Hi everyone. My name is Maury Raycroft, one of the biotech analysts at Jefferies. I'd like to welcome our guest today, Tom Schuetz, the CEO of Compass Therapeutics. Thanks so much for joining us today, Tom. Thanks, Maury. Thanks for inviting us today. We're going to do a fireside chat format. For those who may be new to this story, if you can give a one-minute intro to Compass. Sure. Compass is located in Boston. We're a monoclonal antibody discovery and development company in oncology. So we currently have three drugs in the clinic. Our lead program is a DLL4 VEGF-A bispecific antibody, which we'll talk about some of the data that we've generated to date with that antibody. Our second program in the clinic is a novel CD137 agonist antibody. And we just presented some data at SITC in Houston, where we believe we've identified a biomarker of activity of that drug. And we found some really unique correlations between markers on tumors and response to the drug. And then finally, our third program in the clinic is a novel PD-1/PD-L1 bispecific antibody that came out of a screen for synergy with PD-1 blockade. And we saw some very unusual preclinical activity. And that drug is now in the middle of a phase I dose escalation study. Finally, we have a set of preclinical candidates that we have not disclosed yet. We've been working on VEGF IO bispecifics for about 12 to 18 months. We're very close to nominating a couple of development candidates. We'll talk more about that next year. Got it. It's a great intro, and you're with your 009 bispecific. You're running a BTC phase II, III potentially registrational study with top line results expected first quarter of next year, so pretty big update coming up. Can you provide an overview of the study design and the patient populations that you're enrolling into the study? Sure. Thanks. So as Maury mentioned, our lead program we call 009. It's a DLL4 VEGF-A bispecific antibody. So DLL4 is delta-like ligand four. It's the cell surface ligand for the receptor Notch 1. VEGF-A, of course, is the well-known soluble ligand for the VEGF family of receptors and is the target of very important drugs in oncology such as Avastin. So in a phase II study in patients with advanced biliary tract cancer, we saw something quite unexpected, which is we had a 37.5% combined response rate, confirmed response rate when 009 was combined with paclitaxel or Taxol. So why do I say unexpected? Because other chemotherapy combinations that are used in that patient population include the chemotherapy regimen FOLFOX, which is 5-FU, leucovorin, and oxaliplatin. That three-drug combination regimen has a 5% overall response rate in the same patient population. So we had an unequivocal efficacy signal there. That led to the design of a randomized trial, which is ongoing today in the U.S. That study is a two-to-one randomization of 009 versus 009 plus paclitaxel versus paclitaxel alone in patients with advanced biliary tract cancer who have received one prior line of therapy. This is a pure second line study. Only patients that are treated in the second line are eligible for the study. Part of the reason we made that decision is when we looked at our phase II data, among the responses that we had, the vast majority of responses were in patients treated in the second line setting. In the second line setting in our phase II study, we actually had a 63.6% confirmed response rate, which would be absolutely unprecedented. So the power calculations for that study, the study is 90% powered to detect an overall response rate of 33% in the combination arm and 10% in the paclitaxel control arm. And as you mentioned, we're expecting a response rate readout from that study by the end of Q1 of the coming year. So very, very exciting time at the company for sure. We're within five months today of a readout from a randomized study in a very important patient population. Got it. And can you talk a little bit more about the patients that you're enrolling into the study? So you mentioned second line, pure second line patient population. How else does it compare with the prior data that you've shown? Any other differences that people should know about? Yeah, no, the patient population for the randomized study is virtually identical to the phase II study, with the only exception being we're limiting the patient population to the second-line only. Got it. Okay, and let's see for your discussions with FDA around this study, what's your view on what FDA could view as the minimum threshold for response rate? So it's an important question, of course. So when FDA did not tell us that, they simply said that the regulatory significance of the study depends on the magnitude of difference. Those were their words, magnitude of difference in response rate between the two treatment arms. Got it. They didn't tell us what that magnitude needs to be. We're powered for a 23% absolute difference, so 33 minus 10, of course. A 23% absolute difference in the treatment arms. I think if we hit that, it'll be a potentially transformational result for this patient population. Got it. Okay. And can you talk about how responses are measured? Is it investigator assessed or independent? Responses are read in both ways. They're read by the investigators and by an independent central review. The primary endpoint of the study will be based