another checkpoint called the COM701.
We also have a license to AstraZeneca, the right to use COM902, our TIGIT antibody, as part of their TIGIT PD-1 bispecific, rilvegotomig. It is now in 11 phase III trials. This is very exciting for us. More recently, we also licensed another first-in-class asset that we identified computationally, antibody blocker IL-18BP, a cool and new way approach to harness cytokine biology for cancer. It was licensed to Gilead, and we will now run the phase I dose escalation trial. Behind that, we have a whole pipeline of assets, immuno-oncology focused again, different mode of actions, all derived from our computational platform.
Okay. Can we start with COM701? So this is the anti-PVRIG antibody. You're evaluating this as a maintenance option in patients who are recurrent platinum-sensitive ovarian cancer. Can you talk a little bit about the data that you've generated with this asset and this tumor type specifically today? I think you kind of just presented a synopsis of that at ESMO. And then maybe also you can kind of speak to why you think the biology of PVRIG inhibition, you know, can be clinically relevant, can be clinically relevant as a way to preserve responses in a tumor type where the history of IO therapy is really not that great.
Yeah. I think we can look at it from three different, supportive, arguments. One is the biology. PVRIG, in contrast to TIGIT and PD-1, has a very different biology. In general, it is a unique checkpoint. Without going into too many details, these unique biological properties enable, when you block it, to substantially increase the number of T cells in the tumor microenvironment. We have seen these signals, translational signals, in patients which are PD-1 negative that are not only responding clinically, but we see this massive entrance or proliferation of T cells. This unique biology, which is specifically in ovarian cancer, the pathway is also very dominant. We think that this biology is the reason why we saw the signals, which is the second aspect that I will touch now, in these places which indeed typically checkpoints have not showed much effectivity.
Talking about the clinical signals we have seen. Specifically in ovarian cancer, we had a patient who was PD-1 negative, failed three prior lines, platinum-resistant last line ovarian cancer patient, PD-1 negative, responding to monotherapy of COM701. She stayed on the study with excellent safety for two years. Other patients in monotherapy responded immunologically and were stabilized. Also in combination, where PD-1 or PD-1 plus TIGIT historically, PD-1 negative ovarian cancer, you could expect what we saw in the past, or by others, 0% response rate. We saw responses in PD-1 negative ovarian cancer also in combination. Overall, we have the biology, we have the clinical signals in the last line platinum-resistant settings. An IO-only treatment in this landscape that's starting to be dominated by cytotoxic agents and ADCs, we knew it's not going to be enough.
We did what many IO therapies did, exploiting the signal in the last line and going earlier. We were going earlier into platinum-sensitive settings in patients who respond to platinum, they have low tumor burden, they have more a functional immune system. There is also some evidence that the platinum chemo sensitized the tumor microenvironment to respond to COM701 MOA. Our goal is to give COM701, which only in the last line settings showed durable and safe activity, to exploit that to prolong the progression-free survival of these patients.
Okay. Maybe you can talk a little bit about the design of the trial just from kind of a high-level perspective, patient numbers, eligibility criteria, et cetera. And then also speak to the kinetics of patient enrollment that you've seen thus far and, you know, where you are in terms of your site activation process.
We talk about an adaptive trial design. The sub-trial one arm is taking patients who receive second, third line platinum-sensitive ovarian cancer. They receive the platinum. They are not eligible for BEV and PARP. For these patients, they're really in unmet need. There is no standard of care. Typically, what happens, they receive the platinum, and even if they respond, there is no maintenance treatment approved. Eventually, or actually after a median of 5.5 months, they're going to relapse, they receive chemo again, and eventually they become platinum-resistant. We take these patients. We exclude patients with liver metastasis based on our observation in the last line patients, again, to get the patients which are the most probable to respond to COM701. We take patients who respond, partial response or complete response, and then we randomize them.
Forty patients will receive COM701 monotherapy and twenty will receive placebo. The reality is we're going to expect that the control should be 5.5 months. We expect the results in Q1 2027. Where we are, we opened now most of the sites. Recently we opened the big academical sites in the U.S. We also opened some sites in France from ARCAGY-GINECO, which is a great group that approached us that have experience in the platinum-sensitive setting trial. Most of the sites are open, and now is the time for enrollment revamp. We expect the results in Q1 2027.
You mentioned the five and a half month control arm assumption. What's your level of confidence around that number?
Actually, in these settings, it's quite solid because there are quite a few phase III trials, mostly in the earlier PARP studies, but it's a very similar population. In almost all of them, it was quite stable around five-six months. The historical control in this case is quite stable. Yet we also have an internal control arm of the placebo because we wanted to be certain that the signal that we see is an actual confirmed COM701 monotherapy signal.
