Relay Therapeutics Inc Aktie

Good morning, everyone. It's my pleasure to introduce the Relay management team. Next to me, I have Sanjiv Patel, President and CEO; and Peter Rahmer, Chief Corporate Development Officer. Thank you for joining us.

To start here, you've reported data from your PI3K-alpha inhibitor as a doublet and triplet in breast cancer as well as in rare disease and vascular anomalies in particular. Could you provide an update on the company's strategy and what you plan to accomplish over the next 12 to 24 months, including further data and updates for zovegalisib in that period?

Absolutely. So first of all, thank you for the invitation, and thank you for the grand venue that we're in today. It's been a busy time over the last year for our company. Our mutant selective PI3K inhibitor, zovega, has had multiple data that we've disclosed. The first in second-line hormone receptor positive HER2-negative breast cancer, we showed data in March of this year at ESMO TAT in Paris to show that we had a very competitive 11 months' worth of PFS in these patients, and that's against the 5.5 months of capivasertib, and we are now busily running a Phase III trial there.

So the goal there over the coming 2 years would be to later this year, guide to when we would have full enrollment of that trial. Enrollment is going great at the moment. And then obviously, as we come towards the end of this kind of window that we've talked about, we'd like to file and then start our commercialization efforts. We know capivasertib has been very successful in this field at the moment with almost $1 billion worth of sales there. So we think we can meaningfully differentiate from that and hopefully take market share.

In the second data that we showed, we showed data concerning our frontline hormone receptor positive HER2-negative breast cancer trial. And there, we showed data from third-line patients and showed a very competitive profile. And we announced that we will go forward in the frontline with a combination with Pfizer's atirmociclib, their selective CDK4. And so over the coming 18, 24 months, obviously, we want to start that trial and get well through trying to enroll those patients because we see this as a very big opportunity.

And then the third data drop that we made was in Philadelphia last month now at the ISSVA meeting, which is the largest global meeting for vascular anomalies, where we showed data for zovega for the first time, where we showed a 70% volumetric -- or 60% volumetric response rate, which compares very favorably to the benchmark set by others in the 11% to 20% range. So we believe we have a truly meaningfully differentiated compound there. So over the next 18, 24 months there, obviously, we will push very strongly towards -- getting towards an accelerated approval and then commercializing in this field where there are really no good therapies available for these patients. So we have had a very productive year, but I think we have a much more productive 18 months ahead of us, too.

Great. Starting with zovega and breast cancer for your Phase III study in the second-line plus group, what do you believe is the minimum delta on PFS versus TRUQAP to promote uptake? And based on physician discussions, what is most important to improve upon for tolerability?

Yes. I think, obviously, TRUQAP was launched about 18 months ago now and took significant share from alpelisib, which was the approved nonselective PI3K inhibitor. And that really the share that it took was through a perceived better tolerability profile and that it showed less hypoglycemia, but had significant rates of diarrhea. Now in the real world, I think we've seen that it does have a liability on hypoglycemia as well as diarrhea. So I think our hope would be to differentiate on all kind of PI3K wild-type-related toxicities, so diarrhea, rash, hypoglycemia, stomatitis. And we believe with zovega, we have a very good profile to be able to do that.

In terms of the efficacy bar, obviously, at 5.5 months, we know that from talking to physicians that a 3-month delta to that would be clinically meaningful. And obviously, we sit in our data that we showed earlier this year at an 11-month PFS across all mutations. And so we believe that we should have a meaningfully differentiated profile both on the safety and on efficacy.

And your triplet achieved about a 44% overall response rate in heavily pretreated patients. So based on historical translation to earlier-stage patients, how do you believe this ORR could evolve in the frontline patients? And what do you believe will be competitive there on PFS?

Yes. I think in this profile, what we showed was data in the third line. And obviously, we know that patients do progressively less well as you go further and further through the lines of therapy. And so we think it compares very favorably at 44% with what we've seen with other CDK4/6 combinations, which are in the 20s. So we have a kind of delta of 20% in the third line. And we think that delta should carry forward into the front line. But the most important thing here in these third-line patients is to look at the tolerability profile.

The patients in the frontline will have to be on this triplet for potentially 2, 3, 4 years. And so we know from the Roche INAVO120 trial that adding a PI3K inhibitor in the frontline is additive. They've shown both PFS and overall survival benefit. The challenge in the front line is all tolerability and stacking 3 drugs together as they do with a nonselective PI3K inhibitor then with a CDK4/6 inhibitor, which obviously comes with its own liabilities, can prove a very challenging profile for a patient to take long term.

And so what we saw in the third line was very exciting on the efficacy front, but more exciting on the safety front in that we showed a very nice profile combining Pfizer selective CDK4 dialing out some of the CDK6-related toxicities and then a mutant selective PI3K inhibitor, obviously dialing out some of the wild-type PI3K toxicities. And so we believe this selective-selective profile in the frontline should provide a tolerability profile that would allow patients to stay on therapy for multiple years.

And do you plan to report PFS in the advanced patients in the first half update?

Can you take that one, Pete?

Yes. That will be dependent on data maturity at that time from the dose-finding portion of the study, which was what we reported on a month or so ago. It is possible that we could reach a level of maturity in those patients to start to make a point estimate of median PFS. We've also moved into expansion arms at the potential Phase III dose of 150 milligrams of zovega with 300 milligrams of atirmo. Those arms are enrolling well now and -- but likely wouldn't be mature enough to start to be able to estimate PFS.

And what do you see as the bar here in advanced patients?

Yes. I think we know that our doublet in second-line patients is about 11 months median PFS. These patients are a bit more later line than that. So I think if we started to see a median PFS in our unoptimized dose finding, probably starting to match in and around the doublet, that would be a good sign that as we moved into healthier earlier line patients, that the delta that Sanjiv was talking about would continue to play out.

Great. Could you speak to your confidence in the dose work with zovega here, noting the concentration increases when you've combined?

Yes. So we're using in our go-forward work in the frontline, a dose of 150 milligrams twice a day. And that's compared to our second-line dose of 400 milligrams BID. And so what we're seeing is a drug that increases the concentration of zovega when it's -- but that kind of is a relatively flat relationship. And so at all the doses of atirmo that we've used, we're seeing pretty similar exposure with 150 milligrams BID of zovega. We feel pretty confident that we can go forward with that dose.

And how is your use of atirmo a differentiator versus other frontline regimens, especially Eli Lilly's? And how do you expect atirmociclib development to play out? And what gives you confidence in zovega's ability to compete here?

Yes. I think just on the Lilly compound, it's hard to say. We see our profile is differentiated. They showed data at last week's ASCO meeting. Their doublet response rate still sits in the 30s. As you saw from our disclosure in the year, we're showing response rates in the 40s. And obviously, we've shown PFS at 11 months. We haven't seen the PFS yet reported for that combination from Lilly, even though they have, I think, sufficient follow-up now. And then in the triplet combinations, obviously, we've shown response rate in the 40s. And I think there, again, we were in the 20s. So I think there's still a lot of unknown. It could be follow-up, it could be dose, but we haven't seen the Lilly compound really to be able to get close to us at the moment.

And then on the safety side, I think it's been noted, it does have some -- what we believe to be drug-specific liabilities around the LFT abnormalities that they've seen at the doses. And so I think unknown yet whether the Lilly compound will be competitive.

In terms of the 2 different approaches to take here, I think as we've talked about a little bit earlier, our approach is to use a selective-selective combination. So selective mutant selective compound as we'll both use, both Lilly and Relay. But on the other side, to also use a selective CDK4 because we believe that, that will provide a more tolerable profile -- single piece of toxicity because these patients are going to be on therapy for multiple years versus Lilly's approach of using the CDK4/6 inhibitors. Obviously, they are the owners of abemaciclib and have a vested interest in its use. And so we think a selective-selective combination will be much more competitive in allowing a tolerable profile for multiple years versus the approach that Lilly could take with a CDK4/6 plus mutant selective.

Great. And is your agreement with Pfizer for PI3K regimens exclusive?

Yes. So the -- there is a period of time in which we are exclusive to each other for starting Phase III trials with other -- with us with another CDK4 with them with another selective PI3K-alpha inhibitor. And it gives us both time to get the trial up and going and largely make it so that neither party can move into another Phase III in the very near term with another selective molecule.

And Pete, you do plan to initiate the Phase III frontline study in early 2027. So what are the areas you need to align with on the FDA prior to starting the study?

Yes, it's really just traditional blocking and tackling here. We need to bring -- the conversation will -- as all of these do will revolve around dose, and then the trial design is pretty straight down the fairway. We're largely mirroring the ongoing FourLight-3 study that Pfizer has ongoing in the frontline endocrine-sensitive population. So I don't think there would be any real questions around that.

And then because we are going to be doing zovega plus atirmo plus AI versus CDK4/6 of choice plus AI, we will likely need a contribution to parts arm. And whether we handle that by adding a small arm of atirmo and AI into the study or we leverage the FourLight-3 study and do a referencing of that data of the PIK3CA mutant population. I think that will be another question we iron out with the regulators.

Great. And we're fresh off a slew of medical meetings here where we saw Celcuity's data, which showed approximately an 11-month PFS for both the doublet and triplet. Notably, just speak to your updated thoughts on zovega's commercial positioning in the context of that.

Yes. I mean I think there's been a lot of investor interest in the profile that Celcuity would show at ASCO and its implications on the commercial profile for zovega. As you know, zovega is an oral compound and the Celcuity compound, gedatolisib has a weekly IV regimen. And so this is really going to come down to -- we all know that an oral regimen in this setting, both in the second line and then especially in the front line is going to be preferable. And so the real question was just how much better was the Celcuity data needed to be to justify a weekly IV regimen.

Our market research has shown anything above 15 to 18 months' worth of PFS was what they were going to need to show to get usage in a world where there is an oral therapy with an 11-month PFS. And so the bar is relatively high. These patients do not want to come in once a week and spend the majority of the day at an IV infusion center. And many health systems across the world are just not set up for it, especially if there's an oral.

And so when they came in with their 11-month PFS, it's a great trial result for patients because obviously, the 5.5 months' worth of alpelisib that they were up against shows a true like paradigm shift for patients. I think the conclusion, I think, that both us and investors drew was that once there was an oral compound in the market with 11 months' worth of PFS, it was going to be tough for an IV regimen to be competitive. No patient is going to want to come in once a week for an IV regimen when they can have an oral therapy, especially if both of them provide exactly the same PFS. So I think it's kind of dispelled the competitive threat for us. And I think we're just full speed ahead now recruiting the trial and getting this to patients.

And why do you think the triplet and doublet performed in line here? And what gives you confidence that the addition of atirmo will result in a differentiated triplet for zovega?

Yes, it was a slightly unusual finding. And I think the better conclusion here is to say that their triplet performed as expected. It was almost double what the alpelisib arm doublet produced. The doublet arm, which they put in there, gedatolisib plus fulvestrant was really there just for a contribution of components. And it wasn't adequately powered to show true differentiation.

And so I think if that arm had been slightly more robust, I think we probably would have seen the same separation that we saw in the wild type where they show, I think, 1.5 to 2 months difference between the doublet and the triplet arm. I think we would have seen the same. So I do think we feel that there is a meaningful benefit for adding the atirmo. We see it in our own data where we see in late-line patients, our triplet doing very well versus our doublet. And we think gedatolisib probably would have shown that if it was adequately powered.

From a strategy standpoint, you're sitting with these 2 verticals, this cancer vertical with -- particularly with breast cancer here and the rare disease opportunity. Do you plan to take breast cancer forward independently or partner?

I think we plan to take everything forward. I mean I think we sit on 3 very robust commercial opportunities now, all we believe to be blockbuster in potential. I think we've seen companies like Menarini show that you can commercialize successfully in breast cancer as a relatively small company. And on the vascular anomaly side, it lends itself perfectly to a company of our size, given there's only a small number of vascular anomaly centers in each large country.

So I think we have the ability now to execute. We obviously raised capital a few weeks ago, and as we have a robust balance sheet that should take us well into 2029, executing all of these opportunities, well past the top line data for vascular anomalies and the second-line breast cancer trial. So we feel pretty confident to go forward and execute.

On the vascular anomalies front, you reported strong first data there, including a 100% volumetric response rate at the 300-milligram BID dose, and you selected both the 300-milligram BID and 400-milligram QD doses for expansion. Could you discuss, I guess, firstly, why you believe you're not leaving efficacy on the table with the 300-milligram BID being the highest dose?

So I mean, given the fact we had 100% response rate at the 300 milligrams, we do believe that we have maxed out the efficacy. At the 400-milligram dose, what we saw was a drop-off in efficacy where we saw a 57% response rate. And that really was because of 2 reasons. One, we started to see some of the tolerability challenges that come with a little bit of hypoglycemia, diarrhea, some nausea that doctors were very keen to dose reduce rapidly. And the reason why they were trying to do this rapidly is because they could see the 300-milligram dose was very effective. And so we do think that in the end, at 300 milligrams, we've maxed out the dose.

Now the unknown here is, could you max out the efficacy at an even lower dose than 300 milligrams. And that's why in the expansion cohorts, we'll open up the equivalent dose to 200 milligrams BID, which is the 400 milligrams once a day to test the hypothesis that can you even max that efficacy to a lower dose than the 300 milligrams.

And just further rationale there on taking the 400-milligram QD instead of 200-milligram BID?

We have a very flat PK curve and a long half-life. And so we have tested a lot of different doses in our oncology work. And what we see from that is 400 milligrams once a day has exactly the same PK profile as 200 milligrams twice a day. And so for just the added incremental benefit of having once-daily dosing will take forward 400 milligrams once a day versus the 200 milligrams twice a day.

And mechanistically, why do you believe zovega is able to benefit patients who have failed alpelisib?

Yes. It really comes down to the same logic that ended up showing meaningful differentiation in breast cancer. For the first time, we can actually see what a mutant selective molecule does in this patient population. We have a meaningful window of selectivity between wild-type and mutant and nearly biologically inert against all the other PI3K family members in the other -- the rest of the kinases.

So I think what you're seeing here is that manifestation coming out in the clinical data and seeing for the first time, what does this level of mutant selectivity accomplish for these patients. And it's really exciting because the treatment goal here is to treat -- start treating these patients as early in childhood as you can to -- because these lesions, these noncancerous lesions grow with the child.

And so if you can intervene as early in childhood as possible, you're going to hopefully blunt the growth of that lesion and therefore, alleviate a lot of the comorbidities that these patients have to deal with throughout their lives. And we can't do that today with alpelisib and sirolimus. The way these molecules are being used because of their toxicity profile is short bursts of treatment, maybe for 12 to 18 months. And then the tolerability profile is such that these patients have to come off those treatments, and they have to try to let their health reconstitute and maybe they can go back on or switch on and off.

And here, we're excited because we think the data that we showed a couple of weeks ago really speak to the potential of treating these patients as a chronic disease and as early in childhood as possible and keep them on drug for their life. So hopefully, they can -- the burden of this disease can become as less as we can possibly make it.

What does the trial for accelerated approval look like here? And maybe touch on the spend associated with the program as well as the size of the opportunity you're addressing?

Yes. So we haven't had a regulatory interaction yet. So that the FDA will be the ultimate arbiter of what that path looks like. But we intend to do that this year and come back before the end of the year to speak about the outcome of that interaction. But what we can look at is the precedence here for alpelisib. Alpelisib was approved off of 37 patients of a retrospective chart review of those patients being treated under compassionate use, really speaks to the unmet medical need here in this patient population. And they unfortunately failed their first confirmatory study, EPIC-T2, but their current confirmatory study is a single-arm study that has an N of 104 patients, and that's for full approval.

So if you think about the spectrum of potential obligations on our side for the N to be able to get to accelerated approval, it's reasonable to think it's somewhere in between those 2 bookends, 37 and 104 patients. And the regulatory endpoint here is to -- in that volumetric response rate to exclude 15% in the lower bound of the confidence interval. Our data today excludes the 15% to the lower bound of the confidence interval. So I think when we go to regulators later this year, that -- the nature of that conversation will be what is the end that they would like to see to give us all confidence that these results can be recapitulated, and then also, what's the safety database that would need to be seen to be able to facilitate accelerated approval.

Fortunately, we have about 500 patients' worth of cancer data here to help support the understanding of the safety profile of zovega. So that should be a bit of a tailwind for us in those conversations. The size of this opportunity is very, very large. There's 170,000 of these patients from a prevalence standpoint in the United States today with PIK3CA mutations. And we don't believe all those patients will ultimately seek chronic systemic therapy. But if you get down to the 3 current subtypes of focus for us, which is PROS, PIK3CA-related overgrowth spectrum, lymphatic malformations and venous malformations, there's 100,000 of that 170,000 inside of those 3 subtypes. And then we believe about 25% of those would seek chronic systemic therapy for the disease. And so that represents about 25,000 patients.

And the simple math shows that at the alpelisib-like pricing of $400,000 a year, every 2,000 to 3,000 patients would be $1 billion in peak sales if you can truly provide a therapy in which these patients can stay on chronically. And certainly, if accelerated approval pathway is available to us, which we think it is, the development dollars here are not that large. And if you play that forward to a confirmatory study, which if the alpelisib precedent holds, that what we likely think would be the solution there is to just expand our current expansions further, enroll more patients, and that can act as a confirmatory study. So the total development dollars in comparison to the opportunity size is very small.

And do you believe you could cover both PROS and lymphatic malformation within the same study in application?

That's going to be what we proposed. We believe the biology of the disease here is the exact same across these 2 disease. There's different levels of severity. Obviously, the syndromic presentation brings with it a different level of disease burden than the more severe lymphatic malformation patients only. But in our early data sets, it's very encouraging to see that of the 4 patients that were evaluable for efficacy with lymphatic malformations, 3 out of those 4 had deep durable volumetric reductions for a 75% volumetric response rate so far.

We need to grow that in a little bit further and see that signal stay consistent. But we believe it makes a lot of sense to keep these 2 subtypes pooled, especially because given the frequency of PIK3CA mutations in these 2 subtypes, obviously, 100% in PROS, 80-plus percent in lymphatic malformations, these patients can be diagnosed clinically and treatment can be directed without a diagnostic test in these 2 settings. And so we believe that the consistency of the disease biology, the consistency of our data and the fact that these patients can be diagnosed clinically is all very good rationale to keep them pooled together, but we'll see what the agency says.

