Good morning and welcome to the IDEAYA Biosciences Investor R&D Day. At this time, all attendees are in a listen-only mode. A question-and-answer session will follow the formal presentations. As a reminder, this call is being recorded, and a replay will be made available on the IDEAYA Biosciences website following the conclusion of the event. I'd now like to turn the call over to your host, Yujiro Hata, Founder and Chief Executive Officer of IDEAYA Biosciences. Please go ahead, Yujiro.
Good morning. This is Yujiro Hata, CEO of IDEAYA Biosciences, and I'll be your host today, and a warm welcome to all of our listeners, and thank you for participating in IDEAYA's 2024 Investor R&D Day. Please note we will be making forward-looking statements, and please refer to our SEC filings as appropriate. To kick us off, I'll first introduce our terrific lineup of guest speakers and provide an outline of today's agenda. It gives me great pleasure to introduce our wonderful guest speakers today: Dr. Carol Shields, Chief of Ocular Oncology at the Wills Eye Hospital; Dr. Amy Scheffler, a leading ocular oncologist from the University of Texas; Dr. Kornelia Polyak, a world-renowned breast cancer specialist from Harvard Dana-Farber; and a member of the National Academy of Sciences and the National Academy of Medicine.
Dr. Timothy Yap from the MD Anderson Cancer Center, who is a leading KOL in the area of DNA damage repair and precision medicine oncology, and Dr. Ramon Kemp, Vice President and Head of Early Oncology Development. For today's agenda, I'll first lead us off with a brief summary of our vision and strategy to build a leading precision medicine oncology company. Then Dr. Carol Shields and Dr. Amy Scheffler, hosted by our CMO, Darrin Beaupre, will facilitate a neoadjuvant uveal melanoma roundtable. Next, Dr. Kornelia Polyak will provide a detailed walkthrough of the challenge of tumor heterogeneity, which is one of the central themes of today's Investor R&D Day. Then Dr. Timothy Yap will walk through the mechanistic clinical combination rationale for the phase I PARG inhibitor IDE161 with PD-1 inhibitor Keytruda and with topoisomerase payload-based ADCs.
Next, Dr. Ramon Kemp will provide an overview of IDE275, also known as GSK959, our phase I Werner helicase program. Lastly, Dr. Michael White will walk through our three new development candidates, including PRMT5 inhibitor IDE892, bispecific B7H3/PTK7 ADC IDE034, and KAT6/7 dual inhibitor IDE251. Lastly, we'll finish with closing remarks, preliminary 2025 guidance, and the endless Q&A portion of the webcast. We have several core strategic pillars that have guided our efforts over the last nine and a half years to build a leading precision medicine oncology company. Through these efforts, we have advanced five potential first-in-class agents into the clinic that span a potential registration-enabling trial to early phase I clinical trials and over five high-conviction preclinical programs, including three that are now in IND-enabling studies.
In aggregate, these 10+ clinical and preclinical programs are targeting a large range of patient selection biomarkers from GNAQ/11, MTAP deletion, BRCA, high MSI, 8p11 amplification, B7H3/PTK7, among many others. For today's R&D Day, our primary topic will be addressing the challenge of tumor heterogeneity through transformative combinations. Since IDEAYA's founding nearly 10 years ago, one of our core strategic objectives that connects each of these programs in our pipeline is the goal to address what we believe is one of the greatest challenges we face in effectively treating cancer, and that's the challenge of tumor heterogeneity. One of the earliest reports of tumor heterogeneity dates to 1833, nearly 200 years ago.
Since then, we have learned that tumor heterogeneity comes in many forms, from spatial or structural heterogeneity that can be partly driven by differences in tissues as well as cell types, genetic heterogeneity that has been uncovered from the advent of molecular biology and genetic sequencing, clonal diversity driving selective pressure and resistance to therapies, among many others. This multifaceted nature of tumor heterogeneity is what we believe presents one of our greatest challenges in treating cancer. More specifically, this has hampered our ability to deliver a broad, consistent, and truly deep and durable response in our target cancer patient population. I wanted to provide several examples of approaches we have utilized as an industry to address the challenge of tumor heterogeneity. One is early diagnosis and the ability to deliver the most effective therapy early in a patient's journey.
To put it simply, typically, the more advanced a cancer disease gets, the level of tumor heterogeneity is likely to increase, highlighting the importance of intervening early with a highly effective therapy. Second is the ability to deliver the precise therapy to the right patient at the right time through identification of an actionable cancer driver mutation. And the final example I'll provide in one of our most effective approaches to address tumor heterogeneity is through rational combinations, which has been a core strategic focus since our founding.
We believe IDEAYA has been an early pioneer at enabling several novel rational combinations in the field of precision medicine oncology across multiple clinical programs and biomarker populations, including PKC and c-Met and GNAQ/11-driven metastatic uveal melanoma that is now in a potential registration-enabling trial with our partner Pfizer, MAT2A and PRMT5 to target MTAP deletion lung cancer with our partner Amgen, MAT2A and TROP2 Topo ADC and MTAP deletion urothelial cancer with our partner Gilead, Werner helicase and PD-1 to target MSI-High solid tumors with our partner GSK, Pol Theta and PARP to target BRCA solid tumors, also with our partner GSK, and PARG and PD-1 to target MSI-High and microsatellite stable endometrial cancer with our partner Merck.
Through these rational combination approaches, we believe we have targeted key underlying biology of these cancer pathways to directly take on the challenge of tumor heterogeneity with a singular goal: to deliver a safe and more durable response in patients. To realize IDEAYA's vision and strategy over the last decade, we have worked towards building a world-class R&D and drug discovery enterprise to deliver potential first-in-class therapies across a multitude of novel and high-value precision medicine oncology targets for MAT2A, PARG, Pol Theta, Werner helicase, PRMT5, KAT6/7, among many others, and through the discovery of these novel agents, we have enabled a broad range of potential first-in-class rational combinations across multiple biomarker-defined solid tumor populations, including GNAQ/11, MTAP deletion, BRCA, MSI-High, 8p11 amplification, among many others.
We will now begin the guest speaker section of our webcast with a neoadjuvant uveal melanoma panel, and through the rest of the sections, we will showcase our various potential first-in-class precision medicine oncology programs and do a deep dive on how we are addressing the challenge of tumor heterogeneity across several of our preclinical and clinical stage programs. It now gives me great pleasure to introduce our first guest speakers, Dr. Carol Shields and Dr. Amy Scheffler, for the neoadjuvant uveal melanoma panel. Darrin, please take it away.
Thank you, Yujiro. I would like to welcome everyone today to this roundtable discussion with both Dr. Amy Scheffler from Retina Consultants of Texas in Houston and Dr. Carol Shields from Wills Eye Hospital in Philadelphia, Pennsylvania. These are two well-known and established ocular oncologists who have participated in the darovasertib clinical development program and have attended our recent Type C FDA meeting where we discussed the registration trial of darovasertib in the neoadjuvant setting. The purpose of this roundtable is to highlight some of the key aspects surrounding the management of primary uveal melanoma and to discuss the potential darovasertib has in the future management of this disease. Dr. Shields, I thought we could begin with a very brief overview of the general management of subjects with primary uveal melanoma, specifically subjects who require either enucleation, plaque brachytherapy, or proton beam.
Perhaps you could describe which one of these therapies predominates in the treatment landscape, Dr. Shields?
Yes, thank you, Darrin. Primary uveal melanoma is the most common malignancy or cancer of the eye, and its management varies depending on tumor size, location, and secondary features like retinal detachment, glaucoma, optic nerve involvement, and lastly, patient or family desire, so currently, in 2024, approximately 80% of patients with choroidal melanoma are managed with plaque radiotherapy or proton beam radiotherapy, and approximately 20% of patients are managed with enucleation, that is, removal of the eye. Typically, small melanomas measuring 0-3 millimeters in thickness and medium melanomas measuring 3-8 millimeters or 10 millimeters in thickness can be treated with radiotherapy. On the other hand, large melanoma over 10 millimeters in thickness are generally offered enucleation. Previous prospective studies, namely the Collaborative Ocular Melanoma Study, have shown that plaque radiotherapy is equivalent to enucleation for medium choroidal melanoma.
Generally, after plaque radiotherapy or proton beam radiotherapy, the eye shows radiation side effects that have their onset about one to two years later, and this can lead to permanent vision loss. Of course, after enucleation, the patient has no vision and is fitted with an artificial eye that does not see. So this pretty much sums it up in a nutshell: choroidal melanoma management.
Thank you very much, Dr. Shields. This is very, very helpful. Before we get into the specific details around the key aspects of primary uveal melanoma that darovasertib could address, let me provide a few introductory slides around some of the key observations that have been made with the use of darovasertib in the neoadjuvant setting. This slide provides some data that we've presented previously from the investigator-sponsored trial being run in Australia by Dr. Anthony Joshua, along with data from our own IDE-196-009 study. Shown here are uveal melanoma subjects who either required enucleation or plaque brachytherapy and were treated prior to definitive management with neoadjuvant darovasertib. As shown in this waterfall plot, the vast majority of subjects experienced tumor shrinkage, and in addition, 61% of subjects destined for enucleation had their eye preserved, a remarkable finding.
Focusing now on subjects who require plaque brachytherapy, that is, those subjects with small to mid-sized tumors from study IDE-196-009, we have seen evidence of not only tumor shrinkage, but a reduction in the dose of radiation to key visual structures, including the fovea, macula, and optic disc nerve, and in addition, using a vision prognostication tool, we've been able to determine that a number of subjects have the potential for preserved vision secondary to neoadjuvant therapy. In this slide, I think it's of value to consider the parameters used in this vision prognostication tool developed by Arun Singh and his colleagues at the Cleveland Clinic, which is shown on this slide. Those parameters include tumor size, particularly width, as well as radiation dose to the fovea and optic disc.
Therefore, one would surmise that a therapy that can shrink tumors and reduce the amount of radiation needed to control them could have a significant impact on long-term visual outcomes. So with that, I'll transition back into the questions for our guest. Dr. Scheffler, you and your colleagues are on the front line of treating patients with uveal melanoma, and you see patients frequently who have tumors large enough where their eye needs to be removed. Based on your interaction with patients, could you comment on the benefit of enucleation sparing for patients like this? And also, with darovasertib in the neoadjuvant setting, do you see a day that perhaps enucleation sparing could be a very rare procedure?
Yes, Darrin. I do believe that even today, the majority of patients, if given a choice, would choose not to undergo enucleation. As Dr. Shields said, this treatment sadly is still necessary in some cases, a minority, but some cases, due to the size of the tumor or patient's other concurrent medical conditions. And eliminating the need for this surgery entirely would be an excellent goal for patient satisfaction, no question about it.