on the independent central review. Got it. Okay. And what's your expectation for the performance of the control arm, given that there's not a lot of great references out there for what the benchmark should be? Right. So that's absolutely correct. There have been some small studies published. MD Anderson published a frontline study. It's an older study, to be fair, more than 20 years ago, where they enrolled 15 consecutive patients in the frontline setting and treated those patients with Taxol, and they saw no responses. So the expectation of performance in the control arm is frankly quite low. For the purposes of the power calculations, we assume 10% just to be extremely conservative with sizing the study. Got it. Okay, and from K. Walls' perspective, what's the bar for success in this study? Should we expect similar response rate, PFS, and overall response rate as your prior study, which is predominantly run in Korea? Or do you think that there could be some deterioration in the outcomes that could be acceptable? Oh, sure. I think the one randomized study that was done in the same patient population was the three-drug combination FOLFOX, again, 5-FU, leucovorin, and oxaliplatin versus best supportive care. So in that study, FOLFOX had a 5% response rate. And best supportive care, of course, was zero. And the difference in overall survival was 6.2 months in the FOLFOX arm versus 5.3 months in the best supportive care arm. So a 0.9-month improvement in median overall survival. So in terms of what docs would consider meaningful, I think any statistically significant improvement in response rate would be considered meaningful because there's just really no consensus standard of care in this disease, which is an incredibly common disease. It's one of the only common cancers that does not have a consensus second line standard of care. It's unbelievable, really. Yeah. Yeah. And maybe talk about the unmet need a little bit in current treatment paradigm for BTC patients. Sure. So again, in the second line setting, only FOLFOX really had sort of any difference. You got to remember the control arm in the FOLFOX study was supportive care. So it just tells you something about how sick this patient population is. NCCN guidelines have a number of different chemotherapy regimens. So the patients that get frontline therapy and then progress, which is everyone, unfortunately, they go to second line therapy. And only about 15% total of the patient population is eligible for a targeted therapy, either an FGFR2 small molecule, an IDH1 inhibitor, or a HER2 targeted antibody. So the total of all those is 15%. So 85% of the patient population has to go to physician's choice of a chemotherapy regimen. So it's a massive unmet need. We're positioning the regimen of 009 plus Taxol to become the consensus second-line standard of care in this patient population. Got it. Makes sense. And in your prior study for median PFS, you reported 9.4 months. And then for survival is 12.5 months. And then in the second line patients, that the gap between PFS and OS was smaller too. What are some of the explanations for potentially just the difference between PFS and OS that you're seeing? In between the second- and third-line patients in that study or between OS and? Yeah, between PFS and OS. Yeah. You know, I think we had a PFS at 9.4 months that's actually larger than what you see in the frontline setting with gemcitabine, cisplatin, and a checkpoint inhibitor. Both GemCis durvalumab or GemCis pembrolizumab have median progression-free survivals in the seven-month range. So in the second and third line patient population, we're actually longer than that. Now, one thing in the FOLFOX study, I think something that's very, very important to keep in mind here is the median PFS in the FOLFOX arm was 4.0 months. And the median OS was 6.2 months. So just subtracting that, 6.2 minus 4, the expected overall survival in the third line setting, believe it or not, is 2.2 months. So once you progress on FOLFOX, on average, you've got 2.2 months left to live. It's unbelievable. So once a patient gets to the third line setting, the delta between PFS and OS is quite small. Got it. Okay. Yeah, that makes sense. And then for overall response rate, if you succeed there, is there a path to immediately file for an accelerated approval based on the ORR data? Or what do you need to see in terms of PFS and OS to support the scenario? And what does the timeline look like for these different scenarios? Sure. The short answer to your question is I really don't know. We haven't had that kind of a dialogue with FDA. In the original conversation with the FDA and exchange of ideas around the study design, that's where they used the term magnitude of difference in response rates to convey to us that the data that they would need to see in order to determine the regulatory path forward. Let's maybe just do some theoretical outcomes here. We hit the power calculations: a 23% delta. I think we'd have a conversation with FDA about potentially using that to support accelerated approval and then using the PFS and/or OS data from the study to maybe convert the approval to a full approval. Let's say the data are more of a blowout: 30-plus percent difference in response rate. Then I think the conversation about potential accelerated approval could be more straightforward