And in pediatric patients, how are you expecting growth inhibition to play out, which has been an issue with alpelisib?

We'll measure all the relevant metrics to make sure we track that appropriately. It's too early to tell today and especially because our initial data was in adults and adolescents, so 12 years and up. And we just started the 6- to 11-year-old dose escalation earlier this year. So it's going to take us some time to really track that signal with zovega. But we do believe that, that liability is coming from wild-type inhibition. And again, here is where our selectivity profile should play out in our favor to hopefully minimize or not see any of that growth curve inhibition.

And just describe to us the treatment pathway, including the numbers of centers of excellence and patients per center and where genetic testing stands today?

Yes. So this is an area where we need to continue to do more work as a field. You have to remember that we've only been using systemic therapies in these patients for about the past 10 years or so. And only about the past 15 years have these patients been starting to be diagnosed genetically. And so there's a lot more to learn and figure out. And fortunately, with the data that we are presenting and generating, we'll be able to ask these questions a bit more clearly within the community. But today, we believe that most of the moderate-to-severe patients are making their way to these centers of excellence or the vast anomaly centers.

They tend to overlap with the children's hospitals in both the U.S. and Europe. So probably the vast majority of them are making their way to these 30 to 40 centers in the U.S., probably similar footprint in Europe. But -- and so we think that it is a very attractive commercial model from that vantage point too, to where you can get to a lot of these patients quite efficiently. And I think we'll continue to help raise awareness and push more genetic testing. But again, the PROS and LM patients are -- fortunately can just be diagnosed clinically and don't require a genetic test to direct treatment.

Can you discuss future directions for the Dynamo platform here and your business development strategy?

Yes. So obviously, we've been very successful over the last decade, bringing first-in-class and best-in-class molecules forward. The platform is directed against targets that we don't think others could tackle. And so it's really simple as that. We see lots of talk around the commoditization of research or the offshoring of research. And so that really leaves us with a conundrum of what should you focus on. And so I think what we try to look for is clinically validated targets where the conventional approaches have been futile and that we can deploy our people and platform against trying to solve them. And so they sit across both rare disease and oncology. But there's no point us trying to work on the second-in-class molecule. We know that others are just going to be much better and much more efficient than us at doing that.

The final question here. There's so much focus on AI development in health care. And clearly, you were one of the first here with using this machine-based platform. Is that really the secret sauce that has enabled you to get the products to these stages and to show the efficacy rates that you've shown versus competitors? And how do you see leveraging that on the forward with the next targets?

Yes. Obviously, we've been at this 10 years. We're one of the first companies to use computational approaches in the discovery process. And I think what we've learned is it's not a kind of silver bullet. The ability to make a new medicine is time-consuming, and has many, many thousands of steps in it. And those steps involve machines, people, data, algorithms. And each one of them, the competition approaches can make incrementally better.

And so if you stack the kind of 1,000 processes together, you make the entire process incrementally better. And I think that's what we found. And obviously, we see a lot of companies over the last 10 years kind of launched themselves with our platform is -- can you use the competition approach to revolutionize the entire process? We haven't seen that. We see it makes everything incrementally better and then overall, the whole thing gets better. But you still need humans, you still need labs and you still need the kind of intuition that goes on with kind of long-term drug discovery. And so it isn't going to solve all your problems basically, but it will incrementally solve some of them.

Great. Well, with that, thank you so much, Sanjiv. Thank you, Pete.

All right. Thank you. Thanks for the invitation.

Thank you.

This is our New York City Healthcare Conference, and it's a pleasure to see you all. We have the Relay management team, fresh from their flights from ASCO as I am as well. So they are full of energy and really can talk for days. But no, I'm very grateful to have the team here, especially as we're starting to see, I think, the HR-positive breast cancer space play out and understanding where there's different therapeutic opportunities. So I'm actually quite excited to have this conversation with you. Sanjiv, why don't I hand it off to you to give some intro remarks, and we'll get going.

Yes, excellent. Thanks, Akash, for the invitation, and thanks to Jefferies. We had a very productive year so far. We've shown three data sets in three large commercial opportunities. The first data set we showed was in second-line hormone receptor positive, HER2-negative breast cancer with ESMO TAT. We showed in our pivotal trial dose of 400 milligrams BID, an 11-month PFS with an oral regimen, which we're going up against capivasertib with its 5.5 months worth of PFS. So we feel very confident as we run our pivotal trial that it will be a success and that will offer patients a paradigm shift in treatment. The second data set we shared was in looking at our frontline hormone receptor positive HER2-negative breast cancer trial that we'd like to run, where we announced that we will use a novel selective, selective regimen with Zovega and atirmociclib, Pfizer's selective CDK4. And we showed what we believe to be a very clean safety profile with great efficacy in third-line patients, and we're excited to bring that into the frontline as rapidly as we possibly can.

And then a couple of weeks ago at the ISSVA conference in Philadelphia, which is the largest conference for vascular anomalies, we showed what we believe to be paradigm shifting data for these patients that are chronically underserved, where we showed a 60% volumetric rate of response in a field where the traditional characteristic has been somewhere between 11% and 20% -- and so we're kind of full speed ahead now to try and execute all three pivotal trials as rapidly as we can and get these medicines to patients.

Understood. I'm going to start with the breast cancer opportunity, though I want to acknowledge, I think the orphan story is just as interesting. And frankly, I think it's had an even bigger impact in terms of your near-term stock performance, but we're fresh from ASCO. And I think we've got two really, really interesting data sets. We have finally the persevERA data. And then also we got data from Celcuity, which I think was really interesting. I wanted to just start off a bit with Celcuity. And like, look, this is -- it does look like it was an impressive PFS benefit. No additive benefit on the triplet versus the doublet. And then 11 months, when we think about the absolute benefit on PFS was kind of in the range that your team had shown. I'd love to get your take in terms of how you look at the Celcuity data. What did you learn about your own product and how it gets positioned post that data set?

Yes. I mean, obviously, as those of you who have been close to us, it has definitely been a question that we've been asking and then trying to answer for the last years. It's great to actually have the data and the experiment run in the public domain. And for those of you who didn't follow it, what they showed with their triplet regimen, which is a weekly IV regimen of gedatolisib with palbociclib and endocrine therapy, an 11-month PFS, which is an impressive result, given the standard of care there on the control arm was 5.5 months. And so we think it will be a step forward for patients. I think as we -- as it pertains to us, I think the question was, would it be a commercial headwind if we were to then launch into that market. And I think our belief is having shown now two large cohorts of patients showing with our oral regimen an 11-month PFS that it will not and that we should be able to take share from that market that Celcuity will build very rapidly as patients are going to prefer an oral regimen.

I think the concern was that they were going to show 15, 18 months, whatever the numbers that were being banded around, which obviously would have been more challenging for us to compete against. So I think in terms of just how we feel about the commercial opportunity, I think we feel very confident. In terms of their triplet versus doublet, yes, it was an interesting result to see that in the end, there was potentially perceived no benefit for having a CDK4/6. Now we don't know if that's actually the conclusion. As you know, their doublet arm was there really only for a contribution of components. And so it's hard to kind of statistically be accurate that they actually have no benefit. I think our belief is what we saw in the wild-type arm is probably play out in the mutant arm and that there will probably be a little bit of benefit 1 to 2 months for having the CDK4/6 on board. But really now, it does create some confusion for prescribers, do they prescribe the triplet or the doublet given the data that has been seen there.

Understood. And look, is going to be -- this is a tricky question, and I'm struggling with this as well. How do you -- Forget patients. How about you? I mean you have a partnership with Pfizer with Atirmo and really interesting early data, but early data and a potential to move this into kind of a first-line setting. So has your confidence on kind of the AtirmoPIK3-alpha combo changed at all as a result of the Celcuity? And if not, why? Why does that data point not necessarily -- read across to the trials you're running?

Yes. No, it really hasn't. As we're moving into a frontline setting, we have proof today from Roche's INAVO120 trial of inavolisib in an endocrine-resistant patient population, frontline treated, where we saw doubling of PFS and an OS benefit for the triplet regimen of inavolisib with palbo and fulvestrant versus fulvestrant plus palbo. So we know from that experiment that adding a PI3K inhibitor on top of the CDK4/6 inhibitor in frontline patients can have a very meaningful efficacy benefit.

Now one of the shortcomings of that trial was it was highly contrived, run in a heavily selected patient population to avoid some of the metabolic liabilities of inavolisib. And those data have, therefore, been challenging for prescribers to translate into real-world usage, and we've seen that as a headwind to uptake of inavolisib. But I think when we look at that, when we look at the fact that in endocrine-sensitive patients in frontline, patients with a PIK3CA mutation do not receive the full benefit of the CDK4/6 regimens and perform slightly poor on PFS, about a year poor than patients who are PI3K wild-type, tell us that as you get into early line breast cancer, being able to hit both the endocrine node, being able to hit the PI3K node and then being able to complement that with CDK4 inhibition is really going to be optimal for maximizing outcomes for patients.

Understood. And I do want to kind of stratify that a bit because if I remember correctly, Roche has run, I think endocrine-resistant, that triplet frontline, the sensitive population is still coming out. How do you -- and again, this will be a nice bridge to PersevERA and this topic of endocrine sensitivity. How do you expect the benefit of a PIK3 outlet triplet in a resistant versus sensitive population? I mean it sounds like you also expect the Roche trial in the sensitive population despite the safety liabilities to still read out positively.

If patients can stay on inavolisib, that's going to be the key question because the resistant, they only need to stay 15 months. When you get into the sensitive population, the statistics of the INAVO123 trial are such that the assumption for performance of the control arm, the CDK4/6 doublet in that trial is 20 months. And they're looking for about 30 months in the triplet treated arm. So they'd have to keep patients on a fairly challenging regimen for about twice as long as they did in the INAVO120 endocrine-resistant population.

Understood.

I think part of your question was how do we have confidence that in the endocrine-sensitive patient population that addiction to the pathway holds that we saw in the endocrine-resistant population. And that's probably best explained if you look at the MONALEESA-2 results of ribo plus AI in that study, we saw a good robust subset of PIK3CA mutant versus PIK3CA wild-type, and there's still a clear delta in performance there. The mutant population was about 19.5 months in that study and the wild-type population was closer to 30 months, clearly articulating that there's addiction to the PI3K-alpha mutant pathway in that patient population.

That's a great point. Now heading to the other data set, which was PersevERA. And it's interesting. I think certainly on the pharma investor side, there's this idea that frontline adjuvant for SERD is this kind of $20 billion to even $40 billion opportunity. I can tell you, I'm very much skeptical on that. I don't think we saw necessarily that same uptake with the CDK 4/6 as with the NATALEE trial.

Again, major revenue contributor, but not to that level. And when I look at the Roche monotherapy data, I say, okay, maybe the Stage 2 -- like there's a subset of patients as a monotherapy, okay, that's a group. That's not $20 billion, that's $3 billion to $5 billion, right? Like substantial, but not what I think people talk about. And to me, the uptake is really the combo. But I'd love to get your take now that you've seen the persevERA data, let's start with the first biological question. Do you feel like SERDs work "in first line? Or do they work in a subset population?

So there's no question they work. What the question is, do they work better than our existing drugs, aromatase inhibitors. And if you go back -- so you mentioned the lidERA adjuvant trial that was monotherapy where clearly the SERD was superior to the available investigator's choice endocrine therapy in a monotherapy setting, right, where you're really just targeting the estrogen receptor node and where slightly better targeting in that case translated into a meaningful difference in terms of clinical outcomes.

The challenge when you get into the combination setting is now you've got your CDK4/6 inhibitor on board, which we know from the registrational trials of the CDK4/6 inhibitors versus endocrine therapy monotherapy like transformed the outcome for patients in terms of PFS. That CDK4/6 inhibitor is driving a lot of the benefit. And I think it's challenging in that setting to have slightly better endocrine inhibition and have that translate into a meaningful difference on PFS, which is what we've seen in persevERA. Now there's some suggestion when you look at the PFS curves that there may be some late separation of the curves that could be accounted for by suppressing the emergence of ESR1 mutations with the oral SERD as opposed to the aromatase inhibitor, which would then place more into the SERENA-6 type approach where you actually use the emergence of ESR1 mutation potentially as the trigger for identifying patients who would receive the more broadly potent SERD or SERD.

Understood. So that outcome did not go, let's say, AstraZeneca's planned, but understood. Now when we think about the CDK4 because again, these are all external data sets, but directly apply to your team. We're starting to see frontline data, neoadjuvant data with atirmo. And there's also that even second-line data that they've only top line. But to me, the question is really, can we get separation faster? And then obviously, the safety benefit could that late emergence tail occur as well. What have you seen so far with CDK4 that makes your team confident that this would have a benefit over, let's say, a CDK4 biased drug like Verzenio, right? Again, because I know there's going to be extrapolation, but what data sets are you seeing that's driving your confidence on this partnership?

So, I think the main data is the data that we've generated. We know that in the frontline, tolerability is everything. These patients are going to need to be on drug for multiple years. And we know from the INAVO120 trial that the efficacy is there of adding a PI3K inhibitor to a CDK4/6 backbone. The challenge is just adding a CDK4/6 backbone to a PI3K inhibitor leads to tolerability challenges and patients just can't keep on drug and the dose intensity declines and that what's potentially going to lead to these trials not doing as well as they could. And so our data in these late-line patients, third line on average with both the selective profiles of a CDK4 selective and a PI3K selective has shown a very tolerable safe profile. And so I think in these late-line patients, if we're seeing what we're seeing, I think we feel great that when we bring this to the frontline, you will see both the dose intensity and the longevity of these patients on treatment, which will then inevitably lead to the efficacy. And I think that's what's given us confidence to put this in the front line and run this trial.

So it's really more about how many patients are still on drug right now rather than just looking at the initial response rate. Exactly. Understood. And there's a possibility -- I mean, if you look at those curves, you're at 44, there's a few patients there, which are pretty borderline. And I mean, this is a session we've even had on your vascular malformations where like we haven't seen any patient not responding to the drug. They just may have not been on the drug long enough.

When you think about how that early data set with Atirmo will evolve over time, is there a potential for responses to deepen? Is there any historical precedents we can maybe point to? And in terms of data disclosure on Relay side, when can we expect another update with the Atirmo combo?

Maybe Don can take on the first one and Pete the second.

Yes. Yes. So with regard to responses deepening, I think there could be a possibility of responses deepening. We -- the data we presented or that we disclosed in April was across all of the doses that we had tested for the triplet at or below our recommended Phase II dose. So a number of those patients were treated at dose levels below what we would ultimately take into Phase III. And our experience in the fulvestrant doublet was that at some of the lower doses were active, but sometimes the responses came later. So I think in this patient population with the triplet, we'll just have to continue following the patients to see if some of those near responses on subsequent scans are able to deepen further.

And on the disclosure, we've guided to showing additional triplet data in the first half of next year. I think what to expect with that update would be obviously more follow-up from the dose-finding portion of the study that we've already reported on. But we've also since moved into dose expansions as we've talked about in that disclosure at the recommended -- the potential recommended Phase III of 150 milligrams zovega, 300 atirmo. So we'll show at that point in time, the data we have from some of those expansion cohorts also. So it would be a good level of maturity from the dose-finding data and then early data from the expansion cohorts.

Understood. I mean, could that update next year, could we start seeing signs on PFS? And like, I mean, is that still going to be mostly just focused on duration of response, focused on ORR and -- but are we going to get enough maturity where you could start to see differentiation versus what you would expect in kind of a first-line standard of care setting?

Yes. If we continue to just isolate the dose-finding portion of the study, that portion would have over a year follow-up at that point in time. And so it's possible to potentially start to estimate a median PFS point estimate.

But with the caveat that we're in median third-line patients right now. So there'd have to be a correction. It's...

Understood. Another tough question in terms of how to design these studies. But again, I selfishly ask because we're trying to figure it out. When you think about powering and your expectations on, let's just say, CDK4/6 plus AI in a frontline population, I mean, if we think about it in the broader population, okay, it can be about 25 months, but we've seen that PIK3-alpha mutated patients seem to have worse prognostic outcomes where PFS might actually be more like 18 months. What's the right assumption in terms of the comparator arm for your Phase III design? How do you power that?

Yes. So we have a few different data points to look at. The one I just cited recently was the MONALEESA-2 study where they did good sub-characterization of the mutated population. And in that study, you saw 19.5 months of median PFS. Additionally, in the INAVO123 study, they have posted their statistical plan on CTIS, and they are assuming a 20-month performance in the control arm for that study. And you would imagine they have pretty good real-world information and the fact that they own and Foundation Medicine. And then the most recent data set we could look at a little bit too is in the VIKTORIA-1 data that was just reported. They did give the median duration of treatment on the -- of the first-line metastatic therapy that those patients were just on. And you saw somewhere between 17 and 22 months of median PFS there too -- or median duration of treatment not a perfect one, but again, all as you're saying, kind of triangulating back to about that 19, 20 months.

Okay. And when you think about powering for a Phase III to show kind of a clinically meaningful effect, I mean, we're thinking PFS? Are we thinking OS? Like what are we thinking on hazard ratios for both of those endpoints?

Yes. I think we'll fall short of giving the precise hazard ratios, but I think the way to think about it is we've done some primary market research asking physicians what would be a clinically meaningful improvement against that 20 months in those patients. And generally, the answer is about 6 months. And so I think from a -- obviously, the primary endpoint in these studies is progression-free survival. We will certainly be conservative in how we power this to make sure we have enough powering for OS as a key secondary. And knowing that the threshold for clinically meaningful improvement is about 6 months, all these things will go into our statistical assumptions.

Understood. I mean -- but bottom line, I mean, if in a frontline population that's not PIK3-alpha mutated, it's 25 months, you're saying, look, we're going to fix that problem, and then let's hope CDK4 can add something on beyond. But that's -- is that the right way to think about this?

Yes. Our hypothesis is not at all based upon Atirmo needing to add anything new or different on the efficacy side, simply that it needs to be safer. And I think the data we have to date is very supportive of that.

Now as we think about -- let's go maybe on to the vascular malformation side because I think, again, it's an incredibly important part of that story. And Sanjiv, I want you to talk about this in terms of ROI because I don't think that gets talked about enough. Help us understand how much additional capital will you really need to get this to the commercial stage onto the market when we think about PROS versus the broader kind of population, what is that -- what's the capital cost? What's the time lines as your base case right now?