Thank you, Dr. Scheffler. In the data that's been presented thus far for darovasertib, most patients treated in the neoadjuvant setting have tumors that shrink, which is likely due to the fact that darovasertib targets protein kinase C, which is a key driver for this disease. Dr. Shields, perhaps you could spend a few moments discussing the positive outcomes that could come out of tumor size reduction for subjects planned for plaque brachytherapy, specifically when it comes to things like radiation reduction, vision preservation, and perhaps even development of metastases, since these are all endpoints that are planned to be evaluated in the neoadjuvant registration trial.
Thank you, Darrin. darovasertib could and does reduce tumor thickness by approximately 30% based on the study from Australia. So a 12 millimeter thick melanoma heading to enucleation might reduce to 8.4 millimeters in thickness. And initially, this tumor might have only been treatable with enucleation, but now, after three months of darovasertib, it is amenable to plaque radiotherapy because we commonly irradiate an 8.4 millimeter thick tumor. This is a huge difference, and this is one of the many roles of neoadjuvant darovasertib for uveal melanoma. So regarding radiation reduction for patients undergoing plaque radiotherapy, there are several important publications that have addressed this. In the journal Brachytherapy 2021, Meidenbauer et al. looked at 339 patients from Cleveland Clinic with small and medium melanoma.
The authors studied reduction in radiation dose to important structures because previous teachings have said that any small or medium melanoma must be irradiated up to 5 millimeters. In their study, they actually looked at what the actual apex thickness was, and they didn't irradiate to 5 millimeters. They irradiated to the actual thickness. They noted that there was 30%-40% reduction in radiation dose to the lens, optic disc, and foveola in those who received the actual radiation rather than the previously recommended 5 millimeter thick radiation. They also found that those that received the actual radiation, a vision of 20/50 or better, was seen in 56% compared to 31% in those that received the 5 millimeter thick radiation.
This is an important study because it underscores the point that if we can reduce actual tumor thickness, we can protect visual structures and hopefully improve vision outcomes. Now, there was another report from Investigative Ophthalmology and Visual Science, 2004, by Puusaari et al. They looked at 96 patients from Finland with large melanoma, median thickness 10.7 millimeters and base 16.5 millimeters. Their conclusion was that tumor thickness does predict radiation complications and vision loss after plaque radiotherapy. They said with renovation of their old ruthenium plaque or iodine plaque to a newer plaque that can shift the isodose curves away from visually important structures, they can actually reduce vision complications even more. Now, a comment about vision preservation. Our team from Philadelphia published in Archives of Ophthalmology 2000 an analysis of 1,106 consecutive patients who had good vision and uveal melanoma.
Their vision was 20/100 or better. Following plaque radiotherapy, this group of patients developed poor vision, 20/200 to light perception in 68% at 10 years follow-up. Using multivariate analysis in that study that we published, we showed that factors predictive of poor vision were increasing tumor thickness. With darovasertib, if we can reduce tumor thickness, we may be able to protect vision. Lastly, just a short comment on metastatic disease. Our team from Philadelphia published in Archives of Ophthalmology 2009, "Metastasis of Uveal Melanoma, Millimeter by Millimeter in Over 8,000 Eyes." We found in these 8,000 consecutive patients that the mean tumor thickness on average was 5.5 millimeters, and the five and 10-year rate of metastasis was 15% and 25%. The important finding in that study was every millimeter increase in thickness caused 5% increase in metastatic rate at 10 years.
So reduction in thickness with darovasertib is really an important finding. Thank you.
Thank you very much, Dr. Shields. Another important outcome to be evaluated in the registration trial in plaque brachytherapy subjects with small to mid-sized tumors is the development of macular edema, which we predict would be less common in the darovasertib treatment arm compared to the control arm. Dr. Scheffler, for the audience, perhaps you can describe, you know, what is macular edema? Why do patients get this? What are the consequences? And perhaps you could describe what some of the benefits could be if we could prevent it.
Macular edema is swelling of the retina due to inflammation, which occurs when there is leaking from fluid from the retinal blood vessels. In these patients, this is caused by permanent damage to the pericytes and other cells that are lining the inside of the retinal blood vessels. That damage is caused from the radiation and leads to a cascade of biological signaling that leads to this unrelenting leakage of fluid, extravasation of exudates, hemorrhaging, and so forth, ultimately leading to damage to the healthy retinal cells that are necessary for good vision. Radiation retinopathy leads to vision loss in nearly 50% of patients by just year three after surgery, after brachytherapy, and is the major cause of irreversible significant vision loss in these patients.
In recent years, we and others have performed prospective clinical trials looking at using anti-VEGF medications to help slow down this vision loss, but these medications do not reverse it completely or prevent it entirely. The degree of macular edema and vision loss that patients develop is affected by many factors, including the dose of radiation necessary to treat the tumor, the proximity of the plaque to the fovea, and other patient factors such as diabetes. Nonetheless, nearly all patients develop this side effect eventually. Being able to minimize this or eliminate it by using darovasertib would be an extremely significant result in terms of minimizing vision loss for patients over time.
Great. Thank you, Dr. Scheffler. Dr. Shields, I'll go back to you now. The registration trial that IDEAYA is proposing for darovasertib in the neoadjuvant setting plans to enroll subjects at high risk of metastases. So we're talking about subjects who typically have Class II by gene expression profiling or perhaps monosomy 3 or stage III. And this is simply so that an EFS endpoint can be reached in a timely manner. But what about those subjects who are low risk for metastases, for example, like a Class I subject? If the darovasertib was approved in this setting, do you think that this therapy could be relevant for a patient population like that, the low-risk population?
Absolutely, Darrin. I think that darovasertib would be of interest to both patients with large, medium, and even small melanoma, and for patients with genetic high risk and low risk. So regarding melanoma size, of course, with large melanoma, darovasertib could shrink the tumor so that the eyeball could be saved with radiotherapy. And for small melanoma, darovasertib could shrink the tumor so that a lower dose of radiation to the foveola and optic disc could be achieved. And regarding genetic classification, for high-risk tumors classified as Class II or monosomy 3, of course, darovasertib might reduce the risk for metastatic disease. In an analysis of over 1,000 patients from our service in Philadelphia, those with monosomy 3 carried a 10-year rate of metastatic disease at 40%-60%. Gosh, it would be great if darovasertib could reduce that risk.
Now, for patients classified as low-risk tumors with Class I or no monosomy 3, darovasertib might further reduce the risk for metastatic disease. And currently, again, in that study of over 1,000 patients who had no monosomy 3, the 10-year rate for metastasis was 20% or less. And it would be wonderful if darovasertib could reduce that risk even more. Thank you.
Thanks, Dr. Shields. Dr. Scheffler, sort of in a related question, one of the things that we have not yet embarked on is treating those patients with smaller tumors. Those are tumors that are 3 or 4 millimeters in thickness, although we're getting prepared to do that in our phase II trial. Maybe you could make a few comments about what you think about the relevance darovasertib could have in these patients with the smaller uveal melanomas.
Well, Darrin, in reality, these are typically the patients who present to us with the best vision because their tumors are smaller and they've had fewer secondary effects on their vision from the tumor at baseline. And because, as we noted, the higher the radiation dose the patient receives, the worse their vision becomes. In these patients with small tumors, reducing the dose of radiation required to treat the tumor may truly result in excellent vision in reality in the long run. So in many ways, ironically, these are the patients who stand to benefit the most from a dose reduction in radiation and an improvement in long-term vision.
As Carol mentioned, because on average, patients with small tumors typically have poor genetics in only 10%-15% of cases, the majority of these patients do not develop metastatic disease, and so improving their vision and quality of life will have a very significant effect on their long-term life experience.
Thank you, Dr. Scheffler. As we close, what I'll do is I'll pose one final question for the two of you. You know, based on the data that you've seen thus far for darovasertib in the neoadjuvant setting, what are your thoughts of the degree of impact it will have if it achieves regulatory approval? Could this be practice-changing, for example? Dr. Shields, I'll begin with you.
Thanks, Darrin. Given the data we have with large melanoma, shrinking the tumor enough so that 60% of patients avoid enucleation, I think patients with small, medium, and large melanoma would likely benefit from darovasertib in the neoadjuvant setting. I honestly can say that this could be a game changer, a new way to more scientifically control eye cancer with upfront neoadjuvant darovasertib to shrink tumor and reduce metastatic risk, and then tumor consolidation with radiotherapy. I believe this would allow for less radiation dose for such patients and hopefully protect visual acuity, with the ultimate goal being to reduce metastatic potential. So yes, I believe that all patients with uveal melanoma would want a trial of neoadjuvant darovasertib to shrink tumor, provide less radiation, and achieve better vision outcomes, but most importantly, to hopefully reduce metastatic risk. Thank you.
Thanks, Dr. Shields. Dr. Scheffler?
I echo everything that Dr. Shields just said, and I would add from a practical standpoint, if you have a patient, once you know regulatory approval is achieved, if we have a patient who's sitting in our chair who has one of these tumors, currently, you know, there are a fair number of logistical hurdles we have to pass through in order to book a patient for surgery, plan their radiation plaque, go to the OR, and do it. This is a treatment that can be initiated much more quickly and much more easily, and I think would be not only efficacious in a neoadjuvant setting, but also hugely beneficial for patients' mental state in that it's something they can get started quickly that we know is highly effective at shrinking tumors and hopefully in reducing the risk of metastatic death.
Great. I'd like to take this opportunity to thank both Dr. Scheffler and Dr. Shields for joining us today and providing a little bit of background on uveal melanoma as well as the potential role of darovasertib in the neoadjuvant setting. Yujiro, I'll hand it back to you at this point.
Thank you for the great discussion on the neoadjuvant uveal melanoma panel, Darrin, Dr. Shields, and Dr. Scheffler. Next, I will hand it off to our CSO, Dr. Michael White, to introduce our next set of speakers.
Thank you, Yujiro. To launch this section, we're going to return to the theme of tumor heterogeneity that Yujiro introduced in his opening statements, and there's no better person to do that than Dr. Kornelia Polyak. Her research is focused on tumor evolution, and she's internationally renowned for seminal contributions to the tumor microenvironment and tumor heterogeneity fields that led to major advances in our understanding of breast tumor genesis. Her lab was one of the first to characterize cytophenotypic and genetic heterogeneity in tumors and developed innovative technologies to decode tumor ecosystems. We're delighted to have you join us today. Dr. Polyak, please take it away, and please take yourself off mute.
Yeah, thank you, Mike, and thanks for inviting me to give you an overview of intratumor heterogeneity. If you can go to the next slide. Diversity or intratumor heterogeneity is a hallmark of cancer. What this slide shows is that virtually anything we measure in a tumor, such as mutations, looking at spatial heterogeneity, chromosomal instability, every one of those are heterogeneous in the tumors. And this heterogeneity is what is the main reason why we haven't cured cancer, and it's so challenging to overcome therapeutic resistance. So due to the intratumor heterogeneity, there could be pre-existing clones of cancer cells that will drive resistance, and in the presence of treatment, these will grow out. On the other hand, cancer also continuously evolves, and there is acquired de novo resistance during treatment.
