All of our attendees to IDEAYA's 10-year anniversary R&D Day, today marks an important milestone in the company's history, and we want to thank each of you for sharing this special day with us. I hope everyone will walk away feeling inspired by the extraordinary R&D ongoing at IDEAYA and the strategic vision that will drive forward our growth as a global leader in precision medicine oncology in the coming decade. Please note, we'll be making forward-looking statements today, and please refer to our SEC filings as appropriate. To kick off our 10-year anniversary R&D Day, we wanted to first share a short video to help bring the broader IDEAYA organization and our extraordinary scientists and research labs in South San Francisco to all of you in Midtown New York City today. Please begin the video.
Ten years ago, IDEAYA Biosciences was founded with a vision to discover breakthrough first-in-class targets that will transform the way we treat cancer. Today, we're at a significant inflection point, and our focus on scientific innovation in addressing the highest unmet medical needs in cancer will continue to be our strategic guidepost in the coming decade. We started with a small but passionate team in South San Francisco focused on one goal: to bring first-in-class therapies to patients with genetically defined cancers. Our targeted oncology approach is grounded in a deep expertise in cancer biology, a strategic focus on first-in-class targets and rational combinations, and a biomarker-enabled clinical development strategy. This allows us to advance first-in-class medicines that are specifically tailored to the patient's genetics. In 10 years, we've advanced seven clinical to IND-stage programs across multiple biomarker-defined patient populations, and we expect two more INDs by year-end.
Our pipeline includes differentiated programs addressing high unmet medical needs in cancer, including uveal melanoma, lung, colorectal, and gynecological cancers. We believe d darovasertib has the potential to become the standard of care therapy across the uveal melanoma patient journey. At the same time, we're scaling our next-generation drug discovery and clinical development engine. We are in a golden age of cancer biology. Our molecular understanding of cancer has never been greater, which enables more precise targeting of the molecular underpinnings of cancer and to identify rational combinations to address the challenges of resistance and tumor heterogeneity. IDEAYA has been a research pioneer focused on identifying new first-in-class targets that were once thought to be undruggable.
Second, we believe the next frontier lies at the intersection of therapeutics and diagnostics, particularly in early-stage disease, where intervention in the pre-metastatic setting has the potential to shift the survival curve and to dramatically improve long-term outcomes. Early-stage disease is where precision medicine has the potential to be the most impactful and where we believe the greatest patient breakthroughs will occur in the coming decade. Lastly, we're investing in the future, embracing artificial intelligence and machine learning to drive more efficient and faster drug discovery. Over the past decade, we've built a remarkable organization, a culture rooted in scientific innovation and teamwork. Now we're entering a new era in oncology research. We're at the dawn of an innovation renaissance fueled by modern technology, scientific advancements, and a clear molecular understanding of the drivers of cancer.
IDEAYA will continue to focus its R&D efforts in the highest areas of unmet medical need in cancer, always willing to take on the greatest scientific and clinical challenges to advance patient care. We've only scratched the surface of what's possible in the treatment of cancer. We will now introduce today's speakers: Dr. Arun Singh, the Director of Ophthalmic Oncology at the Cleveland Clinic, Darrin Beaupre, Chief Medical Officer, Dr. Michael White, Chief Scientific Officer, Dr. Jasgit Sachdev, Senior Vice President of Clinical Development, and myself, CEO of IDEAYA, will also serve as your host today. For today's agenda, we have approximately two hours of presentations and 30 minutes for analyst Q&A. I'll kick off to provide a brief introduction and to outline IDEAYA's strategic vision for the next decade to build a global leader in precision medicine oncology. Next, Dr.
Singh will provide a walkthrough of updated clinical data for darovasertib and neoadjuvant uveal melanoma, specifically in plaque therapy-eligible patients. Next, Dr. Darrin Beaupre will walk us through the first-in-human data that was presented yesterday in Barcelona, Spain, at the World Cancer Lung Conference for ID E849, a potential first-in-class and best-in-class Phase I DLL3 TOP1 ADC for small cell lung cancer. Next, Dr. Michael White will walk through the preclinical mechanistic combination rationale for both IDE 161 PARG and TOP1 ADC and the IDE 397 MAT2A and Trodelvy clinical combinations and MTAP-deletion. Next, Dr. Jasgit Sachdev will present the first-in-human IDE 397 and Trodelvy clinical combination results in MTAP-deletion urothelial cancer. Dr. Mike will then present the product profile of IDE 892, our potential best-in-class PRMT5 inhibitor, and provide a walkthrough of our AI-enabled drug discovery capabilities.
I will then provide brief closing remarks, and we will start the analyst Q&A portion of the event. Since founding IDEAYA Biosciences 10 years ago, we are now on track to have nine clinical to IND-stage programs in the clinic by year-end. The key strategic priorities over the past decade have been pursuing first-in-class targets, discovering patient selection biomarkers, and enabling transformative rational combinations. Through this, IDEAYA has discovered and advanced multiple first-in-class targets into the clinic, including Pol Theta helicase, Werner helicase, PARG, and MAT2A, and has been a research pioneer across multiple biomarker-defined solid tumor populations, including MTAP-deletion, BRCA, and MSI-high.
As we look forward into the coming decade, we believe there are three major macro trends that will redefine the way we treat cancer, including, one, a deep molecular understanding of the drivers of cancer that will lead to new breakthrough targets that were once thought to be undruggable, including helicases to transcription factors. Second, the continued growing intersection between therapeutics and diagnostics that has the potential to deliver better survival outcomes in early-stage disease, including in the neoadjuvant and adjuvant settings. Finally, the continued transformation of the life sciences industry fueled by information technology, specifically the application of artificial intelligence in drug discovery. These three macro trends in oncology and life sciences will serve as important strategic guideposts for IDEAYA Biosciences in the coming decade. We have launched two potential registration-enabling trials for darovasertib, and U.S. commercial readiness activities are ongoing.
We believe darovasertib is uniquely positioned to be the potential standard of care therapy across the uveal melanoma patient journey, including in the neoadjuvant, adjuvant, and metastatic settings. We're excited to advance our efforts to bring darovasertib to patients worldwide with our partner, Servier . Beyond darovasertib, we have identified our next two clinical growth drivers in our TOP1 ADC and DNA damage repair and MTAP-deletion clinical portfolios. Our TOP1 ADC and DNA damage repair clinical strategy is supported by two potential first-in-class clinical stage programs in IDE 849 and IDE 161, as well as a third program in bispecific B7H3/ PTK7 ADC IDE034. We are on track for an IND next month. In MTAP-deletion, we believe IDEAYA has built the broadest pipeline in the industry with potential first-in-class MAT2A inhibitor IDE 397 and phase II potential best-in-class PRMT5 inhibitor IDE 892, and a third MTAP program, which is on track for an IND next year. Lastly, in the medium to long term, our strategic focus on early-stage disease and AI-enabled drug discovery will continue to drive our growth for our next-generation pipeline. For our ADC pipeline, we have a three-prong strategy we're implementing, including one, advancing potential first-in-class ADC programs into the clinic that have sufficient single agent activity to pursue a monotherapy accelerated approval path in the later-stage metastatic setting. Second, enabling ADC immunotherapy clinical combinations to pursue earlier-stage disease, including in the first-line setting.
Finally, enabling first-in-class clinical combinations with TOP1 ADCs and novel DNA damage repair small molecule mechanisms such as PARG to enhance the durability response of DNA damage repair mechanism-based ADC payloads. To advance our industry-leading MTAP portfolio, we are pursuing a three-prong combination strategy. First, targeting multiple nodes in the MTAP pathway to address key resistance and bypass mechanisms by taking a two-hit approach with MAT2A and PRMT5. Second, targeting key genetic coalterations of MTAP such as RAS, CDKN2A, and EGFR, which is the strategic premise of our potential first-in-class MTAP program number three, where we are on track for an IND next year. Third, leveraging the DNA damage repair mechanisms associated with the MTAP-deletion pathway, which is the basis for our Trodelvy and IDE 397 clinical combination that we will present later today.
Lastly, as we look to enhance our IND engine capabilities in the next-generation pipeline, we'll continue to invest our growing capabilities in AI-enabled drug discovery. IDEAYA's differentiation in AI-enabled drug discovery lies at the intersection of first-in-class biology, AI-enabled drug discovery, and structure-based drug design. With that, we'll now advance to the clinical development update portion of the event today, and it is now my pleasure to introduce Dr. Arun Singh from the Cleveland Clinic to walk through the darovasertib data in the neoadjuvant uveal melanoma setting Dr. Singh.
Morning. Thank you, Yujiro. I'm here to present background information pertaining to primary uveal melanoma and some of the emerging clinical data from the OPTIMUM-09 trial that tests darovasertib as neoadjuvant therapy in subjects who are eligible for plaque-brachytherapy. Many medical professionals are involved in the management of a patient with uveal melanoma. Most commonly, the patient presents to a general ophthalmologist after having some difficulty with vision. The suspicious findings trigger a referral plan that can involve ocular oncologists, radiation oncologists, medical oncologists, considering any other physicians over long-term follow-up and the management of these patients. Often in the United States, tumor biopsy samples are obtained for genetic testing to identify patients at metastatic risk. The management of uveal melanoma and subsequent outcomes are all size-dependent. If a patient has a small tumor, they may be observed for a period of time.
If tumor is medium-sized, plaque-brachytherapy is undertaken to deliver radiation to achieve local control. If the tumor is large, an enucleation procedure will be required. That is where the eye is removed. Fortunately, based on these two procedures, local relapses are relatively uncommon. However, in the process of this therapy, the patients may either lose an eye or lose their vision from radiation complications. About 40% of patients go on to develop metastatic disease, which has limited treatment options. The current treatment paradigm for localized melanoma, therefore, opens the door for novel therapies that can be implemented in the neoadjuvant setting that can protect vision, preserve eye, and may even save lives. This takes us to the OPTIMUM-09 trial that tests darovasertib in the neoadjuvant setting. This would be the first of its kind.
The vast majority of uveal melanoma tumors have mutations in GNAQ or GNA11 that activate protein kinase C, a key driver of tumor growth. darovasertib is an oral potent PKC inhibitor, which has shown promise, particularly in combination with crizotinib in metastatic melanoma. In this study, darovasertib, provided as a single agent at the dose of 300 mg twice a day prior to primary local therapy, was evaluated to determine its effect on tumor shrinkage, radiation reduction, visual outcome, and eye preservation. In this presentation, we will focus our attention on plaque-bracket therapy cohort, which included 39 subjects that were evaluable for safety and 21 for efficacy. Herein, subjects received over 12 months of neoadjuvant therapy with darovasertib and then were transitioned to plaque-bracket therapy. For those subjects who received benefit from the neoadjuvant therapy, adjuvant therapy was planned for an additional six months, up to six months.