I mean I think in contrast to breast cancer, which is a well-established market where we see capivasertib closing on $1 billion of sales and the trials that we're running both in the second and the front line, it's a very kind of definite kind of path to market. You kind of know what the trials are, they are 500 patients in the second line. You know the kinds of sales forces that are out there in the U.S. So I think that world is relatively clear. This one is a little bit more opaque. But I think what we can tell you is the historical trials here have been relatively small. Novartis retrospective chart review of 37 patients was what they got accelerated approval on and their confirmatory trial was just over 100 patients. And so obviously, we showed 32 patients worth of data last week on the safety, 20 on the efficacy, and we're continuing to enroll at a rapid rate.

So to answer your question, it could be a very rapid pathway to an accelerated approval and then the confirmatory trials are in the zone of 100 patients at the upper end here. So this could move both very rapidly and with relatively modest dollars. And then on the commercial build, it's relatively similar. There are only a small number of vascular anomaly centers in each kind of large European country, and there's probably somewhere between 20 to 30 in the U.S. And so you could probably build yourself a relatively modest commercial presence and footprint here to generate significant revenues. And so all in all, on one side, it seems very attractive in that it's very small numbers of dollars here. The kind of unknown on this side, obviously, is how actually we build this market because it's a nascent market.

It's not a kind of very kind of well-developed market as we see on the breast cancer. And so the dynamics of both are attractive to us. One is well-trodden pathway, very clear, just produce the data, get share of voice, generate market share. And the other one is a huge potential for us with relatively modest dollars. But it may take a little bit of time to develop that to fruition.

Understood. I mean just to put a finer point on that, would it be fair to say the amount of additional capital needed to get from the clinical trials, we're talking about $50 million.

Yes I think that's a fair estimate. It's kind of between $50 million and $100 million worth of investment.

$50 million to $100 million terms of running the clinical trials. And then in terms of sales force build, I mean, if we look at historical precedents, we're thinking about maybe something $100 million in terms of annualized run rate. I mean...

Yes, probably less than that.

Even less than that. Interesting. Now you gave really encouraging early data. There's going to be more cohorts that enrolled. There's going to be more subtypes and then there's also an FDA discussion to really be had. So let's start with kind of where are you in terms of your dialogue with the FDA? Have you had initial conversation after this top line press release? Any updates there?

Yes. So we're currently guiding that we will have a regulatory interaction this year and come back to the Street later this year with the outcome -- output from that interaction. The goal there is to have a conversation around what would be required -- if accelerated approval is available to us, which we think it is and then what would be required to accomplish that. And the two key questions there are going to be around dose and the end required to support accelerated approval. And our going-in assumption there will be that we would pull together both PROS and at least pull together PROS and lymphatic malformations. And so that's where we have a number of topics to try to get through with them, and we would come back and report the outcome of that later this year.

Understood. And can you -- I know -- what I think would be very helpful is, can you kind of prospectively lay out the outcomes because there seems to be kind of three. One is, all right, we have data in PROS. That's what we're going to let you do the expansion in, which seems to be, I think, a bit restrictive. There's the one that I think your team hints is probably most likely, which is we'll go after at least the two subtypes to the two biggest markets, and we'll say, hey, we have enough -- we show consistency of data in both of these malformations and let us run a confirmatory trial. There and then in VM, we may have to run another separate study. And then there's the other outcome where, look, you can run a trial in all three. First of all, is that the right framework? And then when you set investor expectations, what's the reasonable base case right now?

Yes. So the one thing you -- a key point you mentioned there is consistency of data across these subtypes, which is something that the agency will look at. Today, we have since we've reported on four patients in lymphatic malformations, three out of the four patients responded, two of our deepest responses are coming from the lymphatic malformation patients. So I would say today, we're seeing signs of consistency of data between the PROS and the LM patients. We're not seeing a difference in safety profile across those two patient populations. And so we need to now grow that in lymphatic malformations and that consistency has to continue to stay there. But we would -- there's no drug approved systemic therapy approved for lymphatic malformations, and we believe it to be a severe unmet medical need. And so yes, our going-in hypothesis is a base case of keeping those two pooled together.

Now if you landed in just an initial label in PROS and you had to separate our lymphatic malformations into a separate single-arm study for accelerated approval, perfectly fine outcome also. Alpelisib with just a PROS label and all the shortcomings of that drug, it probably treated on the neighborhood of about 5,000 patients or so since they've received accelerated approval. We know in this context, if you could have an agent that could really be used chronically, every 2,000 to 3,000 patients would be about $1 billion in peak sales. So if that's where we land for an initial label and we do sequential close-in sNDAs for lymphatic malformations and venous malformations, that's a perfectly fine outcome also.

Understood. Now I know you have, I think, four -- you had eight LM patients, four that were clinically evaluable. So we may get more data on that subtype as well, and that might feed into the discussion of kind of consistency. Can you kind of remind us what dose those other four patients are going to be on? And really how -- what's the data cut we could get with that other subgroup by the end of the year?

Yes, they're spread across the dose randomization doses, 300 -- 400, 300 and 100. And so those eight patients total from the dose randomization portion of the study, so 25% of the 32 patients fully enrolled. And then we've also started expansions. We started expansions a couple of months ago -- a month or so ago at this point. And so we'll also have patients coming in on the expansion cohorts that could be included in an update later this year. And so we could have more than those eight lymphatic.

Okay. Understood. So it sounds like more updates to come there. Maybe just, Sanjiv, this is kind of the question for you. You're in this kind of unique perspective where, again, you have -- you talk about risk adjustment and spend and then TAM. And both are very, very attractive markets. But the derisking component on the breast cancer side is certainly going to take longer than maybe the orphan side. And you're also, to a certain extent, capital constrained. How do you think about -- I mean, we've seen external financing, royalty agreements, debt agreements. What other sources of capital do you think investors should be paying attention to when you kind of counterbalance both of these opportunities? And number two, is there an appetite to maybe spin off one of these indications entirely because you feel like that could actually be a value unlock?

Okay. So we'll take those slowly. So as you know, we ended the quarter with over $600 million of cash and then we just did a $300 million financing. So we have a pretty robust balance sheet. That should take us through to having top line data on both the breast cancer and the VAs pivotal trials. And we don't imagine there'll be too much difference in the timing of those two approvals. In terms of the frontline trial, we have now money to prosecute that all the way out to the readouts of these two. So in terms of just our ability in the near term to execute and generate meaningful value-creating data, we feel pretty confident. And then I think you're in a great situation because obviously, you sit on two registrational data sets. You sit halfway through a frontline trial. And so I think all of the tools that you just listed, debt, further equity, royalty financing, partnerships, all become available to us. And so I think we feel pretty confident now of our ability to get into the next decade and commercialize in all three indications.

Understood. I'm going to sneak in one more question. Post the ASCO data set, because, again, there's multiple CDK4 players, there's multiple people looking at what to do in HR-positive. Do you feel like there's been kind of this renewed interest externally of your drug in breast cancer specifically?

I think absolutely. If you go back kind of 18, 24 months, there's a lot of unanswered questions. PI3K inhibitors, were they going to be meaningful? What would happen to the oral SEDs? Would they dominate in the front line? What about CDK6 in PI3K-alpha mutated patients, CDK4, CDK4/6 retreatment, all of these things, I think, are starting to resolve themselves. And you come back to the simple basics of 40% of patients have a PI3K-alpha mutation. In those patients, the best way to treat them in any line, early, first-line, metastatic, second line is to have a PI3K-alpha inhibitor. So I think that is becoming clearly understood by everyone in the field, providers, strategics and biotechs.

Understood. On that note, I really do appreciate it. Thanks so much.

Good day, ladies and gentlemen. Welcome to Relay Therapeutics Vascular Anomalies Update Call. As a reminder, this conference call is being recorded. I would now like to introduce your host for today's conference, Mr. Pete Rahmer, Chief Corporate Development Officer at Relay Therapeutics. Sir, you may begin.

Thank you, operator, and good morning, everyone. Thanks for joining us. We are very excited to share our initial vascular anomalies data with you today. You can access the press release from today. The slides we are reviewing and a replay of this call by going to the Investor Relations section of our website.

As a reminder, during this call, we will make certain statements that are considered forward-looking statements within the meaning of the Private Securities Litigation Reform Act of 1995, including expressed or implied statements regarding our strategy, business plans and objectives, and expected therapeutic and clinical benefits of our product candidates, potential of our platform and our product animates progress, timing and execution of our clinical trials. Such forward-looking statements are not guarantees of future performance, and therefore, you should not put undue reliance upon them. These statements are subject to numerous risks and uncertainties that could cause actual results to differ materially from what we expect.

I refer you to our SEC filings and on our website for a discussion of risk factors. Forward-looking statements in this presentation speak only as of the original date of this presentation, we undertake no obligation to update or revise any of these statements.

I'm joined today by Sanjiv Patel, our CEO; and Don Bergstrom, our President of R&D. And now I would like to turn the call over to Sanjiv.

Thank you, Pete, for the introduction, and thank you for all of you joining the conference call listening. Today, we'll share further data on our lead program, Zovegalisib, Zovega for short, our pan-mutant selective PI3K alpha inhibitor, which has the potential to address 3 very large commercial opportunities. The first of these is second-line hormone receptor positive HER2-negative metastatic breast cancer whereas a company we're laser-focused on executing our pivotal trial and ReDiscover-2, which is recruiting patients globally as we speak. The second of these very large opportunities is first-line hormone receptor positive HER2-negative metastatic breast cancer where we plan to initiate subject to regulatory feedback, a frontline Phase III trial in early 2027 with a novel, novel selective-selective combination of Zovega and atirmociclib.

But today, our focus is on enhancing the initial data on the 1/3 of these very large commercial opportunities, vascular anomalies. While the data is early, we believe it shows a potential change that step change in efficacy with Zovega versus what is currently being used as the standard of care, alpelisib. Our initial clinical data shows a 60% volumetric response rate with Zovega versus the 10% to 30% that has been seen in other clinical trials and assessments with alpelisib. We saw the 300-milligram BID dose of Zovega, a 100% volumetric response rate. We believe this to show the full potential of what a mutant selective inhibitor can do in these vascular anomaly patients and believe this leaves very little room for improvement of our volumetric response rate for anyone that will come behind us.

At 100-milligram BID dose of Zovega, we saw a 29% volumetric response rate. We see all of these responses deepen over time, as you will see in the spider plot later, where we show that each patient's response deepened between 12 and 24 weeks. And post data cutoff, we saw is tangibly manifested as we saw an additional response from 100-milligram BID patient who converted from stable disease at the 12-week scan to an unconfirmed volumetric response at their 24-week scan. Taking this into account, our volumetric response rate overall across all doses become 65% and 43% at the 100-milligram BID dose. Based on all of this initial data, we believe we have a very favorable profile versus the current therapies. As we see robust initial response rates across different doses, different types of vascular anomaly, different mutation types, different levels of pretreatment and different baseline lesion volume sizes.

In addition to the volumetric reductions we saw, patients saw symptomatic benefit with 89% of the patients treated by their physicians were assessed to have a clinical improvement of their symptoms at week 12. And so as we said, this is a potential step change in efficacy with Zovega. All of this is seen at a level of tolerability that compares very favorably with the existing treatment option at alpelisib. Across all doses, we saw initial rates of grade 3 plus toxicities at Zovega 15% versus 70% for alpelisib. If we just look at combining the 100-milligram and the 300-milligram BID doses of Zovega, we see the common adverse events associated with wild-type PI3K alpha inhibition were low-grade, manageable and reversible. No rational stomatitis of any grade was observed no Grade 3 hypoglycemia or diarrhea was observed, and there have been no discontinuations to date due to AEs and all patients currently remain on therapy.

Given all of this, we're excited to open multiple Zovega expansion cohorts, and we're taking forward the 300-milligram BID dose which, as we said, today, we report 100% volumetric response rate in this data cut and we believe also has a very competitive safety profile. And the dose just below this 400 milligrams once daily dose, the once-daily dose is equivalent in PK to a 200-milligram BID dose. Both doses will allow us to further explore the target profile and generate clinical data that's needed to move forward this important program towards a potential approval and serve a very large patient population.

Note that the expected TAM of this opportunity is somewhere between $6 billion and $8 billion in the U.S. alone. This assumption is driven on our base of that there are 25,000 addressable patients in the U.S. and a pricing assumption similar to existing treatment options available, which is around $400,000 a year. And so if you just assume that 2,000 to 3,000 patients on chronic treatment that calculates to about a sales opportunity of $1 billion. And remember that we believe there are 25,000 addressable patients in the U.S. So I think all of this explains why we're very excited about sharing the dates today, and we're very excited about the large market opportunity that this provides.

I'll turn it over to our President of R&D, Don Bergstrom, to provide an overview of vascular anomalies and overview of the trial we're running and obviously share the excellent data that we're showing today. Thank you.

Thanks, Sanjiv. Vascular anomalies is an umbrella term for a variety of conditions with a wide range of clinical presentations. There are many separately named syndromes and conditions that fall into this category. Their commonality is that they consist of malformed vasculature and in some cases, other surrounding tissue types. And they share common genetics, of which PI3K alpha Victor CA gene is among the most common driving mutations and vascular anomalies where patients seek medical therapy. Current treatment options today leave large unmet medical needs as local treatments like surgery and topical ointments can only treat symptoms of isolated accessible lesions, and nonselective systemic treatment options carry a toxicity burden that is unsuitable for chronic use and severely limits the ability to inhibit the disease driving mutation.

Zovegalisib is the only mutant selective PI3K alpha inhibitor currently in clinical trials for patients with PIK3CA driven vascular anomalies. It's mutant selectivity allows for greater target coverage while maintaining tolerability, which we believe could lead to improved efficacy. It may also be suitable for chronic treatment, which is important as this disease is known to recur once treatment is paused. Across all of these subtypes, there are roughly 170,000 patients living with PIK3CA driven vascular anomalies in the U.S. Because of the size and concentration of PIK3CA mutations within the subgroups. The initial focus of our programs is in PROS PI3K-related overgrowth spectrum and LM, lymphatic malformations.

PROS with 5,000 to 10,000 patients with as many as half likely to seek a systemic treatment for their condition and LMs with approximately 65,000 patients with about 1/4 likely to seek systemic treatment. That leads, as Sanjiv pointed out, to approximately 25,000 addressable patients in the U.S. alone. Last year, we opened the ReInspire trial of Zovega and vascular anomalies. We leveraged nearly 250 patients worth of experience for Zovega in oncology patients across the doses we would be studying in reinspire to proceed immediately to randomized dose finding in adults and adolescents with vascular anomalies.

In the first part of the study, we randomized patients across 3 doses of Zovega. 400 milligrams twice daily, which is our recommended Phase III dose in combination with fulvestrant in breast cancer, 300 milligrams twice daily and 100 milligrams twice daily in adults and adolescents aged 12 and older. The trial is open to patients with PIK3CA mutated vascular anomalies, including all subtypes of pros, lymphatic malformations and PIK3CA-mutated venous malformations. Patients with PI3K mutation can be enrolled on the study and patients with PROS or LMs without a documented PI3K mutation were also allowed to enroll given the high prevalence of PI3K mutations in these diseases.

The objective of Part 1 was to evaluate the safety, tolerability and efficacy of Zovega in patients with PI3K driven BAs. And for the first time for a PI3K inhibitor in this field, to understand the relationship between target coverage, safety and efficacy with the goal of optimizing a treatment regimen for these patients. Efficacy is assessed in the trial by measuring change in disease lesion volume by MRI scans performed every 12 weeks, read by a blinded independent central review, and by investigator and patient reported changes in disease symptoms.

Here at the ISPA conference, we are sharing initial clinical data with a cutoff date of April 15, 2026, and including the initial efficacy data from 20 Part 1 patients who have reached their first 12-week MRI scan and safety data on 32 patients who have received Zovega including the 12 patients who have not yet reached their 12-week MRI scan. As expected, the patient population consisted primarily of patients with PROS, representing 69% of the patient population. 25% of patients have lymphatic malformations and 6% of patients have venous malformations. The median age of the population was about 25 years old with 2/3 of patients aged 18 or older, and 1/3 of patients aged 12 to 17.

Half of patients had a nonkinase domain PI3K mutation, 31% a kinase domain mutation and 19% had no documented PI3K mutation at study entry. 1/4 of patients were prediabetic at study entry. Most of these patients have previously been treated with alpelisib with some of these patients discontinuing alpelisib only days before signing consent for our trial. And 72% of patients had previously been treated with sirolimus or alpelisib as prior systemic therapy for their disease. Baseline characteristics were generally balanced across the 3 dose cohorts. Although in the small sample size, there were some imbalances, including over half of the 300-milligram BID cohort being prediabetic at study entry.

The pharmacokinetics of Zovega and vascular anomaly patients was as expected based on our experience in solid tumor oncology patients with a very flat PK profile and low peak to trough ratios. With the 3 doses achieving the levels of target coverage we projected, about 90% mutant PIK3CA inhibition for the 400-milligram dose, over 80% for the 300-milligram dose and approximately 60% for the 100-milligram dose. Across the 20 patients who had reached their 12-week MRI scan at the time of data cutoff, 12 patients or 60% and achieved a 20% reduction in lesion volume to be considered a volumetric response. The response rate for the 300-milligram dose was 100% and for the 100-milligram dose, and it was 29%. But as Sanjiv mentioned, post data cutoff, we saw an additional response from a 100-milligram BID patient, who converted from stable disease to an unconfirmed volumetric response at their 24-week scan.

Taking this into account, the BRR overall becomes 65% across all doses and 43% at the 100-milligram BID dose. Even at early time point, a response rate like this as well as the depth of responses has not been seen with nonselective inhibitors. As you can see on the waterfall plot, responses were observed across all 3 doses, across PROS subtypes and lymphatic malformations. Across all PI3K mutations, including in patients with no documented PI3K mutation. And patients previously treated with alpelisib, sirolimus or both agents and were independent of baseline lesion volume, including responses in patients with baseline lesions with a volume of several leaders.

Despite our plus of change in lesion volume highlight the rapid and deep onset of response seen, especially at the 300- and 400-milligram doses. And also show that responding patients who have had a subsequent scan at 24 and 36 weeks, the responses are not only maintained but are deepening. These thoughts also highlight the 3 patients at the 100-milligram dose who have not yet achieved a response, but show a deepening of volume reduction at subsequent scans and our ongoing treatment in the study with one of those patients, as we've mentioned, converting to a volumetric response after data [indiscernible].