The next slide shows that one of the important measures is to be able to quantify resistance in a tumor, intratumor heterogeneity in the tumors. And for this, we and others have used the so-called Shannon Index of diversity, which is basically a diversity index ecologists use in ecosystems. And these little illustrations show us that what are the examples when you have an even, you know, heterogeneous or very skewed ecosystem. Basically, if you have a heterogeneity is high, then you have many different species, and each individual within a species represented about the equal numbers. Whereas if you have a very skewed or uneven distribution, then you have a particular clone or a species represented in a very high proportion, and the others are at lower proportions. So how do we apply this to cancer?
Cancer cells with unique identifiable features can be considered a species. We can also look at diversity of the tumor microenvironment and then the topological distribution of the cancer cells or the species within this ecosystem. Therefore, based on this calculation, we can come up with a number that quantifies how heterogeneous a particular tumor is. The next slide shows some of our early work. We started working on intratumor heterogeneity almost 20 years ago. Initially, we wanted to characterize the cancer cells that are more stem cell-like and more differentiated within a tumor. In breast cancer, those are marked with CD44. Those are the more stem cell-like cells. CD24 are the more differentiated luminal-like cells. On the left-hand side, the heat map shows the gene expression profile of those cells from normal tissue, normal breast, and also breast tumors. They're very distinct.
You see, like the green and red shows the differentially expressed genes. So they vary have different expression profiles. But in addition, we also looked at genetic alterations in the CD24 and CD44 positive cells. And in the middle, you show on a FISH looking at copy number gains on particular chromosomal regions like 8q24 and 1q21. And what we noted that they do have differential, you know, these CD24 and CD44 cells are, you know, transcriptionally different, but also genetically different. For example, a gain on 1q21 seems to be present in only one of the population. Then we extended this further, and we did this so-called ImmunoFISH on the right-hand side. You see an example where we did an immunofluorescence and FISH combined on a slide. And in this particular case, we just highlight a very nice example of a subclonal evolution from a ductal carcinoma in situ.
You see that in a CD44 population, there is a subclone that has a gain of 11q13. Then the adjacent invasive cancer has this clone dominant in all over, in all the both in the CD24 and 44 population. That just shows how much heterogeneity you have in a tumor and how much it changes during progression. On the next slide, we extended our studies to looking at associations between intratumor heterogeneity and therapeutic resistance or response. In this study, we looked at cellular genetic heterogeneity as measured by looking at chromosomal number gains like 8q24, and then looked in a neoadjuvant study looking at before and after treatment. On the left-hand side, you see these are tumors that basically every single little circle there is a tumor cell.
For every cell, we know the copy number gain on the top panel. The intensity of the color reflects how many copies of a particular 8q24 chromosome they have. And then, as you see, we had different regions of the same tumor collected at the same time. And then we had the pre and post-treatment samples. So as you see, based on this distribution of these different colors, even at the same time, different regions of the tumor are very highly heterogeneous. And then the before and after treatment samples, you see like very striking differences. In one tumor, there was actually a loss of those cells with a copy number gain in a triple negative tumor. In a HER2 positive tumor, it was the opposite. Actually, they did not have that gain cells before treatment, and then after treatment, they gained very high frequency of those cells.
On the lower panel, we also see that not just the genetic alterations are different, but also the phenotype for these more stem cell-looking and differentiated cells also very heterogeneous within a different region of the tumor and also before and after treatment. On the right-hand side is a quantification of this Shannon Index that I described to you that basically quantifies heterogeneity in a tumor. And then if we classify patients into who had complete pathologic response or partial response or stable disease, the patients who had the complete pathologic response had the lowest pretreatment diversity, showing that how much the diversity contributes to therapeutic resistance. On the next slide, I will show you another technique we developed that allows us to look at not just copy number alterations within the tumors in situ, in intact tissue, but also point mutations.
We call this technique the STAR-FISH, and it's basically a two-step PCR where we design primers for a particular mutation like PI3 kinase in this case and HER2 amplification. So we basically, you can look at HER2 amplification, PI3 kinase mutation in the same cell in intact tissue samples. And I don't have time to describe all the details of this study, but the next slide shows the important conclusion, which basically we looked at changes in diversity before and after treatment. This was again a neoadjuvant study. And then we looked at long-term outcome of the patient depending on whether they change the diversity or not. And on the top panel, this green line shows that if we're just considering diversity without considering topology in the tumor, then patients who have an increase or decrease, they both did equally fairly well, and there was no difference.
But on the lower panel, as you see in this survival curve, we see a significant separation from the people who had patients who had changed during treatment or did not have change during treatment. Patients who did not have significant change during treatment in the topological distribution of the cancer cells did exceptionally well, like 100% of them don't have any recurrence, whereas the ones that had changes during treatment in this diversity, those do poorly. So this just shows that it's not just diversity, but also the location within a tumor. The topology is also important to consider when we're determining outcome. The next slide shows that we can actually create these nice topology maps of the cancer cell populations within tumors. So the colors here reflect the five different genotypes depending on HER2 amplification, PI3 kinase mutation, single event, or double event.
And then again, the density of the circles reflects how close those cells are to each other. And just, it's a nice picture to highlight the diversity because you see that in this particular tumor, pretreatment, three different areas of the tumor are very heterogeneous. And then you see that post-treatment, also high heterogeneity and also significant change compared to the pretreatment cases. And we can actually calculate on the right-hand side. We calculated clustering score, basically like looking at if the genetically distinct populations are they closer to each other or more dispersed within a tumor. And what we see that the PI3 kinase mutant cells, they tend to be more dispersed within a tumor.
It looks like, you know, like the HER2 amplified cells tend to be tightly clustering, the PI3 kinase more dispersed, suggesting that PI3 kinase mutant cells actually may be more migratory and invasive, which was confirmed in preclinical models. Next slide shows another study we did, and this was actually a neoadjuvant clinical trial designed particularly to address whether intratumor heterogeneity for HER2 would determine response to HER2 targeted therapy. In this study run by my colleague Ian Krop, they did a neoadjuvant treatment with pertuzumab and T-DM1. Both of those are only HER2 targeted therapies. And then we looked at intratumor heterogeneity for HER2 and whether the patients had PCR or not.
And on the right-hand side, you see that it's very striking, like none of the HER2 heterogeneous patients had PCR, meaning like if you're heterogeneous for HER2, then you're not responding to HER2 targeted therapy. The next slide shows we did a follow-up on this study where we actually did a detailed transcriptomic and also spatial profiling of these tumors. And I just highlight one really striking finding from this study, which is this PCA plot of looking at transcriptomic differences between pre and post-treatment samples in patients who were HER2 heterogeneous or non-heterogeneous. And then on the bar graph quantifies the pre and post-treatment difference of transcriptomic profiles of those tumors.
And basically what we saw that the HER2 heterogeneous tumors, they don't seem to respond to the treatment because the pre and post-treatment biopsy were essentially, you know, very similar transcriptionally, which is the reason why it's interesting because in these tumors, not all the cells are HER2 negative. It's only like a subset of them seems to be lacking the HER2 amplification. But despite that, it looks like the whole tumor is not responding to a HER2 targeted antibody therapy. And in more detailed studies in a paper that we described, what we found that these tumors seem to have another way of turning on HER2 signaling pathway downstream of HER2. So they don't care by, you know, treating with an antibody, not even if it's an antibody drug conjugate.
The next slide shows some of our preclinical studies that we use to basically characterize the behavior of heterogeneous tumors. In this particular study, we devised a kind of a subclonal heterogeneous model of breast cancer. On the left panel, we looked at, we had about 18 clones. Basically, we looked at the growth of the tumor when we inject each clone individually or we inject a combination of the tumors that indicated a polyclonal tumor. What you see there, that the purple graph, the purple line, which is the polyclonal tumor, grew the fastest. The yellow and the red line shows some of the clones like IL-11 and CCL5 that was able to drive the tumor individually in monoclonal tumor, but even those were not growing as fast as the polyclonal tumor.
So that shows that you have a polyclonal tumor. It has a phenotype that is not seen by any of the individual clones. So that really suggests that there is some kind of gain of, you know, special function that this heterogeneity produces. On the right-hand side, we also looked at the clonality changes during the tumor growth. And then what was kind of two interesting conclusions, one of them that we found that actually some clones, for example, IL-11 is always remains a minor clone, like never becomes more than like 10% of the total population, but it's still an important driver of the tumor because if we block it, eliminate it, then the tumor doesn't grow. So that just shows that not always the dominant clone is that's really driving the tumor. You can have subclones that actually could be more important.
And then we also looked at here quantify the clonal frequency within tumors when they are monoclonal or polyclonal. And then in a polyclonal tumor, that's the purple bars, what we found that none of the clone can become really dominant in a way that takes over the whole tumor, suggesting that there is some interference among clones within tumors. The next slide shows kind of the explaining like what are the drivers and what is the contribution of the different cell autonomous and non-cell autonomous drivers to intratumor heterogeneity. So a cell autonomous driver could be like a point mutation in a tumor that is conferring the phenotype, you know, the tumor cells grow faster. But in that case, only that cell that has the mutation is growing faster. And then you can have multiple mutations, such mutations in a tumor.
For example, IDEAYA has this, you know, combination treatment when you try to inhibit two different genes within a tumor like PARP and PARG inhibitor. That could be one way to block these cell-autonomous drivers that are within the tumor cells. Then we also have non-cell-autonomous drivers that basically makes not only the cancer cell that has a particular change, but also the surrounding cells proliferate faster. And in that case, you can have some kind of cooperation among these clones. And this can be blocked again if you're targeting both of these mechanisms. And a good example for this is darovasertib and crizotinib treatment in metastatic uveal melanoma when you're basically blocking different signaling pathways in a tumor, including some of the growth factor signaling pathway.
Then lastly, another non-cell-autonomous driver, for example, IL-11 in a particular study is when you have a subclone that is angiogenic or immunosuppressive and in combination with the other clones that have other genetic or epigenetic alterations, they drive the tumor. These tumors tend to be the most heterogeneous ones because there is not really a selection for every single subclone to have immunosuppressive or angiogenic property because the whole tumor benefits from a particular subclone having those features. Then a good example, these tumors are the hardest to treat and a good example for a combination therapy that could be effective in these tumors is this Werner inhibitor and dostarlimab in MSI-High tumors. Basically, this just shows that for heterogeneous tumors, you really need rational combination therapies that target the tumor driving events.
