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

I'd like to welcome everyone to day one company one of the Piper Sandler Healthcare Conference. I'd like to welcome Compass Therapeutics, their CEO, Tom Schuetz. Welcome, Tom.

Thanks, Biren. Thanks for having us at the conference. Really appreciate it.

And I was looking at your stock price year-to-date, up 290%. I mean that's quite impressive compared to XBI of 30%. So the company is clearly doing something right. Can you tell us -- maybe we start off with the two clinical stage programs. And if you could maybe provide us an overview of both programs, target indications and then we can dig in a little bit more.

Sure. Yes, I think to your first point, I think it might be related to our high-quality analyst coverage. Compass Therapeutics, we're located in Boston. We're a monoclonal antibody discovery and development company in oncology. We actually have three drugs in the clinic, two bispecific antibodies and one monoclonal antibody and a third bispecific antibody going to enter the clinic early in the coming year. So our most advanced program is a DLL4 VEGF-A bispecific antibody and that drug is called tovecimig that is currently in a randomized trial in patients with advanced biliary tract cancer. Earlier this year, we announced that we hit the primary endpoint in a randomized trial of overall response rate, more than tripling the response rate seen in the control arm. To my knowledge, we have one of the highest response rates ever seen in patients with advanced biliary tract cancer treated in the second-line setting. In addition, which -- and we'll get into this, I'm sure, there was a substantial difference in the disease control rate as well, suggesting that we are -- could see an important difference in progression-free survival as well. We'll come back to that point in a minute. Our second program that is currently in a clinical trial is a drug called 8371, that is a novel, potentially first-in-class PD-1, PD-L1 bispecific antibody. That came out of a screen for synergy with PD-1 blockade using a technology that our scientists at Compass developed, and our scientists did a screen for synergy with PD-1 blockade and discovered made an important discovery that PD-L1 blockade is the best combination partner for PD-1 blockade in a bispecific antibody. That drug has recently completed a dose escalation Phase I study fascinatingly. Now we saw no dose-limiting toxicities in that study. But more importantly, 15 patients treated in the post checkpoint setting, we already have 3 responses. We have 3 confirmed responses now including a patient with non-small cell lung cancer, triple negative breast cancer. And the third indication we've not yet disclosed but the patient with triple-negative breast cancer is extremely important. So this patient relapsed while receiving adjuvant KEYTRUDA. So technically speaking, this patient is actually refractory, then received TRODELVY followed by two additional chemotherapy regimens, started the study with approximately 9 centimeters of tumor burden, and that is now down to approximately 5 millimeters. So greater than a 90% decline. The patient is now durable continuing on therapy, potentially this patient could convert to a complete response, which would be absolutely amazing. And I just want to make one final point about 8371 and I know I'm repeating myself, but all of the patients were treated in the post checkpoint inhibitor setting. And we already have 3 responses in the first 15 patients treated. So I believe, of course, I'm biased, we have the best Phase I data of any checkpoint inhibitor because drugs like KEYTRUDA, Opdivo, Tecentriq, those were all tested in the checkpoint-naive patient population. This is a little bit of a subtle point but I believe this is extremely important that we're seeing monotherapy responses in the post checkpoint patient population.

That's great. So I mean, both are -- seem like a very exciting programs with a lot of potential. And then I guess maybe if we could start with Tovecimig. COMPANION-002 trial, the Phase II/III pivotal study as you mentioned, you've reported the primary endpoint and achieve that on overall response rate versus paclitaxel alone. PFS, OS end points are expected by the end of Q1. Can you tell us how we should look at the readthrough from overall response rates to PFS and OS and the confidence in the PFS and OS endpoint? And then I think also you had disclosed that the 8 weeks, there were a significantly greater number of patients that progress in the PAC arm versus the tovecimig arm. So maybe can you talk about the rate-throughs from that data point as well?

