IDEAYA Biosciences, Inc. (IDYA) Earnings Call Transcript & Summary

All right. Great. Welcome, everyone, to Citi's 2026 Virtual Oncology Leadership Summit. We have a smattering of small mid-cap and large-cap companies over the next 2 days doing fireside chats. We also had a KOL call this morning on breast cancer. But this session, a very important one, is with the senior management of IDEAYA, and it's my great pleasure to introduce three of the executives at the company. Josh Bleharski, CFO; Michael White, CFO; and Darrin Beaupre, CMO. And just remember, for those listening, if you have questions for any of these gentlemen, please e-mail me at my Citi e-mail address, and I can relay them over to the company

So welcome, everyone. Thank you very much. Josh, maybe we could just start off, if you don't mind, with a kind of a quick level set. You have a lot of programs, but you have obviously a very important catalyst coming up in MUM. So just give us a few high-level thoughts as far as where the company is now and what are some of the key things to look forward to in the next quarter or 2.

Yes, sure. Thanks, Yigal, and thanks to Citi for hosting us today. So IDEAYA is a clinical-stage precision medicine oncology company. We have about nine programs in the clinic today led by darovasertib, which is our lead molecule for uveal melanoma. I think as you mentioned, we have an upcoming top line data release for the combination of darovasertib and crizotinib in the metastatic setting of uveal melanoma. As we noted in our Q4 earnings release, we have triggered the 130 events required for the top line analysis. And we are in the process of collecting that data and analyzing that data and expect to put it out at the end of March. So very exciting data event on the near-term horizon here. Beyond the metastatic setting, we're also exploring darovasertib outside of the metastatic setting in the adjuvant and neoadjuvant settings, we're already enrolling in neoadjuvant study and expect to begin enrolling in the adjuvant study in Q2 of this year. So exciting to see where darovasertib can add benefit beyond this initial metastatic indication. Beyond darovasertib, we do have a deep pipeline, I'd say, next asset is our DLL3 Topo-I ADC that is expected to begin, well, it is in a Phase I study currently in the U.S., we're enrolling patients. We've guided to providing an initial Phase I data update from that trial by the end of this year. And also to initiate a registrational study as a monotherapy in sort of later line small cell or NEC at the end of this year as well. We also have a number of assets focused on the MTAP space, led by IDE397, our MAT2A inhibitor that we've been exploring in combination with Trodelvy for MTAP-deleted urothelial cancer. We'll look to provide an update from that study in 2026 at a medical meeting. And then obviously, earlier in the pipeline, we have a number of assets that are entering the clinic now led by our bispecific B7H3/PTK7, Topo-I ADC that's entering Phase I studies right now in our KAT6/7 dual inhibitor IDE574, which is also entering Phase I studies right now. We ended the quarter with about $1.5 billion in cash, and our runway is into 2030. So I feel like we're well positioned for 2026 and looking forward to the top line release.

Okay. So that's a very good outline for the next 45 minutes. So if we could start maybe with the immediate catalyst. I know there's a limit to what you can say since you just got to lock the database of understanding or in the analysis phase. But can you just kind of at a high level, just remind us what the study -- how the study is designed, what are the key -- what should be the expectation in terms of a positive scenario for the study? What would you expect to see both in your -- in the combo arm, as well as in the control arm. And maybe we could talk a bit about some of the patient demographics as well after that.