because, I mean, a 30% delta in response rate would be something that's never been seen in this disease. If it's lower, let's say the delta is 15%, as an example, we're probably going to need PFS and OS to support that application. Got it. That makes sense. And so if response rate, yeah, if you need to rely on PFS and OS, what could some of the timelines look like there? Right. So that's a complex question because in these studies, the progression-free survival and overall survival are the so-called time to event analyses. Those analyses are triggered by a total number of events in the study. And today, we haven't seen enough events in order to be able to accurately project when that might be. I think in the coming quarter or two, we'll get closer to having a better idea about when those analyses would occur. But I think almost certainly they'll be in the second half of next year. So PFS and OS maybe in the second half of next year. But just to maybe state the obvious here, if the overall response rate, if the overall response rate endpoint is clear, then the longer it takes to get to the PFS and OS readouts, the better it is because that means that patients are not progressing and not dying. So the longer it takes, the better it is. Got it. Makes sense. And can you explain what triggers the first readout? And we're wondering if it's based on a fixed number of patients or if PFS event number also plays a role. So for the first analysis, you mean response rate, right? So the response rate is simply triggered by time. So that's why we've been able to be so accurate or so consistent in our projection of when that's going to occur. So that analysis occurs 28 weeks after the last patient was enrolled. And we announced the last patient was enrolled in August. So that gets you to March. Got it. And so the PFS has nothing to do with triggering the response rate endpoint. The total number of events will trigger the time to event analyses later. Got it. Okay, and in your study, crossovers allowed following progression, you mentioned using the Rank Preserving Structural Failure Time method to account for crossover impact on OS. Can you give us a high-level overview of how this OS adjustment works? You're the best. So that's a complicated statistical test that was used in the ivosidenib approval study. So why does that matter? So the IDH1 inhibitor that was studied in biliary tract cancer did a randomized trial in the second-line setting of the IDH1 inhibitor versus placebo. And crossover was allowed. So progression-free survival was the primary endpoint of that study. But then they used a novel statistical test, as you mentioned, the rank-preserving structural failure time test, which adjusts the overall survival data for the performance of patients after the crossover. So it's simply a way to correct the analysis for the crossover. And we have prospectively defined the use of that method in this study. And I think the fact that it was used once before in exactly the same indication, I think, is relevant and helpful for us. Got it. Okay. And this was something that you got alignment with FDA on as well. Yes. Right. And for patients crossing over to the active arm, are they immediately enrolled in the active arm upon progression? And are the responses from these patients included in the overall safety response rate and PFS calculations? Okay. Once a patient is thought to have progressed at the site, their progression needs to be centrally confirmed. So the patient's progression is centrally confirmed. And then they have to meet a set of inclusion criteria again just to make sure that they're sort of eligible for the study. Of the three things you mentioned, post crossover, none of that data would be included in the original overall response rate comparison. And none of that will be included in the original PFS comparison. But we're also going to do a comparison of PFS in that population post and pre 009. So that's going to be a really interesting and I think unique analysis because that's going to be internally controlled by the patient themselves. So we're going to look at their PFS on paclitaxel alone. And then in the same patients, their PFS on 009 and paclitaxel. So that's a separate analysis. The third thing you asked, safety, yes. The safety will be included as well. Got it. And would that additional PFS analysis, would that be something that you'd probably share that along with the original PFS analysis? Oh, yeah. Okay, and in August, you mentioned that the study is fully enrolled. Can you walk us through just how enrollment's been and some of the, I guess, obstacles or tailwinds for recruiting patients? Sure. So I think it took us a little longer to get some of the academic sites open, which sounds like I'm whining here, right, with contracting and budgeting and all that kind of stuff. So it just took us a little longer. But once our academic centers were open, the study enrolled incredibly rapidly. One of our lead investigators told me earlier this year that the study enrolled in record time. That was his words. And I think it's a reflection of how common the patient population is. And I think one of the things we learned from the enrollment into the study was that the majority of these patients are seen at academic centers. And I think that's important as we ultimately begin to plan