Sure. So we reported the best overall response rate data in April of this year. And of course, best overall response rate can include many different things. It can, of course, include complete responses, and we had a complete response as adjudicated by blinded independent central review, an extremely unusual finding in advanced biliary tract cancer. Partial response, stable disease, but progressive disease can also be your "best overall response". But if PD is your best overall response. What that means is, by definition, it had to occur at the first scan time point, which was at week 8. So if you had stable disease at week 8, you'd be in the SD category. So if PD was your best overall response, what that means is it had to occur at week 8. And in the combination arm, we only had 16% PD, in the combination arm, as you mentioned, we had 42%. So 100 minus those numbers is a maximum of 58% progression-free survival maximum at week 8, 84% in the combination arm. So the progression-free survival curves are probably already significantly diverging. So we are quite confident that we'll hit the PFS endpoint. And I think one of the things you asked an important question about read-through, right? And I think one of the things to keep in mind with the read-through from the response rate to the PFS and OS endpoints is the waterfall plots. So the waterfall plots from the study are fundamentally different. So in the combination arm, the majority of patients had decreases in their linear tumor burden which I alluded to earlier with the disease control rate. The disease control rate was close to double in the combination arm as compared with the control arm. So we believe that will read through directly to PFS. And then as you mentioned, both PFS and OS analyses are triggered by OS events and we reported in August, and then we reiterated in November with our Q2 and Q3 earnings releases, respectively, that we're seeing fewer deaths in the study than we initially projected. So because of that, we've recently updated at 18 months median follow-up when we still had not hit the number of total events, we were well above where you'd be expected based on historical data in this disease. So we're enthusiastic about the OS endpoint. And as you mentioned, looking forward to reporting those data toward the end of Q1 of the coming year.

And you mentioned that the OS event rate has been slower to accrue. This is a 2:1 randomized design. So clearly, with 80% events that are required for the study to unblind, you would need patients in an active arm to have an event. Is that a fair characterization? And I guess, in your view, what do you think are the reasons why the events have been slow to accrue?

Sure. So a 2-part question there. The first part of your question is yes, we need to see 80% pooled OS event in order to trigger the analysis of PFS and OS. So second part of your question, of course, is very important but also, of course, very complicated. So whenever you see something like this, you wonder whether or not the control arm is doing something that you didn't anticipate. We have any evidence to support that? Probably not because in our best overall response rate analysis, the response rate in the control arm was only 5.3%. So paclitaxel is not doing something that we didn't expect. The 3-drug combination of FOLFOX in a randomized trial had a response rate of 4.9% to [indiscernible] so we're not seeing it there. And in terms of the incidence of progressive disease with 42% progression at week 8 and the median PFS in the control arm is probably going to be something like 2 to 2.5 months. It's going to be in that range. I don't know what it's going to be. But based on our overall response rate data, it suggests that it's in that range. So the control arm is not doing something unanticipated. I know for sure that we are not missing data. We've had extremely low numbers of patients lost to follow up in this study. And I think our clinical operations group at the company is just doing a phenomenal job here. We've only lost something like 5% of patients to follow up, which is just staggering low number for the study. So -- and I know today that every patient, except for one single patient that is known to be alive today was known to be alive in either October or November. So we're very up to date on data collection, which then leaves the third possibility that we're seeing a drug effect which is, of course, what we believe. We believe that what we're seeing is fewer patients have died, more patients are alive for a longer period of time because tovecimig is affecting overall survival. That's what we believe.

And in the paclitaxel control arm, there is an option for those patients to cross over to receive to tovecimig. Can you talk a little bit about the percentage of patients that have crossed over and have received tovecimig in the trial? So that's the first question. And the second question is, how are you going to account for those patients that cross over onto tovecimig in your OS analysis?