The next slide shows that we also did mathematical modeling on our data, basically determining kind of predicting in a mathematical way, like what would happen. Would the heterogeneity be maintained if there is cooperation among the clones or not? And as you see on the bar, if there would not be interaction among the subclones in a tumor, the darker color indicates that eventually you would actually decrease heterogeneity because a dominant clone would take over the tumor. But on the top one, the lighter color shows that if there are interactions within subclones, direct or indirect interactions to the microenvironment, that's what's maintaining intratumor heterogeneity. So basically, the subclonal interactions are important to drive the tumor heterogeneity and maintain it over time. The next slide shows another study we did where we looked at not just the primary tumor, but also metastatic lesions.
This is just summarizing the study. Basically, what we found that polyclonal tumors not only promote primary tumor growth, but they also promote metastatic, distant metastatic lesion growth. Surprisingly, even these metastatic lesions could be polyclonal. Even clones that on their own would not form metastatic lesions were able to contribute to these metastatic lesions. What we found that that's because the clonal cooperation within a primary tumor changes both the local environment in a tumor, but also have systemic effect. For example, in this case, those cytokines act on mesenchymal cells in the lung and then recruit neutrophils that form the so-called metastatic, pre-metastatic niche. The cancer cell can basically lodge into those niches and able to grow. That's how you get metastatic lesions.
That's how even clones that would not be metastatic on their own just basically can flourish in these metastatic, pre-metastatic niche. The next slide shows, like I told you, like we see this very high degree of intratumor heterogeneity, and that's bad for therapeutic responses associated with resistance. Then basically, what can we do about it? I already gave you examples that you can have combination therapies. But another example I'd like to give is that basically we know that evolution is driven by phenotypic diversity. Phenotype is actually what we can modulate and we can target. A good example again from IDEAYA is this combination therapy by a MAT2A and the PRMT5 inhibitor in a MTAP negative tumors.
The next slide shows what we found, and this is a complicated slide, but basically what we found that KDM5, which is a histone demethylase, seems to be regulating cell-to-cell distances. In this particular study, we looked at breast cancer cells, MCF7, you see, and we treated them with a KDM5 inhibitor, C70. In the middle, you just compare the blank MCF7 parental 70, and basically we see a significant decrease in cell-to-cell transcriptomic heterogeneity. This was based on single-cell RNA-seq of the tumor. This is not only just observed in a preclinical model in a cell line, but on the right-hand side shows that in an ER-positive breast cancer, the KDM5 expression correlates with the transcriptomic heterogeneity in the tumor, suggesting that the higher KDM5 you have, the higher the transcriptomic heterogeneity.
Next slide shows that this is true not just for KDM5, but for multiple histone-modifying, and I think in general epigenetic enzymes. One of their main functions could be to regulate cellular epigenetic and transcriptomic heterogeneity. And targeting these enzymes in combination with targeted therapies could be more effective because we can decrease heterogeneity and enhance responses to the particular targeted agents. And on the right-hand side again shows that KDMs like histone demethylases or KAT, which is a histone acetyltransferases, targeting these could be beneficial because we're basically decreasing the heterogeneity of the cell population. And IDEAYA has this KAT6/7 inhibitor that could be a good combination with endocrine therapy potentially to improve efficacy of the treatment. So the last slide is just summarizing what I was showing you, like high intratumor heterogeneity in general is a poor prognostic marker.
I gave you examples of different types of treatment, HER2 targeted therapy, chemotherapy. Heterogeneity is bad. Heterogeneity is especially bad for your target that you're trying to inhibit. I also showed you in preclinical models that polyclonal tumors have unique phenotypes that are not seen by any of the individual clones. So that's why it's important to have, I mean, preclinical models that reproduce this heterogeneity of the tumors, and then I also showed you that we can modulate phenotypic heterogeneity by epigenetic regulators, for example, and those could be very good combination agents. They may not always have single-agent efficacy, which has been one of the frustrations we had with epigenetic therapies that in general, drug companies always want single-agent efficacy, and you may not always have it for epigenetic drugs, but they're very good combination agents because they would decrease the risk of resistance and prolong therapeutic responses.
And then I also showed you that you could have a high degree of heterogeneity, both pre-existing, and then these could be driving pre-existing and also acquired resistance to therapeutic agents. And that's why we think that we need rationally designed combination therapies applied at the right time because I also showed you that the heterogeneity can be changing along the treatment. And that's why we need to monitor it and design the therapies at that particular time for that particular tumor. So thank you for your attention.
Thank you very much, Dr. Polyak, for that fantastic overview. As you made very clear, delivering durable therapies to the clinic is going to require strategies to address intratumor heterogeneity. And as you noted, this is very much an IDEAYA priority with respect to both target selection and the underlying mechanism of action that leads to rational combinations.
Continuing on that theme, I'm really delighted to introduce our next speaker, Dr. Timothy Yap. He is one of the world's leading physician scientists focused on targeting the DNA damage response with novel therapeutics. Dr. Yap is, of course, a key partner for IDEAYA's IDE161 PARG inhibitor program. He's going to take you through the mechanistic rationale underpinning our IDE161 Keytruda and IDE161 ADC combination opportunities. I know, Dr. Yap, you're suffering from laryngitis. Really appreciate you taking one for the team for us this morning. Please take it away.
Thanks very much, Mike. Good morning, everyone. Thank you all for being here. As you've already heard from Yujiro, you know, IDE161 is a potential first-in-class small molecule targeting PARG, poly-ADP-ribose glycohydrolase that removes PAR chains from PAR-related proteins.
Now, this activity releases DNA repair protein complexes from chromatin to resolve DNA repair events initiated by PARP. And it's particularly important to maintain DNA replication fork stability in cancer cells with high levels of replication stress. The latter is a potentially widespread phenomenon whereby dysregulation of replication origin firing downstream of oncogene activation leads to replication fork collisions with each other with transcriptional machinery, with chromatin abnormalities, and also with DNA repair intermediates. Now, interfering with PAR-dependent resolution of replication stress results in replication catastrophe and can deliver single-agent anti-tumor activity, as shown in the left panel of this slide. These mechanistic relationships have been extensively evaluated and validated with IDE161, as shown in the right panel of the slide.
For example, in tumor cells with molecular fingerprints of replication stress, here referred to as biomarker positive, IDE161 exposure causes significant DNA replication fork perturbation, including failures in replication fork progression, accumulation of transcription and replication conflict, and ultimately inactivation of machinery that's required for stall replication conflict and restart. This leads to replication fork collapse, pan-nuclear DNA damage, as you can see in the gamma H2AX staining on the bottom right of this slide, and cancer cell death. Importantly, you know, findings like these are really guiding our patient selection strategies and the IDE161 monotherapy trial. These are also delivering mechanistic insights that have led to some really critical and important combination therapy opportunities that I will take you through today. Next slide, please.
On this slide, as shown on the left, genome-wide evaluation of treatment-dependent gene expression changes in tumors in vivo revealed the type I interferon response typically associated with innate immune pathway activation as top gene expression program activated by IDE161 in multiple different tumor types. Moreover, pathway analysis by IPA indicates an activation of genes associated with antigen presentation, suggesting bona fide interferon pathway activation by IDE161. Direct testing of this hypothesis in cell culture models, as shown on the right-hand side of this slide, demonstrated IDE161 dependent induction of interferon beta production and secretion, activation of interferon response transcription factors, and also an abundance of innate immune response protein production. Next slide. Now, as shown in the schematic on this slide, a likely explanation for this result is the strong connection between replication stress and nucleic acid sensing pathway activation.
Collapsed replication forks and associated R loops from failed resolution of transcription and replication conflict result in cytosolic exposure to short double-stranded DNA and RNA-DNA heteroduplexes. These, in turn, then activate nucleic acid sensing pathways through sensors that you're familiar with, such as cGAS and STING. This is particularly important because sufficiently strong innate immune pathway activation in tumors can promote adaptive anti-tumor immunity. Consistent with this mechanism, we observed robust combination benefit in a syngeneic tumor model where IDE161 was dosed together with an anti-PD-1 antibody, as shown on the right-hand side of the slide. These preclinical observations lead us to evaluate potential drug effects on a T cell receptor repertoire from our IDE161 phase I trial. Now, changes in T cell clonal expansion, as detected by peripheral T cell receptor sequencing, can be an indication of adaptive immune reactions.
As shown on the bottom right, T cell receptor sequencing of PBMC specimens from our IDE161 monotherapy patients revealed extensive expansion of T cell receptor clones on treatment versus baseline. This was observed in the majority of patients evaluated after 22 days of therapy and suggests that IDE161 treatment promotes adaptive immune responses in these patients. Next slide. Now, given the potential of IDE161 to indirectly promote anti-tumor immunity as a consequence of its anti-tumor activity, we have established a clinical collaboration with Merck to evaluate IDE161 in combination with pembrolizumab in the ongoing IDE161-001 trial. This study is a multi-center dose escalation dose expansion study that is evaluating IDE161 as monotherapy in solid tumors with HR defects. Based on the preliminary efficacy signal seen during dose escalation, the solid tumor basket cohort in dose expansion will be enriched for endometrial cancer.
In addition, the combination of IDE161 plus pembrolizumab is also being evaluated in patients with relapsed or metastatic MSI-High or MSS endometrial cancer who have progressed on platinum chemotherapy and anti-PD-1 therapy. Both HR-deficient and HR-proficient cancers are being enrolled in the combination cohort, and we dosed our first patient with this combination this month. Next slide. Next, I'd like to move to the second combination you just introduced. A unique mechanistic relationship between PARG and topoisomerase inhibitors that offers a potentially spectacular opportunity to broadly enhance the efficacy of topo-1 ADCs in solid tumors. Taking a step back, the starting premise really arose from work in Yves Pommier's group at the NCI, as shown in the left-hand side of this graph.
And as many of you know, when topoisomerase I inhibitors like camptothecin bind the enzyme, they essentially create covalent crosslinks between topoisomerase and single-strand DNA breaks. If these protein DNA conjugates are not removed prior to arrival of a DNA replication fork, they will essentially cause replication fork collapse. Removal is initiated by PARP1, which PARylates the topo I cleavage complex to recruit DNA repair machinery. The trapped topoisomerase is proteolytically degraded by the proteasome to expose damaged DNA, but this can only happen after topoisomerase is de-PARylated by PARG. PARG inhibition protects the trapped topoisomerase from proteasomal degradation, resulting in an accumulation of unresolvable DNA lesions that will ultimately block replication fork progression and kill the cell. As seen on the right-hand side, consistent with this model, IDE161 significantly enhanced accumulation of camptothecin-induced topoisomerase I cleavage complexes and the resulting pan-nuclear DNA damage.
