Compass Therapeutics, Inc. (CMPX) Earnings Call Transcript & Summary

All right. Thank you, everybody, for joining us. I'm here with Thomas from Compass Therapeutics. Thank you so much for joining me this afternoon or this -- I don't know what day is it?

Thanks so much for inviting us. Happy to be here.

Great. Let's start things off with a minute or 2 from you, introducing yourself, the company. What are you looking forward to going into '26?

Sure. Compass, we're based in Boston. We're a monoclonal antibody discovery and development company in oncology. We currently have 3 drugs in the clinic and a fourth drug about to enter the clinic. So 2026, to answer your question specifically, is going to be an incredibly eventful year for us. We -- for our lead program, Tovecimig, a DLL4xVEGF-A bispecific antibody. We have a readout on the secondary endpoints of progression-free survival and overall survival that will be coming up in the end of Q1. So that will be a very big important update for us. In the second half of this year, 2025, something very, very important was observed in our Phase I clinical trial of our PD-1, PD-L1 bispecific antibody. We have 3 responses in the first 15 patients treated in the dose escalation study, including the first-in-human de minimis dose. So we have some incredible observations of efficacy there, and we're moving toward cohort expansions in patients with triple-negative breast cancer and non-small cell lung cancer based on what we've seen so far, and we should have a readout from that in the second half of the coming year with a presentation at a scientific meeting in the first half of the coming year. Then our fourth drug, a novel PD-1xVEGF-A bispecific antibody is going to enter the clinic at the beginning of the coming year. Obviously, a class of drugs.

Everybody is feeling good.

That you're well familiar with. But we've got a differentiated molecule, we believe, of course, so looking forward to getting that drug into the clinic. And our CD137 agonist antibody, where we've discovered a biomarker of response, we're going to do a biomarker-driven basket study with that drug to begin in the first half of next year. So '26 is going to be an incredibly exciting and eventful year for us.

A lot going on. Excellent. Let's start with that DLL4xVEGF bispecific that you mentioned. Obviously, we had BTC data earlier this year, very nice delta in ORR. Based on the latest look, as we head into PFS data, as you mentioned next year, I think confidence is reasonably high. Can you talk about what's driving that differential response, how you think about the evolution into longer follow-up data?

Sure. So again, Tovecimig, a DLL4xVEGF-A bispecific antibody, DLL4, delta-like ligand 4, the cell surface ligand for Notch-1. So this is a next-generation angiogenesis disruptor. And as you mentioned, earlier this year, we announced the results of the primary endpoint analysis in a randomized study, and we more than tripled the response rate in the control arm compared with the combination arm, including Tovecimig. As part of that analysis, which was best overall response rate as assessed by blinded independent central review, we noticed that there was a substantial difference in the incidence of progressive disease. And that, I think, is quite important. 42% progressive disease at week 8 in the control arm, only 16% in the combination arm. So...

That's going to be a major driver for PFS.

Exactly. So suggesting that the PFS curves are already separating early. So we're enthusiastic to get the PFS readout. The second half of your question, so both the PFS and the OS analyses are driven by 80% OS events. So initially, we did that because in every second-line biliary tract cancer study ever, PFS and OS are basically right on top of each other or they're 2 months apart, some number like that. So we said let's just do the 2 analyses together. But we announced in August that what we are observing, and this is quite a simple disclosure because it's a simple statement of fact. We observed that there are simply fewer deaths in the study than we had projected and which means patients are living longer and more patients are alive than we had projected. So what's driving that, right? So...

On the controlled trial where we would expect at least the control arm to have a reasonably predictable event.

That's right. That's right. So it's -- we -- what we see on a weekly basis are pooled OS events. So I just need to be clear about that. That's what we're -- that's the data that we're receiving. But it's very hard not to conclude that we're seeing some effect on overall survival. And based on our overall response rate data, the progressive disease incidents in the 2 arms, what we believe we're seeing is the formation of a tail on the Kaplan-Meier curve in real time. That's what we believe we're seeing.

Maybe important to note that this is a 2:1 randomized study. So if you're waiting for 80% OS events and they're just not happening in 65% of patients.

That's right. So that's a great summary. I think that with a 2:1 randomization and if the drug is having an effect on overall survival, that certainly could explain what we're seeing, but wouldn't that be great?

It might be great.

Right.

Well, then let's talk about the opportunity in second line. This biliary tract cancer in general, not an enormous indication, second line, obviously, a subset. Can you talk about what the market opportunity is in second-line BTC and where you might go from there?

Sure. Sure. So Biliary tract cancer, I think, is a bit of an underappreciated opportunity. About 25,000 patients annually are diagnosed in the United States. The claims aggregator, Komodo Health, did an analysis that they presented publicly at the Cholangiocarcinoma Foundation meeting a couple of years ago. And in -- we did some third-party market research earlier this year where we updated that analysis. But in 2023, scientists from Komodo reported that there were 23,000 patients annually in the United States, making it a very, very important indication. Our analysis that we did this year suggests that number is closer to 25,000. And there was a JAMA oncology paper published, which suggested that the incidence of biliary tract cancer is increasing.