Right. So I can give an overall summary. In terms of the patients that are enrolled, that we think the patient population is going to be very, very similar to what we've seen in our 01 trial and the TEVI study. So we're talking about now a study in metastatic uveal melanoma, specifically in HLA-A2 negative patients. Where we're asking the simple question is how does the darovasertib crizotinib compare to what would be considered standard of care in this space. And today, it depends on what part of the world you're in, in the United States. First-line therapy typically is immunotherapy, which often includes the ipi/nivo combination. If you're in Europe, they frequently use single-agent checkpoint inhibitors. And if you're not a candidate for immunotherapy, occasionally chemotherapy is used. And so those are what the treatments are -- which are available in the comparator arm. That's the control arm. So what are the expectations for the kind of output you might see? What is -- what are people aware of in terms of benefit from those kind of therapies? So focusing in on the control arm. There were two large meta-analyses that we've presented in our SMR abstract nice curves, both from PFS and OS. One of them was the Rental publication enlargement analysis, thousands of patients, same with the Australian meta-analysis. But both of them show that in general, the expectation for standard of care therapy in this space is a median progression-free survival of about 2 to 3 months. That's pretty standard. And then in terms of overall survival, you're looking at about 13 months. Now if you look at the darovasertib in crizotinib combination, the data that we presented recently at SMR and asked, okay, so what does that data look like? There, we presented that the median progression-free survival in first-line patients was about 7 months in the median overall survival was about 21 months. So significantly better than what's been reported with standard of care agents in terms of progression-free survival and overall survival. In terms of response, again, standard of care therapies delivers a response rate of around 10%. And then that SMR presentation, we showed that we have a 30-plus percent overall response rate. So in all of the major parameters, ORR, PFS, you're looking at, at least a doubling to tripling of what standard of care therapy does and then with OS an approximate doubling. So that sets the stage for this randomized trial, which, again, as Josh has pointed out, we have the events necessary to begin the data cleaning process. The query resolution process and the data evaluation process to generate the data set. Again, success criteria here. We're expecting that the control arm will produce a median progression-free survival of 2 to 3 months. And based on our success criteria of 5.5-month median progression-free survival in the daro crizo arm would equate to success. But we think we're going to beat that. We think we're going to beat it by a lot. What we would implore people to consider as they look at the data as we will share it at the end of this quarter around that period of time is to really look at the hazard ratio, right? Because it's really the comparison between the two arms that really tells you the story. We don't expect some people have are, oh, maybe the control arm is doing better. But you have to remember, we're scanning people every 6 weeks. And so therefore, we're likely to see events earlier than some of the other trials that have been conducted. And so I think it's unlikely that the control arm is going to do much better than what's ever been published before. Very unlikely. And so therefore, this 7 months or so of what you see with darovasertib is likely to be magnified. But again, the data will speak for itself. We'll see it soon. And so what may you hear at the time of the top line, the kind of things that would likely to provide you are, what was the overall response rate for the treatment control and the control arm. What was the durability of that, maybe in the case of duration of response. We obviously will provide what kind of progression-free survival we've seen and maybe a confidence interval. But one of the other things I wanted to point out, if we don't say anything about overall survival, that doesn't mean there's a problem with overall survival. What it likely means is that there'll be -- there's just not enough events to make a call at this time. So again, we're looking at this early. The first interim for OS happens at the beginning of 2027. However, during that period of time, the FDA will have the opportunity to look at overall survival from data generated from the study, not only at the time of the top line analysis, but also, during the filing, as we provide updated safety data, we're likely to provide an efficacy update there as well. So long and short of it is, we think we have a high probability of success based on the data that we showed at SMR. We believe that the control arm is going to perform exactly as what's -- how it's performed in thousands of patients based on two large meta-analyses. We think the data from our SMR publication is very robust. It's got 2 years of follow-up and the data has been very solid through that entire period of time. So really, hopefully sets the stage for a very important study for patients with uveal melanoma, which you know have very few options, only one approved therapy, and that one approved therapy has a response rate of around 10% and a median progression-free survival of 2 to 3 months. So this, we think, will be an improvement on what therapy is out there for this disease.

Okay. So that's very, very helpful. So if I understand correctly, your point is that because you're scanning at this higher frequency every 6 weeks, therefore, you should be able to capture some of the early progressions. And therefore, that pushed this expected range of 2 to 3 months to the lower end?

It might. I mean, I'm speculating. But remember, when you get a scan at 3 months as that's the first time you get a scan, you could have progressed 2 months before that. You never know it because nobody scanned them, right? So -- and with immunotherapy, sometimes they even treat post progression, right? So the progression events could actually happen earlier, not later. But you know what, the data will speak for itself. I don't believe, based on all of the data that's been published that the control arm is going to be much better than what we've seen in these meta-analyses. And if that turns out to be the case, and we're doing earlier scans and we have a rock solid 7 months for progression-free survival, with our daro crizo combination, I just don't see how it could lose. But you know what, the data will speak for itself.

