MacroGenics, Inc. (MGNX) Earnings Call Transcript & Summary

Great. Next panel is ADC, Antibody Drug Conjugates. I'm Yigal Nochomovitz. I'm one of the biotech analysts here. I cover all 3 of these companies, MacroGenics, Zai Lab and Zymeworks. So, Scott Koenig, CEO of MacroGenics, welcome. Rafael Amado, President, Head of Global Oncology and R&D at Zai Lab and Paul Moore, CSO at Zymeworks. So thank you all very much for being here. I appreciate it. Just to start out, why don't we just do a quick lightning around, give us a very quick introduction to the company and then with a specific focus on your ADC programs and just some of the very key highlights, and then we can go from there.

Terrific. I'm Scott Koenig. So MacroGenics is 23 years in operation. We've been focused on developing immune-based therapeutics from the inception. We have 3 validated proprietary platforms, Fc-engineered molecules, bispecifics, which we call DART molecules, and then ADC program. The company has been successful at being involved in the approval of 3 molecules that we had either developed from the inception or had a large part in this development. And right now, we have a very vigorous broad pipeline of molecules, which include ADCs Fc-engineered molecules and bispecific molecules. With regard to the particular ADC technologies, we've been working on ADCs since 2013 at the time of our IPO. The lead molecule right now is a molecule called Vobra duo, which is the shortened version of vobramitamab duocarmazine. This is a molecule targeting a very novel molecule called B7-H3 within the ligand of the checkpoint family, highly over-expressed in most solid tumors. We're currently in a Phase II study right now, which we expect to largely enroll by the end of this year and complete early next year, where we're comparing 2 different doses of Vobra duo and patients with castration-resistant prostate cancer. The plan is to pick one of those doses and then move forward into a registration study. We have a second molecule with a different technology targeting ADAM9 through a 50/50 collaboration with ImmunoGen using their next generation maytansinoid molecule, DM21 and with follow-up data later this year from that. And then in the past 1.5 years, we entered into a broad collaboration with Synaffix, which have a very novel ADC technology platform where they've identified the major glycosylation position in the Fc domain, removed those glycans and conjugated a hydro space linker and then putting on multiple different toxins on that. And we have 7 different slots that we are -- we have licensed from Synaffix and what we have said is that the first of those molecules will file IND later this year. This will be a topoisomerase inhibitor molecule. We have a second next year, and then we're working on targets 3 and 4 going forward. So that's sort of a quick summary on the ADC side.

My name is Paul Moore. I'm at Zymeworks. I'm the CSO there. Zymeworks, you probably aware as a sort of an innovator in antibody-based therapeutics. What maybe sets Zymeworks at par is that in-house, we have capability in both protein engineering, which has been the fundamental of Azymetric platform, which is industry well accepted as a leading platform for heterodimer Fc and really opens up a lot of space for bispecifics. But in addition, we also have in-house capability in generating ADCs. So we have our own chemists and our own capability, which allows us to merge both expertise and antibody engineering with ADC technology. In the ADCs, the first molecule that Zymeworks has developed as an ADC is a molecule ZW49 or zani zo, which was based on our proprietary Auristatin payload. So that molecule is being developed in the clinic. On the front end of that molecule is the bispecific capability, which really powers efficient internalization. So there, we were marrying both bispecific with ADC. Since then, the team in-house has worked a lot on different technologies, different drug specific conjugation, different payloads where we've landed just based on being pragmatic with where the field is, is on the TOPO platform. And so what we -- on topoisomerase inhibitors and so what Zymeworks has developed is its own proprietary TOPO payload. So we have our own camptothecin based molecule from screening hundreds that we selected a lead that we've wired into our molecules. And there, we also think a lot about how we conjugate. So we're a little bit unique in our thinking there on the chemistry and how we do that. So that combined sets us a little bit apart from what others are doing. We have 3 molecules that we've discussed at the preclinical stage, 2 of which we're moving into the clinic in '24 and '25, and we'll talk more about those during the panel.