Primary endpoints for this cohort included safety and radiation dose reduction, and secondary endpoints included vision and relapse-free survival. In the 39 subjects evaluated for safety, the majority of treatment-related adverse events were grade 1 or grade 2, with a low frequency of high-grade adverse events. The most common adverse events included diarrhea, nausea and vomiting, rash, and fatigue. Dose reductions and discontinuations due to treatment-related adverse events were uncommon, as was the frequency of serious side effects. Therefore, darovasertib was well tolerated in this setting. In the 21 efficacy evaluable subjects who had received at least three cycles of the therapy, the baseline characteristics revealed a median age of 56 years, median tumor height of 7 mm, and a median basal diameter of 14 mm. Overall response criteria required a 20% or greater decrease in tumor reduction.
As shown in this waterfall plot, the majority of patients experienced tumor shrinkage, and a large proportion reached the 20% reduction threshold for partial response, providing an overall response rate of 76% in this patient population. Let us review a few case examples. The first case was presented at the International Society of Ocular Oncology meeting in July of this year. The subject is a 62-year-old male who experienced vision changes and was diagnosed with uveal melanoma involving the right eye. The tumor height was 6.2 mm, and the basal diameter was 13 mm. The patient had evidence of retinal detachment as shown on the ultrasound, and after 12 cycles of darovasertib treatment, the subject experienced a 63% reduction in tumor height and 71% reduction in the basal diameter, and the retinal detachment also resolved. This led to improvement in vision while the patient was receiving neoadjuvant darovasertib.
Also, you can see the change in tumor intrinsic characteristics, where it goes from a hollow echogenic appearance to one that's more dense. That's typical of tumors that are undergoing cell death and necrosis. Furthermore, the radiation dose to the optic disc and fovea was reduced. Based on our vision prognostication tool, his risk of 20/200 or worse vision at three years, that is the same as legal blindness, was also reduced. The second case is a 56-year-old female with a tumor that was 4.9 mm in height and 14.7 mm in basal diameter. After 10 cycles of neoadjuvant darovasertib, there was remarkable flattening of the tumor, 77% reduction in the productive tumor dimensions. It's remarkable. Radiation reduction to the key visual structures, such as optic disc and fovea, was near 40%.
Because of distancing of the tumor from these structures, the probability of having 20/200 or worse vision at three years was also reduced. Darovasertib seems to have significant potential in the neoadjuvant setting. Before we transition to additional data from the OPTIMUM-09 trial that Darrin will present, I want to briefly describe two tools that were used in this analysis, and they will be included in the registration trial to support some of the relevant endpoints. In this slide on the left, you can see an example of an eye using eye physics software where the tumor is simulated for radiation treatment prior to neoadjuvant darovasertib. Once neoadjuvant therapy was completed, a repeat simulation is performed to identify the radiation distribution after tumor reduction. In this example, a reduction in radiation dose to the key eye structures, that is fovea and optic disc, is evident.
By comparing pre and post-Darovasertib treatment simulations, one can measure the benefit with respect to radiation dose reduction. On the right is a vision prognostication tool that my team developed. The tool predicts the risk of severe vision loss, that is 20/200 or worse, that is the same as legal blindness, at three years in subjects who receive plaque-bracket therapy. Of the five baseline factors, darovasertib can favorably improve all four modifiable factors by causing tumor shrinkage. Now I'll hand it over to Darrin, who will discuss the additional results from the OPTIMUM-09 study and how they link to endpoints of the registration trial that is being initiated. Darrin?
Thank you very much, Dr. Singh. That was terrific. Based on the safety and efficacy data gathered from the OPTIMUM-09 study, a registration trial testing darovasertib in the neoadjuvant setting has been initiated. What I will do here on this slide is describe to you the overall design of the registration trial, and in subsequent slides, I will provide you data from the OPTIMUM-09 study that supports each of the primary and secondary endpoints chosen for the analysis. OptimUM-10 is a two-cohort study enrolling subjects with small to mid-sized tumors requiring plaque-brachytherapy and also patients with large tumors requiring enucleation. Eligible subjects will be those with a high risk of metastases and also moderate to high risk of vision impairment.
Subjects will be randomized in a 2:1 ratio to either` immediate primary local therapy or neoadjuvant therapy with darovasertib, 300 mg twice a day for up to six months. Patients will be subsequently followed for the endpoints listed on the right of this slide. The primary endpoints include vision preservation measured by the reduction in the proportion of subjects with best-corrected visual acuity greater than or equal to 15 letters lost in the plaque-brachytherapy cohort and eye preservation rate in the enucleation cohort. The secondary endpoints include the proportion of subjects with clinically significant macular edema, visual acuity of 20/200 or worse, and radiation reduction in the plaque-brachytherapy group. Overall response rate and event-free survival will be evaluated in both cohorts.
On this slide, we will address the first key primary endpoint in the registration trial associated with vision preservation, which is supported by data from the ongoing OPTIMUM-09 phase II trial. Data from this study, shown in the left portion of the figure, reveals that after darovasertib neoadjuvant therapy, there is significant tumor reduction measured by both productive diameters and apical height. In addition, this correlates with radiation reduction in the majority of patients. This is further associated with a greater than 20% reduction in radiation dose to at least one key visual structure seen in nearly half of the patients treated. This is significant since a 20% reduction in radiation dose is known to correlate with improved visual outcomes. In addition, looking at the individual vision-associated eye structures on the right, in this patient population, a significant reduction in dose to the fovea, optic disc, and lens was observed.
The goal of the registration trial is to reduce the loss of three lines of vision, that is 15 letters, in the treatment arm versus the control arm in 20% or more of subjects. With this level of radiation reduction, it seems likely that darovasertib, provided in the neoadjuvant setting, will have a relevant impact on visual outcomes, and I will show you more evidence to support this in the next few slides. Focusing for a moment on the enucleation cohort, where eye preservation is the primary endpoint, data from two trials, both the OPTIMUM-09 study and the NAINOM study that tested darovasertib in the neoadjuvant setting, show that most subjects, both enucleation and plaque-bracket therapy subjects, experienced tumor shrinkage, with 61% of subjects scheduled for enucleation able to undergo eye-preserving therapy instead.
Based on this data, darovasertib received breakthrough therapy designation from the FDA for patients with primary uveal melanoma for whom enucleation has been recommended. In the registration trial, a lower confidence interval bound of only 10% needs to be exceeded in order for the study to have a positive outcome, which appears readily achievable. In this slide, we move on to data that supports the secondary endpoints for the OptimUM-10 registration trial. I've already shown you data regarding the ability of darovasertib to reduce radiation dose. This reduction is likely to be associated with a lower level of macular edema, which is relevant since radiation-induced macular edema is the leading cause of post-plaque-bracket therapy visual impairment. Using the vision prognostication tool described by Dr. Singh, data from the OPTIMUM-09 trial shown here reveals that neoadjuvant darovasertib can impact four of the five predictive factors that affect visual outcomes.
In this data set, 2/3 of the patients had a reduction in the risk of 20/200 vision or worse at three years, and nearly 40% had a 20% or greater reduction in this risk. In addition to longer-term visual outcomes predicted to improve, data gathered from the OPTIMUM-09 trial revealed that darovasertib in the neoadjuvant setting could also improve vision while subjects were receiving treatment. In this trial, 65% of subjects had any improvement in visual acuity score. The median number of letters gained was five, and 8 out of 20 subjects had a greater than or equal to five letters gained on two consecutive visits. Other examples of benefit during the neoadjuvant period included a reduction in subfoveal fluid and improvement in retinal detachment.
Finally, addressing the endpoint of overall response from the registration trial, the response rate is predicted to be high based on the waterfall plots you have seen thus far. With respect to event-free survival, the objective of this study is to show that delaying primary local therapy by using neoadjuvant treatment results in no detriment with respect to local or distant relapse. This would seem likely since protein kinase C is the key driver of this disease, and darovasertib is an effective protein kinase C inhibitor. Also, thus far, the relapse rate from the OPTIMUM-09 neoadjuvant trial remains low. Even in those few subjects that progress on neoadjuvant darovasertib therapy, patients are taken directly to primary local therapy as soon as they show evidence of tumor growth.
Therefore, the design of the OptimUM-10 registration study and current supportive data suggest that a neoadjuvant approach should be safe and effective without causing detriment in EFS. In closing, I would like to first take a moment to thank the many investigators that helped us design and execute our clinical studies in uveal melanoma. Some of them are listed here. This specific group worked with us to design and execute our trials in the primary uveal melanoma setting, where a neoadjuvant therapy that can save eyes and preserve vision does not currently exist. We've had extensive discussions with this group regarding the implications of our study results and the probability of success in the OptimUM-10 study.
Unanimously, these investigators have high enthusiasm for darovasertib in this setting, and they believe our trial, as designed, has not only a high likelihood of providing benefit to uveal melanoma subjects, but also has a high likelihood of meeting the primary endpoint. Therefore, in conclusion, for this update, darovasertib represents a novel therapy for primary uveal melanoma. Importantly for patients, it has received FDA breakthrough therapy designation in the neoadjuvant setting for patients who require enucleation, as well as fast track and orphan designation. Data from OPTIMUM-09 has revealed that 3/4 of plaque-bracket therapy subjects treated with darovasertib experienced greater than 20% ocular tumor shrinkage by productive diameters. Half of the patients achieved greater than 20% decrease in simulated radiation doses to at least one key visual structure.
65% had any degree of visual improvement during neoadjuvant therapy, and 40% achieved greater than or equal to five letters gained at three consecutive visits. Results from a vision prognostication tool revealed that 2/3 of subjects received at least some reduction in the risk of 20/200 vision or worse at three years, and nearly 40% obtained a greater than or equal to 20% risk reduction. A clinical data update in greater than 90 enucleation or plaque-bracket therapy subjects who received darovasertib in the neoadjuvant setting will be presented at ESMO next month as an oral presentation. In the phase III randomized neoadjuvant uveomelanoma trial, OptimUM-10 has been initiated. I would like to turn it over to Dr. Singh to provide some of his thoughts regarding the evolving data from the OPTIMUM-09 trial and the implications for the uveal melanoma landscape. Dr. Singh?
To summarize, everything in uveal melanoma is size-dependent. We classify tumors as small, medium, and large, and the things we offer to the patient, the treatment choices, and the outcomes are all size-dependent. For large tumors, we offer enucleation, and they lose their eye. For medium-sized tumors, we offer radiation therapy, and they lose their vision. All of these tumors—small, medium, and large—are causing metastasis, and the risk of survival or metastasis is also size-dependent. darovasertib is reversing this tumor growth. It is converting large tumors to medium-sized tumors, and some medium-sized tumors will also transform or reduce to the classification of small-sized tumors. Therefore, one can derive that the rate of enucleations or globe preservation will go up. Radiation toxicity, which is size-dependent, will also reduce. They will have vision preservation, and all of that will also impact the risk of metastasis.
This is a major change in how we think about uveal melanoma and what we may be able to offer in the near future, and it has not happened since 1915. That is when radiation therapy was started for uveal melanoma. I am excited, and everybody else in our business is excited. Thank you so much.
