All right, we're going to go ahead and get started. I'm Stephen Willey, one of the senior biotech analysts here at Stifel and glad to have with us for the next presentation the CEO of Compass Therapeutics, Tom Schuetz. Tom, you want to make any introductory comments before we jump into Q & A and hammer you with some really probing questions?
Sure, I'm looking forward to it. Steve? No, thank you for inviting us here. Always happy to attend these events. Very, very much appreciate the invitation.
Okay, perfect. Let's start with the topic that seems to be front of mind for most folks, which is Tovecimig and the COMPANION-002 trial. You know, you're now guiding to a top line disclosure of event driven data in late 1Q 2026 that's obviously triggered by 80% death event fraction, which means that at that time, you know, you're going to have almost two years of median follow up on the. I think you'll have something close to 18 months of follow up on the last patient enrolled. These are second line biliary patients that historical data kind of consistently suggest should not be living for longer than six, seven months. It's a long lead in. The question is what is going on?
Right. I think I agree with all of the premises in that question, which is, you know, we're running a randomized trial of Tovecimig plus Paclitaxel versus Paclitaxel alone. That study is called COMPANION-002 and that is enrolling patients who have Advanced Biliary Tract Cancer who have been treated with one prior line of therapy. It is a pure second line study. You're correct, historical studies have generally shown that the Median Overall Survival in this patient population is approximately six months. The three drug combination FOLFOX in two different randomized trials had median overall survival of 6.2 months. You're correct. Also, the last patient enrolled in the study was enrolled in August of 2024. There are 168 patients enrolled in the study. The 84th patient was enrolled in the first week of March of 2024.
We recently updated our guidance, just got a little bit more granular and a little bit more specific in our guidance that we would be completing the analyses of those studies in March of 2026, late Q1 of 2026. You're correct. That's two years after the 84th patient was enrolled, 18 months after the last patient was enrolled, which frankly is a little bit even more remarkable when you think about it. You asked a very specific question. What's going on here? Whenever you see something like this, that you see fewer deaths in a study, which the other way to say that is patients are living a lot longer than we had originally modeled and expected. Whenever you see that, there are all kinds of potential explanations. Is the control arm somehow behaving in a way that we did not expect?
We actually have some information about that because in our Best Overall Response Rate Analysis, you know, 42% of the patients in the Paclitaxel Arm had already progressed by week eight. The response rate in the paclitaxel arm was 5.3%. In a couple of FOLFOX studies, the response rate to FOLFOX in one study was 4.9%, almost exactly what we are seeing with paclitaxel. In one of the randomized trials of FOLFOX, the median progression-free survival was 2.8 months, so almost exactly what looks like we are going to be tracking toward. The paclitaxel arm is behaving just like you would think FOLFOX would behave. There is no plausible conceivable explanation that there would be a delayed treatment effect of a drug like paclitaxel. That is just preposterous. The second explanation: are we missing data? You know, is our follow-up up to date?
Earlier this year we went to. This is called survival sweeps. We have monthly survival sweeps in order to monitor patient survival status. We're incredibly up to date there. Remarkably, we have incredibly few patients lost to follow up. Our total number of patients lost to follow up is only going to be something like 5%, which for a study like this is just a staggeringly low number. The other explanation is we're seeing an effect of the drug and it's hard not to look at these data month to month to month and not see a tail forming on the Kaplan Meier curve in real time. Obviously I cannot prove this, but we believe we're seeing in real time an effect of the drug on overall survival.
Okay, can you say or do you know if any patients are still actively receiving therapy?
I don't know the exact answer to that. There are a couple of patients that are still receiving active therapy, but I think it's a relatively small number.
I guess within the Tovecimig Arm specifically, I know you referenced the fraction of patients who had experienced PD as a best response in the control arm. That was, I think, 16% in the Tovecimig Arm.
Correct.
I believe in the Tovecimig Arm there were another 14% of patients who didn't get a week 8 scan.
Not a valuable or.
Not a valuable.
Yep.
For one reason or another. I guess if you think about that 30%. Right. How does that then fit into your, into this notion that we still have not yet hit the 80% death event fraction when you have 30% of patients in the tubes and beg arm who presumably didn't get maybe an ideal outcome?
Sure. Yeah. I think I sort of look at that data and I take 100 minus that number. So 70% of the patients appear to be deriving some benefit from the drug, which is just a stunning number. If that's true, that that reflects 70% of the patient population are deriving benefit from the drug and now we're seeing fewer deaths in the study. You know, suddenly all of those facts are lining up to suggest that we're seeing a real treatment effect from the drug.