Okay. So the first part of your question, we know that about half the patients approximately half the patients in the control arm crossed over to receive active drug, which also could be contributing perhaps not to the fewer number of OS events that we're seeing. We are accounting for crossover using a statistical technique called the rank-preserving structural failure time, RPSFT. This is a statistical analysis that adjusts the Kaplan-Meier overall survival curves for crossover. And this statistical technique was actually used in biliary tract cancer in a randomized trial that led to FDA approval in the study called ClarIDHy, C-L-A-R-I-D-H-Y, ClarIDHy which was a study of the IDH1 inhibitor, TIBSOVO versus placebo. That also allowed a crossover in that study a little bit more than half, maybe closer to 2/3 of patients actually crossed over and they use the RPSFT to analyze their overall survival data. And it is my understanding, by the way, I don't -- I can't -- I don't know if this is true or not. But it is my understanding from talking to folks who are at Agios at the time that the RPSFT was recommended to Agios by the FDA. So that's what I was told. I assume that's true. So with the RPSFT, the adjusted hazard ratio for OS in the ClarIDHy study was 0.49, and that got that drug labeled in second-line BTC. You could not have a more perfect regulatory precedent for us.

And so this is going to be your primary OS analysis, it's going to be the -- on the rank-preserving structural failure time...

Yes.

And so can you talk a little bit about the stats hierarchy after ORR? What's the next endpoint that you're going to look at? Is it PFS? Or is it OS?

PFS. So the stat hierarchy is important because we're using a technique called hierarchical testing to control for alpha spend. So there are a couple of ways to do that, one commonly used way is to "split" the alpha, but hierarchical testing has been shown to control alpha spending better than alpha splitting. So the way that hierarchical testing works is you do the primary endpoint analysis, if that's positive, which it was with a p-value of 0.031, all of the alpha era then rolls over to the next analysis, which is PFS. If PFS is positive, a full 0.05 alpha rolls over to the OS analysis, then to the DOR analysis.

Got it. And then -- how do you think -- you talked about ClarIDHy and how FDA looked at the rank-preserving structural failure model. Have you had discussions with FDA or other regulatory experts on the acceptance of this model, given the draft guidance on the OS endpoint that was issued by FDA earlier this year?

So in that draft guidance, I think it's fairly clear in that guidance that in indications of tremendous unmet need that crossover is acceptable on to FDA and second-line biliary tract cancer in the United States outside of patients who have tumors with actionable mutations, which unfortunately is a small fraction of the patient population, there is literally no labeled therapy in the United States. So there's -- it's the large -- it's an unimaginable unmet need with no labeled therapy. So in no scenarios, crossover designs are acceptable. And so the RPSFT was in the protocol from the beginning and our statistical analysis plan was submitted well over a year ago, just after the study was fully enrolled, so way, way ahead of time. And FDA had no comments on that.

Great. And so the study is on track for completion in Q1 in terms of the PFS, OS readout. After that readout, what are the next steps on filing the BLA? Should we expect that to occur, I guess, in the second half of the year?

So I think the next step after the PFS and OS readouts would be to complete the full analysis of every endpoint in the study and then have a meaningful interaction with the FDA. So presumably, that interaction would occur in Q2 and which would put us in a position if that interaction went well to submit a license application in the second half of the year. We have fast track designation, so we would certainly get a priority review, which should put our PDUFA date inside the first half of 2027. So we could be -- today, we could potentially be within call it, 18 months of launching this drug in the United States.

That's great. And what's the market potential in second-line biliary tract cancer in the U.S.?

Sure. So biliary tract cancer is much more common than you might think. So about 25,000 patients annually are diagnosed with biliary tract cancer, and there was a recent JAMA oncology article indicating that the incidence of biliary tract cancer is increasing. It is by -- over the next decade, it's projected to become one of the most common malignancies in the country. So 25,000 patients annually, some of those patients are eligible for surgery, unfortunately, only about 10% of patients get surgery. About half of those patients relapse. So we did some third-party market research earlier this year that was commissioned by -- we commissioned a well-known firm to do that market research. 70 physician payer interviews and surveys. And based on all of that work, which is a very large project, we believe, about 15,000 patients annually in the United States alone would be eligible for this therapy. So it's about 3x larger than the platinum-resistant ovarian cancer market. And then if you think about the rest of the world, Japan, which has a smaller population, of course, than the U.S., but the incidence of BTC in Japan is higher. And the total number of patients in Japan is probably close to the total number of patients in the U.S. and the EU, of course, just about twice the size of the United States. So there are probably well over 100,000 patients diagnosed annually just between EU, Japan and U.S. So it's a very, very large market.