But what's duration of therapy for second line currently?

Okay. So interesting. So the duration of therapy for second-line chemotherapy is certainly low single-digit months. So in our Phase II study, which was a single-arm study, just to be fair about that, our median PFS in that study was 9.4 months. So we clearly were seeing something different. Whenever I'm asked this question, I'm always quite conservative, and I use the lower bound of the 95% confidence interval from that study, which was 5.4 months. So if we're seeing that...

It's still reasonable to expect that maybe in the randomized study, you could see mid- to high single digits treatment that would indicate a pretty substantial market relative to where we're currently sit.

Exactly. And we believe that something like 15,000 patients annually -- and again, this is in the U.S. alone, in second-line BTC alone, about 15,000 patients will make it to second-line therapy. Make the math easy, 6 months on drug. As you know way better than I, recent oncology drug approvals are priced significantly high. So this is a multibillion-dollar commercial opportunity.

Bet you, in rare indications. Absolutely. So let's talk about the expansions beyond that. There's an IST going on in first line currently. You're looking at a basket that includes a number of tumor types, where do you think this mechanism, this dual anti-angiogenic mechanism makes the most sense going forward?

Sure. So a couple of different thoughts there in that question. So yes, we have a frontline study ongoing at MD Anderson right now. And that's quite interesting because MD Anderson was the second largest enroller in the second-line randomized study, which is open label at the sites. So it's tempting to conclude that they must have seen something in order to send us a proposal for an investigator-sponsored study. So they're adding Tovecimig to the TOPAZ1 regimen, which is gemcitabine/cisplatin and durvalumab. That study initiated earlier this year and is ongoing now. Should get a -- I'm going to get an update from the investigators in the early next year and might be able to talk about some of that next year. And obviously, moving to the frontline setting, which would be 50% more patients, potentially...

Presumably more duration.

Potentially 50% more time on drug. Now suddenly, you're talking about a multibillion-dollar opportunity in BTC alone. But the second part of your question, I think, is fascinating because there are many different malignancies that are enriched for DLL4 expression. So among those, hepatocellular, gastric, ovarian, renal cell, colorectal. So several different malignancies are enriched for tumors that express DLL4. The way you asked the question, I think, is you had a sort of subtle way that you asked the question, which is how might targeting angiogenesis in these various indications...

In that list that you just mentioned, hepatocellular, colorectal especially feel like really fruitful areas...

Exactly. Right. Because colorectal, the global standard of care in colorectal first line, second line, third line, if you're KRAS mutated is chemotherapy plus Avastin. FOLFOX Avastin, FOLFIRI Avastin, LONSURF Avastin, right? So colorectal would be very fruitful indication to potentially combine Tovecimig with chemotherapy. We're looking at some second-line designs right now, thinking about combining Tovecimig with FOLFIRI.

Better Avastin, better in IDH1 in a molecular regimen.

Exactly. That is exactly right. Better version of Avastin, next-gen version of Avastin. Ultimately, commercially, we would want to think about replacing Avastin in all of the Avastin labeled indications, which then this would be an incredibly successful drug.

That would be -- well, that would be quite a product. Let's talk about the commercial readiness. You've made comments about manufacturing capability in the past. But where are you from a prescriber awareness perspective, especially in second-line BTC. How aware are the docs about this and how ready are they to adopt it?

Sure. So maybe a 2-part question there. So first of all, on the manufacturing side, something that I think is super important, but just always sort of gets swept under the rug. We've developed, I think, some extremely important internal expertise on bispecific manufacturing. So we've got a commercial-ready manufacturing process for Tovecimig, which is very robust and something I'm very, very happy with. In terms of prescriber awareness, so we have begun our prelaunch preparations with some initial pre-commercial work. We'll have more to say about that early in the coming year on the leadership side. So that's something also I'm very, very excited about. One of the things that I'm super happy about in our randomized trial, we've had -- we are very, very fortunate to sort of have a who's who of academic medical centers. And...

You mentioned MD Anderson, obviously.

Yes, MD Anderson, Johns Hopkins, Mass General, Stanford, UCSF, Mayo, Chicago, WashU, it's just a tremendous -- Columbia, a tremendous list of academic medical centers. And I think that will put us, I think, in a good position for expanding prescriber awareness and in 2026, really want to begin our full-on commercial preparations for the potential '27 launch of this drug.

Excellent. Well, I want to make sure we have time to discuss those other assets you mentioned in your opening remarks. There's quite a list of them. But one thing that I wanted to make sure we touch on is that PD-1, PD-L1 bispecific 8371. That's a very interesting molecule, very different from other approaches we've seen in this space. Can you tell us a little bit about the design of the molecule and how you're positioning it in those lead indications you mentioned?