Okay, I know you have this Hierarchical Analysis. I believe it's PFS, OS, DOR, they're all triggered by these OS events that we're talking about. For OS, you've pre-specified this rank preserved analysis in an effort to adjust for crossover. Will you also be showing just a conventional KM analysis of survival that doesn't necessarily do that crossover adjustment?
Okay, so just briefly, just some background on your question. The primary endpoint of the study was overall response rate and we announced that analysis in April. That analysis was statistically significant. We more than tripled the response rate in the control arm. The study by definition is already positive. There are a couple of different ways to control for statistical error. One of them is to so-called split the alpha between various analyses. Another way to do it is to do something called hierarchical testing where you do the analyses in a specific order and if an analysis is statistically significant, you can then roll over the alpha error a full 0.05 into the next analysis. You're correct. The order of analysis of the secondary endpoints is progression-free survival.
If that's positive, 0.05 goes to overall survival and then to duration of response. The pre specified, and by the way, when I say pre specified, this was in the statistical analysis plan that we submitted to the FDA in October of 2024, so more than a year ago, was the rank preserving structural failure time. To your specific question, the intent to treat analysis is a sensitivity analysis of the overall survival. Yes, to your specific question, we will be presenting that analysis as well with our top line data in late Q1.
Okay, interesting. What else should we expect to see within the framework of that disclosure?
We have prioritized these analyses. PFS, OS by the rank preserving structural failure time and the intent-to-treat analysis, top-line safety and demographics.
Okay.
The rest of the analysis of all of the study, which is obviously a mountain of data, would then be presented at a subsequent medical meeting.
Okay, I'm guessing March still gets you into the potential late breaker for ASCO.
I happen to know these dates off the top of my head. The abstract deadline for ASCO is January 27, which includes late breakers. You need to populate late breaking abstracts by March 9. Okay, I do not know today.
Got it. I know there's some. So Tovecimig 's a. It's a Bispecific DLL4 VEGF-A. I think the VEGF-A part of it is very well understood. The DLL4 is maybe a little bit less well understood. I think there are some kind of characteristic DLL4-specific Adverse Events of specific interest related to Pulmonary Hypertension, Hemorrhage, Cardiac Dysfunction, et cetera. Are you seeing safety data on a blinded basis? Is there anything that you can say directionally about those AEs of special, of special interest that may be related to DLL4? Is it just too early to say anything?
Yeah, I think it's too early to say. You know, I also don't believe we would ever be able to split the two ends of the molecule and, you know, determine what is what. So I don't think, I don't think we can do that. I think it's too early to tell. I, I just one thing. The study had a Data Safety Monitoring Committee which met four times during the study and compared the DMC, compared arm-specific safety data and did not come back to us with any concerns.
Okay. Plans to pursue breakthrough therapy.
Sure.
What, if anything, is rate limiting there?
I don't think anything's rate limiting there. That'd be a great thing to do. Once we have these analyses completed, how.
Are you thinking about the total addressable market opportunity in second line BTC? What does a commercialization footprint need to look like for Compass?
Sure. I think Biliary Tract Cancer is a much more common cancer than I think people appreciate. Earlier this year we did a third party market research project looking into the epidemiology. As part of that we did a claims based analysis, diagnoses in the U.S. Ballpark, about 25,000 patients diagnosed annually, which is consistent with the U.S. SEER data, Surveillance, Epidemiology, and End Results data, which indicates about 23,000 patients annually. Komodo Health, the claims aggregator, presented data at the Cholangiocarcinoma Foundation meeting in 2023 and they got to just over 23,000 patients annually in a two year span, 2021 and 2022. About 25,000 patients annually, about 15-20% of those patients are eligible for a Targeted therapy.
In our market research and calculations of the addressable market in the second line setting, we've actually excluded all those patients, which I think is pretty conservative actually because for FGFR2 inhibitors, ballpark response rates are around 30%, which means 70% of those patients are not served well by those drugs. IDH1 inhibitors are less than that, her two targeted drugs a little bit more than that. There is still a fair amount of those patients that I think would be eligible. We ultimately, making some adjustments for the number of patients that would make it to second line therapy, which is based on some of the data from front line studies, get to about 15,000 patients annually. Now that's second line BTC in the United States alone.
It's a very significant market opportunity, easily well north of a billion US dollars annually, something like three times larger than the platinum resistant ovarian cancer market.
Do you think you would possibly gain traction from a commercial perspective in some of these? I guess there would be third line patients who have failed some kind of targeted therapy. I mean I would have to imagine that maybe this combination, if it shows data like we think that it could, would be an interesting salvage regimen for those second.