While the lesion volume measured by MRI is the regulatory endpoint in vascular anomalies, symptomatic improvement as assessed by the investigator and the patient is a major goal of therapeutic intervention. Patients at all doses showed improvement in symptoms, whether assessed by IGIC, which is the investigator's assessment of the patient's general symptom burden, PGIC which is the patient's assessment of overall symptom burden our IADRSS pain scale, which is the investigator's assessment of pain in patients who report pain at baseline. Across all 3 measures of clinical benefit and across every dose, patients experienced improvements and symptoms with continued Zovega treatment with symptom scores at week 20 showing continued improvement relative to week 12. And impressively, 89% of patients showed improvement by investigator assessment and 79% of patients achieved improvement as assessed by patient assessment.

The initial response rate for any dose of Zovega shown today was higher than what has been reported for either alpelisib in either the EPIC T1 or EPIC T2 at any time points. We saw the same response rate of 60% in patients with the diagnosis of PROS versus the diagnosis other than PROS. An 80% response rate in patients with kinase domain PI3K mutations, 55% nonkinase MA mutations and 50% in patients without a documented mutation, we've also seen comparable response rates in patients with prior alpelisib or sirolimus were naive to systemic therapy at 62% and 57%, respectively. Similar response rates in patients older than 18, 64% and younger than 18, 50%, and this coincided with the vast majority of patients reporting an improvement in symptom burden.

Shifting to safety. Zovega was generally all tolerated across all 3 doses with most TRAEs reported as grade 1 or grade 2. The most common AE was headache, all grade 1, typically occurring within the first month or so of therapy then resolving. Other common AEs were fatigue, nausea and diarrhea and decreased appetite, all low grade. We only observed Grade 3 plus AEs at the 300-milligram BID dose. One case of hypokalemia in a patient who is taking a potassium wasting diuretic and 1 case of hypertension in a patient who entered the study with existing grade 2 hypertension. And there was only 1 event of Grade 3 hyperglycemia seen across the study in a prediabetic patient treated at the 400-milligram BID dose. Hyperglycemia was otherwise low grade and typically only seen in the first cycle and the majority of the grade I hyperglycemia seen at the 300-milligram BID dose was in patients free diabetic at baseline, of whom many had recently been treated with alpelisib.

We did not see other AEs frequently observed with nonselective PI3K inhibitors, including rash or stomatitis. This initial safety profile allowed for patients to maintain a consistent dose intensity and a low rate of dose reductions. Across the 100-milligram and 300-milligram BID doses, the median dose intensity was greater than 99% and fewer than 1/4 of patients required a dose reduction. The rate of dose reduction was higher at the 400-milligram BID dose, but as shown in the efficacy data, this dose is a higher dose than needed to achieve maximal efficacy for Zovega in these patients.

And importantly, all patients remain on study with no discontinuations for any reason across all those levels. At a high level, we are seeing clear differentiation compared alpelisib with this initial data. It is challenging to make a direct comparison to the safety profile of alpelisib in VAs as the EPIK-P1 trial was a compassionate use study and did not have prospective collection of routine safety and the EPIK-P2 trial used a 125-milligram dose of alpelisib, which is half the label dose and not the dose that's currently being assessed in the ongoing EPIK-P4 trial. The most robust recent safety data we have for alpelisib in exposure similar to those achieved at the 250-milligram PROS dose came from the PIC-P5 trial in breast cancer, where 71% of patients had a Grade 3 or higher AE. The overall rate of grade 3 or higher AEs observed in Zovega to date compare favorably to any trial of alpelisib and we believe represents a profile that could support chronic use and vascular anomalies, again, with no reported rash or stomatitis which are traditionally challenging AEs for the PI3K inhibitor class.

I will now share some patient vignettes that highlight the potential for Zovega to benefit patients with VA noting that these venues may not be representative of all patient experiences. The first vignette is a 44-year-old patient with a KTS subtype of PROS. His lesion has a non kinase domain mutation in PIK3CA, Q-546K. He has not previously been treated with systemic therapy. He has involvement of the entire left leg with a baseline target lesion of over 18 leaders. He was randomized to the 100-milligram BID dose of Zovega.

The patient experienced rapid symptomatic relief soon after starting Zovega rated as much improved on the IGIC instrument at week 24. Of note, this station does not meet the threshold for a volumetric response. But at week 24 had a 19.75% reduction in this lesion volume, reflecting a 3.6-liter reduction in this very large target lesion. He remains on the 100-milligram BID dose now in his 35th week of treatment. He's tolerated Zovega well and has not required any dose interruptions or modifications, and we were particularly excited to hear from the investigator treating this patient that prior to starting treatment, he couldn't walk further than from the front door to his car.

And within 2 weeks of starting therapy, he was able to walk around the block. The next vignette I will share is from a 12-year-old patient with a CLOVES subtype of PROS. His lesion has a helical domain mutation in PIK3CA, E542K. The patient was previously treated with both sirolimus and alpelisib. He has chest lesions resulting in both pain and difficulty breathing. He was randomized to the 300-milligram BID dose of Zovega. At the 12-week scan, the patient achieved a volumetric response with 49% decrease in his target lesions, which deepened to a 55% decrease at week 24. This was accompanied by improvement in symptoms rated as much improved by the investigator at week 24, and the patient is tolerating Zovega well and remains on treatment at the 300-milligram BID dose at week 32 with no dose modifications.

The last vignette shows the potential for Zovega in DAs outside of PROS. This is a 42-year-old female with a facial lymphatic malformation. Her lesion has a helical domain mutation, E545K. The patient was previously treated with both sirolimus and alpelisib. The patient did not respond to sirolimus and discontinued due to adverse effects. The patient had some symptomatic improvement of balalisib but also discontinued due to AEs. The facial lesion has resulted in ear pain, oral pain and jaw pain at baseline, and the patient was randomized to the 300-milligram BID dose of Zovega.

At the 12-week scan, the patient achieved a volumetric response of 31% decrease in their target lesion, which deepened to a 53% decrease at week 24. This was accompanied by improvement in symptoms rated as much improved by the investigator at week 24 and improvement in ear pain and oral pain, both rated as much improved. The patient is tolerating Zovega very well and remains on treatment at the 300-milligram BID dose at week 24 with no dose interruptions or reductions. The 100-milligram BID dose of Zovega with a 43% volumetric response rate demonstrates an initial profile that it already just 12 weeks shows dramatically greater volumetric response from what alpelisib data has shown at 250-milligram BID at any time point with what we believe to have a markedly improved safety profile with no Grade 3 or higher AEs, no hyperglycemia rash or stomatitis and broad symptomatic improvement as reported by both patients and investigators.

The 300-milligram BID dose of Zovega with 100% volumetric response rate shows an initial profile with what we believe to show the full potential of what a mutant selective PI3K inhibitor can achieve in vascular anomalies with a low rate of grade 3 or higher AEs, no grade 3 hyperglycemia, grade 2 hyperglycemia primarily in prediabetic patients, no rash or stomatitis, and broad symptomatic improvement as reported by both patients and investigators with all patients still ongoing. Therefore, we've decided to move into dose expansion with 2 doses of Zovega, the 300-milligram BID dose and then an intermediate dose between 100 and 300-milligram BID. It's a 400-milligram once-daily dose, and we believe this could show meaningfully better efficacy than a 250-milligram dose of alpelisib, meaningfully better tolerability than a 250-milligram dose of alpelisib with the convenience of a once-daily regimen for chronic use. Patients are currently enrolling at these 2 doses in our expansion cohorts.

Zooming out a bit further on the development and regulatory path, given the existing precedents in this space, we believe the accelerated approval pathway may be available to us, pending additional data maturation and feedback from regulatory authorities. Additionally, weight-based dose escalation is ongoing in pediatric patients aged 6 to 11. And once 2 dose levels are cleared in those 6- to 11-year-old patients, we'll open the 2- to 5-year-old cohort.

I'll now pass it over to Pete to wrap up.

Thanks, Tom. Clearly, these are very exciting times here at Relay. We started the year with the goal of making 3 meaningful disclosures against 3 very large market opportunities: second-line breast cancer with our ESMO TAT disclosure, frontline breast cancer with our recent triplet data disclosure. And now today, with these promising initial data in VAs, we've been able to do just that. We saw meaningfully differentiated initial activity and tolerability in VAs as compared to alpelisib at the very first dose level and across doses, response rates and symptomatic improvements have not been -- that we've seen -- that have not been previously reported by other agents.

And we see this activity regardless of mutation type, prior treatment and most importantly, these early data support our hypothesis that Zovega is active across VA subtypes, showing in lymphatic malformations, 3 out of 4 LM patients are in response and our 2 deepest responses are coming from LM patients. There are about 65,000 LM PIK3CA mutated patients in the U.S. we believe approximately 15,000 of these patients could be addressable with a well-tolerated and effective chronic systemic therapy.

Moving forward with the development ongoing against these 3 pillars of value creation, we believe we are entering an extended period of consistent updates. First, in second-line breast cancer by the end of this year, we believe we will be able to give an update on when we should reach full enrollment for ReDiscover-2. In frontline breast cancer, we will provide a regulatory update for the end of the year and remain on track for our first half of '27 Phase III start. We'll also likely give a Phase I/II triplet data update in the first half of 2027.

And in vascular anomalies, we plan to give a regulatory and data update by the end of this year. As you can see, we have a robust set of updates planned over the next 12 months. We have $642 million in cash, which gets us multiple clear data catalysts into 2029. And only incrementally more capital would be necessary through registrational data and early commercial launch in both breast cancer and vascular anomalies, pending data and regulatory outputs. In addition to continued data readouts from the Phase I/II triplet studies and initial NRAS date.

With that, I will now open it up to Q&A and hand it back to the operator.

[Operator Instructions] And our first question is going to come from Akash Tewari with Jefferies.

This is Amy on for Akash. Congrats on the great data. Just 2 from us. How should we think about the dose response between the 300 and the 400-milligram BID dose? Did patients on 400 have more dose interruptions? Or are there any differences in baseline characteristics? And then just 1 on your subtype data. You're seeing pretty strong early activity in LM. No response is in venous yet. How are you thinking about that venous population? Is that just more heterogeneous? Or do you think that could deepen over time? And then as you think about the initial accelerated approval package, is your base case assumption still a pool PROS LM in venous filing? And what specifically would the FDA need to see from a consistency of benefit standpoint across the subtypes to support a broader label?

Thanks, Amy for the questions. I mean I'll let Don to take the first 1 on the 300, 400-milligram question.

Yes. So I think, Amy, as you hypothesize that the 400-milligram dose we did have some patients who required dose interruption, which compromised their dose intensity and led to not achieving a response at that first 12-week scan. They do all remain on therapy and could certainly, as you see in the slide or plots have deepening of response -- that get them to a response over time.

But given that we are already essentially maxing out our benefit by every metric that we can look at, at the 300-milligram dose, we feel the 400-milligram dose is clearly super therapeutic and have made the decision not to pursue bringing that forward into our expansion cohorts.

On the second question on the venous malformations, where you can take that.

Yes. As we've said from the beginning, we believe that the disease biology here is going to be consistent across these subtypes. We've now demonstrated very clearly across 2 of the subtypes that we are seeing consistency of benefit. Not surprising, the enrollment in this stage of the study is weighted towards PROS and LMs. That's where alpelisib has accelerated approval on the pro side. And both alpelisib and sirolimus have had a history of use -- off-label use in the LM patients. We believe the venous enrollment will come over time. And we do believe that we should see a benefit there consistent with the other 2 subtypes.

And in terms of accelerated approval and how we think about pooling these studies -- these patient populations, -- the current study continues to be open to PROS, LMs and venous malformation patients. The difference between some of these subtypes is an LMs and PROS, can be prescribed therapy without a genetic test because you have 100% of patients with Pros having the piece mutation and 80-plus percent of LM. In venous malformations you're going to need a genetic test to direct therapy because there the frequency of PPC mutation is about 20% to 30%.

As to exactly how many of these 3 get pooled in an initial filing, that will be subject of conversations with the regulatory authorities. But we are going to continue to enroll all 3 sub-types. We think that the strength of this early efficacy certainly sets us up well to contemplate the accelerated approval pathway and that will be the topic of conversation with the regulators later this year.

I think the final point to note there is obviously the PROS and the LMs form the bulk of the addressable patient population. And I think we've got a long way in today's data set to derisk the fact that we believe Savage is going to be efficacious in both of those subtypes, and therefore, unlocking the large commercial opportunity there.

And our next question will come from Yaron Werber with TD Cowen.

Congrats on really terrific data that is much, much better than expected. So I have a couple of questions. The first one is -- there's actually 2. So you had 4 patients that have data out to 24 weeks old responders, they all improved -- they improved actually further when you go to 36 weeks. I mean they're getting to one of them at negative 61 and the other one has an 85% reduction. And then there's 3 other patients, 2 of them are at week 24, and they're essentially on the cusp of responders, they're the 100 BID. They are -- both of them are at 19.5, 19.8, would you expect those to respond when you do the third scan?

And then my second question is Protara announced that they are going to file based on 40 patients for accelerated approval. Obviously, a different therapeutic, it's injected into macro cystic and mixed cystic malformation. So obviously, a lot more topical sort of a different market than what you're targeting. But they're talking about accelerated approval based on 40 patients. How does that factor into your time line and your thinking for accelerated approval?

Thanks for the question. Maybe on the kind of genetics question. Obviously, you can see from the spider plots that the 300-milligram BID dose, you see very rapid declines in lesion size. And obviously, all of the responses that we had at 300 milligrams all happened at the 12 weeks and then they deepened. And so you see this kind of profound responses. And just to note, the responses we have, we have multiple now over 50% reduction. We saw in the alpelisib data, the kind of biggest responses they had were in the kind of 45% response rate. So one of the things is not just the response rate, but it's actually the profound depth of response that we're seeing.

In terms of the kind of 100-milligram BID dose, obviously, you're seeing there the kinetics are slightly different, but we are seeing these patients kind of hovering just below the 20% threshold as well as seeing some responders. And then as we said post the data cut, we saw one of those that we're hovering around 18%, 19% and converted from 18 and 19 at the 12-week scan to minus 35 at the 24 weeks car. So there, the kinetics seem to be kind of smoother declines over time. And so I think what we're excited about is just the profile and the flexibility that allows treating physicians.

In one side, on the 300-milligram dose, we believe that 100% response rate you're seeing profound responses very early. And then the other dose, you're seeing a kind of smoother, more kind of balanced responses over time. And I think patients will want different things and treating physicians a lot of different things. And I think the doses that we have here basically cater to everything. At one end, you're seeing profound efficacy. At the other end, you're seeing a tolerability profile that suits itself to lifelong dosing. So I think we're very excited about what we're seeing. In terms of the Protara question, maybe, Pete can comment on by the 40 patients is going to be enough.

Yes. I mean, look, it's obviously -- all these things are interesting precedents, right? So I publish it was approved off an accelerated approval off of 37 patients. The Protara data set is 40 patients, the Palvella data set that they're filing on also kind of in that ballpark. We need to generate our further data on our own and go have our own regulatory interactions, but these are obviously interesting precedents and it goes to show what can happen when you have profound efficacy that we've been able to demonstrate today.

But like you call out, these are different modalities and address different needs in different patient population, especially the sclerotherapy front, that's a local therapy that is not long term effective. You can see multiple -- or vignettes. These patients unfortunately have to go through multiple rounds of sclerotherapy. You can only address lesion by lesion, there are some lesions that are just not accessible for sclerotherapy. But another good tool and the toolkit for physicians and patients. But we're excited to be bringing forward a therapy that actually goes after the root cause of the disease here. It can treat it systemically.

And our next question will come from Bradley Canino with Guggenheim.

Great data and great progress for the trial. Can we just double-click on the prior question with some more details about how you believe whether you are or leaving efficacy on the table by not bringing forward the 400 mg BID oncology dose? And then second, you talked about cash runway but what about incremental cash needs going forward to contemplate the full development of vascular anomalies here with the plan and the frontline breast cancer study you plan to start next year?

All right. Maybe I'll start the first one and hand it over to Don. I think the 300-milligram BID dose, or 100% volumetric response rate, and we're seeing 80% to 90% patients seeing symptomatic benefit. We believe there is no further efficacy left on the table. And so as you go above that, you're just seeing greater tolerability challenges. And so we don't think there's any room for us to go higher than 300 million or any need for us to go higher than the 300. And that's why you're seeing us exploring the expansion cohorts, 300 BID. And then 1 below that, 400 once a day. So just to be clear, that's slightly confusing because that's the equivalent to 200 milligrams BID in terms of PK, you have a very flat PK profile. And so we've seen that in the oncology side and that PK has been validated.

And so we do not believe that there is any reason to go higher than 300 milligrams. And we believe that, that has a safety profile and an efficacy profile, it is going to be very hard to beat. And therefore, that's why we took it forward.

In fact, as we look across all of our metrics of efficacy, we feel we max out efficacy somewhere between 100 and 300, right? There's nothing we can look at that as going above 300 gets you additional benefit. We're clearly active in 100. So we're actually exploring this intermediate dose because it's possible that you could actually max out efficacy below 300 milligram. And in fact, we project that the 400-milligram once daily dose will give us exposure that lies right on top of the 80% inhibition line for MIK alpha. And I think it will be very interesting. Again, this is the first experiment that's ever really been able to assess the relationship between target inhibition and efficacy. And I think it will be very interesting to assess whether you can even go lower and max out the efficacy. So that's why we've opened up this additional cohort.

But remember, we've seen 11% response rate here for alpelisib for 12 weeks. And so we're seeing 100% at a 300-milligram BID dose. That gives us the confidence to say that I think we're hitting the efficacy in [indiscernible]. In terms of the cash runway, as you know, we have $642 million of cash at the end of Q1. And that gives us enough runway to generate the VAs data that we've just seen, take our second-line breast cancer trial towards registrational data to move forward general operating expenses and running our research team. We will need a modest amount more cash to make sure that we can get the vascular anomalies trial to registrational data.