for potential commercialization. The fact that these patients are seen at a small number of centers, I think, is beneficial for commercial planning. Right. And I agree. And for the readout, is it only an update on response rate? Or could we see an early PFS and OS? At the end of Q1? Yeah. No, it'll only be response rate. Got it. Okay. And the study's being run in the United States. But we're wondering what percentage of the study population do you expect to be of Asian descent? I don't know the answer to that. I don't see any reason to think that it would not sort of mirror the general demographics of the U.S. population. Okay. And is there anything else you can say about patient baseline characteristics and demographics at this point? No. Okay. And you've got an IST trial in frontline BTC in combo with GemCis, cisplatin and durvalumab that's going to be run by MD Anderson. This was disclosed around your second quarter earnings. Can you talk about the study rationale here and just the status of this study? Sure. Yeah, we're really excited about this development because our second line study is open label at the sites. MD Anderson was the second largest enrolling center in the study. Earlier this year, MD Anderson approached us with a proposal to do a frontline study adding 009 to gemcitabine and cisplatin and durvalumab. That study, we're almost ready to go. At the SITC meeting in Houston two weeks ago, I met with the investigators for that study. We're maybe a month or two from that study being ready to go. It's live on clinicaltrials.gov now. It's a 50/50, 50-patient study, 12-patient safety run-in, adding 009 to the frontline regimen, followed by 38 additional patients. Got it. Okay. And so that should be starting soon? Yeah, so within a month or two, I would think. Okay, and then I wanted to talk about colorectal cancer as well. You had a program update in second quarter, and in third quarter, you announced you're going to start phase II in second-line DLL4-positive CRC. Just checking if you've gotten feedback from FDA on this study design? And can you walk through the design details and next steps? Sure. So we have not gotten feedback from FDA on this study. So our phase II colorectal cancer study, which was a monotherapy study, we had two out of 41 patients with confirmed responses. And these patients were treated in the fourth line setting. And in the fourth line setting, tyrosine kinase inhibitors, regorafenib, fruquintinib, those drugs have a 1-1.5% response rate. So we have one of the highest response rates, albeit low, just to be fair. But we have one of the highest response rates ever seen in this incredibly difficult patient population. We also had PFS of 3.9 months, disease control rate of 71%, and a median overall survival of 10.2 months. So we have a real signal there. Now, one of the things we observed in our original phase I study was patients whose tumors were DLL4 positive did better than patients who had DLL4 negative tumors, which, of course, makes sense because this is a DLL4 targeted therapy. Now, what we're doing now, we collected biopsies in our phase II colorectal cancer study. We're analyzing those biopsies now for DLL4 expression. And we're starting to get some hints from that. And we're beginning to plan for a DLL4 positive CRC study, fraction of patients with CRC who are DLL4 positive, probably in the 60%-70% range. So a really significant patient population. We have an unequivocal efficacy signal, I believe, as a monotherapy. And of course, drugs like Avastin, the real therapeutic potential of these drugs was really uncovered once they were combined with chemotherapy. So, planning for a second line combination with FOLFIRI in the planning stages. Probably we'll seek FDA guidance on that, maybe in the first quarter, looking to start that study in mid-2025. Got it. Okay. And we're almost out of time. And we didn't talk about the rest of the pipeline. So I want you to. Whose fault was that? That's definitely my fault. So I'll let you talk about some of the key highlights from that. And you mentioned VEGF, IO, which you haven't discussed a lot about yet. There's a lot of interest in that too. Maybe just cover the key highlights and what investors should be focused on going forward. Sure. Our second drug, 471, the CD137 agonist, we have confirmed responses in the post PD-1 patient population. I think we have one of the highest response rates, I think, almost ever observed in that population, including a complete response in a patient with metastatic small cell lung cancer who today is more than four years out from starting therapy in the third line setting. So just an incredible result for him. His biopsy was really interesting because his tumor cells were covered with NCAM, which is CD56. And CD56 attracts NK cells. So there is a very plausible biological mechanism of action here. So we're going to run a CD56, also known as NCAM, positive biomarker-driven basket study. That study will start mid-next year. Our PD-1, PD-L1 bispecific is in the middle of phase I. And our VEGF IO bispecifics that we've been working on for 12-18 months, we're now very close to nominating a development candidate. And I think early next year, we'll have some guidance on time to IND filing for our fourth program. Got it. Awesome. Thanks, Tom. Thanks for joining us today. Thanks, Maury.