Okay. I know you've kind of refined your guidelines around when you might expect to have an interim analysis from this trial. I believe you're now guiding to the first quarter of 2027.
Mm-hmm.
That interim, I think you have said is going to be triggered by some combination of patient enrollment, some combination of PFS event accrual. How do those two things work together logistically? Will the interim also have kind of a pre-specified futility look that will accompany that as well?
Yes, there is a data monitoring committee which are unblinded, and they have their own statistician, et cetera. Obviously, we're blinded. This is a blinded study. When a certain amount of events are occurring, they are starting to meet regularly. They look at the safety. They look at the unblinded PFS and clinical effects. They obviously have also the ability to decide that the study is futile early if it does not bring value to patients. By the interim analysis, we, I mean, then they recommend us when to unblind the study. This will be at the point in which they think that either the study is futile or the study is most probable to show the benefit we are looking for of the PFS differences versus the control.
Interim analysis, actually at this time point in Q1 2027, we expect to have the meaningful results to show if there is a clinically meaningful effect or not.
Okay. And so will there be any quantitative data for you to share at that point, or will it just be a qualitative disclosure around, "We have received feedback from DSMB. They have told us to continue the trial as planned," or is there going to be a real quantitative disclosure of data associated with that interim analysis?
Real quantitative. This is the plan.
Okay.
To have actual PFS results with the supportive, statistical analysis. It's not a, not obviously, it's not a registrational study, but we do have a statistical plan that incorporates both the historical data and the internal control data. This data should be available to share, and not just qualitatively.
I know you have, you know, a pre-specified stat plan like you just mentioned. You have, you've got your empowering assumptions that inform that, but this is still a 20-patient control arm, right? There's likely to be maybe just some variability-driven wiggle room around that five and a half month number. There's the stat-sig part of it, but there's also the clinically meaningful part of it. I guess how would you characterize the latter in the absence of the former, right? Is there a certain threshold benefit that you want to see in terms of, in terms of above and beyond the control arm that would give you confidence that the signal is real, maybe in the context of not hitting stat-sig?
Yeah. Again, what do we expect to see? Again, it will depend on the actual data and the confidence intervals, et cetera. Assuming we talk about five months PFS of the placebo, like the historical control, also with our discussion with the KOLs who are building the study, we think that three months of improvement is going to be very clinically meaningful. Actually, from our discussion with the KOLs, not with the FDA yet, this could be a start of a conversation with the FDA about accelerated approval. Again, there is really an unmet need, no standard of care. Three months and more are going to be meaningful. Less than that is going to depend on the results, again, standard deviations or confidence intervals.
For us to continue to move forward, we need to be absolutely convinced in the signal. Again, this is not a registrational trial, but again, the study should clearly show a COM701-driven effect in monotherapy for us to move forward.
Okay. If the data does show that, what could be your next steps? Do you reboot this study, upsize it, and pursue that as a registrational effort? Do you do that and maybe start looking at COM701-based combo opportunities in some earlier lines of therapy? What are your options?
Earlier line and later line are always an option. Now we have some initial results of [afshali KEYTRUDA] with chemotherapy in the last line, platinum-resistant. It is an option. Obviously, going earlier in maintenance, but even if we focus on this specific setting of the second, third line platinum-sensitive settings. It is an adaptive trial design, so we have the flexibility. We can either, as just said, enroll more patients, think about accelerated approval if the FDA will accept that, or stop the study and now, knowing the magnitude of effect, start a pivotal trial. If we think that is the right thing to do, also based on competition and emerging landscape, to combine after we know that we have a monotherapy signal, now to end combinations. The combination can be BEV, it could be PARP, it could be Combine O2.
We have the flexibility in the trial design to add more arms. We have full flexibility, again, to see monotherapy effect and then make decisions about the next steps.
Okay. So I believe you are now the only company that has a PVRIG targeting antibody in the clinic. Or is there one more besides Glaxo?
There's BioNTech.
Okay. There's BioNTech.
BioNTech have the, it's, they also have an Fc active, but it's a privileged TIGIT bispecific. It's, they have an interesting study in combination with their PD-1 VEGF in HCC. And there are some other companies maybe in earlier stages. Yeah.
I guess the question is asked in the context of GSK just discontinuing the PVRIG they got from Surface. I think they also purged all their other assets under that CD226 axis. Did you know anything about the Surface molecule just from a competitive intelligence perspective? Was there anything about COM701 that might be different from that version of PVRIG?
Very different. Yeah.
The PVRIG. How so?
First of all, I don't, I'm not aware of any clinical data disclosed by GSK. We don't know the reason for closure of the program, whether clinical data or just because they closed their TIGIT program and then they closed all the related assets, which are CD96 and PVRIG blockers. Regardless of that, we followed quite closely on the Surface Oncology of the days. The first and main difference is that their molecule was an Fc active, probably following the hype of the Fc active TIGIT that was in these days. We think it's critically important to have Fc reduced to block PVRIG and not to deplete anything, especially since PVRIG is not even expressed on regulatory T cells, just on NK and CD8 T cells. You don't want to deplete.