I am Rafael Amado. I am an R&D for Zai Laboratory. And Zai is a 9-year-old company with products in commercial states in China. There are 5 products approved, the latest of which is efgartigimod for myasthenia gravis. And our mission, if you will, to develop products, both for the Greater China region as well as globally. And we have 13 products in development in multiple therapeutic areas, many of them are in oncology, others are immuno-inflammation like efgar, Neuroscience and Infectious Disease. Those are the main therapeutic areas. And in terms of ADCs, we entered the space of ADCs with -- I'm sorry. There's a problem with my microphone. [Technical Difficulty] In terms of ADCs, we entered the space with a collaboration with Seagen, with tisotumab or TIVDAK, which is a tissue factor antibody conjugated with our starting through cleavable linker and this is approved. We received accelerated approval in September 2021 for platinum-resistant cervical cancer. And we've been participating with them in their post-approval commitment, which there was a press release, I believe, it was yesterday, showing that this study was positive compared to dealers' choice or physician choice. So this was a survival trial. So it was positive both in survival as well as the rest of the end points and it should serve together with the PK study as well as the Chinese patients that are continuing to be accrued in China to get TIVDAK approved in China for platinum-resistant cervical cancer. And in addition to that, earlier in the year, we did a deal with a company called MediLink and this is for an ADC, that targets DLL3. And DLL3 is a validated target through bispecific products from Amgen and Bayer. Both have shown that this target can actually lead to responses and prolongation of survival in patients that have received platinum-etoposide in small cell lung cancer. So this is a pre-IND. We plan to launch the IND before the end of the year and started the study very shortly after. And we can obviously talk about the pros and cons of these new generation ones. Both of them are different in terms of the payload and the linker, et cetera. And one makes an ADC, a good ADC, it's a matter of debate, but I'm sure that we'll all have a lively debate about it.

Okay. So that sort of leads into my main thematic question in terms of what makes a good ADC. We've had obviously some very big successes and we've had some notable failures in the ADC space in the last several quarters. So what are the challenges in making a good ADC? And how do each of your programs address those challenges? You've already mentioned some of the details, but maybe go into more detail. Go ahead, Scott.

So what's the starting point on any of these technologies is, first of all, what's the target, where it's expressed, what's the amount of expression and what is the particular epitope that's recognized, particularly when we went through the efforts on targeting B7-H3, we knew this was a target that was overexpressed on multiple different tumors. So number one is that we went and identified different epitopes and different variable domains that had different profiles in terms of incorporation into the cell versus localization on the surface of the cell. And so for example, in our programs on B7-H3, we have one variable domain that targets which is called enoblituzumab, an ADCC type mechanism, because you want that antibody sit on the surface of the cell, and we're pursuing that for a different set of indications. With regard to Vobra duo, we picked a variable domain that had the best incorporation into the cell and made a big difference in terms of both the activity profile as well as ultimately the safety profile. So that was the first starting point. Then the next question is, which linker you're going to use and which toxin you're going to use. And of course, obviously, different tumors will respond to different mechanisms, which are basically chemotherapeutics. So as you heard, our first lead molecule for Vobra duo is a molecule called Duocarmycin, which we have a partnership with beyond us. This is a DNA alkylating agent. We are obviously working now to find the optimal exposure to mitigate side effect profiles that are associated with alkylating agents in this particular case. The second point is, other tumor types may have different responses to different toxins. And then the big question becomes, do you want to have a linker that is stable? Or do you want to have a linker that is actually clean? Because what happens is within the tumor microenvironment will be always be cells that express your target, but others that are either lost it or have lower degrees of expression. So within the tumor niche, you want to have ones that we believe that by having a cleavable linker will be able to kill tumors in the local environment. So it's that balance. And obviously, it's important, obviously, what specific expression of the target on normal tissues, which will also determine which particular toxin you pursue. So those are the things that we thought about in moving forward. I should also point out there is now also a beautiful literature that's evolving that you can start out with the same variable domain for a particular target, but now incorporate different toxins. And so that's one of the strategies that if you think about what MacroGenics is doing is we're building this portfolio, is we're looking at this as a suite of potential opportunities where once we've identified a terrific variable domain, we're going to put in different linker toxins on to go after various different tumor types. And as there's some recent evidence of -- in fact, very recently, within this last week that you can go after a target with one toxin and if a patient develops further disease, you can come out now with a different toxin, but still use the same variable domain.