Now we're going to transition over to a program that probably everybody's been thinking about in the last 24 hours. It's the IDE 849 program for subjects with small cell lung cancer and other neuroendocrine tumors. What I'll be presenting in the next 15 or 20 minutes is data from the first-in-human trial of IDE 849 that was recently presented at the World Conference on Lung Cancer in Barcelona, Spain, sponsored by our colleagues at Hengrui Pharmaceuticals. Let me begin with some background information. Small cell lung cancer is a highly aggressive neuroendocrine carcinoma with a poor prognosis and limited treatment options. Importantly, in small cell lung cancer, DLL3 is a surface protein expressed in a large proportion of these tumors, allowing targeting with an antibody-drug conjugate. IDE 849 is an antibody-drug conjugate comprised of a humanized IgG1 monoclonal antibody linked to a topoisomerase-1 inhibitor payload via a cleavable linker.
In preclinical studies in tumors that expressed various levels of DLL3, IDE849 inhibited growth and demonstrated a bystander killing effect in small cell lung cancer cell lines and was also effective in small cell lung cancer xenograft models. This set the stage for the clinical testing of IDE849 in a first-in-human trial. This slide provides the study design for the first-in-human study, IDE849. It was a multi-center trial that enrolled subjects with small cell lung cancer and other neuroendocrine tumors that expressed DLL3. Patients were required to have progressed or recurrent disease after standard therapy and have a performance status of zero or one. Patients with brain metastases were eligible, and the total number of patients for this analysis was 100, with a data cutoff of June 20th, 2025.
A standard dose escalation design was performed beginning at a starting dose of 0.8 mg/kg every three weeks, with a top dose that was tested of 4.2 mg/kg . During the conduct of this trial, an additional cohort of 3 mg/kg was included. The primary endpoints included safety in defining a recommended phase II dose, with secondary endpoints including efficacy, pharmacokinetics, and immunogenicity. Dose escalation proceeded up to 4.2 mg/kg , and at that dose, a grade 4 decrease in platelets was observed. Therefore, this dose level was discontinued. Due to the fact that the majority of anti-tumor activity occurred between 2.5 mg/kg and 3.5 mg/kg these two cohorts, along with the 3 mg/kg cohort, were expanded. This slide presents the demographics of the patients enrolled in the study. They were 70% male, 30% female. Median age was 60.
Most subjects had a performance status of one, and most patients had a diagnosis of small cell lung cancer with advanced disease. Patients with brain metastases constituted approximately one-quarter of the population. About half of the patients had at least two prior lines of therapy, and immunotherapy had been used in the majority of subjects. Here we provide the pharmacokinetic profile of IDE849. Following a single-dose systemic exposure of IDE849, total antibody and payload increased proportionally with doses ranging from 2.4 mg/kg- 4.2 mg/kg . Across the 2.4 mg/kg- 4.2 mg/kg doses, the mean half-life ranged from about 10 - 11 days. A tumor stasis concentration of 3 micrograms per mL derived from preclinical tumor growth inhibition modeling justified the clinically effective dose range of 2.4 mg/kg- 4.2 mg/kg .
Exposure of free plasma payload was low across these dose levels, with the mean half-life ranging from four to five days. The following slide displays the safety profile observed in the early phase of this trial. The most common treatment-related adverse events were myelosuppression. In the entire patient population, 69% of patients experienced neutropenia. 33% were high grade. Febrile neutropenia was uncommon. This was followed by anemia and thrombocytopenia. Here, the grade 3 or greater adverse event rate was only 6%- 7%. The GI organ class was the next most commonly affected, but the overwhelming majority of these events were low grade. Treatment-related adverse events leading to dose reduction or discontinuation occurred in 15% and 2% of subjects, respectively. Serious adverse events were r eported in 16% of patients. More importantly, at the 2.4 mg/kg dose, the grade 3 treatment-related adverse events rate or greater was below 20%.
A 7% rate of interstitial lung disease was reported across the patients in the study, of which 1% were grade 3. There were no grade 4 or grade 5 cases reported. Tumor response is represented in this waterfall plot and accompanying table of subjects treated with doses ranging from 0.8 mg/kg- 4.2 mg/kg . The vast majority of the subjects experienced tumor shrinkage, with over 70% having a best response of PR. A significant proportion of these were confirmed. In fact, at the 2.4 mg/kg dose, a dose with a highly favorable benefit-risk profile, the confirmed response rate in second line was 70%. Anti-tumor activity was seen in subjects with varying degrees of DLL3 expression and even in subjects where DLL3 levels were below detectable limits. The disease control rate was over 90%.
The efficacy data is expanded in this slide as we look at the response over time. In these figures, you can see that the majority of patients respond within the first several months, and in several subjects, this durability goes beyond the six-month time period. Keep in mind the data is still immature. At this time, the median follow-up of the patient population is only three and a half months. Looking at subjects specifically who had brain metastases upon entering the study, all subjects showed evidence of tumor shrinkage, and a confirmed overall response rate of 67% with a disease control rate of 100% was observed across all dose levels tested. A confirmed response rate of 83% and a disease control rate of 100% was observed specifically in those subjects who received 2.4 mg/kg dose.
As mentioned, although the data remains immature, this progression-free survival curve looking at subjects who were treated at doses of 2.4 mg/kg or greater showed a median progression-free survival of 6.7 months in subjects who received second-line therapy or beyond. The median progression-free survival was not reached in subjects who were treated in the second line only. This is meaningful since with standard-of-care therapy, even in first line, the progression-free survival is typically less than six months. In conclusion, IDE849 demonstrated a tolerable and manageable safety profile in patients with relapsed small cell lung cancer. The most common high-grade adverse events were associated with hematologic toxicity, and the treatment had a low rate of dose reductions and discontinuations due to adverse events. There were no treatment-related deaths thus far observed on the study.
IDE849 demonstrated promising anti-tumor activity with a best response of PR in nearly 3/4 of patients, including patients with brain metastases, with early evidence of durability, and follow-up is underway to evaluate longer-term outcomes. A study of IDE849 has been initiated in the United States, which will include global sites outside of China and will be dosing its first patient very, very soon. With that, I'll hand it over to Dr. Michael White, our Chief Scientific Officer, who will talk to you about some of our combinations that we'll be testing in the clinic very soon or are currently underway. Mike?
Thank you, Darrin. Thanks for saving us some time. I'm not going to have to run through this at 3x ordinary speaking speed. I think, as we are all aware, the clinical experiences like the one that Darrin just described here mean that oncology is enjoying a real renaissance in ADC therapies, and that's driven primarily because of advances in linker payload strategies that are focused on topoisomerase-1 inhibitors. The next challenge is going to be optimizing strategies for sufficient payload delivery to drive robust responses in tumors with otherwise suboptimal antigen density. It is in the context of this challenge that a unique synergy between IDE 161 and topoisomerase inhibitors offers a potentially spectacular combination opportunity to be able to enhance both efficacy and duration of topoisomerase-based ADC.
For those of you not familiar with the program, IDE161 is a potential first-in-class small molecule inhibitor of PARG, a poly-ADP ribose glycohydrolase, that takes PAR chains off of PAR-related proteins in order to resolve DNA repair events that are initiated by PARP. This is particularly important to maintain replication stability in cancers with replication stress. If you inhibit PARG, you get the accumulation of these complexes as chromatin roadblocks on the chromatin, and these cause DNA replication catastrophe and cancer cell death. That mechanism of action is what is offering a potentially stunning opportunity to amplify the potency of TOP1, topoisomerase-1 inhibitors in cells, and that was first reported by investigators at the National Cancer Institute and summarized in this schematic right here. As many of you are likely aware, a topoisomerase inhibitor works by stabilizing a covalent linkage between topoisomerase-1 and the DNA nick.
That means to get out of the way before replication 4 comes through, or you're going to have replication catastrophe and potentially kill the cell. This lesion is recognized by PARP, and it is PAR-related by PARP in order to recruit DNA repair proteins that will chew off topoisomerase-1, this is via the proteasome, to expose that nick to allow it to be repaired. However, before that happens, this complex needs to be de-PAR-related to expose topoisomerase to the proteasome. There's only one enzyme that does that, and that is PARG. If you inhibit PARG, you wind up with this sugar-coated boulder on the chromatin that accumulates and causes cell death through mitotic catastrophe. It's that mechanism that can enhance the cytotoxicity of an otherwise subtherapeutic dose of a topoisomerase inhibitor.
Consistent with that, we see very, very nice combination benefit of IDE161 plus a topoisomerase inhibitor across over 400 cell lines here with responses that are 10 to 100-fold more potent than what you would get with either single agent. We see this across multiple lineages, as shown in this panel. In some of these lineages, the vast majority of the cells are responding much better to this combination than either single agent. As you would expect, when you put this combination in cells, you see a big upregulation of PAR chains because of these cycles of futile DNA repair events, and that comes along with an accumulation of these topoisomerase cleavage complexes. These are those covalent links that I was just telling you about that accumulate because they cannot be de-PAR-related and degraded by the proteasome. Importantly, these mechanisms translate very nicely with respect to efficacy.
Here's a small cell lung cancer model. It is very poorly controlled by IDE161 or a maximally efficacious dose of topotecan, a topoisomerase inhibitor. You can see the combination; we get complete regressions. Here's a high-grade serous ovarian cancer. It is somewhat controlled by topotecan. It does have a response to IDE161, but that relapses over time. The combination, however, gives you complete regressions, and those are maintained off treatment. As you might imagine, this offers a really exciting opportunity to enhance the activity of ADCs that have topoisomerase payloads. We evaluated that here using one of everybody's favorite ADCs and HER2. Here are three different models: non-small cell lung cancer, colorectal cancer that's HER2 low, small cell lung cancer that's HER2 low. Each one of these models is poorly or completely not affected by IDE161 alone.
They have responses to a HER2 ADC, but they can be limited to varying degrees in various models. In every case, when we put that combination together, we see deep regressions that can be durable, that are very, very tolerable. We view this as an important mechanistic substantiation preclinically of the opportunity that this is going to afford in the clinic to make ADCs work more broadly across all tumor indications. Two shots on goal that I'm very excited about that we have in our own portfolio, the DLL3 asset that Darrin just described. As you heard, this is doing very well by itself in small cell lung cancer. We also reported this morning at the World Lung Conference that we see nice combination benefit with IDE 892 together with IDE161 in tumors that have low antigen density.
Here you can see this fantastic combination opportunity or combo activity here. This bodes well for delivering a durable response in tumors with low antigen density. It's actually probably particularly important as we expand beyond small cell lung cancer into other important neuroendocrine tumor types that are DLL3 positive. Second opportunity we have is our IDE 3034, the B7H3/ PTK7 bispecific antibody. We view bispecific antibodies as a particularly important opportunity to combine with IDE161. That's because if they are avidity-tuned to only bind double-positive cells, they can have very specific tumor binding versus normal tissue. That can also come with an efficacy penalty. You could get suboptimal payload delivery because of heterogeneity of the exposure of each of those antigens on the tumor. This is a very, very nice ADC bispecific. You can see here we get good double-positive binding.