As expected, this was accompanied by significant elevation of total protein PARylation and replication fork stalling in a combination setting as compared to either single agent. Next slide. As shown in this slide, this molecular mechanism can translate to robust anti-tumor activity. On the left is a small cell lung cancer model that is poorly responsive to IDE161 given alone and has a limited response to topotecan, a clinical topoisomerase I inhibitor that many of you will be familiar with. However, the combination of both of these drugs delivers complete responses with no additional impact on body weight beyond that observed with topotecan alone, and as you can imagine, these observations open up a potentially game-changing opportunity for ADCs with topoisomerase inhibitor payloads because IDE161 exposure would maximize therapeutic benefit from otherwise suboptimal payload delivery. We tested this hypothesis with an ADC, as shown on the right.
As you can see in three different models that are poorly controlled with either IDE161 or niraparib alone, the combination delivered robust anti-tumor control that was very well tolerated. We view this as strong supportive evidence for the potential of IDE161 as a backbone combination partner for topoisomerase inhibitor ADCs, and we are aggressively pursuing this opportunity through both collaborations and a wh olly owned combination you will now hear more about in a moment.
Thank you so much, Dr. Yap, for that outstanding overview. Very exciting mechanistic relationships, strong scientific rationale for these priority IDE161 combination strategies. It all really adds up to important new opportunities, as you noted, to address unmet clinical need. I now have the pleasure of introducing Dr. Ramon Kemp for an update on our newest clinical program, the IDE275 GSK959 Werner helicase inhibitor. Dr. Kemp leads GSK's early oncology development team and is a champion of both the Pol Theta Helicase program and the Werner helicase program. Dr. Kemp, we really appreciate your participation here today. Please take it away.
Thank you, Mike. I'm delighted to share with you an update on our Werner helicase program we're conducting in combination with IDEAYA. I'll also provide just a brief update, as Mike indicated, on our polymerase theta program as well, so tumor cells with mismatched repair deficiency are classified as MSI-H igh, and they have a series of dinucleotide repeats of thymine and adenine residues. When these TA repeats become expansive, they can form secondary DNA structures referred to here as a cruciform formation, and uniquely in these tumor cells, Werner helicase would naturally resolve these secondary structures to allow the cell to undergo normal DNA replication.
However, with Werner inhibition, these secondary structures persist and ultimately lead to double-strand DNA breaks, cell death as well. In the lower panel, when we considered the development of an inhibitor molecule, we demonstrated that the helicase domain and not the exonuclease domain was critical for activity. And therefore, we designed our molecule around the helicase. These data, again, are depicted here on the bottom left portion of the panel. On the right panel shows some of the early validation experiments that supported the activity of a Werner helicase inhibitor only in an MSI-High setting and not in an MSS setting. And as you can see in the inserts below, Werner inhibition in this case leads to the characteristic chromosome shattering. So this is very important in terms of how we develop our clinical program and patient selections. Next slide, please.
These data from preclinical studies demonstrate a very compelling set of evidence around the activity of our clinical molecule IDE275 or GSK959 across a variety of different settings. In both cell-derived, or the CDX, and patient-derived, or PDX models, you can see that we're demonstrating not only tumor growth inhibition, but regressions, and in some cases, complete responses in endometrial cancer, in gastric cancer, and in colorectal cancer cell models as well. On the right panel shows very strong responses to Werner helicase inhibition from a patient who received platinum-based regimens in first and second line, as well as progression on a PD-1 inhibitor. And remarkably, even after progression on these colorectal cancer standard of care therapies, treatment with GSK959 still significantly inhibited tumor cell growth.
These data provide a very strong and compelling support for us to design a broad clinical development program across several different tumors and in different patient segments. Next slide, please. So in summary, IDE275 and GSK959 shows very robust activity in multiple in vivo models across key tumors in which we see higher prevalences of MSI-H igh, and notably, again, in endometrial, gastric, and colorectal cancers. The consistency of these data suggested in other MSI-H igh tumors will likely see very similar responses to Werner. And we also have a very clear biomarker strategy to identify these patients. And the last slide, please. So we're very excited about moving GSK959 into phase I to see the benefits that we believe it will bring to patients. Our IND has been filed and is active as of October of this year. We anticipate that we will initiate patient dosing imminently.
Initially, we'll begin a monotherapy dose escalation and expansion cohorts. We then plan to combine our Werner helicase inhibitor with our PD-1 inhibitor, dostarlimab, which has demonstrated phenomenal activity in endometrial cancer patients through our phase III RUBY trial, as well as in GI cancers. So stay tuned for more information in the near future about our phase I program. So just a brief pivot again, as I mentioned earlier, to other clinical assets that we're in collaboration with IDEAYA, which is our polymerase theta program. This program continues to progress very well in our phase I study. We're in monotherapy dose escalation, and we anticipate our combination program with our PARP inhibitor, niraparib, will begin in the early part of 2025. Thank you.
Thank you so much, Dr. Kemp, for that exciting update. These programs are exceptional collaborations between GSK and IDEAYA, and we're really looking forward to this new molecule delivering meaningful benefit to patients in the very near future. I'm delighted right now to step in in this next section to represent our discovery team to introduce IDEAYA's 2024 clinical development candidates. We're going to start with IDE034, our B7H3/PTK7 bispecific ADC.
Now, Dr. Yap just walked us through the mechanistic rationale for combining IDE161 with ADCs, namely the unique synergy of PARG inhibition with tumor-specific delivery of a topoisomerase I inhibitor. This is potentially a huge opportunity, and we believe bispecific ADCs are one of the most impactful strategies to leverage that opportunity because an appropriately designed and engineered bispecific can significantly enhance selective delivery of a toxic payload to tumors versus normal tissue. This concept is illustrated on the left.
With a bispecific, antibody-antigen avidity can be tuned to require cooperative bindings such that both antigens must be present on the same cell for effective payload delivery. This can dramatically enhance tolerability by sparing normal tissues, but can also come with an efficacy penalty due to suboptimal payload delivery as a consequence of the cellular heterogeneity of individual antigen expression in double-positive tumors. IDE161 could really shine in a scenario like that, given our preclinical observations that PARG inhibition can amplify payload toxicity and promote a therapeutic response to what would otherwise be subtherapeutic inhibition of topoisomerase I. IDE034 was carefully selected for properties that optimize tumor-specific payload delivery in meaningful patient populations. As shown on the right, this B7H3 PTK7 bispecific antibody displays enhanced double-positive tumor cell binding and internalization versus B7H3 or PTK7 monoclonals.
Proprietary linker payload is very stable in circulation and designed for site-specific cleavage by a protease that releases the payload only following tumor cell internalization. IDE034 has a drug-antibody ratio of eight, which fully saturates the available cysteine conjugation sites. The antibody is fully humanized, and the heavy chains are modified to promote heterodimer formation, all of which makes for efficient quality manufacturing. With respect to the antigens we're targeting, we estimate that B7H3 PTK7 double-positivity occurs with meaningful frequency in lung, colorectal, head and neck, and ovarian cancer. Moreover, PTK7 is enriched in subpopulations of tumor cells that are often referred to as tumor-initiating cells or cancer stem cells, which potentially allows an asset like IDE034 to maximize ablation of a major source of tumor heterogeneity and adaptive resistance to therapy. The preclinical activity of IDE034 looks great.
We see robust regressions in double-positive PDX models, stasis and antigen low, as you can see in model number three, and partial response with relapse in single-positive tumors, all of which speaks to the selectivity of the asset and the combination opportunity with IDE161. Finally, IND-enabling activities are well on track for a potential 2025 phase I start. Now let's turn to IDE892, our MTA-cooperative PRMT5 inhibitor. I'd like to start this section by reiterating a point made by Yujiro earlier. We see MTAP-deleted tumors as one of the most important precision medicine opportunities out there right now. They're common. They're hard to treat. Because of that, we have heavily invested in establishing both depth and breadth of mechanistic insights into this patient population in order to advance multiple new therapeutic opportunities.
As you all know, deletion of CDKN2A tumor suppressor locus is a driving force for loss of MTAP because CDKN2A and MTAP sit right next to each other in the 9p21.3 cytoband. But as you can see, if you look at those bell curves in the panel on the left, there are another 30 or so genes in this region that are commonly co-deleted with CDKN2A or MTAP or both. Importantly, these co-alterations perturb a number of biological systems that can potentially be leveraged for therapeutic benefit. This includes chromatin remodeling, mRNA quality control, mitotic fidelity, and lineage identity, to name a few. Capitalizing on these tumor-specific mechanistic associations is a core component of our strategy to address unmet need in MTAP in all tumors.
This includes our MAT2A inhibitor, our wholly-owned MTA cooperative PRMT5 inhibitor, and our undisclosed third program in this arena that is on track for potential development candidate nomination next year. On this next slide, I'll take you through some of IDE892's key properties. As you can see on the left, this molecule was designed to capitalize on the unique structural templating of the PRMT5 active site by MTA to deliver a robust MTA-dependent target binding, as well as selectivity against SAM-occupied PRMT5. IDE892 is a single-digit nanomolar MTA cooperative PRMT5 inhibitor with exceptional target residency time, over 100 hours, as measured by surface plasmon resonance within a purified system, and it displays cellular target occupancy that is good or better than current clinical compounds.
Moving to the top right, that strong target occupancy corresponds to strong pathway inhibition in MTAP in all cancer cells, as measured by total SDMA immunofluorescence or by direct quantitative assessment of symmetric dimethylation of the PRMT5 substrate, histone H4R3, and the SmB splicing factor. This, in turn, corresponds to highly selective toxicity in MTAP-deleted cells. You can see that on the top right using the HCT116 isogenic pair. You can see that on the bottom row from an evaluation of selective IDE892 cytotoxicity across a panel of over 800 highly molecularly characterized cancer cell lines. Here, if we ask which baseline gene expression signatures correlate with IDE892 sensitivity, the answer by a mile is loss of MTAP, CDKN2A, and CDKN2B. Of course, this is because of gene-level copy number loss, as shown in that middle panel.
Finally, if we ask which genetic perturbations across this cell panel most closely mimic the activity of IDE892, we find depletion of PRMT5 itself or its cofactor, WDR77, also known as MEP50, all of which indicates that we have a very pathway-selective compound. Moving to the next slide on the top left, as expected, this biochemical and cellular activity profile translates to exceptional combination benefit with IDE397 that is as good or better than we have seen with clinical MTA-cooperative PRMT5 inhibitors. Complete regressions are achieved in challenging MTAP-deleted preclinical models using exposures well below those required for monotherapy activity. Also, as expected, the combination is well tolerated and is pathway-sparing in normal tissues.