And you mentioned that you think you beat the -- the 5.5 is sort of the threshold number, whereby if you get 2 to 3 in the control and you hit 5.5% in daro criz, then it works?

Correct.

And -- but you said that you think you could beat that 5.5 by a lot. And that's based on -- tell us more about that, that's based on the Phase II experience? Are there other factors that drive that statement?

Yes. No. So well, the statement is around the 7 months. So 7 months is clearly better than 5 months, 5.5 months, right? So that's what I'm talking about. I'm thinking we're going to be better off than 5.5 months. But again, how much, I don't know. But when the data reads out, that's when it will tell us. But that number has been pretty stable, pretty stable.

Okay. And then there was also some discussion, I think, over the last several months related to the relative proportion of ECOG 0 versus ECOG 1 and that relative proportion may give you some tailwind. Can you just expand upon that?

Well, there was some more patients with -- like if we're comparing across the TEVI study to our 01 trial, there were more performance status in our study relative to the TEVI study, suggesting that the data that we have may actually be better in a similar patient population to the TEVI study. But the reality is, is in the registration trial, I suspect all these things even out. It will be a patient population that's very much akin. Now in terms of the geographic, there may be some variance there because that's another question that comes up a lot. We're probably -- if I had to guess, we're going to have probably about 60% U.S., 40% EU, and it might have been sort of flipped with the TEVI study. So obviously, they didn't have any ipi/nivo in that particular study, but we'll have a significant proportion of ipi/nivo in our study, probably at least half of the patients will have ipi/nivo. So those are some interesting distinctions between the two. But again, most people -- ipi/nivo, this is going to be the study where Marlana Orloff said it herself just yesterday is an interesting study because people are going to also be able to look at how does ipi/nivo compare to single-agent checkpoint inhibitors. And if you -- based on everything we know today, we would expect it to be more toxic. So that wouldn't surprise us if we saw that. And then in terms of efficacy, many people believe especially in terms of overall survival, even in PFS, there's probably no difference between single agent checkpoint inhibitors and the combination. You might have a little bit of a leg up with a response rate with ipi/nivo, but she says, you know what, those are all been small studies. This will be the largest study that actually looks at that. And the way we've designed it, we've designed it purposely to say that, you know what, we're the king, we're the best therapy out there. You can't argue that we didn't have the right controls. You can't argue that we didn't have the right designs, I can't argue that we didn't have the right output with a strong PFS, and it all set, I think, the new standard for what goes on in uveal melanoma. But again, the data will speak for itself, and we'll be hearing about that soon.

Now what you mentioned also that OS, you may say something, you may say nothing, don't interpret saying nothing as a negative in any way. But what would -- what sort of -- what percent of events accrued in OS would be enough to be able to make some sort of statement on a trend or otherwise?

Yes. Like many people think it's around 30% or so. If you're kind of less than 30%, what can you really say? You're starting to get above that 30% threshold, you start getting some confidence. Obviously, the higher you get above that, the more confident you become. So if we're sort of at a threshold where it's still a bit too early, we'll just probably be silent on.

But it sounds like you -- I mean, it's not so far away that it's a definite no. It sounds like you may be getting close to that. Is that fair?

Well, and like I said, we have -- there's this opportunity that we have to look at the data during the filing as well, as well as the time of the first interim. So between the top line -- and the first interim that happens at the beginning of next year, there will be an opportunity to look at the data. And then we'll see how it matures. Remember, an important factor here is this study doesn't allow crossover, right? There's no crossover because we want to preserve overall survival. That's what the FDA wants to do, right? But if there's an obvious trend to improvement in overall survival, and it's unlikely that it's going to reverse. Now we have an ethical dilemma, don't. We have something to think about for patients. How long are we going to wait before we allow them to cross over. And so that's a discussion that has to be had with the FDA as the data evolves. It's a discussion that has to be had with our independent data monitoring committee. And again, it's all about the data and how it evolves. And if it evolves in the positive direction, highly positive, really, these are the kind of discussions we have to have.

Okay. So but for the immediate situation for the end of March, we should not expect anything to be said about OS. Is that the plan?

I wouldn't say that. I would just say that if you don't -- what I said was is if you don't hear anything, don't misinterpret it.