We also see double-positive internalization with the bispecific versus either a single-agent antibody. As we showed you guys last year, we get good durable tumor control in double-positive tumors. If you have tumors with a low expression of one antigen or the other, you have suboptimal control. Here we see multiple dosing with ID 034 and the generation of resistance. IDE161, which doesn't do anything by itself, turns that response into a regression. Opportunity here for us to maximize benefit across all the tumors with topoisomerase-based ADCs. Now I am going to keep you on tenterhooks a little bit before Jasgit's presentation of our next clinical data disclosure, give you the science behind our exciting collaboration with Gilead looking at IDE 397 together with Trodelvy. MTAP deletion. You drove introduced that. This is lost because you lose CDK and 2A. 15% of solid tumors.
Very important therapeutic opportunity because homozygous loss of MTAP installs two distinct vulnerabilities that can be targeted with a MAT2A inhibitor. First, you have accumulation of MTAP's substrate, MTA. MTA is a SAM competitive inhibitor, a PRMT5. You partially reduce the ability of tumor cells to maintain high-fidelity alternative splicing that's required for them to grow and divide. Second opportunity, when you lose MTAP, you lose its products, and its products contribute to nucleotide synthesis. You have reduced refilling of the nucleotide pools that are required for the high demands of DNA replication and repair in tumors.
This generates a very strong dependency on MAT2A to produce sufficient SAM to be able to overcome MTA inhibition of PRMT5 and to be able to refill the folate cycle to make up for de novo purine synthesis and pyrimidine synthesis so that you can keep your nucleotide pools intact to be able to repair your DNA. As you can see here, IDE397 can give you two hits in the MTAP null setting. Inhibition of PRMT5 together with loss of MTAP causes increased DNA damage. I'll tell you why in a second, at the same time as it reduces DNA repair capacity. You can see all of that in action here. From global untargeted metabolic profiling, we see strong suppression of nucleotide pool in the MTAP null setting as compared to the MTAP wild type. You can see that in this Cytoscape plot focused on nucleotide subcluster.
These are all nucleotides, nucleosides, the building blocks required for nucleotide biosynthesis that are lost selectively in this setting. Why does that happen? It's exactly as you might expect. If we look at targeted metabolomics, we see that in the presence of MTAP deletion, IDE397 causes the loss of homocysteine because of insufficient SAM to maintain that SAM conversion to homocysteine. This results in a block of the ability of 5-methyltetrahydrofolate to cycle back into tetrahydrofolate to now fuel this cell cycle and produce nucleotides. You can see that here. Loss of homocysteine breaks the connection between the methionine cycle and the folate cycle. The accumulation of this metabolite can also back inhibit the enzymes in this pathway to further reduce folate cycle production of nucleotides and inhibit DNA repair capacity.
Those mechanisms underpin what we see as a very high association of cell lines that are sensitive to IDE397 as being the same cell lines that are also sensitive to topoisomerase inhibitors or DNA synthesis inhibitors or pyrimidine synthesis inhibitors that we saw across 700 cell lines using the DepMap informer data set. The presence of topoisomerase at the top of this list really underpins this mechanism here that's related to perturbation of RNA splicing. If you have perturbed splicing, the RNA polymerase senses this and stalls as part of a checkpoint repair mechanism. That stalling creates the formation of these R-loops, which are RNA-DNA hybrids that can be detected, as you see here, induced by IDE397, specifically in MTAP null cells. These R-loops need to be resolved before a replication fork comes through.
Otherwise, you're going to have replication fork collapse, DNA damage, deactivation of a DNA damage response. That's exactly what we see with IDE397 in tumors here, read out by phosphorylation of CAP1 and ATM substrate. The way that these get repaired is through topoisomerase, which releases RNA polymerase in order to allow replication through that region that was a former R-loop. You can imagine the consequence of putting a topoisomerase inhibitor together with IDE397. These relationships were substantiated in kind of a spectacular fashion, in my opinion, through multiple genome-wide genetic enhancer screens using CRISPR technology.
As you can see here, the top biological pathways, the top gene enrichment that was observed to be synthetic lethal with IDE397 across multiple MTAP cancer models were components of the DNA repair machinery, components that are involved in the activation of a DNA damage response, and components that connect the methionine cycle to the folate cycle. You can see the GSCA plots down here. If we map each of the individual top enhancers to biological pathways, here's where we line up. We see knockouts that enhance MTA production through activation of polyamine synthesis. We see knockouts that block the connection between the methionine cycle and the folate cycle by blocking the synthesis of homocysteine or blocking the activity of methionine synthase. We see knockouts that directly interfere with the import of the building blocks that are required for nucleotide synthesis.
We see knockouts that take away the ability of cells to repair their DNA or to resolve RNA-mediated replication stress through resolution of R-loops. All of this translates preclinically as shown here. This is an aggressive MTAP null bladder cancer model. It is poorly controlled by IDE397. It is somewhat controlled by irinotecan, a topoisomerase inhibitor, which loses control over time. The combination produces deep regressions, which are quite durable. This is all the way out till day 75 before Claire stopped the experiment, and they're very well tolerated. To sum up here, we see a very strong preclinical mechanistic opportunity to be able to combine IDE397 together with Trodelvy in the context of MTAP-deletion because this combination here does two things. It reduces repair capacity together with triggering genomic instability. These are synthetic lethal with topoisomerase inhibitors.
Put the topoisomerase inhibitor on an ADC to maximize the benefit, and you have something that might be very effective for patients with MTAP tumors who often have shorter PFS and OS with standard-of-care agents like EV and CPI. With that, I will turn the floor over to Jasgit, 2 minutes and 35 seconds under the time that I was allotted, to tell you the important part of this story.
Thank you, Mike. Hi, everyone. I'm pleased to share with you today the preliminary safety and efficacy results of the first-in-human evaluation of the combination of Trodelvy plus IDE397 in subjects with MTAP-deletion urothelial carcinoma. Recently, the incorporation of IO agents in the maintenance setting or as part of first-line standard-of-care regimens has improved the overall survival in patients with advanced or metastatic urothelial cancers. As such, enfortumab plus pembrolizumab, or EV plus P, has become the standard first-line regimen for these patients in the United States. However, patients with MTAP-deletion urothelial carcinoma are less likely to benefit from these treatments, given the cold tumor microenvironment associated with these tumors. Moreover, there are no targeted therapies approved for MTAP-deletion urothelial cancers.
The combination of Trodelvy, a TROP2-targeting TOP1 inhibitor payload ADC, plus the oral MAT2A inhibitor IDE397 that is currently being evaluated in the dose escalation and expansion study IDE397-001 represents a novel targeted treatment regimen that has the potential to improve outcomes for this poor molecular subtype of urothelial carcinoma. Monotherapy IDE397 at the recommended phase II dose of 30 mg daily in MTAP-deletion non-small cell and urothelial cancers showed promising efficacy, and this data has previously been presented at the ENA meeting in 2024. Today, we'll focus on the combination data for Trodelvy plus IDE397. Based on the overall safety observed during dose escalation, two expansion doses have been selected for further evaluation.
As a background, it is important to note that Trodelvy as a single agent produced an objective response rate of 23% in the randomized phase III TROPiCS-04 trial in a patient population where more than 90% of patients had no prior enfortumab exposure. A more recent real-world data set showed an objective response rate that is about half of that at 11% with monotherapy Trodelvy in patients who had prior treatment with EV or enfortumab. This slide describes the patient demographics for the safety population treated at the two expansion dose levels for the combination. As you can see, this is a late-line population with the majority of the patients having received two or more prior lines of therapy that included standard-of-care treatments like platinum and IO agents. A 1/3 of the patients had received prior enfortumab.
This represents a higher proportion of EV-treated patients compared to the TROPiCS trial population. Presented here are the most common all-grade and grade 3 or greater treatment-related adverse events among subjects treated at dose level one, which is 10 mg/kg of Trodelvy with 15 mg of IDE397. This AE profile reflects the known safety profile of both agents. However, it is noteworthy that the rate of neutrophil count decrease and neutropenia reported thus far in this trial is lower than what has previously been seen in a urothelial cancer population treated with Trodelvy. Proactive measures to reduce infectious and other complications from neutropenia in high-risk patients are incorporated in the adverse event management guidelines in the trial. As such, there's only been one case of febrile neutropenia reported at this dose.
The rates of GI adverse events, especially high-grade events, likewise are also low, thereby making this a feasible dosing regimen from a safety standpoint. This dose level is being expanded further and continues to enroll patients. A second dose level that combined the recommended phase II dose of IDE397, that is 30 mg once a day plus 7.5 mg/kg of Trodelvy, was subsequently explored. As noted, this dose level delivered an even more favorable safety and tolerability profile for the combination, as evidenced by the very low rate of grade 3 or greater treatment-related adverse events, particularly hematologic events. No grade 3 or greater neutropenia has been reported so far at this dose level. Additionally, no treatment-related serious adverse events were reported at this dose. This dose level has been selected as the second expansion dose for further evaluation of safety and efficacy of the combination.
We now come to the preliminary efficacy observed with the combination regimen in this heavily treated population of MTAP-deleted urothelial carcinoma patients. The best overall response for subjects who were treated at one of the two expansion doses and had at least one evaluable post-baseline scan available at the time of the data snapshot is shown here on the waterfall plot. There are seven observed partial responses among 16 evaluable subjects, of which six are confirmed responses. If we now look at the breakdown of these responses by dose level, we see that the objective response rate for dose level two is impressively higher than that for dose level one. Four of seven, or 57% of the patients at dose level two, have had a partial response, three of which were confirmed on a subsequent scan, and the fourth is pending confirmation. This highlights the important contribution.
of IDE397 to this combination, whereby despite a lower starting dose for Trodelvy, the efficacy at this dose level appears to be more favorable. In addition, as shown, the safety and tolerability of this dose promises to provide an ideal risk-benefit ratio for this patient population. As noted before, the efficacy of Trodelvy in a post-EV setting, as reported from the real-world dataset, shows an objective response rate that is in the low double digits, about 11%. The median progression-free survival and duration of response have not yet been determined since several patients are still on treatment and follow-up is short.
What we can say, as shown in the spider plot, is that the time on treatment without progression in several patients has already exceeded the historic benchmark of about two and a half months reported in a late-line urothelial cancer population with standard-of-care treatments. What is also striking is the rapid onset and depth of responses seen with the combination, which also compares favorably to that seen with monotherapy IDE397 in this trial. I am now going to share a couple of case studies with you from among the participants treated with the combination regimen in the IDE397- 001 trial. The first one shown on this slide is a 78-year-old patient with a good performance status who enrolled in the trial after having experienced progression on standard-of-care treatments.
His prior treatments included a combination of cisplatin plus gemcitabine, followed by maintenance with an approved anti-PD-L1 agent, avelumab. Progressive disease on avelumab then prompted a switch to an ADC targeting NECTIN-4, the same target antigen as enfortumab . This patient experienced further disease progression with a heavy burden of disease in the liver. Some representative liver lesions on his baseline scan are shown in the image on the right. This subject was enrolled in dose level 2 and by week 12 experienced a dramatic reduction in the liver metastases consistent with a partial response. This partial response was subsequently confirmed on the week 18 scan. The patient continues on treatment at this time. Shown here is another example of a patient who derived clinical benefit from this treatment after his disease had stopped responding to the best available standard-of-care therapy.