And I think as we've seen from some of these questions, we don't believe that's going to be a large amount and cancer patient numbers here we're talking here are also very modest. And then to fund our frontline trial, to the point where we have registrational data set for both second-line breast cancer and the vascular anomalies and bring forward our NRAS program to having some early clinical data. So we think that it's a modest amount of money that we will need, but that will give us runway toward some very significant catalysts, 2 registrational data sets and another early stage data set from our NRAS program. And that should give us a cost of capital at that point that is significantly better than it is today and also give us significant options to raise financing using things like debt or royalty. So we feel like we're in a very strong position on the cash front. And any money that we do need to raise will be modest.

And our next question will come from Salveen Richter with Goldman Sachs.

Could you speak to the eligible populations within Pros and LMs and essentially on the demand side, how to think about where use will play out here.

Yes. The ultimate treatment goal here is to get to patients as early in their life as possible where the disease has not gotten too severe, and initiate treatment then stay on treatment chronically for the rest of their lives. Today, there is a -- there's an adult prevalent population that is probably underdiagnosed and undertreated, that will certainly be a near-term opportunity for us as you think about the early commercial launch in addition to continuing to build education and awareness around the use of systemic therapy as early in childhood as possible. So I think the way we will think about it in the early stages of launch are helping the undertreated and underdiagnosed adult population and really expand the market meaningfully there in addition to penetrating deeply the pediatric population. And so those would be the 2 approaches. And I think that's the type of kinetics we would see early on in March.

And the next question will come from Sean McCutchen with Raymond James.

Could you speak to the trajectory of lesion reductions in those patients that were dose reduced? And then any commentary and as much as you have this on the composition of lesions for patients in the study, fatty fibrotic? And any commentary on robustness of response data to that end?

Yes. Yes. So I mean I think as we've discussed a little bit already at the 400-milligram BID dose in that first 12 weeks, we did have patients who had, I think, critically longer dose interruptions, right, which compromise the ability probably to see a first scan response, but those patients remain ongoing at a lower dose. And I think we'll continue to track those patients over time, but their lesions are still shrinking even with the compromised intensity in the first cycle. In terms of lesion competition, these are complex lesions. I think you could see from our vignettes the nature of the lesion vascular components, fluid components, fibrotic components, fatty components.

And in general, I think what we're seeing is we're seeing broad changes in the MRI across the spectrum of lesion types. So you clearly see the reduction in the lesions in terms of the size, but you can also, for some of the more tortuous veins and tortuous vessels that are a hallmark of some of these diseases you actually start seeing -- in some of the scans, you actually start seeing this qualitative, of course, but you start seeing correction of those tortuous vessels. So I think it gives us encouragement that we're not just shrinking lesions, but by actually hitting the underlying biology of the disease or actually modifying the disease with Zovega.

And the next question will come from Andrew Berens with Leerink Partners.

Obviously, got better than we had anticipated. The first one, I guess, is on -- is there any way to remove PI3K testing as part of the payer requirements to access. We've heard from some of the Medicor docs that could be very difficult logistically. Is there any way to offset that or help with that requirement? And then the second is a strategic question, other strategic question for Sanjiv just wondering on how you see this opportunity versus breast cancer. Would you consider divesting or monetizing breast cancer and using the funds to focus on this? Because it certainly seems more bite-sized for SMID Cap biotech than breast cancer.

Pete, you take the first one.

Yes. Look, I think the best thing we can do is to generate the strongest data possible as we move into regulatory conversations, potential filings and then market access conversations. The -- clearly, a genetic test is not necessary to direct treatment for both PROS and LMs. And I think that will be the position we take. We have multiple patients in this data set with unknown mutation types and not -- that came on the study without a genetic test, and we're seeing clear clinical benefit. So I think that will be the way we approach it, and we'll continue to work with the -- some of the commercial providers on making sure that the access and availability of genetic tests when needed are up to the technical standards to be able to be useful in this patient population.

In terms of the strategic question, I mean it's an interesting one we think about a lot. And I think we just see it in a slightly different way. I mean your point is very valid in that this is going to be a commercial opportunity that is very management. As you know, there are only a small number of centers in each country in the U.S., that's kind of 20 to 30 centers. And so in terms of the commercial model, it is very manageable -- and obviously said the market size here is very significant. So it is a very kind of a sweet spot for a small biotech to commercialize.

And so today's data significantly derisk that opportunity, and we'll move very rapidly towards trying to get to an accelerated approval and generate revenues. But it doesn't change the fact that our breast cancer side is still as attractive. We see capivasertib in the second line moving towards $1 billion worth of run rate, and they've done that very rapidly. And obviously, we know it's going to share of voice marketing. If we can show that we could almost double the median PFS. Companies like Menarini have shown that you can very successfully commercialize as a small company in breast cancer, especially in the U.S.

And so walking away from over $1 billion of revenue in the near term is not something that we want to do. And given the balance sheet that we have, I think we have the opportunity to do it. My own background is I come from a commercial way background, I spent 10 years in commercializing therapies and so I'm excited about going out and winning in both.

And then on the other side of this, you see a frontline trial that will also be there. And so that is a huge opportunity. I mean those -- you can see it from the CDK4/6 sales that are out there in the tens of billions. So we feel that we have 3 very large opportunities. We obviously today of derisked 1. We're very excited about commercializing in the second line. I think it's very manageable. And then I think in the long term, we could really change the kind of revenue trajectory of our company in the front line. So we think all 3 are manageable, and we look forward to executing all 3.

And our next question will come from Silvan Tuerkcan with Citizens.

Congrats on this very strong data. My first question is maybe big picture. I know it's very early here, but how would the how you see this treatment regimen involved? Would there be eventually maybe treatment-free windows, considering you have such fast onset of responses, could patients cycle on and off drug? And then the second question is, is the response rate tied to the baseline lesion volume, so meaning 2 larger lesions responses as well as small regions?

Yes. So I think we can take them one at a time. So on the lesion volume, maybe Don, you can take that one.

Yes. So on the lesion volume, we've done several cuts of baseline lesion volume versus response rate and don't see significant differences between the probability of achieving response with a large lesion versus a small lesion.

And then in terms of the cycling, I mean, this is a comment that we hear clinically that people are cycling on and off alpelisib, and this is because of the fact that they're getting better. And the answer is no. It's because of the toxicity. Patients do not want to tolerate it. We were here at ISSVA Conference in Philadelphia. We're talking to the patient advocates it's very clear that, that is the challenge for patients, which is the toxicity. And so I think what Zovega allows us to do and what you see is in the data is that for the first time, maybe there is a potential for a chronic therapy, that allows patients to stay on therapy chronically.

And now we know that you need to put pressure on the PI3K inhibition constantly. If you stop, the lesions will occur. And so I do think it allows us, for the first time, to allow patients to potentially have a chronic solution here. And as we talked about earlier, the different kinetics of the doses is very interesting. It allows flexibility for prescribers. Obviously, that 300 milligram, you're seeing a 100% response rate and you're seeing very rapid and deep responses. And that may be what prescribers what.

On the other side, you're seeing the 100 milligrams, very much an AE profile that is very, very manageable. We're not seeing any grade 3 toxicity. We're not seeing any hypoglycemia. And so that could be what some other patients want. And so I think what we're seeing here for the first time is profound efficacy or prefinance safety that will allow us to be the kind of solution for everyone here. And so we're very excited about taking it forward.

And the next question will come from Matthew Biegler with OPCO.

I just had a clarification on the patients with unknown PIK3CA status. Is it just that they weren't tested for -- or is there a downstream AKT or maybe P10 over activation? Or just like anything else that would explain the activity that you showed there because there was at least a few patients that did have a nice volumetric response with that.

Yes. So I think the patients with unknown status reflects 2 challenges with testing these patients. One is accessing tissue for performing the diagnostic tests on the lesions. It's not medically feasible, to be able to access the tissue or it's challenging to access the tissue. And then the other is the sensitivity question. This is a disease of somatic mosaicism the variant allele fractions are not as high as what you see in solid tumors. So a lot of those oncology diagnostics don't have the necessary sensitivity to be able to pick up these mutations.

So we have patients who either haven't been tested because of the challenges of getting tissue or has been tested and the diagnostic sensitivity of the assay has not been able to detect the PIK3CA mutation we don't think that these patients have downstream mutation AKT,PTEN. We think, again, in LMs, the vast majority of patients are driven by PIK3CA. So I think these patients in whom we're seeing benefit or picture PIK3CA-driven anomalies, and it's largely technical or medical reasons why you don't have a detectable mutation.

The next question is going to come from Boris Peaker with Jones Trading.

I just wanted to question on safety. You mentioned that 23% of patients had dose reductions. I just want to understand what these primarily at the 300 mg BID or the 100 mg and also just understanding the timing of these dose reductions? Were they typically at the start of therapy or maybe scattered throughout the observation period?

Thanks for the question. Maybe before I hand it over to Don. I'll just make one point around the fact that, obviously, we have a lot of information around safety that came from our extensive experience in oncology. We now have several hundred patients worth of data in the doublet. Now we have much more now in the triplet. And so as we kind of started out this program last year, the ability to share that safety day to both the regulated and with prescribers was quicker for us to launch this and go at the pace that we have.

And I make that point because I know that we'll get a lot of competitive questions going forward and not to underestimate the inherent advantage of having such an extensive multiyear database of entering this field. Because remember, this is a very different feel from oncology and the threshold for folks to try these new therapies are sourced in human is very high. And so maybe with that, I'll hand it over to Don.

Yes. So the reductions, we saw more reductions at 300 than we did at 100. We only had a single patient with a modification in a patient who had some low-grade nausea, decreased from 100 BID to 75 BID, but then actually came right back to 100 BID. It continues to be treated at 100 BID. At the 300 BID dose when we have had reductions, it's generally been relatively early in the course of therapy. And we've been able to successfully be able to manage those reductions, still being able to maintain the dose intensity and still being able to achieve the 100% volumetric response rate.

That does conclude our Q&A session. I will now turn the call back over to Sanjiv for closing remarks.

Thanks for attending the conference call this morning. And hopefully, you believe the data is just as exciting as we do. We're going to go over to the ISSVA Conference now and share the data with physicians, patient advocates and make sure that we can rapidly enroll these expansion cohorts as possible solve to the regulator and hopefully get this therapy to the patients that need it as rapidly as possible. Thank you.

This concludes today's conference call. Thank you for participating, and you may now disconnect.

Good day, ladies and gentlemen, and welcome to the Relay Therapeutics Frontline Breast Cancer Update Call. As a reminder, this conference call is being recorded.

I would now like to introduce your host for today's conference. Mr. Pete Rahmer, Chief Corporate Development Officer at Relay Therapeutics. Sir, you may begin.

Thank you, operator, and good morning, everyone. Thanks for joining us. We're excited to share our frontline breast cancer update with you today. You can access the press release from today, the slides we are reviewing and a replay of this call by going to the Investor Relations section of our website.

As a reminder, during this call, we will make certain statements that are considered forward-looking statements within the meaning of the Private Securities Litigation Reform Act of 1995, including expressed or implied statements regarding our strategy, business plans and objectives, the expected therapeutic and clinical benefits of our product candidates, the potential of our platform and our product candidates, and progress, timing and execution of our clinical trials.

Such forward-looking statements are not guarantees of future performance, and therefore, you should not put undue reliance upon them. These statements are subject to numerous risks and uncertainties that could cause actual results to differ materially from what we expect. We refer you to our SEC filings and on our website for a discussion of our risk factors. The forward-looking statements in this presentation speak only as of the original date of this presentation, and we undertake no obligation to update or revise any of these statements.

I'm also joined here today by -- with Sanjiv Patel, our CEO; and Don Bergstrom, our President of R&D. And with that, I'll turn it over to Sanjiv.

Thank you, Pete, for the introduction, and thank you all for joining the call at late notice. Today we're going to share further data on our lead program, zovegalisib, zovega for short, our pan-mutant selective PI3K-alpha inhibitor, which has the potential to address 3 very large commercial opportunities.

The first of these opportunities is second-line hormone receptor positive, HER2-negative metastatic breast cancer. Last month at ESMO TAT in Paris, we shared data with our pivotal trial dose in this indication, which increases our confidence that we will be successful in showing clear differentiation from the current standard of care, capivasertib. And as a company, we're all laser-focused on executing our pivotal trial, ReDiscover-2, which is recruiting patients globally as we speak.

Today we'll share data on the second of these large commercial opportunities in frontline hormone receptor positive, HER2-negative metastatic breast cancer. And we've been pleased to announce today that we've selected atirmociclib, Pfizer's selective CDK4, as the CDK combination partner for our first-line regimen with zovega.

The data we'll share today will share why we're very excited about this choice, as in heavily pretreated patients, the early efficacy and tolerability of the zovega-atirmo triplet in median third-line patients is approaching being comparable to the standard of care in frontline patients. We know patients generally have worse outcomes as we move into later lines of therapy. So this data gives us great confidence we'll be able to show clear differentiation when we move our triplet from these later-line patients into frontline patients, and present the 40% of frontline patients who have a PI3K-alpha mutation with a much better treatment option. We plan to initiate our Phase III frontline trial in endocrine-sensitive patients in early 2027, obviously, subject to regulatory feedback.

And we're also pleased to announce today that we have a supply agreement in place with Pfizer, to supply atirmociclib for our Phase III trial and palbociclib for part of the control arm.

Finally, in what's a very busy time here at Relay, in a few weeks' time, at the ISSVA Conference in Philadelphia, we'll share data from the third of these very large commercial opportunities, vascular anomalies, where we'll show approximately 20 patients' worth of efficacy data. And our hope is to show zovega could offer a differentiated option for these patients.

Right. Let's get into the detail of our disclosure today and focus on the frontline breast cancer space. This is a very large commercial opportunity, and there are approximately 35,000 frontline metastatic breast cancer patients across the major geographies that have a PI3K-alpha mutation. And they're served today by some very large billion-dollar brands.

The recent Roche INAVO120 trial and its subsequent approval has shown that, in PI3K-alpha mutated patients in the frontline, adding a PI3K-alpha inhibitor to the standard of care, CDK4/6 plus endocrine therapy doublet, allows you to achieve greater efficacy, manifested in both greater PFS and greater overall survival. However, the issue is that the tolerability profile of stacking a nonselective PI3K inhibitor with a nonselective CDK inhibitor leads to a profile that could be challenging for patients to tolerate for multiple years, and has led to the Dear Doctor letters being issued due to potentially life-threatening safety events.

So the goal of a next-generation triplet is to maintain the increased efficacy that has been seen for a PI3K inhibitor in the frontline by having a better tolerability profile that allows patients to maintain their dose intensity over multiple years. We hope to do that in Relay by combining multiple selective next-generation agents together that dial out the off-target toxicity.

Our triplet will be anchored by a mutant selective inhibitor in zovega and a CDK4 selective inhibitor in atirmo. We believe this selective-selective profile should drive greater tolerability, leading to greater efficacy. The data we report today from the triplet of zovega, atirmo and fulvestrant in median third-line patients gives us great confidence that we can hit this profile. And we're pushing to initiate the frontline trial in early 2027.

We've deprioritized the other triplet cohorts we were testing given the step-change nature we're seeing in the safety and tolerability profile of our atirmo triplet. And this long-term tolerability that we're seeing of the zovega-atirmo triplet will enable us to maximize efficacy in these patients.

To cover the data and next steps in more detail, I'll hand it over to Don Bergstrom, President of R&D.

Thank you, Sanjiv. We initiated dose-finding of zovega combined with atirmo and standard dose fulvestrant. We did observe an effect of atirmociclib on blood concentrations of zovegalisib. Atirmociclib increases the concentration of zovega by about 2.5-fold. So a 150-milligram dose of zovega combined with the atirmo leads to blood concentrations for zovega that approximates the exposure we achieved at the 400-milligram dose of zovega combined with fulvestrant that we were testing in the ongoing ReDiscover-2 trial. Zovega does not appear to impact blood concentrations of atirmo.

And we will prioritize moving forward with the 300-milligram dose of atirmo, which is the dose Pfizer investigated in the positive FOURLIGHT-1 trial and is currently investigating in the ongoing FOURLIGHT-3 frontline trial. We will disclose data from 62 patients treated with the triplet of zovega plus atirmo plus fulvestrant, with median follow-up of 7.4 months.

Patients were dosed with doses of zovega between 100 and 200 milligrams BID, and doses of atirmo between 100 and 300 milligrams BID, all with standard dose fulvestrant. We are not reporting data on patients treated at the 200-milligram dose of zovega combined with the 300-milligram dose of atirmo as that dose had a safety profile that, while not a formal MTD, did not meet the safety profile we are targeting for chronic treatment of a frontline population. We intend to bring the 150 milligrams zovega dose forward as the Phase III dose, pending regulatory feedback.

Dose escalation was performed in a heavily pretreated population. As with our prior disclosures on zovega doublet therapy, all patients were required to receive at least 1 prior CDK4/6 inhibitor therapy, and we're allowed to receive more than 1 CDK4/6, multiple endocrine therapies for advanced disease and prior chemotherapy or ADC for advanced disease. But unlike prior disclosures that we've made, this cohort was also allowed to have received prior PI3K pathway directed therapy. And 10 of the 62 patients had received a prior PI3K pathway inhibitor.

Consequently, these 62 patients were third-line or later patients on average. More than half had already been treated with a SERD and more than 1/4 had been treated with chemotherapy or an ADC for advanced breast cancer. The patients also had notable features with nearly half of the patients being prediabetic at baseline, over 60% of patients having visceral disease and over 40% having a co-occurring ESR1 mutation at baseline.

As I described earlier, atirmociclib increases the blood concentration of zovega. All 3 zovega doses: 100 milligrams, 150 milligrams and 200 milligrams, exceeded our goal target coverage of 80% PI3K inhibition sustained for 24 hours a day. The 150-milligram dose of zovega in combination with atirmociclib and fulvestrant approximated the zovega exposure we achieved at 400-milligram BID in the doublet combination with fulvestrant, the dose we're testing in the ongoing ReDiscover-2 Phase III trial.

Even though this cohort was treated at unoptimized doses, with all patients being treated at or below our recommended Phase III dose, the triplet of zovega plus atirmo plus fulvestrant was highly active in this median third-line patient population, with an overall response rate of 44% in patients with RECIST measurable disease, which begins to approach the ORR observed for CDK4/6 plus ET doublet therapy in first-line patients, which has ranged between 53% to 55% in registrational clinical trials.