Also, I think that their antibody was not maybe, their focus was more NK activation. They had this unique epitope. We'll not go through all the details, but I saw less T cell activity, which is critical to solid tumor. I would say that from what we and no one understand, their antibody is different in a few ways that we see it as inferior to COM701.
Okay. So within that CD226 axis, TIGIT is also another target of interest. You mentioned COM902. I know you kind of paused clinical development of that for now. I know you're obviously still very levered to the biology of TIGIT through your partnership with Astra, who is now pursuing 11 registrational studies.
Exactly.
With rilvegostomig. This is the PD-1 TIGIT bispecific. Just maybe some thoughts on some of the updated volrustomig data we saw at ESMO, and really just kind of gives you, or what gives you, and I guess AstraZeneca as well, you know, confidence that this asset can maybe change the narrative around TIGIT a little bit.
Mm-hmm. So first about COM902, I mean, COM902 is, currently we're not opening new trials because of the big TIGIT dispute. We believe that probably there need larger trials to convince some of the community that the Fc-reduced TIGIT antibodies are actually clinically active. But we see, especially with the readouts and the few readouts in the next year's Arcus phase III and others that could inform on Fc-reduced TIGIT antibodies, we see COM902 as a great opportunity for us once the community will believe again in TIGIT. It's not exactly paused, but let's put that aside for a second. What was nice that AstraZeneca have shown in two oral presentations in ESMO is that they showed promising results in non-small cell lung cancer and bladder cancer, one in monotherapy, one in combination.
It was really nice to see the safety profile, which is one of the main differentiators between the Fc active and Fc-reduced. It was very important to see clinically meaningful readouts, even though there were not randomized studies, not many patients, everything is very clear, but in readouts which are more long-term readouts. I think the initial hype around the Fc active TIGIT, a lot was around overall response rate, which was very promising, but not a lot was translated into long-term PFS, OS readouts. Now, looking at the AZ results, they had, in bladder cancer, PFS that is double compared to historical control of pembro plus a TROP2 ADC. The non-small cell lung cancer data was also 21 months in treatment-naive patients compared to less than half of historical control.
It was good to see promising data, even though it's single arm studies in this longer-term effect. If we put this with the Arcus data of doubling of OS versus historical control, we start to see here a pattern, at least this is the way we look at it, that the Fc-reduced TIGIT antibodies are actually showing things that were not so often shown with the Fc active TIGIT in the early days where the hype was.
When do you think the TIGIT debate finally gets answered? Do you think it'll be in 2026 with Arcus data, or do you think it's going to take the totality of the Astra data beginning in 2027 and beyond to, to really kind of understand, what TIGIT is?
In 2026, obviously there is the Arcus data. It's a phase III readout. It's the exact same settings in which there was the overall survival benefit shown, now in the phase II. This is very, of course, important. One, across 2026, there are also multiple readouts from AZ, which are again, not phase III readouts. I think that if the accumulation of phase II readouts, as we see it today, will continue to show the same direction of durability and safety and ability to combine, probably at least some of the community will understand that this is where it goes. Yes, some will want to see a few phase III successes, and this will happen for Arcus in 2026, from AZ, and probably 2027 and on with the multiple phase IIIs they are running.
Can you just remind us the economic leverage that you guys have currently to the Astra asset?
Today we received $30 million in milestones and we're eligible for an additional $170 million.
Mm-hmm.
In regulatory and commercial milestones. Not less importantly, we're eligible for mid-single digit tiered royalties. If they speak about volrustomig, the new IO backbone, and they speak about $5 billion of potential future sales. Actually, looking at the way they combine it with everything they have in the pipeline, again, as the backbone to combine with ADCs and chemotherapies, the potential is even greater. We see these potential royalties as a major source of financial stability for Compugen in the future and to be able to drive further assets from our computational pipeline by having this catch to move forward aggressively in assets.
Okay. Yeah. We just hosted a long panel, Doctors Marin and Sabari. And Doctor Marin, still enthusiastic about rilvegostomig.
Rilve g.
Yeah.
Easier to say rilveg.
Maybe a little bit less of the ADC combo approach, but I think he still has some interest in TIGIT. Maybe we can shift gears to the other clinical stage asset that you have partnered, which is GS-0321. This was previously COM503. Kind of a really interesting approach here, to a cytokine that I think is starting to get a lot more attention from a development perspective. Maybe you can kind of speak to some of that biological and mechanistic differentiation that's been built into this asset.
