Everyone, my name is Maury Raycroft, and I'm one of the biotech analysts at Jefferies. It's with great pleasure that I'd like to welcome Tom Schuetz, the CEO of Compass Therapeutics. We're going to do fireside chat format, so I'll pass it over to Tom to provide an intro to the company and programs, and then we'll switch over to a fireside. Tom, I'll let you start off.
Thanks, Maury, thanks for inviting us to present at the conference. I really appreciate it. I'll go through a few slides and then, as Maury mentioned, we'll have some Q&A following a brief presentation. Just one quick reminder, I will be making forward-looking statements today, and I will refer you to our regulatory filings for descriptions of those. Compass is located in Boston. We're a monoclonal antibody discovery and development company in oncology. We currently have four drugs in the clinic. Our most advanced asset is tovecimig, a DLL4 VEGF bispecific antibody. We recently read out the results of secondary endpoints for that study. The study hit the primary endpoint of overall response rate and hit the key secondary endpoint of progression-free survival. Next steps there will be a meeting with FDA in advance of preparation for the submission of a Biologics License Application.
Later this year, we're going to initiate a phase II study for our second program, a next gen CD137 agonist in patients with NCAM-positive tumors. We also have a PD-1/PD-L1 bispecific antibody, potentially a first-in-class asset. We presented data at ASCO this past weekend in which we showed that we have confirmed responses in patients with triple-negative breast cancer and Hodgkin's lymphoma, as well as non-small cell lung cancer in a phase I post PD-1 study. We also have a PD-1 VEGF-A bispecific that has recently entered phase I testing. Let's talk about tovecimig, and we'll talk more about this, of course, in the Q&A. We recently completed a randomized study of tovecimig plus paclitaxel versus paclitaxel alone in patients with advanced biliary tract cancer who had received one prior line of therapy.
This was a two-to-one randomization, 111 patients randomized to the combination arm, 57 patients randomized to the control arm. Crossover from the control arm was allowed following centrally confirmed progression, a very large number of patients crossed over to the active arm, 54% of patients. 142 or 85% of patients in the study received tovecimig at one point in the study, either at randomization or upon crossover from the paclitaxel control arm. As I mentioned, we hit the primary endpoint in the study of overall response rate. Importantly, all responses in this study were assessed by blinded independent central review. The overall response rate in the paclitaxel arm was 5.3%, very consistent with what you see with three-drug chemotherapy regimens in the second-line setting. FOLFOX, for example, in a randomized study, had a response rate of 4.9%.
We more than tripled the response rate by adding tovecimig, that was statistically significant. In the bottom of this slide, we also had a significant improvement in disease control rate, 61.3% in the combination arm, 36.8% in the pac-only arm with a p-value of 0.0027. That disease control rate translated into a significant improvement in progression-free survival. Here are the PFS curves for the study. There was a very significant improvement in PFS with a hazard ratio of 0.44 with a p-value less than 0.0001. A very important improvement here. I would argue that this is a clinical endpoint because in this disease, where complications of biliary tract obstruction, any extension or prevention of progression in this study is very clinically meaningful. Recall I mentioned to you that more than half the patients in the control arm crossed over to receive tovecimig.
In the intention-to-treat analysis, there was no difference in overall survival. Remember, this is sort of a confounded analysis because the patients who crossed over to receive tovecimig are allocated to the control arm here, and more than half the patients in the control arm are actually not, quote, unquote, "control." One of the prospective secondary endpoints in the study was something we called PFS2. This also indicates that tovecimig has significant treatment effect in the study. PFS2 is an analysis of the patients who crossed over, comparing their progression-free survival on paclitaxel alone to their subsequent progression-free survival on tovecimig plus paclitaxel. The patients who crossed over, their paclitaxel median PFS was 1.9 months. When adding tovecimig, that improved to 3.5 months. A very significant difference simply with the addition of tovecimig plus paclitaxel.