I completely agree, I completely agree. And you know, we also. One of the things we also learned in our market research is this patient population is relatively concentrated in academic medical centers, so. And the Cholangiocarcinoma Foundation, which is the major patient advocacy organization, has also done that analysis. I think part of the reason for that I believe is that these patients have anatomic complications like, you know, blockage of their biliary tract, you know, biliary stents that require changing. That is just not the kind of thing that is done at community hospitals. Many of these patients are seen at academic medical centers. We believe with our relatively targeted sales and marketing footprint in the United States, we can launch this drug ourselves into second line BTC. We are preparing to do so.
Okay, how quickly after 002 data disclosure, would you look to initiate a frontline trial? I know you have the MD Anderson sponsored frontline trial that's ongoing. Would you need to see some of that data before you settled on a strategy?
You know, I think I'd love to just simply see the Tolerability data. Right. Because I think one of the really, I think good things about Avastin is, you know, I think Avastin basically can be combined with almost any Chemotherapy regimen. You know, in Colorectal cancer, Avastin is combined with FOLFOX, FOLFIRI, Lonsurf. That's five different drugs right there. Combined with Paclitaxel, Checkpoint inhibitors, et cetera. I think ultimately we'd like to believe that Tovecimig can be combined with many different regimens. Currently at MD Anderson, frontline studies ongoing there, adding Tovecimig to the frontline regimen of GemCis and Durvalumab, one of the two standard frontline regimens. Would love to just check the box. That combination is well tolerated. Seeing that data, I think we would then prepare full on to go into a frontline study.
Is there an opportunity for you to maybe exclude the Durva or Pembro part of it and maybe think about one of the other earlier stage pipeline assets?
That you have now. Now you're talking, buddy. Yeah, that is a great idea. You know, we have a PD-1, PD-L1 Bispecific Antibody that we call 8371. We have recently disclosed that in the first 15 patients enrolled in a phase one study, including the first in human dose, we have three responses in three different indications and no dose limiting toxicities observed at any dose level. I realize fully that we have a small data set with that drug, but we have the best phase one Checkpoint inhibitor data ever, period. There's not even anybody that's close. Could we swap 8371 for durvalumab? Wouldn't that be amazing? That would be a very cool study to do. GemCis, Tovecimig 8371 versus GemCis Durvalumab. Let's go.
All right, maybe you can talk a little bit about what makes 8371 so mechanistically unique, at least relative to all the other Checkpoint inhibitors.
Yeah, I love that phrase, honestly, mechanistically unique because I think that's very accurate. Standard Checkpoint inhibitors, either a PD-1 blocker or a PD-L1 blocker, will bind to the respective ligand or receptor and block signaling between PD-1 and PD-L1. When we were first thinking about developing a next-gen Checkpoint inhibitor, we needed to first develop a technology. We developed a screening tool that we call Stitchmabs, which is a tool that allows us to combine monoclonal antibody fragments together in a very quick and efficient manner. We used that technology to build a small library. Now, look, when I use the word library, I'm not talking about a million small molecules in a library. You know, this is a small library of biologics. We screen that library.
The combination of PD-1 and PD-L1 blockade was uniquely synergistic. Because it was so unique, we spent a long time investigating the mechanism of action there. It is not just a dual ligand and receptor blocker, which it is. It is also clearly a cell engager. It brings effector T cells or NK cells into proximity with a PD-L1 positive tumor. I think the most unique aspect of this drug is it exposes a metalloproteinase cleavage site on PD-1 and induces the cleavage of PD-1 off the surface of effector T cells. It actually converts PD-1 positive T cells into PD-1 negative T cells. It is a really, really unique mechanism there. Finally, PD-L1. Gordon Freeman at Dana-Farber showed this in his lab.
PD-L1 inhibits CD80 in cysts, and our PD-L1 blocking antibody frees CD80 from cyst suppression by PD-L1 in order to agonize CD28. All those things combined creates a very unique drug. Frankly, to your specific point about a unique mechanism, I think 8371 is something a heck of a lot more than just a Checkpoint inhibitor. It is really a next generation immunomodulatory drug that started out on a path to identifying a next gen Checkpoint inhibitor. It is so much more than that. Amazingly, we are seeing that in the clinic.
Are you able to measure this PD-1 cleavage? Is that something that you can actually detect in the peripheral blood?
In mice and monkeys. In mice, we can definitely do it. In monkeys, we were able to observe the loss of PD-1 from the surface of effector T cells. We had a little bit harder time detecting the fragment. We'll see in our human studies. We've got a little bit more work to do there, but we have not yet been able to detect it.
Okay, so you said dose escalation's complete. What are the next steps and I guess how much additional data should we expect to see in conjunction with the next update you have planned?