In this case, the first-line regimen selected for metastatic urothelial cancer was enfortumab plus pembrolizumab, followed subsequently by cisplatin plus gemcitabine. The tumor had metastasized to both lungs by the time the patient enrolled in our trial. The baseline scan shows a lobulated lung mass in the right lower lobe. This patient was enrolled and treated at dose level 1. As you can see in the follow-up scan, there's complete resolution of the metastatic lesion in the lung with therapy. Tumor shrinkage was also noted in the other target lesions, resulting in a 65% reduction in the sum of target lesions, corresponding to a deep partial response on the first on-treatment scan. This patient's confirmatory scan was performed after the data snapshot, and this response has now been confirmed.
In conclusion, we're very encouraged by the preliminary safety and efficacy of this first-in-class targeted combination for the treatment of MTAP-deleted urothelial cancer, which comprises about a quarter to a third of all urothelial cancers. Additionally, this regimen is also being evaluated in MTAP-deleted non-small cell lung cancer, which represents about 15%- 20% of that indication. What you've seen today is the feasibility of two dose levels for this combination, with the responses noted at both doses. The objective response rate with the combination is higher than that reported with single-agent Trodelvy in an EV-naive or EV-pre-treated population, as well as that reported with IDE monotherapy in this trial. In particular, dose level 2, which combines the optimal dose of IDE397 at 30 mg daily with an efficacious dose of Trodelvy, is trending towards being the optimal dose for the combination.
Further evaluation of both doses is underway, with an aim to define the recommended phase II dose for this combination by year-end, which will be determined by the long-term safety and tolerability, as well as ongoing demonstration of favorable efficacy with a target objective response rate of 40% or greater and durability that exceeds six months. With that, I'm going to have Dr. Mike White come and talk to us about IDE 892, our best-in-class PRMT5 inhibitor.
Okay, so that's all the clinical data. You can all relax. We'll talk about what's coming next IDE892 potential best-in-class PRMT5 inhibitor. As Yujiro noted, a key priority for our MTAP enterprise is to bring IDE397 together with a PRMT5 inhibitor that is MTA cooperative. That's because preclinically, this combination routinely over and over again delivers complete durable responses at a fraction of the dose required for maximum monotherapy activity with either a PRMT5 inhibitor or IDE397 and that monotherapy activity is often inferior than what you can see with that combination. There are really two aspects that are driving this combination benefit that I'm going to take you through that have really led us towards what our target profile is for an IDEAYA PRMT5 inhibitor that's MTA cooperative.
The first is obviously maximal PRMT5 pathway suppression to deliver deep regressions. The second is reciprocal inhibition of bypass mechanisms that occur on monotherapy to be able to deliver this durable response. With respect to the first, as you all know, MTA is a SAM-competitive PRMT5 inhibitor. The ability of an MTA cooperative PRMT5 inhibitor to block this enzyme is completely dependent on the MTA-SAM ratio. This can be very variable across tumors depending upon the extent of MTA synthesis and the extent of MTA export. As we showed earlier this year and then also last year at the two different AACR meetings, this ratio can be optimized by inhibition of SAM production by an allosteric inhibitor of MAT2A IDE397 to really drive things towards locking PRMT5 in the inactive conformation.
You can see this directly here with an assay that is a nanobread assay that we developed in collaboration with Promega that can show a direct target occupancy in cells. The yellow is 100% occupancy of PRMT5. In these dose matrices, we only get that in the combination setting, and we don't see that in the wild type setting. It is very important to optimize the MTA-SAM ratio in order to be able to get maximum PRMT5 pathway inhibition. That is not enough. What about these durable responses that we are seeing? To uncover the biology behind this, we simply molecularly profiled the tumors that were resistant or becoming resistant to monotherapy doses that are maximally effective. This is using the BMS Mirati PRMT5 inhibitor or IDE397 maximal effective doses, establish resistance, maximal effective dose, establish resistance.
Combination at 1/10 of these doses, we get durable responses. When we profile these, the first thing that we found was that these tumors that relapse on the BMS compound do not lose control of the inhibition of splicing. That BMS compound maintains pathway inhibition even after relapse. You can see that here with SDMA IHC from the end of study tumors. You can see that here. These tumors are thriving despite the fact that the pathway is inhibited at day 35 just as much as it's inhibited at day 7. They just don't care anymore. If we look at the whole genome transcript profile from these end-of-study samples and compare them to each other and to day 7, we find that IDE397
has the ability to maintain control of epigenetic alterations of chromatin that can establish resistance and to inhibit MYC pathway activity, which you do not see with the PRMT5 inhibitor. In contrast, the PRMT5 inhibitor seems to have the better capacity for persistent inhibition of mRNA splicing fidelity, as you can see here, which is consistent with its alternative splicing analysis that we see over here on the left. In aggregate, we see a situation that in combination, we are able to interfere with reciprocal bypass mechanisms that would overcome either a PRMT5 inhibitor or a MAT2A inhibitor to deliver a durable response. Observations like those guided our plans for an MTA cooperative PRMT5 inhibitor that is designed really exclusively for combination benefit with IDE397
and our MedChems strategy to be able to deliver this was heavily influenced by a series of mechanistic studies together with our lead discovery team that I'm going to summarize here. First, our lead discovery team identified a previously unreported mechanism by which SAM, SAW, and MTA metabolites can exchange in the PRMT5 active site, even when it's blocked by an MTA cooperative PRMT5 inhibitor. That is because at the back of this pocket, there is a dynamic interface that can create a channel to the cytosol, to the solvent-exposed milieu, such that SAM and MTA can dynamically exchange with each other in that pocket.
This is important because that means that the extent to which an MTA cooperative inhibitor can block that enzyme, its residency time, depends upon how antagonistic it is with respect to SAM, when to keep that out of the pocket versus cooperative with MTA to pull that into the pocket. We really want to solve for that and make sure that we evaluate MTA cooperativity and SAM antagonism at the same time when we build our IDEAYA molecule. Second, the group designs a very nice novel surface plasmon resistance protocol that allows us to directly quantify target occupancy of PRMT5 in a purified system and how inhibitors can bind that in the presence or absence of distinct metabolites. Here's an example. Here, we start out by flowing DMS control or the metabolites or metabolite mix over a chip that has immobilized PRMT5.
We add an inhibitor for a certain duration and measure its occupancy on the protein through the release of the plasmons that are associated with that differential size. As one example here, this is an MTA cooperative PRMT5 inhibitor profile in the presence of DMSO. We add the inhibitor. We get a little bit of binding, but not a lot. This is kind of still in the, shall we say, baseline range of the molecule being able to bind the apo form. In contrast, here is a 1:1 ratio of SAM-MTA metabolites added ahead of the inhibitor, and we can enhance the binding. Over here on the right, you're seeing the aggregate PRMT5 occupancy data, IDE892 our clinical development candidate.
Here you are seeing how it binds with pure SAM versus a SAM-MTA ratio that is equivalent to the average MTAP wild type cell versus pure MTA or a SAM-MTA ratio that is equivalent to your average MTAP-deleted cells. As you note, there is a very, very evident SAM antagonism here. We have less binding of our inhibitor to the protein in the presence of SAM than we would see with no metabolite at all. In contrast, we see very, very nice MTA cooperativity. Over here with this epizoome compound, as you would expect from a SAM-competitive inhibitor, we see very high binding in the presence of SAM, no binding at all in the presence of MTA. This molecule is kind of in between the two. It has strong cooperativity with MTA and is also SAM-cooperative, which is something that we're trying to avoid with our molecule.
The BMS compound, Mirati compound, has a profile that's very, very similar to IDE892 . As you would expect from a biochemical profile like this one, IDE892 has absolutely no activity in HCT wild type xenograft models. This is a HCT MTAP wild type xenograft model. This dose with IDE892 at a maximally effective dose as compared to what you would see in the MTAP null setting. We see very, very strong single-agent and combination activity in multiple CDX and PDX models, including in the setting of relapse. The IND is filed, and we're excited to bring that combination to patients very, very soon. Okay, AIML drug discovery, what are we doing as a company in this space?
We often get asked by some of you, in fact, how it is that we maintain a level of productivity that is perhaps somewhat more than you might expect from a company of our size. I would just like to point out that from a discovery perspective, a major strength of IDEAYA is structure-based drug design, together with the physics, the structural biology, the MedChem expertise that is required to apply structure-based drug design to novel targets. For us, AI and ML, that big explosion of capabilities in that space, is best deployed by integrating that with our in-house capabilities in computational physics, our battle-tested in-house capabilities in computational physics and biophysics, and then to use that integration as a force multiplier in order to accelerate sustainable delivery of new first-in-class opportunities, as well as pursue exceptionally challenging activity profiles.
I'll walk you through some cases here, some use cases that really point out why it is that we have prioritized investing in building these capabilities in-house. I'm going to start by calling out the most daunting challenge that our industry faces with respect to delivering durable therapy to cancer patients, and that is massive mechanistic heterogeneity and adaptive cellular phenotypic plasticity. Those things limit the number of patients that will respond to a given therapy and limit the amount of time a patient will respond to therapy if they're lucky enough to be an initial responder.
Some strategies to deal with heterogeneity include attacking truncal vulnerabilities, things that occur during tumor initiation and therefore are more likely to be carried forward during tumor evolution, or disrupting epigenetic modulators that allow for that phenotypic diversification in the first place that can cause resistance, and also eliminating these tumor persistence cells and pluripotent progenitor cells that serve as incubators for acquired resistance over time. How do you do that? We're kind of excited about the fact that lysine acetyltransferases are gaining recognition as important players within these biological systems and therefore may offer themselves as targets to be able to intercept these things and deliver a durable therapeutic response.
The lysine acetyltransferases are enzymes that function by acetylating the tails of histones at specific locations in the chromatin in order to be able to relax it enough for transcription factors to bind and induce expression of the adjacent gene. Where that happens is specified by these multi-protein complexes that bring the enzymes to the appropriate locations in order to be able to serve specific enhancer and promoter regions. The key piece I want to remind you guys here is, as we showed at the AACR earlier this year, KAT6 and KAT7 are key epigenetic modulators of cell identity and lineage specification that get corrupted during oncogenic transformation to support transcription factors that act as lineage survival oncogene network activators, so a truncal vulnerability.
They act to maintain tumor initiating cell identity, and they act to maintain or allow for a switch of cells into this drug tolerant or persister cell state. These associations offer an important therapeutic opportunity if you can deliver a KAT6/7 dual inhibitor, but that comes with a very, very substantial discovery challenge, both because of the difference of the size and sequence in the pockets of KAT7 and KAT6, and also the need to be able to avoid the close-in lysine acetyltransferase family members, KAT5 and KAT8 that perform essential functions in wild type cells. How did we get around this?