Okay. And then -- but the -- but then you have a formal interim OS at the beginning of 2027, which would be triggered by what percent?

Yes. So that number of events will be north of 30% for sure.

Okay. Got it. And then, maybe we could just turn to like assuming you hit as well it seems very, very, very possible with the numbers that you've cited. What would the filing time lines look like for accelerated approval? What would you -- how would you approach the HLA-positive group, for example? Can you speak to those questions?

Yes. So standard metrics just in terms of filing an FDA review in a case like ours, again, we're going to try to compress this as much as we can, but it's about 6 months to file. We'll try to compress that. And then obviously, the review by the FDA will be 6 months at least. So that's the time period. And then in terms of the HLA-2 positives, we're very excited about that group. The biology says there shouldn't be any difference. We presented data at ESMO in 2023 to show that HLA-2 site patients behaves the same as HLA-2 negative at least when they get treated with the daro crizo combination. We have a data set that's evolving from our Phase II trial, the 01 study that is accumulating about 100 patients' worth of data who are HLA-A2 positive, some of which are treatment naive, but I think a significant proportion of them will be heavy exposed. And we hope over the next year to be able to present that data in two chapters. Chapter 1 will be maybe the first half of the patients because there we'll have about 6 months of follow-up data from that first group of patients that was enrolled. Where we'll be able to share the overall response rate, progression-free survival and perhaps even some OS data depending on how mature the data is. Again, what the idea is showing that the combination is very active in this patient population. And then around the time of the filing, around the time of the approval, what we want to do is get this publication out on HLA-A2 positives that will be that whole 100 patient sample size where at the minimum, we can publish that data provided to the NCCN group to evaluate to consider as part of the NCCN guidelines for therapy with patients -- for patients who are HLA-A2 positive. As you know, their treatment options are very limited. With the only approved systemic therapy being TEVI, so it will be an opportunity for patients in the United States to get access there. But we're also discussing ways we could have a discussion with the FDA about how we could potentially get this on the label. As you know, we're talking about an orphan indication with a large unmet medical need with a very short overall survival in general. So it will be interesting to have that discussion with the FDA around their thoughts. Potentially real-world data could be interjected in the discussion. But also, there's something to be said about just the fact that the biology says that this combination doesn't work any different. We'll have data to support that it doesn't work any different. And so talking to the FDA about the opportunities for the label will be something that we'll be doing over the next -- over this year as we get into the launch phase.

Okay. Well, let's move on to quickly talk about the other -- some of the other lines you're looking at. Adjuvant and neoadjuvant. What are the -- remind everyone what the time lines are there for getting to the critical readouts for those populations?

So just recall, the registration trial for neoadjuvant darovasertib single agent and primary uveal melanoma has two cohorts, right? The first cohort of patients with large tumors those who are destined for a nucleation where the primary endpoint is saving the eye. And then the second cohort are patients with smaller to midsized tumors where plaque brachytherapy is the therapy of choice. And there, it's preservation of vision as the primary endpoint. An important endpoint for both cohorts though is event-free survival. We need to show that by providing neoadjuvant darovasertib in delaying primary local therapy, we don't provide any detriment to patients. And so what we'll do that is show that by using neoadjuvant darovasertib, we provide no detriment with respect to local or distant relapse. So that's the tumor coming back in the eye or the tumor metastasizing. What we have to show is there's really no detriment to subjects when we give this neoadjuvant therapy. In terms of enrollment, it's going to take probably a year to 1.5 years to enroll subjects and the readouts will take about 1.5 years from the time of primary local therapy. Now with respect to the readouts, it could be that the enucleation cohort actually reads out earlier than the plaque brachytherapy cohort. You can imagine that being the case when you're only giving up to 6 months' worth of neoadjuvant therapy. And so within that period of time, you'll know whether you save the eye or not. So the rate-limiting step really will be the event-free survival readout. But again, we may not have vision data from the plaque brachytherapy cohort at the same time that we have all of the nucleation data. So therefore, it remains possible that we read out on the nucleation first. We present that to the FDA, followed by the plaque brachytherapy data. However, if the two data sets come out closely opposed in timing, we may try to file them both together. It really just depends on the timing, but our suspicion is we'll be able to read out a little bit earlier with the nucleation cohort relative to the plaque brachytherapy cohort.