Yeah. I concur with what Scott said. From our perspective, what we do is we think about the ADC from antibody to conjugates or antibody drug conjugate. So starting off even with the antibody itself, I think you have to think not to assume that every antibody will work well as an ADC. So we do spend time looking at that and comparing against benchmarks. So our first molecule that we're putting for -- our first of our TOPO platform of our next-generation ADC is targeting folate pullet receptor. So there, as we thought about how do we design the optimal folate receptor molecule ADC, we thought about the payload, we thought about the linker, we thought about the antibody. And all of those we've optimized. We've thought about what has been successful previously for other ADCs. So that we bake into our thinking on the linker and we do have designed instability in our linker. We thought about the payload, the balance of potency on the payload, there are different flavors of topoisomerase I. A lot of people use exatecan, which is kind of off-the-shelf molecule that you can plug into an ADC. We've taken a little bit different approach there, where we've gone with something a little bit more modest and potency because we believe having that balance between potency and safety is important. And then what we also do believe in is designed instability. We are -- there is some release of the payload that can give you an effect of the chemo independent necessarily of purely on target efficacy. You again have to balance that with unwanted toxicity. But so far, our preclinical models, where we pushed the doses of these molecules up, has supported pursuing and going forward with that design and that molecule. We do think about the DAR. Is it a DAR 4, or is it a DAR 8? We can do that. We can evaluate both. That's also an important factor to consider for our folate receptor molecular lead, first molecule going in, that will be a DAR 8. Then what I also mentioned at the beginning was then -- also then thinking about the antibody. There, then we spent time screening different folate receptor antibodies, benchmark them against MIRVs antibody against other folate receptor antibodies. And there collectively -- for consistently, we see better internalization payload delivery. So that's our thinking. And when we designed it, obviously, in the space, we want to also differentiate. Our goal here is to help patients get more durable responses. And so our strategy, while overlapping with the target and the payload, but there are certain features we feel that do distinguish where we are based on just the collective design of the molecule. So then behind that, we've got other targets that we're pushing forward. We do believe in this, which we can plug in, again, the payload and the linker technology, but then we changed the target, say we move forward into different targets. Our second target that we're going to be pushing forward will be an NaPi2b, which you're probably aware of. Others have tried to push that forward as an ADC, had some reports of clinical responses, but have the responses been durable enough there. We feel, again, with our design, we can overcome some of the limitations of those molecules. Another important feature that Scott had mentioned that we truly believe and is bystander activity, which is something that wasn't necessarily in previous NaPi2b focused ADCs. So that's kind of just a sort of an overview of our approach.