One of the important pieces of IDEAYA 's enterprise in the context of going after new targets are physics-based tools like free energy perturbation that, together with machine learning-based training for parameter optimization, can allow you to predict the affinity of a molecule that has never been synthesized before to a specific domain in a protein that has never been drugged before. We apply this for late lead optimization in the KAT6/7 program because our lead series had an issue with exposure, based on its polarity properties. In order to get around this, the MedChem team penned over 500 innovative design ideas as a way to try to get around this issue.
In order to sort through those, free energy perturbation calculations were used to assign predicted binding affinities in a distributed manner across these molecules and landed the team in a particular space such that out of those 500, a very small number were synthesized, which eventually led to our clinical development candidate. The fact that we were able to prioritize 20% of these design ideas to have success meant that we shaved at least a year off of this program as compared to what we otherwise would have done if we didn't have these predictive tools to be able to sort things based on their predicted affinity. Why should you care? That delivered what we believe to be the world's first KAT6/7 selective inhibitor. You can see that it binds KAT6 and KAT7 very nicely, avoiding KAT5 and KAT8, both in vitro and in cells. This modulates the pathway appropriately.
We see nice inhibition of lysine-23 acetylation and lysine-14 acetylation, the substrates for KAT6 and KAT7. This profile was able to substantiate a number of the mechanistic hypotheses that we had on the table, including the facts that KAT6 and KAT7 are stronger modulators of chromatin architecture than KAT6 inhibition alone. Here you're seeing a comparison of our dual inhibitor versus a KAT6 selective inhibitor for chromatin modulation. This is the clinical compound from Pfizer that is seeing some success in metastatic breast cancer in combination with estrogen receptor inhibitors. What you see in the red box here are individual genomic loci that are much more strongly suppressed by the combination of the dual inhibitor versus KAT6 inhibition alone, both K23 and K14.
Down here, you can see at the resolution of transcription factor start sites, superior modulation of H3K9 by dual inhi bition of KAT6 and KAT7 versus KAT6 inhibition alone. This translates to very, very strong sulfate modulation by the dual inhibitor versus a KAT6-specific inhibitor. Here we use gene expression analysis that was summarized and integrated using OpenAI large language models, where the output was tuned by a custom retrieval augmented generation to be able to deliver a line of sight in an unbiased way on the modulation of lineage specification and epigenetic regulators that are superior by dual inhibition versus single inhibition, then validated by ATAC-seq-mediated enumeration of transcription factor motifs. Here we're seeing in breast cancer cells, GATA3 is very strongly suppressed. This is a lineage specification gene. In lung cancers, FOXA1 is strongly suppressed. This is a pioneering transcription factor.
These are some highlights from breast cancer and lung cancer models that are going to be described in detail at an international cancer conference coming up in about a month. The other thing that you might remember about KAT6/7 is we showed at the AACR earlier this year that metastatic breast cancers treated with KAT6 are associated with both intrinsic and acquired resistance, which can be bypassed by dual inhibition of KAT6 together with KAT7. Here you're seeing an RNA-seq experiment where we are evaluating the individual contributions of these inhibitors to cells, looking at the individual gene expression programs and seeing how they relate to each other. You can see emergence of a specific clone in the KAT6 inhibitor-treated cells with a distinct phenotype that happens in that arm and not the others.
That clone, the gene expression programs that are associated with that include endocrine therapy resistance and components that are associated with drug tolerant persister cells. This is a drug tolerant persister cell population that emerges upon KAT6 inhibitor treatment but does not emerge in the context of KAT7 inhibition together with KAT6 inhibition because a dual inhibitor kills these drug tolerant persister cells and also blocks the maintenance of the pluripotent tumor initiating cell population. Remember, these are the cell types from which resistant populations are derived for acquired resistance. Those phenotypes come together to deliver very, very strong control of preclinical models. This is a PDX model, luminal breast cancer model that has an estrogen receptor mutation that makes it resistant to fulvestrant. This model cannot be controlled by a monotherapy KAT6 inhibitor at any dose. What you're seeing right here is a clinically relevant dose.
At a fraction of the clinically relevant dose of a KAT6 inhibitor, this dual inhibitor IDE574 gives you complete regressions, which are durable off treatment. We also see activity beyond breast cancer, as we had hoped, in a biomarker-specific fashion. If you profile hundreds of cell lines with IDE574 you can see that we have potentially broad opportunities for monotherapy activity across meaningful patient populations, including breast, lung, ovarian, and esophageal gastric cancers. In order to be able to apply this kind of delivery of first-in-class molecules, we are marrying AI and physics-based capabilities across our pipeline from hit ID through to hit to lead and then lead optimization. I'll give you three examples of those. MTAP program three, this is a program that was really enhanced by an in-house built property prediction tool called Harmony.
This is a property prediction tool that is dynamically trained on measured data to be able to predict solubility, permeability, or stability phenotypes of molecules. As you can see here for this program, it did a very nice job of predicting designs with respect to those that would be soluble versus not, those that would be permeable versus not, or those that would be stable versus not. We can go very, very quickly and avoid designs that would be ineffective because they lack these drug-like properties. From an early program one, this is a program that saw a big step change in potency through AI-mediated generative design. This is one that I'm very excited about because this kind of approach is usually limited by the need to weed through an incredible amount of noise, which is indicated by all these gray dots.
The way the team solves for this problem is through a distributed ranking across chemical spaces. It's a two-dimensional projection of a 200-dimensional Euclidean space to evaluate the predictive binding in these different areas using free energy perturbation. That allowed the team to go in a single hop from their early lead into a compound that has very potent activity as well as nice drug-like properties and is exceptionally different from the early hit. Very nice advancement in this space and something that we're looking forward to applying to many more programs. Early program three, sorry, early program two, this is a program where the initial feedback from the MedChem team and the lead discovery team when we were going to launch this program was that it's probably impossible but exciting enough to try.
This is a situation where image enhancement, machine learning-based image enhancement of very, very large crystallography fragment screens allowed us to find a toehold that would have otherwise been invisible. Putting that together with ligand-induced protein dynamics helped us land on a molecule that has both a goal-forward potency as well as a spectacular selectivity between our target. Target one is a therapeutic target and our anti-target. Target two is the anti-target. Moving forward now in a very solid position to be able to do hit-to-lead finding for early program two. I'm going to stop with this slide. Virtual high-throughput screening is something that everybody wants to work. This is one of the most frustrating aspects of small molecule discovery, and that is finding some starting matter against targets that have never been drugged before.
Current technologies seem to be performing adequately for previously drugged targets or for some target classes that are very, very well understood, but they are not performing for targets that have never been drugged for first-in-class targets. We are excited to try to help overcome this limitation by bringing the accuracy of absolute binding free energy to the picture such that we can, at scale, make the calculations that are required to be able to decode intermolecular interactions inside reasonably sized domains of targets that have never been drugged before. We have generated a collaboration in order to build the computational infrastructure required to do this at scale, as well as introduce the machine learning cycles in order to decode these interaction networks. It's picking up steam. We're getting some key deliverables in place, and we hope to tell you more about that soon.
With that, Yujiro, I'll turn it over to you for closing remarks.
Great. Thank you, everybody. We're going to move here into the analyst Q&A portion here in a moment. Really, thank you to Michael, Darrin, and Jasgit. Really, just terrific presentations. It's been really just great going through all of this new clinical data and the various updates that we're working on across the portfolio. In the last 10 years, as you've just seen, we believe we've built one of the broadest and deepest pipelines in precision medicine oncology today. As you heard earlier in the presentation, darovasertib is now launched into multiple registrational trials across the melanoma patient journey, including in the metastatic and neoadjuvant setting. We're targeting to start a registrational study with our partner, Servier, next year in the adjuvant setting.
Next, as you heard today, several key clinical data updates, including first-in-human data across our IDE849 - DLL3 program that was presented yesterday in Barcelona, and also the first-in-human data on the IDE397 Trodelvy combination and MTAP urothelial cancer. As you heard from Michael just earlier, we're continuing to make significant investments in the area of AI-enabled drug discovery that we believe will continue to fuel our future pipeline. We're not done for the year. We've got a lot of catalysts that we're focused on and that we're very excited to be able to present here in the next few months. The next step is at ESMO. As noted earlier, we have over 90 patients in neoadjuvant uveal melanoma data that's been accepted as a proof-of-paper oral presentation.
In addition, fairly recently, we also noted we have an oral presentation accepted at SMR that will also get presented in October. This will be our first reported survival data in the frontline metastatic uveal melanoma setting. Lastly, our PFS results to hopefully enable our first accelerated approval filing. With that, that's the end of our prepared remarks. We'll now move to the analyst Q&A portion. I'm going to ask several members, Daniel, Stu, Mike, Josh, Darrin, and Jasgit to please come up. I will work with the LifeSci team here to hopefully facilitate the analyst Q&A portion. We'll just ask folks to raise their hand. Ideally, try to keep it to one question per analyst. Please just say your name, who you're with. We'll try to do this within 30 minutes. I know everybody has a busy morning schedule.
We'll try and do it as orderly as possible. I will call the name of each analyst. If you could just raise your hand so we could easily find you with the mic, that'd be helpful. We'll begin with Anupam Rama from JPMorgan. Please limit yourself again to one question, one follow-up.
Hi, this is Priyanka here for Anupam at JPMorgan. For the IDE849 update, can you help us understand what you are seeing from the safety perspective by doses that are 2.4 mg/kg or higher? How does that play into your dose selection?
Darrin, you take that.
Yeah, as shown, we've tested a number of doses with IDE849 This is a Chinese patient population. We're getting ready to enroll our first patient in the U.S. patient population. Ultimately, the 2.4 mg/kg dose looks like an attractive dose in terms of benefit-risk. Whether that'll be true in the U.S. patient population, we're going to know soon. Seemingly, that's a really great guidepost. Long and short of it is, I think if you look at our data and you compare it with anything out there with respect to ADCs in this space, I won't mention any names, but suffice it to say at that dose level, we compare very favorably.
Based on the efficacy data there, as I mentioned, we're looking at a drug that has the kind of activity that some in this space would call maybe the most active agent in small cell lung cancer. Remember, in that group, patients get combined chemotherapy. They get a checkpoint inhibitor up front. They have a median progression-free survival less than six months in first line. We're talking about relapsed patients, a very high response rate. It looks seemingly durable. We have a dose that seems extremely tolerable at this point. We're very, very excited to move the program forward in the U.S. The data is still evolving. It's early days, but we're obviously highly excited, especially for patients with small cell lung cancer, really bad disease.
Great. Next question from Maury Raycroft at Jefferies.
Thank you. Thank you for hosting this event. Congrats on all the progress over the years. I'll also ask one, this is Maury from Jefferies. I'll also ask one on DLL3. The reported 6.7 months PFS in the second-line plus population looks good. What's your early perspective on how PFS might evolve, specifically in second-line patients, acknowledging current follow-up is limited at this time?