Importantly, the ORR was comparable in patients with both kinase and nonkinase domain mutations. And responses were observed in patients who had received a prior SERD and/or a prior PI3K pathway inhibitor. Of note, 1 patient, who initially was a confirmed CR, has now converted to an unconfirmed CR.

With the acknowledgment that these are cross-trial comparisons, we can compare the 44% ORR observed for zovega plus atirmo plus fulvestrant with other PI3K triplets tested in later-line ABC patients. In December, Roche disclosed the triplet of inavolisib with 600 milligrams ribociclib or abemaciclib with fulvestrant. And the ribociclib triplet achieved a 33% ORR and the abemaciclib triplet achieved a 28% ORR, both numerically inferior to the ORR achieved with the zovega plus atirmo plus fulvestrant triplet.

And we can also compare our triplet data to ORRs reported in frontline trials, where ribociclib plus ET and abemaciclib plus ET doublets achieved 53% and 55% ORRs, respectively, in endocrine-sensitive patients across both PIK3CA-mutated and wild-type patients. And the inavolisib plus palbociclib plus fulvestrant triplet showed a 58% ORR in endocrine-resistant patients.

Given the historical meaningful increase in ORR moving from later-line to frontline patients, this gives us the confidence that our ORR will likely increase as we move into earlier-line patients at optimized doses. And with the tolerability profile we were observing for the zovega triplet, we expect this will translate into PFS benefit for frontline patients. And therefore, it gives us the confidence to take the zovega plus atirmo plus ET regimen into a frontline trial.

For a triplet regimen that we will move into frontline development, tolerability is key as the objective will be to be able to treat these patients for 2.5 to 3 years, and to be able to treat a broad spectrum of PIK3CA mutated patients in the front line, not just endocrine-resistant patients who were treated in the INAVO120 trial of inavolisib. We are very encouraged that the tolerability profile we are seeing with the triplet of zovega plus atirmo plus ET is consistent with achieving these goals.

Only 40% of patients experienced any Grade 3 or higher treatment-related adverse events. And most of the Grade 3 or higher events were neutropenia, with no cases of febrile neutropenia. Overall rates of hyperglycemia were low with no Grade 3 or higher hyperglycemia, despite nearly half of patients being prediabetic at study entry.

Only -- or less than 10% of patients dose-reduced zovega and 16% dose-reduced atirmo, with the combined rate of reduction of either drug being 23%. And there were very few discontinuations due to TRAEs, with 2 patients discontinuing zovega and 4 patients discontinuing atirmo. Of note, 2 of the patients who discontinued atirmo remained on study, receiving the zovega plus fulvestrant doublet.

We can contextualize the overall tolerability profile we are seeing relative to the other PI3K inhibitor triplets and standard of care CDK4/6 inhibitor plus ET doublets in frontline patients. Our 40% rate of Grade 3 AEs with less than 10% zovega reductions compared favorably to the recently reported data for inavolisib in the MORPHEUS trial in late-line patients, which showed a 68% rate of Grade 3 or higher TRAEs, with 47% inavolisib dose reductions in combination with ribociclib, and 92% rate of Grade 3 or higher TRAEs and 33% inavolisib dose reductions in combination with abemaciclib.

In frontline regimens, CDK4/6 plus ET doublets on their own have shown 81% and 55% Grade 3 or higher AEs for ribo and abema, respectively. And in the inavolisib plus palbociclib plus fulvestrant triplet tested in the INAVO120 trial, a 91% rate of Grade 3 or higher AEs.

The response rate and broader tumor reductions in combination with a quite favorable safety profile is leading to very encouraging durability in the zovega plus atirmo triplet. After 7.4 months of median follow-up, 48 out of 62 patients or 77% remain on therapy. The median PFS has not yet been reached.

This slide shows the proposed Phase III trial we plan to initiate in early 2027. We will focus on endocrine-sensitive, HR-positive, HER2-negative PIK3CA mutated patients who are 12 months or later from completing their adjuvant endocrine therapy or have been diagnosed with de novo metastatic disease. The trial will randomize patients to the triplet of zovega plus atirmo plus an aromatase inhibitor versus a CDK4/6 of investigator's choice, ribociclib, abemaciclib or palbociclib plus an aromatase inhibitor.

Given we are testing both zovega and atirmo in the experimental arm, we anticipate we may need to account for the contribution of components. We are preparing to discuss the design of this trial with health authorities and anticipate being able to provide an update on final trial design and timing before we start the trial.

This is just one opportunity amongst many that we can pursue in the coming years based on the differentiated profile of zovega, which includes moving into early breast cancer in HR-positive, HER2-negative PIK3CA mutated patients, as well as exploring opportunities in other breast cancer segments. And as you've seen from today's data, we are moving zovega to where the field is moving by combining with an emerging selective CDK4 inhibitor rather than existing standards of care.

And we can also explore other classes of combination partners, including oral SERDs. We've assembled an advisory board of leading global breast cancer investigators to guide us to the development of zovega.

I'll now pass it over to Pete to discuss the Pfizer supply agreement details and wrap-up.

Thanks, Don. We initially entered into a supply agreement with Pfizer to be able to explore what bringing together 2 novel selective inhibitors could provide patients, and are excited to announce today we are extending that relationship to a Phase III supply agreement.

Under this agreement, Relay sponsors, fully operationalizes and funds the frontline Phase III trial and retains full global rights for zovegalisib. Pfizer supplies atirmo for the experimental arm and palbo for the part of the CDK4/6 of choice control arm.

With these promising data in hand, we will now over the coming months conduct regulatory interactions to confirm the Phase III design and dose and rapidly prepare to start the Phase III trial by early next year. We believe we have shown you over the last month the potential for zovega to address 2 very large commercial opportunities in breast cancer, confirming the second-line promise with the ESMO TAT data, and now initial data demonstrating the potential in frontline patients. And we look forward to sharing initial vascular anomalies data next month at ISSVA for the third pillar of the large commercial opportunities that zovega could potentially address.

Thank you all for taking the time this morning to join us. And with that, I will open it up to Q&A and hand it back to the operator.

[Operator Instructions] And our first question coming from the line of Brad Canino with Guggenheim.

It's great to see the data and the progress towards the frontline trial. A couple of questions for me, if I may. First, you talk about the benefit of the novel selective-selective approach versus the combo with CDK4/6 inhibitors, which you also tested. I guess, what were the safety data like for the palbo combo? And how do you think competitors will fare if they are potentially going to go down the CDK4/6 triplet route?

I think we know from the safety profile that we showed what a CDK4/6 plus endocrine therapy looks like in the front line. And obviously, as you know, these patients need to be on therapy for multiple years, and so tolerability really is critical. And so what we saw was a step-change in difference with atirmo, which is it's driven by its selective nature of it only targeting CDK4 and dialing out the CDK6 toxicities. And then obviously, combining that with zovega, another mutant selective inhibitor.

And so for us, both agents together led to a tolerability profile that we think would lend itself to a multiyear frontline treatment that a nonselective PI3K inhibitor or nonselective CDK inhibitor just cannot lend itself to.

Okay. And second, you build this approach as a competitive advantage for Relay with this Pfizer supply agreement. But I guess, what's to stop Pfizer from supplying drug to other companies with mutant selective PI3K-alpha inhibitors to do the same thing for Phase III trials?

Pete, do you want to take that one?

Yes. Thanks for the question, Brad. Yes, there are limitations for both us and Pfizer moving forward in the near term with other PI3K or selective CDK4 molecules into a Phase III trial.

We believe that we will definitely be first to market with this selective-selective approach.

Okay. And then last for me. It looks like the rash for the triplet is a new signal compared to the doublet data. Could you comment on the presentation and management of that AE? And then that's it for me.

Yes. So what we've seen in rash has been primarily low-grade, well-managed with antihistamines. It is something that we've not seen at high rates before with zovega. And I think we'll wait to see more robust disclosures from atirmo, including the upcoming FOURLIGHT-1 disclosure to understand what the potential contribution of atirmo could be.

We are in later-line patients. We are in patients who have seen prior PI3K inhibitor pathway agents, many of which carry a rash AE. And there is some trends towards some of the rash-getting patients who have seen those prior PI3K pathway inhibitors. But I think we need more experience to fully understand it.

Our next question coming from the line of Akash Tewari with Jefferies.

This is Amy on for Akash. So first of all, would love to get any color on your decision to go into ER-sensitive patients instead of ER-resistant, like some of your peers where the bar on PFS and OS is lower, what are you seeing on early durability in those 77% patients that are still continuing on trial, and safety to give you confidence that you could superiority in the setting? I believe MONALEESA has shown around 25-month PFS. So would love to kind of get your view on the confidence to beat that.

Thanks for your question, Amy. I'm going to hand it over to Don.

Yes. So I think your point, Amy, with tolerability is key here. I think what we're seeing in our patients, and again, these are third-line and later patients, is at 7.5 months we have over 3/4 of patients still on study. As you can see in the swimmers plot, patients whom are achieving responses are maintaining those responses with long durability. So we haven't been able to really meaningfully calculate the ORR yet since all of these responses for the most part are ongoing.

With 7.5 months follow-up, we're still too early to calculate a PFS. When we look at the Kaplan-Meier curve, it's a relatively flat curve, as you can imagine, with a lot of censored patients, those censored patients, again, representing our patients who are ongoing on study who have not yet had a progression event.

So we're very encouraged in this late-line patient population, both by the efficacy we're seeing, but really the durability of benefit and the durability of being able to maintain dose intensity, with 3/4 of patients still receiving the dose of drug that they started with when they started -- when they came on study.

With regard to what we think we'll need to see in a frontline trial, as you point out, MONALEESA showed overall, in both PIK3CA mutant and PIK3CA wild type patients, PFS in the mid-20s. But the subgroup analysis looking at PI3K mutant versus PI3K wild type showed a 19-month PFS in PI3K mutant patients versus 31 months in PI3K wild type.

So these data would suggest that the PI3K mutant patients may not fare as well with frontline CDK4/6 doublet therapy. And in fact, we've been able to see from some recent protocols that have been published on the European CTIS website that there are assumptions that the PI3K mutant patient population in this setting would have about a 20-month PFS, which would be where our assumption would be for what we would need to be.

And on the -- your endocrine-sensitive -- why go to endocrine-sensitive, I think that it's the place where everyone would like to go if you had a regimen that afforded to you the tolerability to get there, and just most can't. And so what we've demonstrated today is that we clearly have a safety and tolerability profile to bring to these patients.

It's the largest portion of the frontline patients, over 37,000 of them in major geographies throughout the world. And the ability to go prosecute this trial where there's no other pathway inhibitor approved is quite attractive from a probability of success standpoint. So we will be running this trial, again, it's just the doublet standard of care without the involvement of a pathway inhibitor. That in combination with the supply agreement where we get atirmo and palbo supplied for free allow us to run a very manageable frontline trial from a cost effectiveness standpoint.

Excellent. And then just another one, how much do you think atirmo is contributing to your ORR? I know Pfizer reported, I think the last cut was around 32% in a more refractory setting. So would love to get any color here.

Yes. So I think what we're seeing really here is the triplet activity. First of all, we're coming in at 44%. We haven't really seen a large data set of the atirmo doublet in the later-line patient population. So that 32%, I think, was based on 23 or 24 patients who had measurable disease. And specifically in PIK3CA mutant patients, they only had 10 what they call PIK3CA pathway altered patients. So that was a patient population that included PIK3CA, P10 and AKT-mutated patients. And those 10 patients were treated at both 300 and 400 milligrams of atirmo and treated with both fulvestrant and letrozole.

So small data sets, very heterogeneous. I think as we look at the data we're seeing and given that they're exposures that we know are active exposures to zovega, we think we're really seeing the activity of the triplet.

Our next question coming from the line of Yaron Werber with TD Cowen.

This is Jaena on for Yaron. Congrats on the data. I wanted to double-click a little bit also on the attributions of atirmo versus zovega. When we look at your prior doublet data, obviously, this is at the pivotal dose and in a more refractory -- in a less refractory population. But overall ORR increased, 44% versus 43% for your fed data. Can you give us any kind of -- any more detail on how the ORR breaks down by different doses for zovega and how you might expect -- what degree of improvement you expect to see as you move to 1L additionally with an aromatase inhibitor and fulvestrant?

Okay. So maybe we'll break the question down. We'll just finish off the last question around the contribution of atirmo. I think the most kind of powerful thing is, obviously, the fact that Pfizer has entered into a supply agreement, so there's definitely a belief. They have perfect information on the FOURLIGHT trials, that there is going to be an additive component for zovega.

And so then if you come to your question around what is the comparison between our doublet data where we're sitting in the kind of high 40s with kind of close to second-line patients, and this triplet data in much more heavily pretreated median third-line patients, maybe I'll hand that over to Don.

Yes. So I think we are -- obviously, there's a difference in pretreatment, including we've got patients here who have seen prior PI3K pathway inhibitors. And we are seeing activity of the -- with responses of the triplet in those patients, which to us suggests that we're seeing triplet benefit in those patients.

In terms of the dose responsiveness that we've seen, we've reported out these data with pool data on every patient treated at or below the recommended Phase III dose of 150 mg/200 mg, because we've seen activity at every dose that we've studied, including starting at the 100-milligram/100-milligram dose.

So this is a regimen that's broadly active across the dose range that we studied. I think as we get more data at our optimized doses, we expect we should be able to see the data continue to remain where it is, if not improve, given that we would be at more optimized doses.

And then as you've seen other agents go from later-line settings to earlier-line settings, into CDK4/6-naive patients, whether you're talking about the CDK4/6 inhibitors themselves with CDK4/6 retreatment, or the experience with inavolisib moving from Phase Ib in CDK4/6 or second-line and later patients in the front line, you're generally seeing a 20% to 30% increase in ORR as you move into the frontline.

Great. And if could follow up with one more question. Your ORR data excludes kind of, as you mentioned, the 200 mg BID zovega and 300 mg BID atirmo due to safety not being quite as optimal, even though it didn't quite reach the level of MTD. Can you give us any more details on what exactly you observed here to warrant this de-prioritization?

Yes. I mean we just had a slightly higher rate of higher-grade AEs that led to dose reduction, right? And our goal here is really to have a regimen that is giving us the target coverage you want with the optimized AE profile, low rates of dose reduction, that allows patients to remain on therapy. So given that we were seeing activity at several lower doses, we knew that we could move on from that highest dose that we tested.

Tolerability is going to be the critical part of this triplet.

Our next question coming from the line of Sean McCutcheon with Raymond James.

Maybe could you provide some detail on any additional work, if any, that may be required to determine the profile of zovega in combination with aromatase inhibitor ahead of starting that Phase III study?

Thanks, Sean. We have ongoing expansions with both fulvestrant and aromatase inhibitor. Our early experience with the aromatase inhibitor is that it should not -- it's an exercise of just getting a little bit of experience there to be able to move forward and take that to the regulators. And then we'll go have a regulatory interaction and then be able to move quickly towards initiating this frontline trial.

So just some box-checking exercises to go through, and the regulatory interactions, are the main gating items before moving on.

Our next question in queue coming from the line of Salveen Richter with Goldman Sachs.

Just a bigger picture question here about how you think of positioning versus competition such as Celcuity and others in the marketplace and how you think about the cadence of additional updates on this program over the year?

Look, I think the bigger picture, thanks for the question, Salveen, is clear, which is this is a very large patient population. And as you can see, it's served by some very large billion-dollar brands. But they do come with these toxicity challenges and the PI3K-alpha mutated patients do worse. And so we think that this approach allows us to solve both problems, which is adding a PI3K-alpha inhibitor should drive to greater efficacy. And then using atirmo in combination with zovega should allow us to get the tolerability.

And obviously, it's an all-oral regimen. And so as you know, these patients are on therapy for multiple years and they're trying to live as normal a life as they can. And so we think this is the best approach. And we've talked to a lot of patient advocacy groups and they're very supportive of this approach.

Now you mentioned there are other agents in this field. Obviously, there's some of them will be using the traditional CDK4/6 backbones, and we think that they will have increased tolerability challenges. And obviously, we see that with the uptake of inavolisib not being as robust as I think folks expected post the approval.

And then obviously, you mentioned gedatolisib. As you know, gedatolisib is a weekly IV regimen. And so for frontline patients who are trying to live their lives as normally as possible, to come in every week and have an IV infusion and then continually have daily or multiple times a day dexamethasone mouthwash and potentially then other concomitant medications for the AE, leads to quite a significant disruption in their lives. And so we don't think that is a viable regimen in the front line.

And so that's why we're really excited about this oral-oral, selective-selective regimen in the frontline that should allow these patients to live as normal lives as possible.

Yes. On the updates from here question, the key one we can point to is coming back later this year on the other side of our regulatory interaction to confirm both trial design and dose before we set off to start the study in early next year.

Our next question coming from the line of Etzer Darout with Barclays.

This is Luke on for Etzer. Kind of similar to the previous ones, curious to see how you're benchmarking this. Because you are in your Phase III [ design that is ] a CDK4/6 and an aromatase inhibitor as a comparator arm. Talking to KOLs, is there anything that they are looking for out of your triplet that we should be putting attention to that may kind of come right out of that Phase III trial?

Yes. I think it's kind of more of what we talked about. I mean, INAVO120 trial showed you that by adding a PI3K-alpha inhibitor, you clearly saw greater PFS and OS. So that is a given here. The fact that if you're going to run a triplet versus the doublet, we've seen that you're going to be successful. The issue is, do patients want to stay on therapy? And is this actually going to be commercially viable? And I think inavo is also starting to give us the answer there.

And so I think what the KOLs, that we're dealing with, and we have a panel of some of the leading lights across the world advising us, the thing that they're most excited about is the tolerability profile. And I think you see it in the data here. I mean this is a heavily pretreated patient population. We're seeing low rates of neutropenia, which is obviously being driven by atirmociclib's profile. You see no Grade 3 hyperglycemia here, which is obviously being driven by zovega's profile.

So the things that could potentially disrupt these patients in the frontline on chronic treatment are you're seeing low rates of. So I think that's the thing that people are most focused on when we talk to KOLs. I don't know, Don, if I missed anything.

No, I think you really hit. And the other part of that is not just keeping patients on the triplet, but being able to choose to use the triplet in a broad range of patients, as opposed to your very narrowly selected patients. And it comes back to the profile that we're looking to achieve, one where the overall toxicity burden on the patients and the physician for the triplet can be comparable to but not worse, potentially even better, than the existing doublet study we use in the frontline. And we think that's really the profile we -- to get both broad utilization and that shows superior benefit.