As shown on the bottom panel, the combination of IDE397 and IDE892 inhibits SDMA accumulation in the tumors, but does not inhibit circulating concentrations of SDMA in the blood beyond that which is tumor-derived. As we described in the Triple Meeting this year, we believe that an important aspect of combining IDE397 with an MTA-cooperative PRMT5 inhibitor is the resulting durability of the anti-tumor response due to blockade of acquired monotherapy resistance mechanisms. Consistent with this, shown on the right, we see remarkable reversal of acquired resistance to PRMT5 inhibition by addition of IDE397. So we have high conviction for the therapeutic mechanism. We have a great molecule in IDE892, and IND-enabling activities are well on track for a potential 2025 phase I start.
Okay, we'll round out this section with a walkthrough of our third 2024 development candidate, IDE251, a dual KAT6/KAT7 inhibitor targeting tumors with 8p11 amplification in breast and lung cancer, as well as a potentially broad opportunity to target lineage survival oncogenes. As Yujiro alluded to in his opening statement, a fundamental IDEAYA priority is continued innovation within our discovery pipeline that builds on the momentum we have established with our clinical assets. You heard us describe how we're doubling down on addressing the MTAP patient population. We are expanding that playbook to broadly identify drug targets that are synthetically lethal with common somatic structural variants that promote cancer, and we have built a computational pipeline to find and prioritize those opportunities. You also heard about a potentially game-changing neoadjuvant treatment paradigm for uveal melanoma.
We're taking a page from that playbook to prosecute targets that support lineage-specific tumor initiation and progression. These are targets where the indication is the patient selection biomarker, and the mechanism is robust against defeat by tumor heterogeneity. A big opportunity here is lineage addiction, a truncal tumor survival mechanism that relies on cell lineage-specific transcription factor activity. It's a concept that's been on the table for over two decades, as noted in this review by Levi Garraway and Bill Sellers, with blockbuster clinical precedents in breast cancer and prostate cancer. A practical challenge confronting this opportunity is that the majority of lineage survival factors are transcription factors and are hard to drug. But as Nelly noted in her talk, transcription factors can be attacked indirectly by inhibition of epigenetic enzymes that establish the chromatin architecture required for transcription factors to work.
In support of an approach like this, we find that distinct epigenetic enzymes are often amplified in tumors that depend on lineage-specific transcription factor activity. A particularly striking signal is the lysine acetyltransferase KAT6A, which is present on the 8p11 amplicon in breast and lung cancer, as shown in the middle panel. Clinical proof of concept for this target in the setting of lineage addiction has been established by the Pfizer KAT6A/B selective inhibitor in estrogen receptor-positive metastatic breast cancer. However, as I'll take you through in a moment, we believe the larger therapeutic potential of a KAT6 selective molecule is limited by paralog bypass mechanisms within the KAT family.
By inhibiting the main culprit, KAT7, with a dual KAT6/7 inhibitor, we believe we can deliver robust single-agent activity in luminal breast cancer, in 8p11 amplified lung cancer, and in additional meaningful indications dependent upon KAT6 and KAT7 to support lineage-specific transcription factor activity. As shown in the next slide on the left, there is a large body of work indicating KAT6 and KAT7 are mechanistically intertwined epigenetic modulators of cell identity and lineage commitment programs corrupted by oncogenic transformation. Lysine acetyltransferase works by acetylating histone tails to locally relax chromatin enough to allow transcription factor binding and subsequent gene expression. Precise control over where that happens in the genome is mediated by multi-subunit protein complexes that specify KAT enzyme localization. KAT6 and KAT7 share binding partners that guide that process, such as BRPF1, and they share lysine substrates, such as histone H3K23.
Furthermore, H3K14 acetylation by KAT7 can prime nucleosomes for acetylation of H3K23 by both KAT6 and KAT7. These overlapping activities support cancer stem cell identity and renewal, senescent versus apoptotic responses to stress, and lineage-specific transcription factor activity. A fair amount of functional genomic evaluation has further substantiated this KAT6/KAT7 partnership in tumor cells. For example, as shown in the top row of the middle panel, depletion of KAT6 versus KAT7 reveals highly correlated changes in the genomic expression of profiles of estrogen receptor-positive breast cancer cells, and KAT6 and KAT7 are within each other's CRISPR co-dependency networks in specific tumor lineages. Consistent with this, BRPF1, the scaffold protein shared by KAT6 and KAT7, is selectively required for survival of cancer cell lines with high baseline expression of lineage survival genes, including ESR1 in breast cancer and SOX2 or NKX2-1 in lung cancer.
Furthermore, co-deletion of BRPF1 and the KAT7 selective cofactor BRPF2 display a strong paralog relationship in subsets of tumor cell lines. Finally, these genetic relationships can be recapitulated with a KAT6 selective and KAT7 selective tool compounds that display a synergistic interaction in cell viability assays, as shown on the far right. These collective observations motivated our pursuit of a dual KAT6/7 inhibitor. This is a very challenging profile to achieve given substantial differences in their active sites, including size, shape, and confounding residue differences in otherwise conserved regions, as you can see in the structural overlay on the bottom right. On the next slide, as indicated on the left, we have achieved our desired profile with IDE251. This compound is a low nanomolar to subnanomolar inhibitor of KAT6 and KAT7 with at least 100-fold selectivity over other structurally similar KAT family members.
This corresponds to robust pathway inhibition, as indicated by loss of histone acetylation on the KAT6/7 shared site H3K23 and the KAT7 exclusive site H3K14. As anticipated, dual KAT6/7 inhibition leads to strong induction of apoptosis in estrogen receptor-positive breast cancer cells as compared to the clinical KAT6 selective inhibitor, which primarily promotes senescence. When we evaluated global molecular responses, as shown on the right, we found that IDE251 exposure resulted in significant chromatin compaction around transcription start sites, as indicated by ATAC-seq, with over 3,000 differentially accessible regions affected beyond what is achieved with the clinical KAT6 selective molecule. You can see a couple of nice examples of differential peaks in the promoter regions of WNT3A and CD44 there below the tornado plots.
Finally, on the far right, the global gene expression response to IDE251 exposure versus the KAT6 selective compound is consistent with coordinated gene regulation by KAT6 and KAT7, with greater effect sizes observed upon dual inhibition, including key luminal breast cancer lineage dependency genes like ESR1 and GATA3. As shown on this next slide, we see very concordant results in vivo. Superior inhibition of H3K23 and H3K14 acetylation in tumors with IDE251 versus the clinical KAT6 inhibitor, even when the latter is dosed up to exposures tenfold higher than those used in the clinic. This translates to robust single-agent anti-tumor activity with IDE251 in an 8p11 amplified estrogen receptor-positive breast cancer model and what appear to be complete responses in an estrogen receptor mutant PDX model, something that cannot be achieved in this setting with selective KAT6 inhibition at any dose.
As we saw in culture, global gene expression profiles of tumors from treated mice were consistent with enhanced chromatin modulation by dual KAT6/7 inhibition versus selective KAT6 inhibition. And as shown on the bottom right, this included gene expression programs associated with luminal cell lineage identity. With this compound in hand and having established preclinical proof of concept in the setting of hormone receptor-positive breast cancer, we're now focused on prioritizing additional tumor types and biomarker populations responsive to dual KAT6/7 inhibition. As you can see on the left, we're using a data-driven approach to do this through a combination of chemical genetics, functional genomics, and drug response profiling. I noted earlier our interest in 8p11 amplification beyond breast cancer, which was validated by these studies. And as you can see on the right, IDE251 delivers substantial anti-tumor activity in an 8p11 amplified lung cancer model.
In addition, as expected, we're detecting new lineage dependency opportunities, including potentially TCF7L2 and gastric and colorectal cancer and SOX17 in ovarian cancer. Finally, we're noting multiple opportunities to address CDK4/6 inhibitor resistance mechanisms, including PI3 kinase mutations and estrogen receptor mutations. So to sum up, we view IDE251 as a new and significant development opportunity with exciting preclinical proof of concept studies that will form the foundation of our initial clinical development strategy, as I've bulleted here. With that, I'll turn the floor over to Yujiro for closing remarks.
Great. Thank you, Mike, for that terrific walkthrough of our next generation development candidates representing our targeted sixth, seventh, and eighth potential clinical program. To drive forward our vision and strategy to build a leading precision medicine oncology company, a world-class R&D and drug discovery enterprise will continue to be one of the centerpieces that will drive our strategic value in both the medium and long term. We anticipate delivering eight clinical programs by year-end 2025, and we are not slowing down, with multiple preclinical programs also advancing with the goal of growing our clinical pipeline further in 2026 and beyond.
Next, over the next several quarters, watch for IDEA to continue to be a trailblazer in both the discovery and clinical development of additional novel monotherapy precision medicine oncology agents and multiple new potential first-in-class rational combinations, including through pharma partnerships. We believe we'll continue to position IDEA as an industry leader across several potential first-in-class targets and biomarker-defined solid tumor populations.
As we look forward to next year, we wanted to provide our preliminary 2025 guidance. We believe 2025 is set up to be one of our most transformational years for IDEAYA since our founding, with multiple potential catalysts across the clinical pipeline, including a potential first accelerated approval readout, a median overall survival update, and a large neoadjuvant uveal melanoma readout on our lead program, darovasertib. Next, for the IDE397 MTAP deletion program, we believe 2025 will be the year we will make important progress across several clinical combinations, including a targeted clinical efficacy and safety update on the Trodelvy combination and MTAP deletion urothelial cancer, where we are already observing multiple RECIST responses in the dose escalation. Next, with our partner GSK, we anticipate presenting our potential first and best-in-class Werner helicase inhibitor, IDE275, at a major medical conference in the first half of 2025.
What supports all of this is a strong balance sheet with a recent quarterly reporting of approximately $1.2 billion of cash on the balance sheet that we believe provides a runway to at least 2028 and the potential to further extend the cash runway through our GSK partnership for the Werner helicase and Pol Theta programs that are both advancing rapidly in the clinic and have aggregate potential milestone payments of approximately $2 billion. Lastly, we are on track to deliver three INDs in 2025, resulting in eight potential clinical programs next year, and what we believe will represent one of the most diverse and compelling precision medicine oncology pipelines in the biotechnology industry. We have now completed our prepared remarks, and operator, you may now open the line for the analyst Q&A portion of our webcast.
Thank you, Yujiro. Please hold for a brief moment while we pull up questions. Our first question comes from Anupam Rama at JPMorgan. Please go ahead, Anupam. Anupam, you might be on mute.
Oh, can you hear me, guys? Sorry about that.
Yeah.
Yes. Okay, cool. Hey, thanks so much for hosting this R&D day, guys. Super helpful. Just a quick one from me. So with IDE892, your internal PRMT5, strategically, should we be focused now moving forward on an all-internal program, particularly given sort of the preclinical combination data you gave with IDE397? And what are you looking to now learn from that Amgen combination and partnership with IDE397? And it's thinking about the individual components, right? Both IDE397 and IDE892. Thanks so much.