We are hoping we're going to stay well north of six months, and only time will tell. Even if we were to stay where we're at now, we look like maybe the most active agent in this space. We want to play a role in small cell lung cancer as early as possible. Also, we're thinking about other neuroendocrine tumors. There are a number of other high-grade neuroendocrine tumors that could benefit from this kind of therapy, as not only a monotherapy, but as Mike pointed out with IDE161, there's a tremendous opportunity to make tumors that perhaps have lower levels of DLL3 expression also be sensitive. It could actually expand the number of tumors for which this may be relevant.
Just to follow up to the prior question related to dosing, for the combination with PARG, you've got monotherapy data with PARG already. Maybe talk about how doses could work for both of those drugs together and just thoughts on potential overlapping toxicities as well.
Yeah, I would say that the good news is with 161, we've really come up with, we think, a treatment schedule that's highly favorable in terms of a safety profile. That being the case, it makes us much less concerned about overlapping toxicity, quite frankly. I think the way we're delivering 161 is good for patients. It's allowing us to maintain dose intensity. I think we'll still have to do a standard dose escalation sort of design in order to put the two drugs together. I think we're very, very bullish that we're going to be able to deliver the appropriate dose of our DLL3 ADC along with 161 based on what we've seen recently.
Thank you.
Yigal Nochomovitz, Citi .
Hi, thank you. Could you just explain a little more detail how you adjudicate the events in terms of AEs between 397 and Trodelvy given you showed the table? Could you just clarify how you parse those out, please.
Sure. Jasgit, you take that.
Yes, so these are investigator-adjudicated, as you know, for most trials. This is an investigator determination. In a combination regimen, sometimes it's hard to tease out one agent versus another. We know that there are some specific toxicities that may be linked more likely to be from Trodelvy versus 397. That's what the adjudication represented. If there was a toxicity more likely to be related to one agent versus another, that's how it was attributed.
Yeah, maybe just on that point, in many cases, I would suspect the adjudication was to both molecules. I don't think you would just add the two numbers together.
Okay, I was curious on the, I don't know if the expert on the oculars is here, but regarding the prediction tool, is that more of a static tool that's kind of forward-looking, or is there any iterative component, AI-driven aspect to how that's done? Obviously, as you collect more data, you could potentially refine the algorithm and have it do a better prediction job. I was curious about that.
Dr. Singh.
This tool was developed 10 years ago, and there wasn't much of AI then. Nowadays, we are actually treating patients with radiation retinopathy, so we do not have data on untreated cases. It's very difficult to develop this model. There is a current trial going on, DRCR trial, Protocol AL, that has untreated component in it that gives us data on untreated patients. That will be used to refine this predictive tool. It'll happen, but it'll be another maybe two or three years from now.
Gotcha. Thank you very much.
Tyler Van Buren, Cowen?
Great. Thank you very much for the very interesting presentations. Related to the neoadjuvant vision data, can you tell us what proportion of patients had a greater than or equal to 15 letter loss in the phase II and what you would expect in the control arm in phase III? Maybe just a follow-up for Dr. Singh, what is your confidence in the phase III being successful in this endpoint given the tumor reductions and reduction in radiation observed?
Dr. Singh, you want to go first?
It's hard to be highly predictive, but we know, like I emphasized before, there are five things that drive vision after brachytherapy, and four can be modified, and that are being modified with this drug. We do therefore feel that it will help not all patients, but a vast majority of them, because some vision loss is due to tumor location itself. For example, if melanoma is in the macula where the vision is, you can reduce it to anything, but the patient will never get any better vision. There will be some percentage of patients where this will not work, but in the rest of it, 80%- 90% of it, there will be an impact on vision.
Now, with respect to your question about 15 letter loss, there's a couple of issues here. One is the follow-up is relatively short. Also, as you know, for 09, it's a proof of concept study, so there's no control arm. Even if we had that data, it would be hard to sort of make a whole lot of sense out of it. That is where the registration trial comes in. It's a well-constructed, very, very well-designed study to have a control arm, a treatment arm. We know pretty well, based on what's been published, what our expectations are. What do we know? We know that in that patient population, they'll be coming in, they'll be patients with moderate to high risk of visual impairment, which means, based on published data, at about three years, nearly half of them will have 20/200 vision or worse.
That's like 35 letters lost in the ETDRS chart. Now, we're only looking for 15 letter loss, so that's less vision loss. There'll be a control arm. In the treatment group, they'll be, we predict, based on the data that you saw today, significant tumor shrinkage, radiation reduction. That degree of letter loss, 15 letters, we think, based on the data we presented to you today, seems, to be quite honest, in my mind, maybe I'm biased, but readily achievable based on what we've seen. Again, that list of investigators that I mentioned, they're very bullish about the study and the outcomes based on the data that we've shown so far. I don't know, Dr. Singh, if you wanted to add anything to that.
No, I think you just said it correctly. There's not much to add.
Great. Li Watsek, Cantor.
Hey, great. Thanks for the great update across the pipeline. I guess on darovasertib, just curious for patients with less than 20% radiation reduction, is there any benefit to vision loss risk? How is this 20% threshold identified? Also, curious if you can expand on how you determine if a patient responds or not and how many cycles do you typically wait before you make that determination?
Jasgit, you want to take that?
Yeah, I think to answer your first question, we do have data that suggests that if you have a 20% or more reduction to radiation dose, for example, tumor apical dose, that does translate into less visual complications. Dr. Singh can probably corroborate that data. We've also shown that in our data set, 20% or greater reduction in tumor size actually correlates really well with the likelihood of patients being able to get eye preservation. The same group of patients are also more likely to have radiation reduction, as well as eventually a lower risk of 20/200 or worse vision. You had a second part to your question as well, right? Did I address both of those, or?
Yeah, how do you determine a patient response or not?
That's right. In the trial, patients are followed every cycle with tumor measurements. The treatment is allowed up to a maximum of 12 cycles in the neoadjuvant setting, with the response assessment or tumor shrinkage assessment ongoing before every cycle. For most of the patients, the sharpest decline that we see is in the first three to four cycles. However, response assessment is ongoing. What you saw today was their best overall tumor shrinkage at any point during neoadjuvant therapy. On the flip side, if there's any evidence of tumor growth, that's defined as having reached maximum benefit, and then they can go on to local therapy after that.
Just to add on to your point, remember, currently, a response criteria for this neoadjuvant approach doesn't exist. Why? Because there's no therapies out there that can shrink tumors like this. This is number one. Where that 20% comes from is a bit empirical. What we've seen is, we're trying to correlate tumor shrinkage with clinical outcomes that matter. Like, is their vision predicted to be better in three years? Is the radiation predicted to be reduced? Is their eye likely to be saved? It looks like this 20% is the crux when you take a clinically relevant outcome that seems to be associated with that number. That's where that number comes from. I think we're going to be working on a publication with people like Dr. Singh and his colleagues to actually define a response criteria. It'll all revolve around the data that's evolving from this study.
Maybe just a quick follow-up on the DLL3. Wondering if you can just comment on the dose response here. It doesn't seem like there is a very clear relationship. There seems to be a pretty steep dose, I guess, step up from level one. Can you just comment on if you might need to explore lower doses in the U.S.?
Yeah, maybe I could take that. So Li, I think once we kind of get to the 2.4 mg/kg dose, I think we're at a very active dose. As we scale up beyond that, 3.0 mg/kg, I believe it was 3.4 mg/kg, 3.5 mg/kg, we went up to 4.0 mg/kg and above. I think there, as you saw, the activity was fairly consistent. I think here, you know, relatively still small numbers. At this point, I think we're probably orienting towards that 2.4 mg/kg, 3.0 mg/kg dose.
Yeah, the only thing I'd add to that is, remember, it's about durability too. Being able to keep the drug on board for patients, we want to push that six months out as far as we possibly can. Delivering a dose that allows us to be effective with a high response rate and being able to deliver it consistently is important. This is a rapidly growing tumor. You don't want to be off the target for very long. We are going to figure this out when we get into the clinic in the U.S. patient population. The idea would be to find that dose with the best benefit-risk ratio to keep patients on therapy throughout therapy to make sure that their responses are durable.
Corinne Johnson, Goldman Sachs.
Thanks. Maybe a specific question and then a higher-level strategic question. On the darovasertib side, I'd love if maybe Dr. Singh could chime in on how he thinks about treatment duration in the neoadjuvant setting and sort of what he expects to see there, how he'll make decisions about moving on to enucleation or radiation, as it were. Maybe from a strategic perspective, I'd be curious, like we spent a lot of time today talking about the discovery engine, but two of the premier assets were licensed. How do you think about allocating resources and time and prioritization across your drug discovery versus licensing efforts? Thanks.
I'll begin, then I'll hand it over to Dr. Singh. With respect to the duration, in the registration trial, we're looking at six months in the neoadjuvant setting. The idea here is really the maximum amount of tumor shrinkage happens typically within the three-month, four-month time period, tailing off at six months. You do see anti-tumor activity after six months, but most of it has happened in the first six months, which is sort of where that empiric six months has come from. I don't know, Dr. Singh, if you have anything to add.
Also, the magnitude of reduction is anything from 30%- 50%. If you think of a 10-mm tumor, which is a cutoff from medium to large, you're talking about the tumor becoming, let's say, 5 mm, which is easily treated with radiation standard of care. We are converting, I think, a large number of enucleations definitely to brachytherapy. Within brachytherapy, say, if it's a 6-mm tumor, you expect it to become, let's say, 3 mm. That's going to have a very good outcome.
The other thing, too, is when you're trying to preserve event-free survival, you know, getting at the tumor right when you've blasted it as much as you need to is really a great idea, which is another reason why we're staying within the six-month time period.
Josh, do you want to take that question on the investment side?
Yeah, I think the question is around resource allocation across the portfolio. Our goal here, as we're bringing forth these assets, is to get them to as many patients as we can, because we think they're really differentiated. In some cases, that may be doing it ourselves. In other cases, that may be better done with a partner. We're very open to that. I think it will be driven by the data we see, the interest we receive, and clearly, as things play out, we'll keep you all apprised. I think we all are very focused on making sure we remain focused on the assets in the pipeline and the things that we choose to do ourselves, that we do them well. We're also very aware of the resources that it will require to stay in this pipeline.
I think going forward, business development will certainly be a big part of our strategy. I think we've demonstrated internal capabilities along those lines to do smart deals with good partners. That will continue to be a big part of the strategy as we think about taking programs forward in the future.
Thanks.
Charles Zhu, LifeSci Capital.
Hi, good morning. Thanks for providing this update and for making the trip all the way to the East Coast to give it. I got a couple of questions on an MTAP-deleted bladder cancer here. One thing we noticed in their baseline characteristics, about one-third of your patients were treated with enfortamab. Is this a proportion that you would expect in a real-world setting where if you were to run a bigger trial in this particular setting? Can you also provide some color, not only around how these MTAP-deleted patients tend to do on enfortamab with or without a PD-1, but also how much more data you would need to kind of decide whether or not you pursue your doublet in a pretreated setting or maybe in a frontline setting against EV302? Thanks.