When we saw PFS2, that led us to do a subset analysis where we looked at the patients in the control arm specifically. This slide shows the overall survival analysis of patients who crossed over to receive tovecimig versus those who did not. You can see from this slide, the patients who crossed over to receive tovecimig did extraordinarily well in this study with a median overall survival of 12.8 months, which is actually what you see in the frontline setting. These crossover patients, of course, were treated in the second-line setting with paclitaxel and then tovecimig/paclitaxel in the third-line setting, and their median overall survival was similar to what you see in the frontline setting. The patients who received paclitaxel only had a median overall survival of 6.1 months.
If you got tovecimig at any point in the study, either at randomization or at crossover, the median overall survival of those patients was 9.9 months compared to 6.1 with paclitaxel alone. Obviously, with an analysis like this, you can question whether or not these patients would have done better anyways, but they did not. Keep an eye on the red versus blue line on this slide. When you look at PFS, it actually reverses. The patients who crossed over actually did worse with paclitaxel than patients who didn't cross over. Interestingly, these patients also had their performance status deteriorate while on paclitaxel. So on the last slide, what we're seeing here is a clear effect of tovecimig following crossover. On the safety side, most common AEs that we saw on mechanism for VEGF blockade were hypertension. Obviously, the bone marrow suppression here is related to paclitaxel.
Here's a summary of the COMPANION-002 study. Met the primary endpoint with a significant improvement in overall response rate. We had a very significant improvement in progression-free survival with a hazard ratio of 0.44, with an extremely low p-value, representing a 56% reduction in the risk of progression associated with tovecimig. The OS ITT analysis was confounded by crossover. We know those crossover patients drove the control arm ITT analysis because those patients lived a median of 12.8 months, again, clearly indicating that tovecimig has had an effect on overall survival. Next steps here are to meet with FDA and discuss these data in advance of a Biologics License Application. Clearly, when you compare what we saw with tovecimig paclitaxel to regimens that are used in this disease, we clearly are seeing something different.
In the bottom of this slide, FOLFOX/ FOLFIRI, 2.8-month median PFS. We saw 4.7 months, 6.2-month median overall survival. Again, spot on for what we saw with paclitaxel. When you got tovecimig in this study, your median OS was 9.9 months. In the last minute or so, I'll just talk briefly about 8371. 8371 is a novel PD-1, PD-L1 bispecific antibody that emerged from a screen we did for synergy with PD-1 blockade using a technology we developed at Compass called StitchMabs . We recently completed a dose escalation phase I study. This was a 3+3 design at five different dose levels. We had no dose-limiting toxicities in the study, we presented at ASCO this past weekend. We have two at the top dose, two dose levels. Out of six patients treated, we have two very important responses.
These are PET scans from a patient with Hodgkin's lymphoma. You can see the PET-positive tumors circled in blue here that nearly completely disappear. This patient on the bottom was post-transplant and post-nivo. This patient continues on study more than 10 months out. Perhaps the most important patient in the study is this patient. This is a patient with metastatic triple-negative breast cancer who relapsed while receiving adjuvant KEYTRUDA. Patient then received TRODELVY, followed by two additional chemotherapy regimens. This patient had about nine centimeters of linear tumor burden at baseline, two-centimeter lung metastasis on the top that disappeared. On the bottom, a 5.2-centimeter pericardial metastasis that has completely disappeared. This patient we presented at ASCO was in a durable PR at week 48. I can tell you today that this patient actually is now out more than 13 months with a durable PR and continues on therapy.
In summary, four drugs in the clinic. Lots of upcoming milestones this year. FDA meeting for tovecimig. We're also going to start some additional phase II studies with tovecimig, probably combining that with paclitaxel, because we believe we might have discovered something interesting with the combination of tovecimig and paclitaxel. NCAM positive biomarker study for 471. For our PD-1, PD-L1 bispecific, we've initiated cohort expansions in the phase I study in patients with non-small cell lung cancer, triple-negative breast cancer, and Hodgkin's lymphoma. That study's enrolling extremely well. We expect to report cohort expansion data in the second half of this year. Finally, our PD-1 VEGF-A bispecific antibody has initiated phase I testing, and we should be in a position to have some preliminary data from that study later this year.