Our next question coming from the line of Chi Fong with Bank of America.

Thanks for the update. First one, maybe a follow-up. I believe you started exploring the palbo and ribo triplet earlier before atirmo came into the picture. I'm curious, can you talk about how the atirmo triplet data compared to palbo and ribo triplets? What are you most impressed with the CDK4 combo data in comparison, among efficacy, safety and dosing consistency or target coverage? Ultimately wondering to what extent you can claim that CDK4 can add to zovega triplet more so than a CDK4/6 could as we think about the Phase III comparison.

I think it's kind of more of the kind of high level that we've discussed, which is we're very familiar, there's a lot of data in the public domain around the CDK4/6 plus endocrine therapy in the front line. They have high rates of neutropenia, some of them have diarrhea. And so we know what that profile looks like, and we know what adding a PI3K inhibitor to that profile looks like because we've seen it done with palbo in the INAVO120 trial.

So I think coming back to what are we most impressed about, we're impressed with atirmo's ability to dial down the CDK6 related toxicities. And you see it in all the data that Pfizer has shown. It's an impressive compound.

And so then you stack that with our data, and you've seen plenty of data from our ReDiscover data that -- and you've seen it from our doublet data, that we can dial down rash, diarrhea, stomatitis. And also, most importantly is we have very low Grade 3 hyperglycemia. So you stack the 2 together and you get a profile that, for the first time, you could potentially put 3 drugs together and be comparable and in the zone of what you see with the doublets of CDK4/6 plus AI.

And so I think what we're looking for is comparable efficacy -- or sorry, comparable tolerability, but superior efficacy with this profile. And we think the way to get there is a selective-selective approach.

Got it. And my second question is I believe you have tested the -- both the 150-milligram, 100-milligram dose of zovega. How do the response rates compare between the 2 dose ranges?

Yes. We're still sitting at small ends, which is why we're pooling all the data together. As Don said, we saw responses starting at the lowest dose and then straight through to the higher doses, ultimately, even -- especially at the 150 mg. So everything has been relatively consistent to date. And we imagine as we get more experience at the 150 mg specifically over time and likely more pretreated patients, we'll continue to see that increase.

Just one last one from me. Does the first-line breast cancer development plan have any impact on your current cash guidance?

No. I mean, clearly, we -- over time to fund the entirety of the study, we will need more cash. But as of now, our current cash guidance contemplates the starting of the study.

Our next question coming from the line of Eva Fortea with Wells Fargo.

Congrats on the progress. A quick one from us. Can you provide more context on the Grade 4 events of neutropenia and hypokalemia seen in the study so far?

Yes. I mean I think the Grade 4 events have been very rare and well managed. So in general, Grade 4 AEs have not been an issue.

Our next question coming from the line of Silvan Tuerkcan with Citizens Bank.

Maybe a quick first one, what can you tell us about, at this point, about the nonkinase and kinase activity of this triplet? It seems like it's across both patient types. Is that true?

And then maybe a second question, just maybe big picture. Pfizer is developing this CDK4 in the FOURLIGHT-3 trial in the frontline. Did you think would there be only one regimen in the frontline, or would there be options between a doublet for Pfizer and the triplet? Or how could we envision this coming to market at some point?

So I think to take your first question around kinase and nonkinase, we see basically the same response rate across the mutations. And so it's a real true pan-mutant inhibitor that we have. And obviously, that goes on the back of the data we showed at ESMO TAT where we showed very similar PFS in the kinase and the nonkinase. So hopefully, we are very clear that this is a pan-mutant inhibitor.

In terms of your question around the broader strategy of Pfizer's commercial, I'll leave that to Pfizer to comment on. But all we can say is in the PI3K-alpha mutated patients, if we can show the data that we hope to show here, and obviously, both sides are excited about running the trial, it should become the standard of care for mutated patients.

Our next question coming from the line of Boris Peaker with JonesTrading.

Great. Just a question on dosing. So you have a 2.5x exposure increase from atirmociclib, and it looks like you took your 400 mg dose and you divided it by 2.5 to get about to 150 mg. I'm just curious, how consistent is this 2.5x PK interaction across different patient weight or metabolic profile? And also thinking like if a patient dose-reduced atirmociclib for some tox reasons or whatever, is there a protocol adjustment to increase the dosing of your drug to compensate for that?

It's very consistent in all the elements that you said. So that answers your first question.

On your second one, it's not really that dependent on the atirmociclib dose. And so we would not need to dose-reduce if the atirmo dose came down. So it should lead to a pretty simple dosing regimen for patients.

Got it. And maybe another question is on the medium follow-up that you've had so far is about 7.5 months. When will we see the mature PFS for the triplet combo?

At some point in the future.

And I'm showing no further questions at this time. I will now turn the call back over to Mr. Sanjiv Patel for any closing comments.

Thank you. Hopefully, you can see today, we're excited about using this selective-selective next-generation combination to provide a much better solution for the 40% of patients that are mutated in the frontline.

And so we look forward to talking with you over the coming days and hopefully seeing you all in Philadelphia at the ISSVA Conference on May 20. Thank you.

This concludes today's conference call. Thank you for your participation. You may now disconnect.

Hello, and welcome to today's Relay Therapeutics Corporate Webinar. My name is Don Bergstrom, and I'm President of Research and Development at Relay Therapeutics.

And I'm Brenton Mar, Head of Clinical Development.

And today, we're going to cover in detail the disease landscape of vascular anomalies where Relay is currently developing our mutant selective PI3K-alpha inhibitor, zovegalisib.

As a reminder, during this call, we will make certain statements that are considered forward-looking statements within the meaning of the Private Securities Litigation Reform Act of 1995, including expressed or implied statements regarding our strategy and business plans, the expected therapeutic and clinical benefits of our product candidates and progress timing and execution of our clinical trials. Such forward-looking statements are not guarantees of future performance, and therefore, you should not put undue reliance upon them. These statements are subject to numerous risks and uncertainties that could cause actual results to differ materially from what we expect. Please refer to our SEC filings on our website for a discussion of risk factors. The forward-looking statements in this presentation speak only as of the original date of this presentation, and we undertake no obligation to update or revise any of these statements.

Vascular anomalies is an umbrella term for a variety of conditions with a wide range of clinical presentations. There are many separately named syndromes and conditions that fall into this category. Their commonality is that they consist of malformed vasculature and in some cases, other surrounding tissue types. And they share common genetics, of which PIK3CA is among the most common driving mutations in vascular anomalies where patients seek medical therapy. Current treatment options today leave a large unmet medical need as local treatments like surgery and topical ointments only treat symptoms for isolated accessible lesions and nonselective systemic treatment options carry a toxicity burden that is unsuitable for chronic use and severely limits the ability to inhibit the disease driving mutation.

Zovegalisib is the only mutant selective PI3K-alpha inhibitor currently in clinical trials for patients with PIK3CA-driven vascular anomalies. Its mutant selectivity allows for greater target coverage while maintaining tolerability, which we believe could lead to improved efficacy. It may also be suitable for chronic treatment, which is important as this disease is known to recur once treatment is paused.

Let's start at the beginning with some basic biology. Vessels can be characterized as arteries, capillaries, veins or lymphatics. For the purposes of today's webinar, we'll focus on lymphatics and veins as they are most commonly affected by PIK3CA mutations. Lymph vessels, like blood vessels, branch through the body and collect lymph fluid from different parts of the body and return it to the blood. The lymph system is an important part of the immune system to protect the body from infection and disease. Historically, there have been observations of disorders involving these types of vessels. They are diagnosed according to their clinical presentation and visible phenotype and in many cases, treated as disparate conditions.

However, over the last 10 to 15 years, along with the rise of genotyping, it was determined that there are underlying genetic commonalities. Many genes in common were known for their roles as oncogenic drivers of cancer. It turns out that amongst the mutations that are driving the anomalies in veins and lymphatic vessels, PIK3CA is the most common driver mutation. Like in cancer, these mutations cause aberrant growth signaling, leading to disorganized vasculature and tissue overgrowth, which can include other tissue types that surround the vessels, ultimately culminating in the development of several types of vascular anomalies. To note, unlike in cancer, PIK3CA is the single somatic mutation in the affected cells. So features of cancer like genomic instability or acquired resistance mechanisms are not features of vascular anomalies.

To explain in more detail, in vascular anomalies, somatic mutations occur in utero, impacting different parts of the vasculature. These mutations are said to have somatic mosaicism as the body becomes a mix of cells that carry the mutation and cells that have an unmutated PIK3CA. The timing and location of the mutation event are very important for how the disease manifests and persists throughout a patient's lifetime. Mutations earlier in embryonic development can lead to broader involvement with multiple lesions across tissues and organs. Where the mutation occurs also impacts the disease. Even a single lesion occurring close to an essential vessel or organ can have a significant impact on organ function and cause serious clinical symptoms.

And ultimately, the phenotype depends on all of these underlying mutational dynamics and determines the symptoms that are experienced by the patients and the types of therapy that can be considered for treatment. These symptoms occur with a range of severity from more superficial cosmetic abnormalities to symptoms that impact daily life like pain and reduced mobility, up to life-threatening risks like bleeding, infection and impact to critical organ function. Again, these are all part of the same underlying disease, PIK3CA-driven vascular anomalies.

Today, we group these anomalies into the following buckets based largely on the type of vessel involved. The most common are isolated lymphatic and venous malformations. Then there are the syndromic conditions that impact multiple tissues, which are diagnosed clinically based on presentation and include: KTS, CLOVES, FAVA, MCAP and others. Collectively, these syndromes are called PROS for PIK3CA-Related Overgrowth Spectrum and unified genetically by having a PIK3CA mutation. And finally, cerebral cavernous malformations, which occur in cerebral blood vessels and are generally diagnosed later in life. We'll talk about these subtypes in more detail in the following slides.

Starting with the PIK3CA-Related Overgrowth Spectrum or PROS. PROS is an umbrella term for a collection of syndromes with vascular anomalies and tissue overgrowth caused by PIK3CA mutations. Symptoms can vary widely, but include bleeding, infection, loss of mobility and pain. The conditions are syndromic and are often more severe cases. These syndromes are primarily diagnosed clinically based on disease presentation. Studies have shown that essentially all patients with these syndromes carry a PIK3CA mutation. A confirmation of the mutation is generally not required for medical decision-making once the clinical diagnosis is made. Currently, alpelisib is the only approved systemic therapy for PROS, which is marketed under the brand name Vijoice. It has accelerated approval, which we'll cover in more detail later in the webinar.

Sirolimus, an immunosuppressive drug developed 30 years ago to help transplant patients to avoid organ rejection by suppressing a patient's immune system has also been used off-label in these patients for the past 15 years or so. Next is lymphatic malformations. Lymphatic malformations can occur throughout the body, but often occur in lymphatic-rich areas such as the head and neck region. Symptoms can include swelling, infections, pain, compromised airways, difficulties with swallowing, difficulties with speech and other functional impairments that can greatly affect patients' quality of life. They can be characterized as either complex, which are rare, but are systemic and usually affect multiple organs, bones or tissues or more commonly simple or isolated cystic LMs, which are, as they sound, usually confined to one area of the body. They are further characterized by the size of the resulting cysts, either micro, small cysts or macro, large cysts or mixed if there are both.

Our work with experts in the field suggests that the majority of LMs have mixed presentations, consisting of both micro and macro cystic disease. The vast majority, about 80% of lymphatic malformations are driven by PIK3CA mutations, leading to an estimated 65,000 PIK3CA-driven patients in the U.S. with lymphatic malformations. There are no approved systemic therapies for LMs, but our understanding is that alpelisib and sirolimus are used off-label in some instances.

Next is venous malformations. Like lymphatic malformations, these conditions are defined by abnormally formed vasculature. The lesions often have a bluish color and are comprised of distorted vein channels surrounded by disorganized extracellular matrix. Being part of the circulatory system, blood clots can occur, including pulmonary embolism, which can lead to significant medical complications and consumption of clotting factors, which in severe cases, can lead to bleeding complications. Pain and swelling can occur ranging from localized to diffuse. About 20% to 25% of VMs are driven by PIK3CA mutations, equating to about 25,000 patients in the U.S. There are no approved systemic therapies for venous malformations.

Finally, we have the PIK3CA-driven cerebral cavernous malformations or CCMs. These are abnormal clusters of dilated blood vessels in the central nervous system. These typically are diagnosed later in life when patients develop neurological symptoms secondary to the malformation. Depending upon the severity of symptoms and the risk of progression, primary treatment may include local control of the lesion with surgery or interventional radiology procedures. CCMs are not a current focus in the development of zovegalisib, but can be addressed in future trials. Across all of these subtypes, there are roughly 170,000 patients living with PIK3CA-related vascular anomalies in the United States. Because of the size and concentration of the PIK3CA mutations within the subgroups, the initial focus of our development program is in PROS and LMs. PROS with 5,000 to 10,000 patients with about 1/3 of patients seeking systemic treatment for their disease and LMs with about 65,000 patients with about 1/4 of patients likely to seek systemic treatment.

So with that, I'll hand it over to Brenton to talk about the patient journey.

Thanks, Don. Now let's take a look at the patient journey. I know It can sometimes get pretty complicated to get a proper diagnosis and a good care team. And I want to go over this, this is a generalization and not necessarily representative of every patient's health care experience. And in reality, because this disease is not well understood, patients tend to have a wide range of paths leading to their current therapy. It depends on the disorder, but generally signs and symptoms typically present in the first years of patient's life often at or even in utero, sometimes found on ultrasound. These are typically treated by local health care providers such as primary care pediatricians, dermatologists or even surgeons when intervention is required.

And in these settings, treatment is usually focused on the isolated symptoms. The correct diagnosis can often be delayed or misdiagnosed as many providers will not recognize these symptoms as part of a larger syndrome or broader underlying condition. Therefore, proper diagnosis can sometimes 6 years, often requiring referral to an expert provider who is very familiar with vascular anomalies to make a clinical diagnosis.

Over the past 10 years, there have emerged a number of vascular anomaly centers or VACs, which feature a colocation of experts and capabilities, collaborating closely across the entire diagnosis and treatment spectrum, focusing on both clinical care and research. The VACs include a multidisciplinary team with specialties like hematology, oncology, surgery, interventional radiology and dermatology, focusing on comprehensive integrated care for genetic testing, access to clinical trials and reimbursement support, which are not a given in this nascent nature and space. These VACs are often part of major children's research centers in the U.S., Europe and around the world.

Treatment of vascular anomalies varies based on the severity and the availability of effective treatment options. The vascular anomalies driven by PIK3CA, isolated LMs, VMs, PROS and CCM tend to have a more progressive lifelong symptoms and complications. Care teams may take a watch-and-wait approach initially and hopes that lesions will not show progression or become symptomatic, but this is usually only for milder presentations. Supportive treatment is used when symptoms develop. For example, compression garments can manage the venous or lymphatics following anticoagulation for blood clots or pain medication may be used for head and neck lesions, sometimes a feeding tube or even a tracheostomy can be placed to protect the airway.

In other cases, symptoms are localized and a local invasive therapy might be offered such as a sclerotherapy in which an interventional radiologist introduces sclerotic agents via catheter directly into the malformed vessel, in order to kill the abnormal cells where surgery can just frankly debulk or try to resect the lesion. But unfortunately, lesions often regrow, sometimes multiple times and multiple recurrent procedures may be needed. And also some lesions are just inaccessible or too extensive or in places that are just too dangerous to attempt the procedure. So for these reasons, many patients may be better served by a systemic therapy targeting the PIK3CA, mTOR pathway, the driver of these diseases.

As mentioned, more than 10 years ago, sirolimus, and mTOR inhibitor was first used off label here. More recently, alpelisib, a PI3 kinase inhibitor was approved in the U.S. for the treatment of PROS. And these first systemic therapies have showed proof of concept that the PI3-kinase, mTOR pathway inhibition could lead to benefit for patients and patients could take them chronically for years to control their symptoms. But because these agents nonselectively inhibit the wild-type PIK3CA and mTOR pathway, they must be given at low doses chronically to avoid unacceptable long-term toxicities in this population. Ultimately leading to suboptimal efficacy and leading patients to seek more effective alternatives.

Our goal is to provide a systemic therapy zovegalisib that selectively targets the mutant driver of the disease. Mutant selective targeting of PIK3CA may offer the powerful inhibition of the abnormal vascular cells at chronically tolerable doses, potentially leading to robust activity on lesion size and substantial improvement on patient symptoms. We believe that this has the potential to change the way vascular anomalies are treated with systemic therapy being considered earlier in the disease severity spectrum in concert with localized therapeutic options like topical surgery or sclerotherapy. And hopefully provide a safe and effective chronic therapy for as many patients as possible that could benefit from a systemic therapy for their disease.

To put in perspective, let's review a couple of case studies. Starting with a patient born of CLOVES syndrome, a subset of PROS. She presented at birth with multiple vascular anomalies and affected body parts and was initially misdiagnosed. She had multiple rounds of surgery and sclerotherapy in her first year of life with multiple recurrences, and she began sirolimus at age 6, which unfortunately was ineffective and ultimately discontinued after infection occurred, a common side effect of the immunosuppressant nature of the mTOR inhibition. Shortly after, a PIK3CA mutation was identified in the kinase domain, leading to the use of alpelisib, which resulted in a partial regression of the lesion.

In another case, we have a pediatric patient with a lymphatic malformation of the face and mouth, which was identified at birth. The symptoms severely reduced the quality of life in the first year, dealing with breathing and feeding difficulties and multiple rounds of hospitalization. Around 1 year old, sirolimus was used and showed some reduction in lesion volume, but ran into toxicity issues like stomatitis, infections and triglyceride elevation. And over the next 4 years, sirolimus was used intermittently to try and manage the lesion size while avoiding the worst of the expected side effects. While undergoing local treatment, including a partial surgical resection of the tongue, a PIK3CA mutation was identified in this patient.