Okay. And so that would put -- if they are kind of proximal to one another, that would be roughly when in a calendar time frame?

Well, again, as I mentioned, 1.5 years to enroll and then 1.5 years from the time of the primary endpoint. So a minimum 2.5 years. We just started dosing.

Yes, Yigal, what we've guided to is being done on enrollment for neoadjuvant in the first half of next year. So that's all.

About 1.5 years from there.

All right. And then what about the adjuvant study?

Adjuvant study, great question. So that one is on the slate. We're actually putting it together now, plan a discussion with the FDA coming up Type C meeting. We have a trial design in hand. Again, there, we're going to take patients who are at high risk of metastases, which is at least half of the patient population. They're going to be randomized to your standard observation, which is what we're doing today, just monitoring patients, waiting for them to relapse. And again, the median time to relapse is about 3 years. So there's a huge proportion of patients at least in this group, probably 70% to 80% that will metastasize by the 3-year time point. And so what we'll be looking at is a comparator of daro crizo for up to 1 year versus observation -- and so what we'll be attempting to do is show that by providing the darovasertib crizotinib combination that we can reduce the number of metastases down the road in patients who are high risk for metastatic uveal melanoma. So that trial is sort of in swing. It's underway, and we hope to get that launched by midyear.

Okay. All right. Let's move on and quickly to some of the other important topics, and we covered a lot on uveal melanoma. So for the ADC program for the DLL Topo-I ADC, maybe you could just talk a little bit about the time lines there on small cell and neuroendocrine and when we could expect the next readouts?

So that -- so you've seen, obviously, the data from China, it's presented at the World Lung Congress highly exciting, very high response rate with this molecule, 70% with progression-free survival, 6 months or better based on some of the early reads. So it looks very exciting. Our U.S. study based study has just launched just starting to enroll. We think by the end of this year, we're going to have safety data. The nice thing about this study is we were able to start the dosing at a therapeutic dose level. So there really isn't any subtherapeutic dosing envisioned for the trial. So right from the get-go, right from the first patient, we can start expecting some efficacy data to come in. So there will be a dose escalation that goes on in U.S. patients with our antibody as a monotherapy, but in addition, we'll also be combining it with 161, which will be triggered in the second quarter as we get the first initial data from the first cohort of enrollment for this study. And in addition, what we'll be enrolling is both extensive stage small cell lung cancer patients, but also neuroendocrine carcinoma subjects. So by the end of this year, what we hope is we'll have maybe two handful or more worth of patients where we'll be able to provide some preliminary response data, maybe some durability of response data. We'll be able to provide safety data both as a monotherapy and in combination with 161, and we'll be able to provide some early data in some subjects with neuroendocrine carcinoma. Again, we won't have a lot of follow-up because, again, we're just starting to enroll now, so it won't be long-term follow-up, but the initial safety and efficacy data is what we hope to put out sort of by the end of this year. That's sort of where we stand with the Phase I currently.

Maybe I can ask Michael this one or you, Darrin, the combo you mentioned with the PARG, the 849 plus 161, what's the -- just remind us why that is potentially synergistic. What is the biological rationale there?

Yes, absolutely, Yigal. So first of all, DLL3, it's an amazing target. This asset is doing really well as a monotherapy. But as you know, the value proposition is also around durability. And with respect to ADCs with topoisomerase payloads, the ability of the payload to kill those cells is a function of DNA damage as well as DNA repair. And if you get enough payload in, you kill the cells, if you don't, you get progression. And so what we're looking for with this combination is to get durability by maximizing the therapeutic benefit of the topoisomerase payload. The mechanism of action behind that is a very special role that PARG plays with the resolution of topoisomerase I cleavage complexes. So the way these inhibitors work is they bind topoisomerase I, they clamp it on the DNA together with a break in the DNA. That complex gets correlated by PARP. And then it becomes a substrate for DNA repair. However, in order for that to be repaired, it has to be deporelated first so that the proteasome can come in and chew off the topoisomerase. There's only one enzyme in the cell that does that, PARG. So PARG is essential for the resolution of topoisomerase I cleavage complexes. So whatever topoisomerase inhibitor gets in there, we are trapping that cleavage complex on the chromatin replication for comes through and it collapses. So we are amplifying the therapeutic benefit. If we have suboptimal payload delivery, we can get now a more optimal therapeutic response. Very exciting mechanism of action. So we are really looking for enhanced durability beyond what you would be able to get with a monotherapy opportunity. which leads us, by the way, into our selection of IDE034, B7H3/PTK7 bispecific. We have another opportunity there that's distinct, if you want to talk about that at some point.