Great. Thanks, Tom. That was a good highlight and summary of the company. Maybe starting off with tovecimig, which you talked a lot about there. You're definitely seeing a signal in the second-line BTC study. There's been debate around the crossover subgroup, where the crossover patients appear to do better than those initially randomized to active, and so the 12.8 months versus 8.9 months. As you're doing additional work on the data, I guess, what are some observations or nuances about that finding? Is it related to baseline imbalances, timing of crossover, subsequent therapies?
Sure. Thanks again, Maury. Yeah, I agree with the way you frame the question. We unequivocally have evidence here that the drug is working. We have tripling of overall response rate and just an unequivocal and very important improvement in progression-free survival. Yes, the patients who crossed over from the control arm to receive tovecimig plus paclitaxel, those patients did extraordinarily well. You're correct, they did slightly better than the patients who were initially randomized to the combination. We don't know yet what the explanation for that is, but we do know what it isn't. There's no clear imbalance in prognostic factors. It is definitely not post-study therapy. One of the things is very topical, obviously, but when I first saw that, I wondered if those patients all went on to receive RAS inhibitors, right?
That is absolutely not the case. There are more patients with intrahepatic cholangiocarcinoma that crossed over. Again, given the fact that those patients did worse with paclitaxel, I'm not sure what to make of that. We're left with sort of two explanations, right? One, rather unsatisfying, which is small n, right?
Yeah.
It's 31 patients. Something much more important, which is does paclitaxel prime the tumor microenvironment to make tovecimig more effective? That's a very provocative question, and we've started some preclinical work to begin to try to address that question because it's an important question, because it would influence our plans for studying tovecimig in multiple other indications.
Yeah. Yeah, it's an interesting one, definitely worth looking into. For next steps with FDA, I'm not sure if you have any more granularity on the slide up there, but just walk through the next steps for the timing.
Sure
for FDA interactions, what type of meeting you want to have, and the key topics you expect to address there.
Sure. The meeting will be in approximately the first half of August. Backing out from that, approximately July 4th-ish, we would submit the full briefing package to FDA. Then going forward from early to mid-August, approximately Labor Day, we would expect to have minutes from that meeting that we would then, of course, discuss publicly. The real goal of that meeting is to align on a pathway for submission of a Biologics License Application for tovecimig. If we align on that, we could begin the BLA submission process later this year. We would look to do a rolling submission and initiate that this year, finish that in the first quarter of next year, which should put our PDUFA date for tovecimig in the second half of next year because we have a Fast Track Designation.
In this indication that has no labeled therapies for patients without an actionable mutation, I'd be shocked if we didn't get a priority review.
Yeah. Okay. Yeah, it makes sense. Given you hit stat sig on ORR and PFS, do you think that can be sufficient from FDA standpoint, even if the ITT OS signal's confounded by the crossover?
Sure. You just described something like 80% of oncology drug approvals in the United States over the past 15 years or so. We're squarely in the middle of the fairway, as it were, hitting ORR, hitting PFS. Again, I think this is a very important point. Unlike many other diseases where you can wonder about the clinical significance of altering progression, in biliary tract cancer, these patients have anatomic complications in the biliary tree. When these patients get obstruction of their biliary tree, frankly, that is fatal. Halting progression in this disease and potentially delaying the time to biliary obstruction is, I would argue strongly, is a clinical endpoint.
We believe that this could support full approval, and I think importantly, at the end of last year, there was a large meta-analysis published, which showed that the best predictor of OS in studies of patients with biliary tract cancer is actually PFS.
Got it. Okay.
That's a very important paper.
Yeah. Okay. That's helpful. Going into the FDA interaction, what scenarios are you planning around? Do you think FDA could require a confirmatory study? If so, what would your views be on that?
Sure.
Would you do first line or second line?
Sure. You're asking whether or not the data support full approval or accelerated approval? Both of those are approval.