Sure. I think a couple of years ago, scientists at Compass decided to tackle a very challenging problem, which is the design of a next-gen angiogenesis disruptor as a bispecific antibody. And our scientists developed a technology that we call StitchMabs. That technology allows us to covalently link any 2 monoclonal antibody fragments together in a 15-minute room temperature incubation. And what we did with that is we made a small library, okay, which is not something that is commonly done. You don't commonly hear about libraries of biologics because it's so difficult from a molecular biology point of view. But we made a small library, did a screen and what came out of that screen was absolutely fascinating that the best combination partner for PD-1 blockade was actually PD-L1 blockade. Really interesting. I, frankly, was quite surprised...

Very surprising, yes.

To see that. And because it was so unexpected, frankly, we spent probably a solid year investigating the mechanism of action, and that drug is extremely unique. In addition to being a standard ligand and receptor blocker, it is unequivocally a T cell engager. In addition, it does something incredibly interesting, which is the bispecific exposes a cleavage site on PD-1, leading to the cleavage of PD-1 off of effector T cells. So this drug converts PD-1 positive T cells into PD-1 negative T cells. It's an incredibly unique mechanism of action.

Three things going on at once really.

Yes. So we took it into a Phase I study last year, a standard '3+3' design, 5 different dose levels ranging from 0.1 mg per kg to 10 mgs per kg. And we had no dose-limiting toxicities observed at any dose level. suggesting that we could actually have a differentiated safety profile because this drug might be anchored in the tumor microenvironment via PD-L1 to provide local high concentration blockade of PD-1.

But you're not -- so this is a 1x1...

No, it's a 2x2 valency. That's very important.

And even so, even at 10 mgs per kg, I would expect you to be well below doses where you were saturating these receptor systems.

Most PD-L1 -- so most PD-1-based drugs are dosed in the 1 to 3 mgs per kg range. Most PD-L1 drugs are in the 10 mgs per kg range. So we believe at 10 mgs per kg based on that we are saturating. So fascinatingly -- and again, I count everybody when I report the data. So 15 patients because we had no DLTs, it's only 3 patients at each of the 5 dose levels. So 15 total, including the low doses. Among those 15 patients, we already have 3 confirmed responses. So we've got a near CR in a patient with metastatic non-small cell lung cancer treated in the fourth-line setting.

And these are all PD-1 exposed post-checkpoints...

These patients are all post-checkpoint patients, a super important point. That patient, by the way, had squamous histology with low tumor mutational burden, just a fascinating observation. Our most important development at the company, in my opinion, in the past 6 months is a patient at the 3 mg per kg dosing level with triple-negative breast cancer. This patient relapsed while receiving adjuvant KEYTRUDA. So technically, this patient is refractory, exactly, then got Trodelvy and 2 salvage chemotherapy regimens, 9 centimeters approximately of linear tumor burden at baseline, which has completely disappeared, now durable through week 16. This patient could convert into a CR, started out with 87 millimeters of tumor burden, now down to about 5 millimeters. So it's just an incredible observation. Based on that, we're going to do cohort expansions in patients with non-small cell lung and triple-negative breast. And at the last -- at the final dosing level, we have another response in a third indication. That was just confirmed last week, and we'll probably announce that indication early in the coming year and potentially move directly to a Phase II study there. So we're incredibly excited about that drug. And if we have discovered a next-generation checkpoint inhibitor, obviously, that would be transformational for the company.

Yes, those are really very interesting results and obviously, a very unique molecule there. In our last minute or 2, I would -- well, I do want to ask you a follow-up question about that PD-1xVEGF that you mentioned, you said you expect this to be differentiated. Obviously, we've seen a ton of these come out of the woodwork in the past year or 2. What makes yours different?

Definitely, I -- the premise of your question, I think, is unassailable. We've got a ton of these out there now. So one of the things that makes ours different is we have better PD-1 blockade than other molecules that are in development. And I -- this is my own opinion, I don't believe these drugs are going to be able to be differentiated based on the VEGF end of the molecule. It's possible for sure. But I think VEGF capture is quite -- is not is not really that challenging in a therapeutic. So I think these drugs are going to be differentiated on the PD-1 end.

Are you using a similar epitope to the PD-1, PD-L1 bispecific or something?

Yes, yes, that's important. So the PD-1 end of our PD-1xVEGF came from the PD-1, PD-L1 bispecific.

Do you observe the same PD-1 cleavage activity in the VEGF bispecific?

Don't know yet. I don't know yet. So that is a fascinating question. I don't know the answer to that yet. So we are fivefold more potent than other drugs in the class. We presented data at SITC a couple of weeks ago that in head-to-head preclinical studies, take those for what they are. In head-to-head preclinical studies, we're superior to drugs like ivonescimab. So we're going to go into our Phase I indications are going to be gastric cancer, hepatocellular, renal cell and endometrial cancer, 4 indications where both PD-1 blockade and VEGF signaling disruption are effective either as monotherapies or in combination. So those are going to be our 4 Phase I indications, which I think should also provide a way for us to differentiate the development of that drug.

Fabulous. Very active '26 coming up for you guys. We didn't even get a chance to talk about the CD137.

Next time.

Next time. Thank you so much for joining us, Thomas.

Sure. Thanks, [ John ].