And how does the company think about the ex U.S. opportunity? Would you commercialize it yourself? Or would you potentially seek a partner?

So I think in the United States, we're preparing to commercialize this drug ourselves. I think biliary tract cancer is an indication where patients are mostly seen at academic centers. so a small targeted sales and marketing force, I think we can launch this drug ourselves and we are preparing to do so. I will have more to say about that in terms of our commercial infrastructure in the beginning of the coming year. I think one of our corporate goals for 2026 will be to initiate regulatory interactions in Japan and the EU. Specific question, we have said publicly that we're having some conversations with potential strategic partners about ex U.S. geographies. Of course, we'll do what's in the best interest of our shareholders. So we'll see.

And we've got a few minutes, I do want to touch on 8371. So there are 5 dose cohorts that you value in the dose escalation. In the next phase of the study, I believe you're going to evaluate two of those cohorts. Have you decided which two cohorts, which two doses you're going to evaluate for the next phase of the study?

Yes. Because we've seen this truly, again, in my view, a remarkable efficacy signal in Phase I dose escalation. We now are going to move to cohort expansions. So we're going to enroll patients with non-small cell lung cancer and triple-negative breast cancer into two cohorts of 28 patients each, so a total of 56 patients. Within those 28 patients within each indication will be randomized to one of two doses, 3 mg per kg or 10 mg per kg. So we've decided on the two doses just based on what we've seen early on and enrolling those patients to do some small project [ optimus ] work in terms of dose finding, identify a Phase II dose, confirm the efficacy signal we've seen in dose escalation. And I think non-small cell lung cancer, of course, is a more complicated indication for sure. But triple-negative breast cancer is not. So post TRODELVY, if we're seeing responses in that patient population, I'm quite confident that, that would be a single arm response rate study to approval. And in the first half of next year, if we confirm the efficacy signal in patients with triple-negative breast cancer, we're going to move directly to a potential approval study next year. The third indication, and by the way, we enrolled patients with 1 of 5 tumor types into the study. In addition to non-small cell lung cancer and triple-negative breast cancer, we enrolled patients with melanoma, head and neck cancer and Hodgkin's lymphoma. So we've had a third response at the highest dose level in one of those other three indications which we have not yet disclosed waiting for some more data from that patient. But that indication is very, very interesting, and we're now working with some KOLs on potentially going straight to a Phase II study that also, again, in the post ADC, post-PD-1 patient population would be a potential approval study. So I think 2026 is going to be an incredibly exciting year for 8371 with moving to two potential approval trials with that drug next year. So that's a drug we're incredibly excited about. And I think if we have discovered the first next-generation checkpoint inhibitor, we'll be off to the races.

That's great. I mean so a lot of exciting updates over the next 12 months. But you also have a third program that you're potentially moving into the clinic with the PD-1 VEGF. And I think the IND is expected to be filed soon. How does that fit in given the excitement of the first two programs? Do you plan to carry that on yourselves? Or would you potentially partner it at some point?

Sure. So our fourth program, as you mentioned, a drug we call 10726, a PD-1 VEGF bispecific antibody. This is now a new class of drug. Many companies are now working on this class of drugs, and we have a very novel drug candidate, and we presented some data last month at SITC, suggesting that in certain preclinical models, our drug is superior to ivonescimab. So you're correct, IND filing followed by Phase I initiation in the first half of next year. Our Phase I population, we've defined as patients with hepatocellular cancer, gastric cancer, renal cell cancer and endometrial cell cancer, four different indications where both PD-1 blockade and VEGF blockade are effective. So our plan is to initiate that Phase I study and get clinical data next year. But we have stated publicly that there's tremendous interest not only in this field, but in this asset. And again, what we do with this drug moving forward, we'll do what's best for our shareholders.

Perfect. I think we're about out of time. Really appreciate you coming to Piper Sandler Healthcare conference and sharing insights on the story.

Thanks, Biren, and thanks so much for having us.