Yeah.
I think with accelerated approval, I think if we had to do a confirmatory study, one potential study design that you alluded to would be a frontline study. Adding tovecimig to GEM/CIS/DURVA, which is a single-arm study that's ongoing now at MD Anderson. We're seeing some very interesting things from that study that make us very confident that we could add tovecimig to that regimen. Obviously, that's a large randomized study that would be expensive and take a long time, but it would lead to a frontline label. Another study we could do conceivably, if asked for a confirmatory study, would be tovecimig paclitaxel versus investigator's choice of chemotherapy. That's probably a study we'd do in Europe, so that we would not have the control arm be confounded by crossover to commercial tovecimig pac.
Got it. Okay. Going back to the biliary tract obstruction, how many of those events did you see in your study? Is that something you can.
We have not disclosed that, but that's something we're looking at closely.
Okay. Got it. For a confirmatory study, if there is some sort of priming with paclitaxel, is that something you could try to incorporate into the study?
That's a very interesting question. I think we would need more scientific data.
Yeah
in order to do that.
Yeah. Makes sense. Going back to the data, if we focus only on the patients who did not cross over, OS looks at approximately 8.9 months versus 6.1 months. Do you view the 2.8 months delta as potentially supportive enough for FDA to take a positive view of the data?
Yeah. For sure. I think FOLFOX, compared to best supportive care, had a 0.9 month improvement in median overall survival, and that was thought to be good. That is a very important, what is that? 40% improvement in overall survival. Yeah, I think that would be very meaningful.
Got it. You mentioned that just in oncology you see a lot of approvals based on just ORR and PFS 80%. What's the closest analog or example for your guys' situation?
There are so many. Actually, FDA just published an analysis of, help me, the selpercatinib, one of the targeted therapies in non-small cell lung. They published that in JCO in April, that's a very important paper. ORR, PFS, clearly with a crossover and a hazard ratio of 1.26, right? FDA published a full analysis of that. FDA also published a couple of years ago, the first author in that paper is Marino. Rick Pazdur was actually a coauthor on that paper. There was five or six different approvals in that paper where the hazard ratio is above one due to crossover. Copiktra, actually, didn't have a crossover and straight up lost to one of the CD20 antibodies, ofatumumab, that drug still got approved.
Interesting. Okay. On the safety front, you reported approximately 52% Grade 3 hypertension.
Yep.
Were there any Grade 4 hypertension events, and what's the standard mitigation strategy in practice?
I think there was maybe one, I think, but I'm not certain on that. Hypertension is an on-mechanism AE associated with VEGF blockade, right? Whether it's a monoclonal antibody targeting the ligand, the receptor, VEGF kinase inhibitors, hypertension is the most common on-mechanism AE. There are now published algorithms for managing the hypertension associated with VEGF blockade.
Okay. Yeah. There's nothing surprising in the study.
No
on that front.
We use those algorithms in the study.
Got it. Okay. Are you planning any biomarker analyses to compare crossover patients versus those initially randomized to active? If so, what biomarkers are most relevant?
Sure. We don't have enough time to talk about that question. The short answer to your question is for sure, yes. We collected a whole series of specimens to assess for biomarkers, and we're going to full court press that. We had a very nice investigator meeting at ASCO this past weekend, and one of our lead PIs is going to take the lead on helping us with those biomarker analyses. We're going to do everything that everybody could possibly think of, because again, trying to figure out why those patients did so amazingly well, I think is a very important question.
Are there any hunches? Do you think it could be target expression related, maybe ctDNA?
All of those would be on the table.
Yeah. Makes sense.
We collected ctDNA, so we'll take a look at all that.
As you continue to analyze the data, I guess, what are going to be the key crossover adjustment approaches that you think could build a strong case for FDA?