That's a great question. We suspect going forward, the vast majority of patients are going to get EVP in the front line for bladder cancer. That's the patient population that we're going to be dealing with. You mentioned it is true, and the UNITE study laid out some of this data with enfortumab in patients who are MTAP deficient, showing inferior outcomes. That's been published in several indications also for checkpoint inhibitors where they do worse with MTAP deficiency. With respect to where to go in bladder cancer, which was kind of your third question, obviously, the relapse setting is wide open right now. Everyone's wondering what the heck are we going to do in second line? Even cisetoposide in second line doesn't have a tremendously long progression-free survival, and its response rate is definitely less after EVP. That space is wide open. Second line is clear.
In first line, we'll have to think about that. We'd have to think about how effective we are in an MTAP deficient population. Again, it's going to take some time. We're obviously showing you data to date. It's very exciting, but the durability data is going to make the big difference there. We have to see how durable this is to know exactly where we can penetrate. We've even thought about things like the neoadjuvant setting and other things. EVP may be coming into the neoadjuvant setting sometime soon. If it takes over there, then it may be that the first line opportunity in the metastatic setting opens right up for something like us. That's another thing that will sort of play itself out over the next year or so.
Got it. Maybe if I could ask one quick follow-up on that, especially with patients with prior enfortamab. I think case study one that you presented presented this possibility as well. Are you able to comment how many patients maybe received a NECTIN-4 ADC that was not enfortamab in your patients?
I think the majority of the patients that received that category of ADCs are enfortamab treated. This is a global trial. As you know, that uptake of enfortamab and pembrolizumab is increasing. By the time we get into a registrational strategy with that, we'll probably have more, like what Darrin described, a real-world population, the majority of them should have had enfortamab ahead of time.
Thank you.
Robert Driscoll, Wedbush.
Thank you. Thanks for hosting today. Congrats on all the progress. It was really interesting to learn about the differentiated mechanism for 892 today. Can you tell us what you learned from the AMG-193/IDE397 combination that kind of gives you confidence for this mechanism going forward?
Yeah, thanks for that question. I think the key piece for us there is to really optimize efficacy together with specificity. Right? I mean, MTA is not a tumor-specific metabolite. You really want to be able to control for not getting on the target too hard in a wild type setting. We do want to optimize that MTA/ SAM ratio. In the context of selectivity, SAM antagonism, I think, is very important. Being able to bring that into the profile is something that we think is going to work quite nicely in combination with IDE397. That's something that we see happen very, very nicely also with the BMS Mirati compound.
Got it. Maybe just one science question for you, Mike. Have you looked at whether the combination of the PARG inhibitor plus the TOP1 antibody-drug conjugate leads to the accumulation of R-loops?
Can you ask that last part? I didn't hear that last part.
Sorry, leads to the accumulation of R-loops?
Have we looked at the accumulation of R-loops? We haven't looked directly at that, but we can. I think that that's what you would expect to be able to see in that situation. That would be another mechanism by which you could potentially have a combination benefit. I think that that's a good thing to look at. We haven't specifically evaluated that internally.
Thank you.
David Dai, UBS.
Ava Fortejo, Wells Fargo.
Matt Biegler , Opco.
Hey, this is Marcus on for Matt. You mentioned the importance of hitting KAT6 and KAT7. Can you maybe talk more about why only targeting KAT6, like other approaches too, may not be sufficient? Thank you.
Go ahead, Mike.
Thank you for that question. I like that question. One of the key aspects of that enterprise is the observation that KAT7 acts as a parallel for KAT6. If you inhibit KAT6, you actually bring KAT7 in to be able to compensate. One key piece there is to be able to intercept that mechanism that would otherwise perturb the efficacy of KAT6 inhibition alone. Let's not let KAT7 come in and compensate. That is why we see very nice suppression of K23 lysine acetylation, also K9. Those are both KAT6A/B substrates. They're not normally KAT7 substrates until you inhibit those, and then they become KAT7 substrates. Let's get KAT7 out of the picture. KAT7 brings on additional opportunity with respect to K14. K14 is very, very important because that's a lysine acetylation mark that can precede the ability to switch cell fates.
Let's take that out of the picture because that's how you acquire resistance. That specific combination gets rid of that parallel issue that interferes with the ability of KAT6A/B inhibition alone to have maximal benefit, and then bring on the added benefit of being able to kill those drug-tolerant persister cells and block the maintenance of the pluripotent tumor-initiating cells to prevent the emergence of resistance on treatment. Thank you for that.
Thank you.
Justin Zellen, BTIG. Peter Lawson, Barclays .
Thank you so much. Just on your thoughts around the new FDA and your most recent conversations with them regarding trial designs and endpoints, whether it's for daro or DLL3.
With respect to daro, we've had a number of discussions with the FDA. I assume you're focusing on the OptimUM-10 registration study, which we've been talking about for some time. The discussions are over. We're getting ready to initiate the study. We're in great shape. We've come to an agreement on what that design should look like. I sort of described it in the slide without getting into too many details. We're ready to launch the ship at this stage.
We've got about five minutes left in our Q&A session. I'll get through as many names as we can. Sudan Loganathan, Stephens.
Great. Thank you. I really appreciate all the details in the presentations. I have a question on darovasertib. When designing the trial for the darovasertib neoadjuvant study, there was mention as part of an answer to another question how the location of the tumor may have an effect on the visual outcomes, you know, going towards the end of the study or the ability to improve those outcomes. With this in mind, will there be any inclusion or exclusion criteria to hone in on patients that may have the best outcomes in the neoadjuvant study with daro, or more take an all-comers kind of approach for patients that have tumors at any sites of the eye?
It's a great question. That's kind of why, as I was describing the trial, I said we were going to be taking patients with a risk of moderate to high risk of visual impairment. As Dr. Singh pointed out, the location of the tumor matters. If you have a moderate to high risk of visual impairment, that means you're a patient who's got a tumor in a location that's likely to get radiation to cause some issues. That sort of controls for the patient population who will be enrolled, which is the majority of patients if you include both moderate to high visual impairment folks. It does allow us to now have a control arm, which we know is going to have moderate to high vision problems down the road.
We'll be able to compare that to the six months of neoadjuvant darovasertib where the tumor shrinkage happens, the radiation reduction happens. That's part of how the trial was put together to specifically focus and make sure we don't miss that endpoint, which is there ought to be a significant improvement in vision in those folks on the treatment arm relative to the control arm.
Darrin, I'll add one thing. We're excluding patients that are subfoveal, which means if their tumor is right under the fovea and macula. That's a comment Dr. Singh was making. Those patients, regardless of their size, actually are at very high risk for vision impairment, unfortunately. That is an exclusion both for our ongoing study as well as the registration trial.
May I add the baseline vision also, there's eligibility criteria of better than 20/200, right?7
Patients who have poor vision, like hand motion vision, will not be included. That's the other way to make up for it.
Silvan Türkcan, Citizens.
Thank you for taking my question and congrats on all the new data. I had a question also on the neoadjuvant setting. Maybe for Dr. Singh, on the BCVA test, could you maybe help us understand what a 15% letter loss means? Given that there was 65% of patients achieving actually five letters gained, what does a 20-letter difference mean for the outcome of patients here, giving the enrolled patients that are at high risk of vision loss?
I can just give one comment. 15 letters is what the FDA has established to be a meaningful, significant, or clinically relevant vision loss. Five letters is maybe perceived the difference, but FDA does not consider it to be clinically significant. Therefore, 15-letter loss is considered as a good endpoint for vision preservation or vision loss.
Yeah, I would just say that's kind of, you know, during the discussions with the FDA, it was clear that they wanted, you know, a number that was high enough where the impact would be very, very significant. We've obviously seen a trend based on everything that we've seen today, presented today, that suggests we're headed in the right direction. We'll be able to do that in a significant proportion of patients. Remember, like I said, I think the key here to think about this is based on the patient population that we will enroll. At three years, we're talking about half of the patients being legally blind. That's a long way. That's 35 letters on an ETDRS chart. That's a long way. All we're talking about is trying to do a 15-letter improvement. We know the control arm is going to be doing really bad.
Half of those people will be sort of, you know, legally blind. The other half, we think, will be able to do much better than that. 15 letters is a small bar relative to 2,200 vision.
Brian, I think we're over time. Maybe we take two more.
Let's go. Graig Suvannavejh, Mizuho.
Thanks. Two questions from me, and thanks for taking the questions and the presentation.
First, on the daro study, as you look to do the plaque brachytherapy group, I think the prior guidance was maybe data in 18 months after you enroll. I just wanted to make sure if I had the timing correct, so that's the first question. The second question, obviously a lot of excitement about your DLL3 ADC. Could you just talk about the commercial opportunity, given a sense that I think in the first line small cell lung cancer setting, we're seeing increased use of immunotherapy, and then we've got Amgen's IMDELLTRA in the second line setting as well. How do we think about that commercial opportunity?
Yes, maybe let me take the first one. In terms of visual acuity, actual visual data post plaque brachytherapy, that's corrected by 18- 24 months, I think we'll have a good sense of where we are. For the commercial questions, Stu, why don't you take that small cell? What are your thoughts on the commercial opportunity?
Sure, from a small cell perspective, obviously there's been a lot of activity in the DLL3 space, but as you look at the patient populations right now, there is a meaningful opportunity in the second line plus setting in small cell. The much bigger population is in front line. As we're looking at development opportunities, trying to think through how do we get into those earlier lines of therapy. At the same time, as Darrin referenced, we are not missing looking at the opportunities in neuroendocrine as well. The high-grade tumors, they're a huge, huge unmet need, and that translates into a meaningful commercial opportunity as well.
Okay, we can take our final question from Leonid Timashev from RBC.
Hey, thanks for squeezing me in, really appreciate it. I wanted to return to DLL3 and maybe just on the pneumonitis, ILD. I've talked before about how the LINKA technology might reduce the rates of ILD. I guess I'm wondering if you still feel that way based on the data that you're seeing, if you can talk maybe about the doses that you saw the ILD on, and then as we look ahead to the U.S. study that you're running, if there's any differences in baseline patient characteristics or how they're treated in the U.S. versus ex-U.S. that might impact the ILD rates that we'll see going forward. Thanks.
Yes, I'll take that one, and thanks for the question. In terms of ILD rates, we think so far a 7% ILD rate, which is about mid-single-digit per cent, is actually quite good. In terms of grade 3 or higher, there was one patient, so 1%, no grade 4 or 5. Even when you look at checkpoint inhibitors in lung cancer studies, ILD rates are typically about 5%. When you look at ADCs and lung cancer in general, the ILD rates are about 10%. We think here we're very much in range. We need to start obviously monitoring ILD, of course, but grade 1-2 is actually quite common in lung cancer. I don't know, Jasgit, anything else you'd add here?
I think we have several ADCs out there. They're very effective, right? The ILD rate balances out if efficacy is profound, and as you saw, it's trending towards being something that's probably the most efficacious agent we've seen in the small cell lung cancer space in a pre-treated population. Now it's just balancing out that toxicity, which over time people learn how to manage, and you know, if the outcomes are improved, then the risk-benefit ratio becomes more favorable over time.
Great, thank you so much, everybody. This concludes the analyst Q&A portion, and really thank you so much for attending today.