Sure. The immediate next step is, two of the responses that we saw in the dose escalation were in patients with triple negative breast cancer and non small cell lung cancer. We are now, this quarter, initiating cohort expansions at two different doses in patients with those two diseases. We are going to enroll 56 more patients, 28 in each indication, and within each indication, two different dose levels, 3 mg per kg and 10 mg per kg, the two highest dose levels. That is ready to go and we will begin enrolling patients. I would love it if we could present the dose escalation data along with some early cohort expansion data at an important medical meeting in the first half of next year.
Now, last week we also disclosed that at the highest dose level, the 10 milligrams per kilogram dose level, we have a new response in a third indication. As I'm sure you know, PD-1, PD-L1 blockers are approved in something like 20 different indications. We have treated patients with five different tumor types and we have three different responses. We are going down that path right now. We have not disclosed the third tumor type. Not to be obtuse about it, but just waiting for that, waiting for another scan in that patient.
We're not going to include those patients in the cohort expansions because the observation that we have so far is very unique and could potentially lead straight to a phase two study, maybe depending on how the data plays out in that patient and others, of course, frankly, a conversation about going straight to approval in that indication.
Interesting. Okay. Again, these are responses that you are seeing that are A, in checkpoint-experienced patients and B, are, I believe, on the order of magnitude of at least the first two, 80-90% reductions in baseline tumor burden.
Two very important points. Thank you. The patient population is all post Checkpoint inhibitor, so that's incredibly unique. The depth of responses that we've seen in these patients is really remarkable. The patient with triple negative breast cancer started out with nearly 9 cm of linear tumor burden and has had greater than a 90% decline in her linear tumor burden. Now we recently disclosed, also now with some, the beginning of some durability and recently some additional deepening of that response. That's now straight up a near CR. The non-small cell lung cancer patient also had a significant decline in linear tumor burden, also, you know, quite large tumor burden. You know that, you know, these are not small little 1 cm tumors that are going to 7 millimeters. We're like calling that a PR. No, we're seeing something much more different.
Okay. Maybe sticking on the earlier stage pipeline, you've thrown your hat into the VEGF-A PD-1 ring.
Yes.
We just saw some preclinical data for 10726. I know you're guiding to an IND submission before the end of this year. Maybe you can just kind of highlight what you just showed us at CITC. And you know, how that either how that informs, how you think this molecule could be different relative to some of the other iterations that we've already seen some preclinical data for and clinical data for it.
Sure. A couple things. I think what we presented at CITC last week was just the completion of a full, you know, preclinical characterization package for that drug. You know, just a reflection of sort of a very thorough, you know, drug development approach. I was very, very pleased with that. We reported some in vitro data where we showed that we have more potent PD-1 blockade and more potent PD-1 binding than you see with other drugs in the class. That observation led us to do some head to head in vivo studies in mouse models, which, you know, I take that for what it is. You know, these are mouse models, but in a non-small cell lung cancer xenograft model, our drug CTX-10726 was superior to ivanezumab and bevacizumab at controlling growth of that tumor.
Clearly we have an active drug in this class, I think, like with PD-1 blockers and PD-L1 blockers, of which I think there are now eight approved in the United States, I think. If you think about, if you go back in time to the lung cancer efforts with pembrolizumab and nivolumab, some of those differences came down to clinical trial design and indications. We're trying to think about some indications where both PD-1 blockade and VEGF blockade are effective, ideally as Monotherapies. Our phase one population are going to be patients with gastric cancer, Hepatocellular cancer, Renal cell cancer, and endometrial cell cancer. Either PD-1 or PD-L1 blockade is effective in those four indications and VEGF targeting, be it with drugs like bevacizumab or VEGF kinase inhibitors, are effective. That's our phase one population there.
I know this landscape is getting a little bit competitive. Right. I would have to imagine there's still some strategic FOMO that's lingering out there. I would imagine some of the 8,371 data you've generated to date has probably captured some eyeballs. How are you just thinking about BD as a potential mechanism for accelerating the value that you can create with these two molecules?
Sure. We have said publicly we are having some conversations around that molecule. I do think that there is some FOMO out there for sure. I think folks are thinking about not only missing out on this class of drugs, but did they miss out on the entire PD-1 wave to begin with. I think there are companies like that. I think we are having those conversations. I like the way you said it. What a great way maybe to leverage the frontline labeling of Tovecimig in BTC.
Lastly, balance sheet and just what it allows you to execute on here.
Yeah, we've been very fortunate. I have a very supportive set of high quality investments. Investors ended Q3 with about $220 million in cash, which has runway for us into 2028 executing on now all four of these clinical programs.
All right, Tom, always appreciate it. Thank you for the time.
Thanks so much, Steve.