Yeah, no, thank you for the question, Anupam. Yeah, so I would say at a high level, our continued partnership with Amgen continues to be a significant priority for us. As you know, Anupam, we've been in the dose escalation for roughly a year, and at least our current guidance right now is targeted expansion in the tumor type of non-small cell lung cancer. You know, at the end for us, we believe IDE397 has the opportunity to be a backbone in the area of MTAP deletion.
As you know, there are several PRMT5 inhibitors that are out there, and we know there are several groups that have high interest enabling this combination. For IDEAYA, strategically, IDE892 provides the ability to wholly own what we believe has the opportunity to be a potential not only first-in-class combination between these two mechanisms, but also a potential best-in-class combination for these two mechanisms.
So our plan at this point is to push both opportunities forward. But as you mentioned, yes, IDE892, IDE397, with our second molecule anticipated to be in the clinic by next year. So we think we'll be well positioned. And as Mike noted earlier in his section, we would also highlight to folks that we have a third program, an MTAP deletion that's distinct from MAT2A and PRMT5, that we're planning to deliver our third development candidate in MTAP next year as well. And so hopefully people will see that, as was walked through by Dr. Polyak, we are taking a very extensive and broad approach to MTAP around specifically rational combinations. So Mike, anything else you'd add here?
Sorry, Yujiro, yeah, I was on mute there. Yeah , no, great interview. Sorry, great overview. And yeah, we're very excited to proceed exactly as you described.
Thanks so much for taking our question, guys.
Yep.
Thanks for the question, Anupam. Our next question comes from Maury Raycroft at Jefferies. Please go ahead, Maury.
Hi, good morning. Congrats on the progress, and thanks for doing this presentation and taking my question. I was going to ask one on a Werner helicase. So just wondering if the preclinical data you're showing is translatable. Can you contextualize how these MSI-H igh data could compare to standard of care treatment for MSI-H igh? Or what is the bar for success when you get into the clinic? And in the combo study, would you include patients who already have had anti-PD-1, or will these patients be naive to checkpoint inhibitor?
Yeah, let me start, and I'll pass it to Mike here. So I think high level here, Maury, at least our anticipation based on the profile we have observed preclinically is that we will have a monotherapy path forward here, as I think demonstrated by the CRs that we've seen as a single agent across multiple tumor types, as Dr. Kemp walked through. Second, as part of this presentation that's noted on this conclusion slide, which we'll be presenting at a major medical conference the first half of next year, we do believe we have a differentiated molecule from the existing phase I Werner inhibitors.
And we will profile what those differences are, why we also, to my first comment, why we continue to believe we should have hopefully a path forward in terms of monotherapy. Second, as Dr. Kemp noted, we do have a significant focus on implementing the combination with GSK's Dostarlimab. We do believe they have a very good PD-1 inhibitor. So when you look at it from that perspective, we think we're positioned extremely well. Mike, do you want to also maybe just go through some of the work that we've done with the Sanger as well just around PD-1 resistance and continued activity with a Werner helicase inhibitor and MSI-High or anything else you want to add there?
Yeah, absolutely. Yeah, thanks Ray, for that question. So one of the things that I think is also important to note is downstream of checkpoint inhibitor therapy, we think that that is a space where this synthetic lethal relationship remains. That's very important. So that was shown by Mathew Garnett's group from a genetic standpoint.
And as Ramon pointed out in his presentation, that has been recapitulated in preclinical models with the tumors, PDX-derived tumors from patients that had relapsed from those immune checkpoint inhibitors. This is a synthetic lethal relationship with a therapeutic window that you can drive a battleship through, but is also, we're expecting to see strong single-agent activity in those patients downstream of those other treatments. The other piece that YuJiro was referring to is we have a long experience in trying to get a molecule on top of this target in a very effective fashion. It's not an easy thing to do. And one of the reasons is the flexibility of this D1, D2 domain in the helicase. It can adopt multiple conformations.
And one of the things that Matthew's group has been showing quite nicely is that the alterations of amino acids within these domains can have a significant consequence on the ability of a molecule to bind to the target. And we're expecting to see that that is going to arise as an issue with these inhibitors and their ability to engage the helicase domain. But one of the things that we see with our molecule is a distinct binding pose that we believe is going to be very important with respect to getting on top of those acquired resistance mechanisms as they occur with point mutations with compounds that are out there in the clinic right now. So very excited to get this molecule into patients and watch what happens in the MSI-H igh setting.
Got it. Thanks for taking my questions.
Thanks for the questions, Maury. Our next question comes from Gregory Renza at RBC. Please go ahead.
Hey, hey, thanks, team. This is Supriya on for Greg. Congrats on the progress. I have one more question on 892. Could you perhaps double-click on its potential to be best in class? What aspects of its chemical design or perhaps the preclinical data that you may have seen that give you such confidence for it to be better than competitors? And as a follow-up, how do you envision its development path and strategy in light of existing partnership with Amgen? How do you plan to explore its combination with 397? Thank you so much.
Yeah, sure. Thank you for the question. So maybe I'll take a first shot at that in terms of target TPP. And obviously, Mike, please step in here. First, I would say there's obviously been some questions about brain penetrance, yes or no, against this target. Our team's perspective is that we believe brain penetrance is more of a liability than a benefit. And that's based on our interpretation of this target and also some of the clinical data that's out there on some several brain penetrance molecules.
That's the first. Second, as it relates to the balance, I would say, between potency and selectivity, we've done a lot of chemistry here looking at a lot of different chemical series. And our perspective, the objective is not to just keep driving the potency down. We believe at some point, you start breaking the selectivity barrier. And obviously, there's been a significant focus on MT cooperativity. But as we highlighted in Barcelona at the Triple Meeting, there are other types of cooperativity you have to look at, including SAM cooperativity as well as others. So we believe we've struck that fine balance. I think here our molecule probably looks closest to the BMS compound.
Finally, the next third is around DMPK properties. And we do think here IDE892 is a standout molecule. As we know, there's different clinical molecules out there that might have to dose at certain ranges to have efficacy. So we feel quite good about that. And then I would just say, finally, last on this one is IDE892 was handpicked to be combined with IDE397. And sometimes this is an empirical exercise. We've looked at a fairly large number of compounds, including the clinical compounds. And we think IDE892 is truly a standout. So hopefully that directly answers your question. And then, Mike, here, anything else you'd add?
Great, great summary, Yujiro. I think it's really important to recognize the importance of selectivity in the MTA bound form. That's really important for us with our combination of IDE397. So we're bringing a molecule here that is very, very precisely designed in order to be able to occupy the MTA bound form of PRMT5, not the free form, not the SAM-occupied form. That's key. And then, obviously, combinability with IDE397 has been evaluated preclinically extensively. So we see very good opportunity here for promoting this combination therapy.
Yeah. And then also just on your clinical development question. So here, what I would just highlight specifically, hopefully people appreciate our strategic decision to press forward IDE892 towards the clinic is a very clear decision that we are doubling down on this combination with MAT2A and PRMT5.
As Mike very nicely highlighted, we have continued conviction that this mechanistic combination can be absolutely potentially critical to deliver not only more responses, but more optimal durability, specifically in the setting of MTAP deletion lung cancer. Second, as it relates to the clinical development of this, here, we would anticipate being able to integrate the combination fairly early as part of the clinical development of this asset and start integrating the combination with IDE397. And obviously, I think it's pretty safe to say at this point, we are the world leader on this combination. We have a lot of insight on this combination in the clinic already. So Darrin, I know you're on. Anything else you'd add on that?
No, I mean, I think you've hit the nail on the head. The only thing I would add is that it does increase the flexibility. Of course, we're leading the way in non-small cell lung cancer with our Amgen partners. We're really excited about that. But there are additional indications to consider as well. So it just allows us the opportunity to, as you said, really go after the MTAP deficient space, not only with a PRMT5 combo, but with other combinations because we believe it's a very, very large opportunity. These patients have a huge amount of unmet need, generally resistant to other standard of care therapies. And so bringing forward novel combinations in as many indications as possible will be highly relevant for the oncology space.
Got it. Thanks so much and congrats again.
Thank you for the questions. Our next question comes from Charles Zhu at LifeSci Capital. Please go ahead, Charles.
Hello. Good morning, everyone. Thank you for hosting this event and for taking our questions. I'd like to ask two questions, please. First one, a quick, quick one. Any comments on why certain preclinical MSI-High models seem to not respond particularly well to the Werner helicase inhibitor from the GSK section? And for the second question regarding the KAT6/7 inhibitor that you've discussed today, eventually, for some of these indications like ER-positive breast cancer, especially, or even lung cancer, how will you evaluate combination partners, especially given the ongoing shifts in that landscape across CDK, PI3K, you name it? And is this something where perhaps similar to MTAP deletion, you could advance something forward with a fully in-house combination? How are you thinking about that? Thank you.
Yeah, thanks for the questions there, Charles. So first, let me maybe take the first one as it relates to several of the models. I mean, again, if you look at that plot that Ramon went through, again, these bars that are going down versus up, I mean, if you look at that general range in terms of reduction of tumor types, and you sort of see that dotted line of where we would anticipate sort of a partial response if that data mimics, we think that result is quite robust, right? Where you're almost seeing about half of the models.
So if we can replicate that, I mean, that would imply at least the potential based on that data set, roughly a 50% type response rate. Again, that assumes translation. So we think that's quite an incredible result for monotherapy and would position this program extremely well from a monotherapy development perspective. The second, as was noted in Dr. Kemp's slide, that's also the purpose of obviously looking at a PD-1 combination. We're already seeing CRs as a monotherapy, but obviously, going into combination, at least we would anticipate based on this concept of tumor heterogeneity, we should have the ability to hopefully flip some of those other models. So Mike, anything else you'd add here?
Yeah, great, great questions, Charles. So that first piece, one of the things that I think is important to note is we see incredibly strong concordance with the genetics. So preclinical models where you knock out Werner and it kills them, those are the same models where we put our compound on and it kills them. And they line up with each other. Those models where a Werner knockout doesn't kill, we don't kill either. So I think it's partially with respect to the definition of the MSI- high and what is really triggering that response to Werner and getting over that threshold.
So that's that piece. With respect to your second question, the KAT6/7 and what we're going to be doing in that space, I'm really excited about this program because a lot of what you heard about today with respect to tumor heterogeneity in an asset like this one, we think can really gut that, can really stop it at its source. There's a reason why I pointed out CD44 to you there with respect to promoter region that gets compacted when we take this off. So we think there's lots of combination opportunities here for us, both with respect to tumors that are driven by lineage-specific transcription factors. We hammer, obviously, the estrogen receptor response, even in the context of estrogen receptor mutants. So we're excited about that. You probably noted on the CRISPR data that the sensitivity alterations look like they are pointing at some pretty meaningful patient populations.
Some of these can be in combination with assets that we have. Some of these can be in combination with assets that are coming up soon. Thanks for those questions. I think you're spot on in how you're thinking about this.