Yeah. A lot has been said. Clearly, there's been a renaissance in this field. The first one was in 2000, and then there was a lull in terms of products coming through until a lot of what's been spoken about was optimized. The antibody is really important. The differential expression. Antibody by itself can be responsible for some of the effect either through ADCC or CDC or by blocking signaling. The linker is also really important. Most people are using covalent linkers now that are cleavable either chemically or through proteases, internalization or now that the antibody that's important and also processing of the antibody where the enzymatic activity can take place in the lysosome and release of the payload. Normally, you want linkers that have low hydrophobicity and the payload should be lipophilic. So that, it can go into our targets that do not actually -- or cells that do not actually express the target. And obviously, the other thing that has been mentioned is the drug antibody ratio. If you look at some of the most successful recent products, they tend to hide DARs and that may also be important. But there is a price to be paid. Obviously, there's resistance that occurs, and we can talk about the mechanism services. There's often a problem of combining these products with traditional chemotherapy. So it's hard to test them in frontline, where chemotherapy tends to be the standard of care. That could be tested with immuno-oncology products, but those require kind of convoluted development paths. And then there's toxicity. We all know about Stevens-Johnson and ILD and profound neutropenia, et cetera. And one can pretty much say what the dose is going to be depending on what the payload is and the DAR and the ratio, because there's been so much experience already. So all this optimization will continue to occur. And I think that perhaps what has made the biggest difference is the ability of these toxic moiety is to actually permeate into cells that don't express the target because the antibody often can get in the periphery of the tumor. So penetration of the antibody into the tumor is important. But if it doesn't, at least the toxic moiety can, have so much smaller molecular weight and can penetrate into the tumor and cause antitumor activities. So these are some of the concepts that I think have improved the field, but there's still a lot of work to do to get rid of some of these toxicities.

Just a follow-on point. Again, as we think of the evolution of this field, where we start out with chemotherapeutics, just given broadly, we're now giving targeted therapy based on ADCs. It's still a chemotherapeutic. It gets back to the issue is how do we continue to improve the safety profile and which in large part depends on where expression of the targets are on normal tissues. So for instance, one of the things that we are pursuing in addition to just having a conventional antibody, with having a bispecific and multispecific molecules, our DART technologies, we can go after, for instance, different targets which will have a different profile in terms of both expression on tumors and normal tissues. And so by incorporating some of these bispecifics and actually now making these ADCs, you can heighten the ability for particular targets that are expressed on tumor cells, both those targets and mitigate some of the binding to normal tissues to now take targets that historically were not addressable, but now can be addressable by having this bispecific technology.

How do you think about the interplay or the trade-off between the DAR, the potency of the payload and the stability or cleavability of the linker? Like how do those things intersect in terms of the design? I mean, if you have a high DAR, does that mean that you need to have more cleavable linkers or not? Does it mean -- if you have a high DAR, does that mean that you can get away with a lower potency on the payload? How do those 3 critical design elements play against each other?

Well, I can at least take that one from the approach that we've taken. So I think we have some fundamental belief in the linker stability so that we kind of -- we believe in some engineered linker in stability. Then on the payload, we focus on our proprietary payload. But then on the DAR, what we do is we take both through. We take the DAR 4 and the DAR 8, so a more modest DAR and a high DAR and then we just take that through the testing and really evaluate that through pilot talks and a primate. And so there, we can get a read on any issue on tolerability that we would want to avoid in the clinic. And then we also then think about the target. Where is the target expressed? Is the reliability on the target, would that point you towards a lower DAR? Do we have the efficacy, though, with that lower DAR? Those are the kind of things. So it's a little bit -- it's based on the molecule profile, but then also on the target that you're going after and where it's expressed, you sort of factor that into the equation.

We don't obviously have the best rules yet on that. There's a lot of empiricism and is still going, going out here. We've gone through various waves where historically, having a more potent toxin was the most important thing as we got into the TROP2 and the Trodelvy era. We saw that even dirty targets and less potent toxins could be quite effective across various tumors. So I think we're still trying to figure out the rules. They probably determines by which target you're going after, the ability for those targets will play a very important role on what is the necessary range of the DARs they use. My sense is it's somewhere probably in the middle because if you have too high DAR, you may get too much toxicity. But again, I think it will take time to work that out.

Yes. I mean I think a lot of it is empirical. I think the antibody is important, specificity, avidity of the antibody. I think where and how the linkers fleet is important. Some technologies utilized non-natural amino acids, so they know exactly where it's cleaved. And also, the number of toxic molecules broadly makes a difference. I mean, there's -- that's one of the reasons why perhaps HER2 is better than T-DM1. And some of these ADCs are actually better in tumors where the target is not expressed a high level. And this is not something that is well understood, actually. So I think it depends in the patents on the tumor. It depends on the payload, but I think everybody understands that one of the most important thing is the differentiation of expression of the target between normal tissue and tumor. And also I think the DAR is important and maybe that there's a limit as to how many molecules you can put before you end up having toxicity. But with the topoisomerase inhibitors, one can increase the number of DARs and perhaps lead to a better outcome. So I think there has to be more innovation on this so that we can avoid some of these toxicities. And also there's a dearth of targets as well that are only expressed in tumors. I think if we could find targets that were expressed only in tumors then perhaps we could avoid this off-target toxicity that we see often.