Let's do that right now. Yes, tell us what's the synergistic thinking there?

Yes. So the point for a bispecific for us is to be able to maximize tumor-specific delivery of the topoisomerase inhibitor. And we pick B7H3/PTK7, these are both targets that have been validated in the past. Both of these targets make biological contributions to the tumor itself. PTK7 is particularly interesting because it is enriched on the surface of tumor initiating cells that are the cells that produce diversification and tumor heterogeneity that leads to adaptive bypass mechanism. So very nice if we could get rid of those cells within ADC. The point of the bispecific is we have an asset that has been built to have enhanced binding to double positive cells, B7H3/PTK7 double-positive cells. There are abundant evidence using protein for large populations of tumors in important indications like lung, colorectal, triple-negative breast cancer, ovarian, prostate, head and neck, that have double positive tumors. And the only normal tissue where we're seeing double positivity is endometrial, otherwise is single positive. So bispecific, much more enhanced binding to tumor versus normal, get good therapeutic window. That can come at a cost for payload delivery, the amount of payload that you get in, we are going to maximize that with IDE161. So we're expecting this ADC, IDE034 to be really exciting for monotherapy. And then in combination with IDE161, PARG inhibitor, will maximize the therapeutic impact of that payload that's delivered precisely to the tumor. So we're expecting nice response rates with durable responses in some really important tumor types.

Yes. I think that's the part, Mike. It's going after the big tumor types, too. So the potential here is very broad, and it's within the big tumor. So the commercial opportunity, this drug is active could be quite large to say the least.

So the ones that express B7H3 and PTK7 would be what you mean like lung, breast....

Yes, all the big ones, all the good ones.

Significant populations of lung adenocarcinoma significant populations of BA triple-negative breast cancer, HRP, ovarian cancer, prostate, head and neck.

And have we seen anything yet clinically for that?

Just starting to dose now, just starting the dose now, getting ready to dose.

All right. So then moving on, it's a lot to cover in 40 minutes. But so the MTAP is a huge topic, obviously. So we're expecting an update, I believe, for the MAT2A 397 and Trodelvy. So you've gotten some good data so far, and I think you've arrived at a dose. So can you just talk about what the -- what you envision seeing there with this combo in the next update?

Yes. So just from a high-level point of view, as everyone knows, we've been working on this Trodelvy 397 combination in MTAP deficient bladder cancer. We've presented that data previously. So essentially, that data set has grown since the last you've heard and the data continues to look quite promising with a high response rate and the durability continues to mature. I think what it's done for us is it's really helped us prove the principle of if you get in there and you inhibit -- you get in there with a MAT2A inhibitor and you add in a toxin like Trodelvy carries, you see some special activity. We think we're clearly doing better than Trodelvy could do in this patient population by itself. And so from that perspective, we think we've generated some proof-of-concept data that's exciting about other -- that raises the potential about other ADCs that could be used with a similar toxin. Again, as you know, we're expanding out into lung also. So in lung cancer, that data is still evolving. We're still early days needs more data maturity. But suffice it to say, I think we've kind of proven the point that we have an active regimen here, clearly in bladder cancer, still evolving in lung and we're still discussing exactly how to proceed going forward. But we're really, really encouraged by what we've seen so far. I don't know, Mike, if you want to speak anything to just sort of the proof of principle of the biology here.