Yeah. Maybe just, Charles, add to what Mike mentioned. I think we're hopefully people got a sense based on the data we shared today. We are a research leader in the KAT area and specifically around this KAT6/7 dual inhibition. We do anticipate, based on the growing preclinical data that we're generating, that this area will be a significant franchise for us moving forward, similar to Daro, similar to MTAP. We have very, very broad chemistry we've done here. We think we're very, very well positioned for both the medium and long term here.
Go t it. Great. Thank you very much again for taking the questions and congrats on the progress.
Thank you.
Thanks for the questions, Charles. Our next question comes from Yigal Nochomovitz at Citi. Please go ahead, Yigal.
Hi, thanks. Can you guys hear me?
Yep, we can hear you, Yigal.
Can we just go back to the endpoints for the phase III for the neoadjuvant? And I would love to hear if Dr. Shields, if she's still on, could just explain to everyone why she believes and why we should believe that the EFS endpoint has, well, not precedence, but it has a good chance to succeed. I guess there really isn't precedence per se. The question being, obviously, if you enucleate, then the chance of metastases is presumably zero. If you keep the eyes, which obviously you want to do, you may still have some low chance of metastases. So it's not completely clear that the EFS endpoint is going to hit, I guess, although one would think so. So I just would be really curious to hear her perspectives on that.
Yeah. So, Yigal, let me start and also tag team here with Darrin. So first, all our guest speakers are offline now. It's just management for the analyst Q&A portion. I think there's several pieces I would highlight that. And I think Dr. Shields noted several of the key publications. So first, remember here, we're in patients, at least we're proposing to study patients that are at the higher risk population. And so it's very possible that at that point, many of these patients have already had micrometastasis for some time. So I think that's the first. Second, based on several of the papers that Dr. Shields referenced today, including the key paper around tumor shrinkage and the risk of metastatic disease.
And as she noted, the specific stat around, with every millimeter, an ocular tumor shrinks that your risk of metastatic disease changes by approximately 5%. And so I think that's the hope here. If we can have consistent shrinkage of these tumors, which we are demonstrating in the clinic, that hopefully we would have a good trajectory, at least the potential for a good trajectory on EFS. And as a reminder, the endpoint that was discussed as a secondary here was on no detriment to EFS versus needing to show superiority. And then maybe for the next on the endpoints, Darrin, do you want to maybe just high-level go through that and sort of what we're thinking here? I know on the plaque side, we've started to integrate, for example, macular edema. And we will have a type B meeting here coming up. Darrin, do you want to summarize?
Yeah. I just wanted to clear up one fact that I heard that was not correct, which was if you have a large tumor and you're scheduled for enucleation, you get the eye removed, you have no chance of metastases or a low chance. That's not correct. And so just to be clear, the biggest predictor of metastases is what your profile is. So are you a Class 1 or you're a Class 2? Do you have Monosomy 3 or not? So those are the key factors. And you can have a small tumor that has Monosomy 3 and have very high risk of metastases. You can have a large tumor that happens, although uncommonly, to be Class 1 and can have a much lower risk of metastases.
So the risk of metastases is driven by factors that are not necessarily size. Although clearly, as Dr. Shields pointed out, the bigger tumor you have, the higher likelihood you are to metastasize. But the molecular profile is also extremely important as well. In terms of the endpoints for the study, remember, there's two different groups, right? There's the group with the very large tumors that are going to be enucleated. And the primary endpoint there is save the eye. And we've already shown, provided data that shows that more than 50% of their patients are having their eyes saved who are getting neoadjuvant therapy. So we think the likelihood of success there is good. And as Yujiro pointed out, the objective is to show that we can do that without causing any detriment in event-free survival.
In other words, by converting someone from enucleation to plaque brachytherapy, we're not increasing their risk of developing metastases. That's the objective. And then for the plaque brachytherapy group, so this is the folks with the smaller tumors. Those are the ones who get the radiation therapy. The idea is, Dr. Shields pointed out, would be to shrink the tumor so they can get less radiation. And then the primary endpoint there is vision preservation, which in our case can be measured in two ways. One, we can actually look at the vision using best-corrected visual acuity 15 letters. Or the other alternative is if the patient develops macular edema, which has been clearly linked, as Dr. Scheffler pointed out, to poor visual outcomes. So an event in this study would be either if the patient lost 15 letters in their vision or had developed macular edema.
We would expect in the neoadjuvant treatment group, the group where the tumor shrank, the group where the patients presumably had less radiation, they're going to have less vision loss and they're going to have less macular edema. Once again, the key secondary endpoint there is event-free survival. Again, what the FDA wanted to know is, "This is terrific. This is wonderful. The data is outstanding. You're shrinking tumors. You're saving eyes." Based on a prognostication tool, it looks like you're preserving vision likely due to this reduction in radiation. We want to make sure when patients get this therapy and their primary treatment is delayed, such as enucleation and plaque brachytherapy, we're not increasing the risk of metastases. The study will be designed to show that.
Great. Thank you, Darrin. Also, Operator, I know we just hit top of the hour. I think we're going a bit over here. So why don't we take two more analysts?
Okay. Sounds great. So our next question comes from Ben Burnett at Stifel. Please go ahead, Ben.
Hey, great. Thanks very much for the pipeline overview today. I also have another question about the neoadjuvant study. I think some of the KOLs and you talked about some of the historical relapse rates or event rates in high-risk patients. I guess just given this, just curious if you have any updated thinking about how long it might take to conduct the neoadjuvant study, either the enucleation cohort or the plaque brachytherapy cohort?
Well, I can just tell you this, that the patient population has been chosen so that the median time to metastases is approximately three years. So it's not very long. So this is a very high-risk patient population. As we've pointed out before, though, to the credit to the great people we work with at the FDA when we had this discussion, we're enrolling a high-risk patient population in order to enhance our ability to be able to detect whether there is a detriment in event-free survival, which we suspect, if anything, there may be a benefit as opposed to a detriment.
But they even pointed out that if you weren't creating any harm in a high-risk patient population, this would suggest that even a low-risk patient population would have no harm. And so that's why they were open-minded about considering a broad label. The fact of the matter is these patients are going to relapse relatively quickly. The endpoints were chosen in order to be able to detect the difference between the treatment and control arm. I think it puts us in a strong position to be able to do that in a relatively short period of time.
And also, Ben, I would just highlight, and Darrin mentioned the three years, just as a reminder, because the EFS endpoint is no detriment. We don't believe the full timeframe to get, I think, Darrin, we've been saying internally, likely 30%-40% of the event. So that's what we feel comfortable in that two-year timeframe from that first patient enroll. We should be in a good position on the EFS piece as well.
Okay. That's super helpful. And just to clarify, are those estimates kind of around both cohorts, or should we expect that the cohorts could evolve kind of on different timelines?
Well, I would say it's going to be very similar. Again, I don't want to get into the details because we're still having a discussion with the FDA, the actual final protocol design. But suffice it to say, I mean, it's clear that saving the eyes could happen much quicker, but we're trying to be efficient so that we can line up both cohorts in a timely manner. And if done appropriately, we may be able to read them both out at the same time. But it's easy to predict that the enucleation cohort data would come faster because you could learn about saving the eye relatively quickly. But we're trying to set it up in a way so that right from the beginning, we can get a broad label to cover both patient populations.
Wonderful. Thank you.
Great. Last analyst here, Tara.
Yes. So our last question comes from David Dai at UBS. Please go ahead, David.
Hey, great. Thanks for squeezing me in. And I want to add my congratulations on the amazing progress here. Just a quick question on the IDE161, your PARG inhibitor, as the monotherapy in combination with Pembrol. As possible, we think about the safety of index for IDE161. And then if you're planning to add the B7H3 Topo ADC, how should we think about balancing the efficacy and safety, especially given some of the safety issues we've seen from other B7H3 ADC from a competitor?
Yeah. So, David, let me take a shot at that. And then obviously, Mike and Darrin chime in. So first, David, we have been doing dose optimization with IDE161 PARG. We have an initial expansion dose that we moved forward with, which we recently announced in a very specific priority solid tumor type. We're not disclosing that for competitive reasons, but that's been done.
I think we have a fairly good handle on that dose optimization where we see optimal PD and still see an acceptable AE profile. And just so you know, we've even dose escalated now beyond that expansion dose. Second, as part of the combination with the Topo ADC side in particular around our bispecific ADC, as Mike went through, we are seeing greater selectivity of the bispecific versus the mono B7H3. And that's one of the reasons why we pursued a bispecific because, at least theoretically, the opportunity to have a wider therapeutic window versus a mono B7H3. But as you can appreciate, David, there are several pieces involved there. It's also about stability of the linker, obviously, specifically the payload as well. So I think that will be ultimately something we'll test in the clinic.
The last piece I'll just mention as it relates to Topo ADC and the combination with PARG IDE161 in specific as it relates to safety. Ultimately, that's going to be a whole part of dose optimization in the clinic, right, is to find a combination dose that hopefully we can move forward with where we see higher efficacy than the ADC a lone. So, Darrin, do you want to add any comments there or Mike?
No, I think you hit it on the head. I mean, obviously, the therapeutic windows, that's the whole point of the bispecific, right? So I think that'll open up additional opportunities for us to test in a number of indications. So we're just excited to get it into the clinic as quickly as possible. And I think the potential is great. Let's face it, the ADC area is a very hot area right now. People say, "Geez, what's next with ADCs?" It's bispecific ADCs. That's obvious. That's going to be the next big wave, I think. We're going to be right in the middle of it with a way to actually deliver a more potent punch, which would be exciting.
The only thing I would add to that is with respect to the Keytruda combination, we're really excited about what we're seeing in the patients, right? That peripheral event that is indicating that this mechanism is likely really in play with respect to provoking an anti-tumor immune response. So getting that combo in place is very, very exciting, in my opinion. Then back to the ADCs. There's more than one kind of ADC out there, as you guys know. There are some that are very, very specific, delivering the payload to the tumor. There are others that kind of release the payload on the way to the tumor, kind of like a sustained release mechanism. We want the former, not the latter. And that is really guiding the choices with respect to what we pursue in the combination setting.
Yeah. And maybe, David, just as Mike mentioned, the other piece just to note around our bispecific ADC is we were evaluating both DAR4 and DAR8. And based on what we saw, and maybe perhaps not surprising, hopefully added specificity of the bispecific, we are able to proceed with a DAR8. So we're obviously thrilled about that. And we'll be looking forward to that, hopefully getting into the clinic next year. Got it.
Thank you so much.
Yep.
So, Yujiro, that concludes the Q&A session. I'll turn it to you for closing remarks.
Great. Thank you so much, everyone, for participating at this year's 2024 Investor R&D Day. Wish everyone happy holidays. Thank you for the time today.