Sure. You alluded to statistical methodologies. There are numerous statistical methodologies that are available as sensitivity analyses to look at crossover. We used in the prospective design. We use the RPSFT, which stands for the Rank Preserving Structural Failure Time. Unfortunately, the statistical assumptions for that methodology were not met. We couldn't use that. There are multiple others that are actually, those analyses are ongoing right now. I want those analyses to be available for our FDA meeting package. Just again, to try to help understand the effect of crossover. There's the inverse probability of censoring weights, IPCW, which is one that's commonly used. All of those analyses are ongoing right now.
Got it, you'd want to have those done by the July 4th time frame.
Yes
that you included. You wouldn't share those with the public beforehand, really wouldn't be appropriate.
Probably not. It would probably be part of our disclosure around the outcome of the FDA meeting.
Got it. Okay. There are options.
Yes.
I think that's the key .
Yes.
Yeah. Okay, for the MD Anderson frontline study, just a status update on there. How's that going?
Yeah. I met with the investigator last Saturday at ASCO. That's enrolling well. Because we're adding tovecimig to a new regimen, there was a safety run-in as part of that study that's now completed. We're now going to expand that study outside of MD Anderson proper, into the MD Anderson network in the coming months. Looking to just enroll more patients and get more data from that study. We're really confident and excited about what we're seeing there.
Got it. Could that be an update later this year, potentially?
Maybe. Yeah, maybe.
Okay.
Yes.
Okay. For next steps for tovecimig, so you get the meeting request in, meet with FDA, do some sort of disclosure, probably, just make sure I got it right, first half of August, or no?
No, the disclosure would probably be around Labor Day.
After you get the minutes.
That's right.
Yeah, around Labor Day.
Yes.
Right. Those are the kind of key events there. Anything else that would happen in between as it relates to the program?
Probably not.
Yeah.
We're really focused on that, and then that will help us sort of guide our next step.
Are you going to publish the data or have that at a medical conference later this year?
We're hoping to present the data at a medical conference later this year.
Yeah.
Yes.
Okay. Okay. Why don't you talk about 8371 briefly, too. You mentioned the poster at ASCO. What are next steps for the program, and when should we expect the full expansion cohort data in?
Sure. We're really excited about what we saw in the dose escalation portion of the study. This is a next generation checkpoint inhibitor. In 15 patients treated, six patients at the two highest dose levels, we've got three responses in the study, two out of six at the highest dose level. Patients with non-small cell lung cancer, Hodgkin's lymphoma, and triple-negative breast cancer all had very, very important and meaningful responses. Again, all post-checkpoint inhibitor. I think obviously, as you know, I'm biased, but I think to my knowledge, we have the best phase I checkpoint inhibitor data ever. Right?
Yeah.
We've moved to cohort expansions. We're doing cohort expansions in those three indications. Top two dose levels, 3 mgs and 10 mgs per kg. We're doing 14 patients in each dose level, so 28 total, with triple-negative breast and non-small cell lung, so that's 56. At the top two dose levels for Hodgkin's lymphoma, six and six, so another 12. We're adding 68 patients to the 15 patients enrolled in the dose escalation portion of the study. The dose expansion is enrolling extremely well. That just got opened in Q1, and we're already north of 10 patients in that study. I would expect that we could be able to present an important data update from the cohort expansions later this year.
Got it. Okay. That could be a big update later this year. You've also guided to having the 10726 data second half of this year, too. How are you setting expectations for that? Could we see early durability?
Sure. 10726, our own PD-1 VEGF bispecific, we presented preclinical data at SITC about 18 months ago demonstrating that that drug was better than other PD-1 VEGF bispecific antibodies in preclinical studies. We're now in phase I. We selected indications based on indications where checkpoint inhibition and VEGF blockade are effective. The phase I indications are hepatocellular, renal, gastric, and endometrial. First cohort fully enrolled. Should get to the second dosing cohort in the coming month or so, which would put us in a position, I think, if it continues at this rate, later this year, we can present some early data from that study. Would love to have some efficacy signals, just like with 8371. Probably not durability, just probably initial efficacy data from that study.
Got it. Makes sense. Look forward to the updates. Thanks so much for joining us today, Tom.
Great. Thanks, Maury.