Yes. I think I'm right with Darrin on this one. I think it's very exciting that we essentially have proof of concept of the notion that we can amplify replication stress with a combination like this one and do something special. This is something that we've been looking at from a mechanistic perspective for quite some time. alternative splicing, you mess that up, you cause pausing of RNA polymerase that gives you transcription replication conflict, replication stress and then IDE397 because of its role in the methionine cycle is reducing the capacity of these tumors to be able to produce the nucleic acids by the de novo synthesis and salvage pathway that they need to be able to deal with repair. So we're increasing damage, we're reducing the ability to repair and we're seeing that work in patients. So very excited about that. And don't forget the magic here is the ADC with the Topo-I payload. We've got proof of concept with urothelial with Trodelvy. I think there's no reason for us to think that this would not work with other ADCs and other indications. So a lot of room here for us to think about additional opportunities with this combination in MTA all tumors.

Okay. And the thinking with the other combo with 892 and PRMT5 is totally different combo, but what's -- how is that -- what are the advantages of that one versus Trodelvy? It's obviously a totally different thesis biologically.

Yes, totally different mechanism of action. And this one is also very important, very exciting, particularly, we think in the context of lung cancer. Lung cancer is a very mechanistically heterogeneous disease. It has lots of adaptive bypass mechanisms. We know that tumors on PRMT5 inhibitors have acquired resistance. We know that PRMT5 inhibitors cause alterations in chromatin architecture that enhance phenotypic plasticity that leads to resistance. And we know that IDE397, an MTAP inhibitor -- sorry, an MTA, a MAT2A inhibitor will intercept that modulation of chromatin in order to be able to give us a durable response. So that combination, we can inhibit bypass mechanisms of PRMT5 inhibitors, but we can also maximize the ability of perturbation of alternative splicing to work in MTAP null cells because the whole synthetic lethal principle for MTAP loss is MTA accumulation. MTA accumulation allows MTA cooperative PRMT5 inhibitors to work. They allow MAT2A inhibitors to work, but MTA accumulation in tumors is variable. It depends on the extent of polyamine synthesis. It's a byproduct of the production of polyamines. And that combination we have shown can maximize the MTA/SAM ratio to be able to maximize the synthetic lethal effect. So we think we will get deeper responses in the combination setting because we extinguish the pathway in the MTAP in all cells, and we think we'll get more durable responses because a MAT2A inhibitor intercepts adaptive bypass mechanisms that otherwise occur with PRMT5 inhibitors.

Okay. And then there was a third one you just nominated, the CDKN2A candidate. How -- that's also sort of in this pathway broadly. How does that play into the biology? And what additional tumor types could you access by attacking that?

Yes, that's a great question, Yigal, because the 15% of the tumors have MTAP loss, maybe 15% of lung cancer, 20% to 30% of pancreas, 25% to 40% of urothelial, 20% of esophageal, 17% head and neck, big populations. Those get -- those happen because actually CDKN2A is lost. So CDKN2A loss is the most common co-alteration with MTAP loss because CDKN2A harbors two of the most important tumor suppressors known to mankind. You lose those things, you release cell cycle progression. We have a way in to be able to attack that co-alteration and use that as another mechanism of action to deploy with PRMT5 inhibitor, MAT2A inhibitor or actually maybe even a triple combo to be able to maximize the insult on the tumor that has these 9p21.3 regions. So CDKN2A loss drives MTAP loss. So everywhere you have MTAP, pretty much you have loss of CDKN2A. So we're excited about this combination. And as long as we're on the topic of combos, we're really trying to do as much productive damage as we can in this MTAP space and in pancreas cancer, where the frequency of MTAP loss, CDKN2A loss is upwards of 30% CDKN2A deficiencies are even higher, perhaps the majority of pancreas cancers. The big co-alteration there would be KRAS, right? So KRAS is mutant in pretty much all pancreas cancers. And so we have also been looking into that combination. We've been talking to some folks who are excited about doing that. We're in a unique position in pancreas cancer potentially for KRAS PRMT5/MAT2A triple combo. That's a very, very hard space to play in pancreas cancer. So you need a very effective therapy with a good therapeutic window.

Okay. Well, we're not able to get to everything because you have a lot, and we're kind of out of time. So we'll leave some of these other topics on KAT6 and Warner for the next meet up. But I appreciate all the details and the thoughts on melanoma and MTAP and the DLL3. And we're, of course, looking forward to the readout at the end of March, as you point out. So thank you again very, very much.

Thank you, Yigal.