Yes. This is another asset we identified computationally. Now, as you probably know, cytokines are very powerful immune stimulants, but they are terrible drugs. Even though they are approved, cytokines are extremely toxic and people have been trying for ages to engineer them in different ways to overcome the challenges of giving cytokines just to the blood of patients. Normally you get terrible side effects in the periphery. Sometimes you get some cytokines into the tumor to mediate some anti-tumor effects. The cool observation we had computationally is that in a tumor macro environment, you actually have a naturally occurring cytokine induced by inflammasome called IL-18, but it is not active. It is very high in the tumor, very low in the blood, but there is a soluble factor called IL-18 binding protein that binds firmly to IL-18 and prevents its activation.
You have this complex floating in tumor microenvironment, non-active. And nobody ever thought to actually use this observation to treat oncology. We developed a half-unity blocker antibody, not a cytokine, that inhibits the inhibitor, displaces away IL-18BP, and by that enables a tumor-localized IL-18 activity. We think this approach, and actually it was a competitive process, and we licensed it to Gilead, but got a nice preclinical deal because of this competitive process. They were excited like us from the ability to harness cytokines to manipulate tumor microenvironment without seeing anything preclinically for now in the periphery, and this is moving now in phase I. We are running the phase I. We started the phase I in the beginning of the year. It is almost starting to be one year since then.
Yeah, it will be interesting to see how this translates clinically.
Okay. I know you're leading the phase I development effort here. I think it was a trials in progress poster that you presented. It said to you, but were there any incremental details that came out of that poster with respect to either trial design, dose escalation schema, et cetera?
What we showed there is that it's a monotherapy dose escalation and combination with PD-1, ZIM of Gilead. We also have some backfill cohorts, which is, I think, becoming more and more typical these days in oncology, dose escalation to have a bit, in different doses, also in line with project Optimus, some testing early on, some expansion cohorts in specific doses, what we call a backfill cohort. I would say this is most of the details we had there. Yeah.
When does the handoff to Gilead occur?
We did not disclose that, but overall, I think the rationale of this collaboration was that since we know the biology, we know the target, it is a first-in-class approach for the best interest of the target that we will take it into early stages of clinical development. Then after sufficient proof of concept and pharmacodynamic and pharmacokinetics analysis, the big pharma can take over and move it aggressively in the more advanced trials.
Okay. What is your freedom to operate just around being able to communicate progress from this study? Should we think about the phase I as potentially yielding some meaningful news flow and/or catalysts, or is Gilead kind of controlling the narrative here and it's all TBD?
We cannot, we did not give any guidelines, but it is not TBD. I mean, we are Gilead, all the decisions are made in excellent collaboration. They know that we are biotelecinetically poor data. Also, typically when you move from phase I to phase II and you want to increase enthusiasm, especially if you have good results and you want to enhance enrollment, also big pharma report results in these stages. I think a combination of our relationship with Gilead and also their interest, I think we should expect that there will be data out before the phase III readouts. Again, we cannot commit or give exact guidelines for now.
Okay. And presumably some even more significant economic leverage to this asset in terms of milestones, in terms of royalties.
Yeah. We're talking about, we received to date $90 million of milestones. We're eligible for additional $758 million. Here the milestone, the royalties are up to low double digit tiered royalties.
Okay. Are there meaningful milestones? So of that $700+ million , are there meaningful milestones that could be triggered kind of pre-regulatory, pre-commercialization?
We did not give guidelines, but looking at other similar deals that we and others have done, normally with every stage that you move, phase I-A, phase I- B, phase II, all of that, you get the accumulation now of a nice amount of milestones. Again, we do not have specific guidelines, but this is not an atypical deal that we made with Gilead on that. It is obvious.
Okay. And then what should we expect just from a pipeline expansion perspective? How big is the internal effort that you have ongoing there? I mean, you guys were kind of first with TIGIT, first with PVRIG, first with IL-18 binding protein. I think you've shown an ability to discover and develop novel targets. Is there more of that to come?
Absolutely. I mean, actually the biggest group in Compugen is the early pipeline group. This engine is validated and we continuously work to bring more assets out of this engine. Yeah, and we work on different modes of action. We started by checkpoints. The COM503 is already a different approach and we are also working now on different approaches that utilize this know-how that we have in the computational engine to build our internal pipeline.
Okay. Maybe last question. Current cash runway and just what it allows you to execute on.
Our current cash runway is roughly two years, up into Q3 2027. This is assuming no milestones, no incomes at all, a very conservative approach. This allows us to conduct all the trials that we are doing, all the early pipeline. Without any, I mean, without, it has enabled us to pursue all our plans without any income, roughly two years, Q3 2027, into Q3 2027.
Okay. If there are no questions in the room, Eran, thank you very much.
Thank you.
I don't think we have exit music for you, but.
No, the mic just went.