This is a bit of a different question, but we've seen the sort of renaissance of the radiopharmaceuticals with PLUVICTO and others and supply to a lot of companies now, private and public that are pushing in this area. Curious, how do you see that in terms of a competitive threat to the ADCs? Or do you not really see them as playing in the same space?

I basically view it as opportunity for the patients. I really think that there's no single technology. I mean, just in the history of therapeutic interventions for cancer, it's rare that a single agent will cure cancer. And we're always going forward in terms of looking at modalities that have orthogonal mechanisms that can actually complement each other to ultimately have more effective outcomes. I was interested, for instance, in just this week, with even in the radiopharm space, obviously, everybody ran some beta emitters for lutetium. We saw some nice data finally on the alpha emitters coming through Actinium this week. So the point is that having more different modalities to these targets is going to be beneficial for patients. What we have seen, for instance, in the ADC space, combining immune checkpoints with ADCs looks like a very good way to go to get complementary mechanism of action. I think the same story will come out from the radiopharmaceuticals or being able to combine other immune-based therapeutics there. So we see this as opportunities, not threat, at least we do.

Yes, I agree. I think they're both using obviously antibodies to deliver either a chemotherapy or a radio label. So there is some overlap there on the concept, but I think there's nuances there that are different. And I think there's certainly room and need for better treatments for patients even with the advances that have been made, some of the overall responses are still we want to improve. We want to be better. And I think both ADCs and radio labeled antibodies can provide that. We are more familiar with the ADCs, with the process from making ADCs. I think over the years, the radio label that's been more of the challenge of delivery and how do you get the drug to the center that the hospital that's delivering it. That seems to have improved, but I think that still seems a challenge, maybe sitting from the ADC side and looking over. But I think it's great that there's been advances and maybe there will be learnings. And as Scott alluded to, ultimately, a lot of these drugs could be used in combination. I mean, that may be in the future, but the mechanisms, the killing mechanisms are different. You're using a chemo or you're using a radio label that gets released and maybe has more broader activity. The ADC, although you'll have bystander is more focused on getting into the cell.

And the toxicities are different. Obviously, with a lot of the radio label drugs, you're seeing marrow toxicity. So there will be a certain subset of patients based on the historical therapies that they've gotten that may be less able to take a radiopharmaceutical or may be able to take an ADC. So...

I mean, I view this question as a broad question because on the delivery side, we're using antibodies, but there are already technologies out there where people are using polymers where they can get DARs of 60 or enormous numbers or they're using very small peptides or bispecifics. So there will be other delivery systems. And likewise, there would be different payloads as well. There could be oligonucleotides, it could be hormones, it could be small molecules, and they could be radio ligands. And that's not new. I mean, Bexar was in the market. It came out in 2013. I was involved in that Zevalin as well. Then there's been naked beta emitters, which have been pretty effective and there will be radio ligands that are linked to antibodies as well. I think there are some drawbacks with that. One of them is you need a specialized system, setting a hospital where there's a nuclear medicine department where that can be administered. Also, the shelf life of these products is relatively brief. So logistically, maybe a little difficult to do, but it can be quite effective, because these beta emitters, especially the alpha emitters, I mean they can really cause DNA damage, like 30 cells apart. So they can be quite powerful, but at the same time, there could also be toxicity, either the case is not fast enough. So I think this will occur with time, both on the delivery side and on the payload side.

Are any of your pipelines looking at the immune stimulating payloads in addition to cytotoxic? Is that an area of interest or potentially at least or even in the discovery phase for any of your companies?

Yes, we have looked at that, looked at TLR7 agonists and made some progress with that. Certainly, it's an area that has got the excitement, it can stimulate the immune system, get things moving. I think there the challenge has been translating that preclinical. Having the confidence in the safety profile and the efficacy window when you move forward with that, but I think it's definitely an exciting area. It's just right now, I think with the cytotoxic payloads there's a little bit more defined, a little bit more mapped out, how they work, and that's the opportunity that we think we can build upon in the near term. But certainly moving forward, thinking about these alternative mechanisms to overcome limitations for other tumors that don't respond to that type of toxin or require an immune stimulation, I think there's a lot of opportunity there still.

We've considered that in the past. We are not actively pursuing as Paul was pointing out, the TLRs or cytokines as conjugates there, we feel that by developing the immunostimulants through our bispecific DART molecules going at that mechanism and then combining those molecules, developing that is a better way of at least initially.

Yes. Likewise, not much to add. We haven't considered it. We know that they are in development, but I think it's still early to know. I think TLRs are probably the ones that are the most advanced. And time will tell whether it's better to deliberately deliver it to the right cell type versus use some other IO systemically.

All right. Let's spend the last 10 minutes just going through some more company-specific questions. So you already mentioned Vobra duo, Scott, that the B7-H3 ADC, and you're in the Phase II TAMARACK trial. This is in metastatic CRPC. Now given the landscape has changed a little bit, as we all know, with PLUVICTO, you've made some adjustments, as I understand, to the trial. Can you just remind us what those changes were? And what do you need to see to take this program into a pivotal trial?

So if you go back a little bit more than a year ago, we were designed this TAMARACK study as a continuous Phase II/III development where we had a 3-arm study of 2 different doses of Vobra duo versus a control population. And what we found was that we were challenged because at the time, we selected a second antigen receptor targeting agent as the control arm. And as many of you know, a JCO from particularly the European groups they were arguing that maybe this is not the best control group anymore, even though many trials were still ongoing, using that as a control therapy. So we took the decision to amend the trial. We removed the control group. We have a 100-patient study, 50 per arm, where patients have been treated with either 2.7 mg/kg Q4 or 2 mg/kg Q4 of Vobra duo. As I said earlier, that is what we expect that most of the patients will be enrolled this year. We'll finish it out early in '24. This is an open study. We will get to very quickly the decision with regard to both improvements in the side effects, which we were -- was our objective here as well as obviously the activity. But once we have that insight, which I say will come very soon, we will then move forward into picking one of those doses to move forward into a controlled study going forward with the appropriate control population. At this point, the control agent, we have not decided. I would guess that at that time, it most likely will be a dealer's choice on a select number of control agents, but that is our plan going forward with success in the study.

And then, Paul, you mentioned ZW49 earlier, the HER2 ADC, the R-statin molecule. And you also have said that you're prioritizing the HER2-positive non-small cell lung cancer for the lead indication. And this is in the post in HER2 metastatic -- over the post in HER2 metastatic breast setting. So can you just talk about the reason for prioritizing non-small cell lung cancer for the lead indication for this?

Sure. So, for lung cancer -- so for the molecule itself, obviously, we published the day we presented data last year at ESMO. We know we've landed on a recommended Phase II dose. And there, we see evidence of activity at that dose. We have a very tolerable safety profile there. But I think really to move things forward there and get meaningful benefit, we believe that combination is needed there. And so there, when you look at the space where HER2, we could go, we're not ruling out going to the post in HER2. But right now, we see the opportunity in the non-small cell lung cancer with the combination with PD-1 is a very obvious place to go in a measurable clinical trial where we can see the efficacy of the drug. So our thinking there is to have a size clinical study that will give us a sort of a readout on that drug at a certain point. And if we see that activity, that will allow us to sort of expand into other indications and particularly, but then really move into lung cancer, where we think we can be very competitive in that space compared to the other competition in that HER2 overexpressing lung cancer.

Is there a hurdle that you're defining for what you'd want to see in combo with the PD-1 to take it forward?

Yes. We are -- we have in our mind, like a Simon 2-stage kind of clinical design that we will be incorporating into the design that will be sort of a bar that we will want to get to support further investment and continuation of the clinical study. But that will come out as we reveal the study design later this year.

Okay and Rafael, at the beginning, you mentioned the DLL3, ADC, the ZL-1310. I believe that's for small cell lung cancer and neuroendocrine tumors with MediLink using this TAMLIN platform. Tell us more about the TAMLIN platform? What is that exactly? And how did that help design the DLL3 molecule?

Well, it's a platform whereby the linker is a covalent linker. The antibodies internalized the moieties are very lipophilic and they kill the cell and exit the cell right away. And so they cause bystander effect which is superior, at least in in-vitro studies to what we see within HER2. So it's the position and the design of the linker the DAR 8 and is a topoisomerase. So the expectation is that we would be able to escalate at least to 7 to 10 mg/kg. And again, it's a target that is being validated. So if we don't see much toxicity and we can get to those levels, we would be able to move the product relatively quickly. So we are pretty eager to get started. And this is a platform that can generate other antibodies, against other targets that we may have an interest in the future. So again, the main property being widespread that I think that tumor, even if the antibody doesn't penetrate sufficiently into the tumor bed.

Okay. And you're going to -- which tumors are you going to go into in the Phase I?

We'll start with small cell lung cancer and then other neuroendocrine tumors, possibly prostate, maybe GI.

And the Phase III that you're running in China for the second line cervical cancer with the TIVDAK? Or can you just provide an update on that study?

Yes. So we participated in the pivotal trial. That trial was -- had an interim analysis, which was released yesterday. It was a survival trial. We entered the tail end of the study, but we managed to put patients in China and then continue to enroll patients on an extension trial. And so together with the PK study, the original study that led to the accelerated approval in the U.S., the global study, including the Chinese patients as well as the extension study. We believe that should be sufficient to get full approval in China. And so that's we're going to do consultations with CDE to ensure that, that package is sufficient and move forward with the submission.

That I think earlier you mentioned the ADAM9 ADC, 936. Why is that an interesting target? And I think you're going to have some non-small cell lung cancer data in one of the cohorts towards the end of the year. What do you need to show there to take that program forward?

So this is a 50-50 collaboration with ImmunoGen. They are conducting the clinical study. There was a dose escalation study and then expansion, particularly as you point out into non-small cell lung cancer. We're very excited about this particular target. This is a metalloproteinase that is highly expressed on lots of different tumors. Obviously, certain tumors will respond to maytansinoid, others won't. What we have to see is that identifying a dose that's having sufficient number of responses in late [indiscernible] line patients. Final patients are being followed and they were likely to be a report from ImmunoGen that we will participate to discuss that data, but we're not ready yet to talk about it.

And then in the last minute, Paul, I think you briefly mentioned NaPi2b, but it's probably worth emphasizing again, obviously, there were some setbacks with Roche and Mersana for that target, but you're pushing forward and you've developed an ADC for that target. Tell us the rationale, you believe you have a better molecule, as I understand, but tell us why you believe you want to take that forward?

Yes. I think there's multiple reasons. I think there is some clinical precedent for that being a target that can be hit within ADC, just that it wasn't done enough with the molecules that were tested. So there, we think, particularly on these indications, again, the payload is important, the topotecan for the indications that we're going after there, weren't a TOPO based molecules. And then we've also incorporated the bystander activity. The antibody is unique. And then our profile in preclinical where we've looked at both DAR4 and DAR8 looked good. We decided to go with the DAR4, we think that lower DAR here, considering the potential lung expression there warrants that approach, but our preclinical deal looks very encouraging. And we're excited to have a potential best-in-class molecule there.

Perfect. All right. Well, that wraps it up. So thanks again. We'll do this again next year.