So now we'll spend some time going over the current systemic therapy options. We've already mentioned a few of them in these patients as well, but I wanted to put them into context right now. The earlier generation of these therapies include the mTOR inhibitor of sirolimus, which targets mTOR, which is downstream of the PIK3CA mutation. And there are also nonselective PI3 kinase alpha inhibitors, serabelisib or KP-001 as well alpelisib. These drugs actually have been around for decades. Originally, they were both studied in patients with solid tumors in cancer. And they more recently got into looking at vascular anomalies. Only alpelisib is ultimately approved, with an accelerated approval in the U.S. for the limited indication of PROS. And we'll go through each of these in details, but they all share this common shortcoming, as I mentioned, about being nonselective and having toxicity related to the wild-type inhibition.

So starting with alpelisib, it was first approved in breast cancer in 2019. And shortly thereafter, a single investigator in France initiated a compassionate use study in PROS patients called EPIK-P1. This was not a typical registrational study by any means, a retrospective chart review from 8 sites with the majority of patients from his 1 site. No true dose optimization was done, and they used 250 milligrams once daily of alpelisib, which is slightly lower than the 300-milligram label dose for oncology. It supported a conditional approval for PROS in the United States in 2022. They then started the confirmatory trial in 2021 called EPIK-P2. Here, the PIs in the study lobbied for using a lower dose of alpelisib given the lack of previous dose optimization and toxicity concerns. And so the dose was halved from EPIK-P1 to 125 milligrams. The study read out last year and unfortunately failed to meet its primary endpoint of a lower bound of the response rate confidence interval, excluding 15%. The response rate in adults was only 17%.

The next confirmatory study, EPIK-P4, initiated late last year and has reverted back to the 250-milligram dose given the lack of efficacy for EPIK-P2. Vijoice remains on the market in the U.S. given the clear evidence that it has some impact on the disease. It's worth noting that Novartis is running another trial studying alpelisib in lymphatic malformations called EPIK-L1, which initiated in 2023.

Let's dig a little bit deeper into the results of the data from the EPIK-P2 study in which, as I mentioned, adults were started at 125 milligrams once a day and allowed to escalate. 54 patients were in this cohort. Initially, it was a randomized study up to week 16. And at the 16-week time point, you can see 11% of patients had a volumetric response, 0 on the placebo control. After 16 weeks, everybody crosses over to week 24, but we're just following the alpelisib patients right now. And you can see that it marginally came up a little bit to 17%. And then over time, they've updated the data. Again, it's still at 17% at week 48. So it has been relatively stable at that point. And as noted before, because of the confidence interval did not clear the prespecified 15%.

Looking at the maximum reduction observed at any time point on the EPIK-P2 trial, this is data that was presented last year at ASPHO, the American Society of Pediatric Hematology/Oncology. And you can see from the waterfall plot there, the median reduction was about 11% and the max reduction seen was about 46% in the trial.

Moving on to safety. Because the study was run at half the label dose, which clearly impacted efficacy, the reported safety profile is difficult to assess. We can, however, look at the safety data at the 48-week time point, where about 30% of patients stayed at the 125-milligram dose and about 70% of patients escalated to 200 or 250 milligrams once daily. Based on those numbers, you can approximate to an average dose of about 180 milligrams a day. And even with this average dose, we see the rate of severe adverse events at 30%.

Recall that the second confirmatory trial, EPIK-P4 is dosing back at the full label dose of 250 milligrams higher than the dose used in EPIK-P2. We don't have very much data from a prospective trial and how alpelisib performs in vascular anomalies patients at these doses. However, while it's a much different patient population, we can look at the AE profile of alpelisib in EPIK-B5 in breast cancer, where the starting dose was 300 milligrams and the estimated average dose was roughly 220 milligrams. Here, there's clear evidence that higher doses of alpelisib bring a difficult tolerability profile with high rates of Grade 3 hyperglycemia, diarrhea and rash. And importantly, in this metastatic breast cancer population, where tolerance for toxicity is presumably much higher than it is in a chronic use setting. In sum, we feel there is meaningful unmet need for these patients. Zooming out to all systemic therapies for which we have data in this disease space, when we look at the efficacy of all these therapies, it leads us to believe, like we did in oncology, that there is an unmet need for a more efficacious therapy.

Now let's move to a summary of zovega and why we believe it could address the unmet need for these patients. Zovegalisib is the first mutant selective PI3 kinase alpha inhibitor to enter the clinic and was discovered at Relay using our Dynamo platform. Relay scientists discovered a novel allosteric pocket that was key really in being able to gain not only a mutant selectivity for PI3 kinase alpha, but also exquisite selectivity over the rest of the kinase as you see in the first figure. Zovega has been studied in hundreds of breast cancer patients at this point and has the selectivity profile that has resulted in encouraging efficacy and safety. And we hypothesized similar potential in vascular anomalies as you can see in this in vivo model on the right, which I'll go into more detail.

Here is preclinical data, which we have previously presented, supporting our hypothesis that greater target coverage can lead to greater lesion shrinkage without impairing glucose regulation. The graph on the left compares the size of PIK3CA-mutated vascular malformations in mice that received either a dose of alpelisib approximating the approved dose in pink or 2 doses of zovega in blue, one approximating the oncology Phase III dose and half the dose. The alpelisib treatment modestly decreases the malformation size. But on the other hand, each dose of zovega reduced the size of lesions more than alpelisib, which suggests that it can potentially be more effective in people.

Now shifting to the graph on the right, which looks at insulin induction in this model, zovega induced insulin to a lower degree than alpelisib did even when a higher dose was tested. So the ReInspire Phase II clinical trial was initiated in Q1 of 2025 and is evaluating zovega in eligible individuals living with PIK3CA-driven vascular malformations, which includes PROS, lymphatic malformations and other subtypes. Part 1 of the study works for 3 different doses of zovega randomized in patients 12 years and older, including adults, and that's 400 milligrams twice a day, which is the oncology Phase III dose and 2 biologically active lower doses, 300 milligrams and 100 milligrams twice a day.

The initial clinical focus will be in adults and adolescents 12 years and older. However, upon reviewing the safety, tolerability, PK and initial clinical activity, we are excited to announce that as of this quarter, we have opened up a pediatric cohort for individuals 6 to 12 years old. As we gain more experience in this patient population, we'll then select one or more doses per cohort to bring into Phase II expansions. And the protocol is also written to contemplate a randomized placebo-controlled study in Part 3 should we choose to pursue a study. The regulatory endpoint used in alpelisib accelerated approval and being used in other studies as a response rate defined as a 20% or greater volumetric reduction at week 24. Additionally, patient and investigator-assessed outcomes will also be evaluated, and Relay is currently in the process of validating a fit-for-purpose patient-reported outcomes tool for use in the ReInspire trial for this patient population.

And with that, I'll pass it over to Don.

Thanks, Brenton. Vascular anomalies have a very concentrated prescriber base with less than 100 vascular anomaly multidisciplinary centers across the U.S. and Europe. Most systemic therapy is expected to take place within the coordinated care of these vascular anomaly expert centers. The ReInspire study is potentially registration enabling. Currently, the initial clinical focus is on randomized dose finding and then additional data at recommended Phase II doses will be generated in expansion cohorts. Given the regulatory precedent with alpelisib and the fact that no systemic therapy currently has full FDA approval, there may be potential for an accelerated approval pathway.

So in summary, there are approximately 170,000 U.S. patients with PIK3CA-driven vascular anomalies. Current treatment options are limited with local treatments only addressing a subset of symptoms and systemic therapies with off-target toxicities that limit efficacy. Zovegalisib selectively targets the PIK3CA mutation that drives the disease and therefore, may provide a more tolerable and efficacious chronic treatment for patients with PIK3CA-driven vascular anomalies. Zovegalisib entered the clinic for vascular anomalies in early 2025 and continues to actively enroll patients. Thank you for watching.

Hello, again, everyone. My name is Etzer Darout, one of the senior Biotech Analysts at Barclays. It's my pleasure to have our next fireside chat with Relay Therapeutics. With us this morning, we have Sanjiv Patel, CEO of Relay; and Peter Rahmer, Chief Corporate and Development Officer. Thank you so much for joining us this morning.

Maybe, Sanjiv, just to get us started, maybe a brief overview of Relay Therapeutics for those maybe less familiar with the story.

Yes. Well, first of all, thank you, Etzer, for the invite and thank you to Barclays. So it's always a pleasure to be here. So the company has been around now close to 10 years. We're a company that's focused on trying to drug precision medicine targets and we have a very exciting year ahead. It's been set up because over the last couple of years, like every company in our industry, we've weathered the very tough environment biotechs have been through, and we weathered in our own way by taking our lead program, which is an FGFR2 inhibitor and out-licensing that.

We very much focused our research organization down. We staggered some of our preclinical programs, and we focused the majority of our resources on our PI3K-alpha inhibitor, which we believe is the first mutant selective inhibitor to enter the clinic, zovegalisib on 3 big opportunities. Those 3 opportunities are hormone receptor positive, HER2-negative frontline and second-line metastatic breast cancer and then PI3K-alpha-driven vascular amalgamations. All 3 are, we believe to be very large commercial opportunities. And over the coming year, we're going to show 3 data sets.

First of those data sets is in second-line metastatic breast cancer, which next week at ESMO TAT, we'll show data from our pivotal dose, 400 milligrams BID fed. All the data we've shown to date in the public domain has come from the 600 milligrams BID fasted cohorts. And so this will be the first time we'll show this data. It's 57 patients worth of data. There should be enough follow-up to assess both response rate and PFS.

And the goal is to show consistency. We've in the past shown 10 months' worth of PFS in this patient population, where the standard of care is at 5.5 months of capivasertib. So if we can go some way towards showing consistency, that should give investors great comfort that we will be able to run a very successful Phase III trial that's ongoing with ReDiscover-2 trial.

The next area is frontline metastatic breast cancer, where in 2026, we'll show data on which frontline regimen we want to take forward. Currently, we're testing 3 different regimens using both palbociclib, ribociclib and Pfizer's selective CDK4 atirmociclib. And so we will declare which of these regimens we want to take forward.

Our prioritization is going to be made on tolerability. We're looking to get the most tolerable regimen we can. The efficacy, we believe, has been shown by previous trials like Roche's INAVO-120 that a PI2K alpha inhibitor in the frontline is going to add efficacy. They've obviously shown overall survival benefit. The challenge really is getting a triplet regimen that shows tolerability.

And so we hope for sharing which regimen to go forward with and what that trial design is. And then the third area is vascular anomalies. This is an area that's very poorly kind of understood by most investors. But it's a very large commercial opportunity that's unserved. It has a few approved therapies, most notably alpelisib from Novartis has an accelerated approval in a sliver of this patient population PROS.

And what we would like to show is in the first half of this year, 20 patients' worth of efficacy data. And that efficacy in this population is measured through volumetric MRI reduction of lesions. To get a response, you have to have a 20% reduction in lesion volume. The bar has been set by Novartis in their accelerated approval trial at 27% response rate in their confirmatory trial and in the -- obviously in the accelerated approval trial in their confirmatory trial was at 17%.

We'll show data at a slightly earlier time point, 12 weeks in these 20 patients, and we hope that we can show proof of concept here and then move rapidly to next steps. We have about $550 million of cash at the end of last year. So we're well set up over the coming years to generate value for shareholders in these 3 very large commercial areas.

Great. Maybe just for our listeners to just maybe kind of put into context the role of mutant PI3 kinase in breast cancer. Obviously, it's a space that's rapidly evolving and where you sort of see the value creation for having a mutant PI3 kinase selective PI3 kinase in these populations?

So as you know, the hormone receptor positive HER2-negative breast cancer market is very large. 40% of those patients have a PI3K-alpha mutation. And it's a foundational mutation, it's truncal. It's present at the diagnosis. It's not something that emerges through therapy like ESR1 mutations.

And so we've always believed that the Precision Oncology has shown us as well that if you just target the driver, which in this case, we believe the driver is PI3K alpha, you should be able to see efficacy. And we've seen that time and again in successful medicines in our industry. For decades now, our industry has tried to make a mutant selective inhibitor. And the challenge of the previous generations is that they either hit the isoforms of PI3K alpha, gamma delta and just alpha or they hit the wild type as well as the mutant.

And this leads to a range of off-target toxicities, most notably hyperglycemia, diarrhea, rash, stomatitis and so although the efficacy is clear, unfortunately, the tolerability of these prior generations has been challenging. Patients just can't stay on therapy, and that leads to reduced dose intensity and that unfortunately leads to poor efficacy.

And so the goal of a mutant selective inhibitor is can you hit the target hard and make sure that you get the efficacy that patients need, but remove some of the off-target toxicities to make it a tolerable regimen. And we believe that's exactly what we have achieved with zovegalisib.

And so the data we've shown to date shows that we have very low grade 3 hypoglycemia, diarrhea, rash, dermatitis. And so we've kind of moved away from some of the previous generations challenges. And that in the second line has shown a PFS in the data that we've shown of almost 2x what we see in the standard of care. And we believe that will also translate into a frontline regimen that should be tolerable because you're removing some of the toxicity.

And obviously, in the vascular anomalies as well, these patients are going to be chronically on therapy. And so again, the vascular anomalies is driven by PI3K-alpha mutations. So if you can cover the mutation above the IC80 or even close to the IC90 for 24 hours, you should be able to see the efficacy.

But given these patients are going to be chronically treated for many, many years, the tolerability is paramount. So that's why the industry has always wanted a mutant selective inhibitor, and I think we are the first to deliver it.

Great. And I think the data abstracts, I believe, will probably be published tonight for your upcoming presentation. What can we expect, I guess, would we get more data at the presentation relative to what we see at the abstract? Maybe if you could also remind us what we can expect to see during the presentation.

Yes. So you're correct. The abstracts are slated to come out tonight around 6 o'clock or so. The abstract will contain the top line data from an earlier time point. So we will have PFS response rates and the PK data because recall, what we're largely doing here is trying to show this translation of the 600 bed dose, the 400 -- the 600 faster dose, 400 fed dose. So it will contain all of that.

And then the data presentation itself on Monday will have a later time cut, but a lot of the same headline numbers. I think you'll see the key takeaways. And again, the goal of these data is to show consistency from the 600 faster data.

Right. Great. And obviously, we feel a lot of questions around comparisons with Celcuity's molecule. And maybe just your thinking about sort of the differences between those molecules, those programs and then also the data that they presented to date relative to what they could show from that new population, which people are anticipating greatly from them.

Yes. I mean, as we talked about, we have a very kind of targeted therapy, which hits PI3K-alpha and has selectivity between the mutant and the wild-type and essentially leaves everything else alone. I'm not saying the biology of the cellcuity molecule, gedatolisib, it has broader coverage and hits multiple nodes in the pathway.

And they have obviously shown great data in the wild type and obviously moving very fast towards an approval there. And that's great for the field because in the wild-type, therapies just aren't providing the efficacy that's needed, and this will be a step change for patients.

I think in the mutant data that's upcoming, the real challenge is just going to be -- it is an IV therapy weekly. It does have a range of toxicities that do come from having multiple nodes hit in the pathway. And so the challenge is just going to be how good does that data have to be above both the standard of care today.

And so 5.5 months is the capivasertib bar, but beyond an oral therapy that could emerge in the field like us in and around the 10-month PFS and how high does the efficacy have to be above that to justify the IV and the toxicity that goes with this.

And I think, obviously, the commercial markets will provide the answer to that over time. I think what we can do is provide a clean as we can, tolerable therapy with as high PFS as we can that's convenient to patients that has an oral route of administration. And we hope that all the work that we've done with both physicians and patient groups, that's a very attractive profile to have.

And in terms of the enrollment of your Phase III ReDiscover-2 trial, maybe any commentary around how enrollment is progressing? And perhaps is the trigger for sort of the data timing would be maybe a completion of enrollment? How are you kind of thinking about when you disclose sort of the data timing?

Maybe I'll take the first part and then maybe, Pete, you can take a bit about guidance. We're very happy with how the trial is going. And so obviously, we started it late last summer. We're setting up sites across the world. We're enrolling patients now in every region in the world. There's great enthusiasm given both the strength of our data and the trial design itself. Obviously, we're using an active comparator in capivasertib, which given the strength of the launch of capivasertib, it's a very attractive proposition, especially in countries where capivasertib is not reimbursed.

And so -- and given the kind of various trials that are out there, most of the trials in this field in the mutant population in the second line have kind of completed their enrollment. So we're not kind of up against headwinds in terms of other trials. So we feel very positive about how things are going in terms of the guidance, maybe I'll hand over to Pete.

Yes. I think once we get the -- all sites up and going globally and have a good sense of the steady state enrollment curve and pace of enrollment, we'll endeavor to try to give maybe some guidance to when we could reach full enrollment or a good range of when that might come. And then these are event-driven studies, PFS is the endpoint.

The control arm does in around 5 to 6 months. So typically, in precedented studies, the top line data comes not too far after that full enrollment period. But I think that's kind of the kinetics of how we'll try to keep folks updated on the study.

Great. And then for the front line, maybe the types of data disclosures we can expect to ultimately help decide, right, which is the right CDK4 selective CDK4/6, how are you thinking about that and what that data set could ultimately look like that help drive that decision?

Yes. Obviously, as we talked about, we are doing 3 different triplet ongoing work. The majority of that work is happening in second-line patients because that's the patients that we could most rapidly access. And obviously, we are looking for the tolerability profile above all else. We know that the efficacy of the triplet will play through. And so the disclosure that we'll make at some point in 2026, we'll probably focus on the go-forward regimen and it will show the -- obviously, the demographics of the patients that we have. These are going to be heavily pretreated patients with multiple lines of therapy, the tolerability profile and whatever follow-up we have at the time.

It's probably going to be too early to assess PFS in these patients. But our thesis is very clear, which is that side of it has already been proven. We're trying to figure out now which is the most tolerable regimen that patients could take 3 to 5 years' worth of therapy. And then obviously, we will share the next steps in terms of the go-forward trial design that we hope to run in the front line. But there's a lot to try and communicate here to investors when that disclosure is made.

Obviously, we need to show all the kind of precedent comparative data that we're going to try and make the decision against, have investors kind of understand how we came to the decision and then obviously share the next steps. So it will be a fulsome disclosure with lots. And so we'll probably have several bites of it over time.

Yes. And that will be the first step is to disclose those 3 key pieces and give a sense of timing, and then we can come back at it as we get closer to the launch of that actual study. And then that data we're generating in the Phase I/II setting with the triplets, that will mature over the year -- over the next few years as we get the frontline trial up and running. And so that will be able to produce folks continued insights into the profile of the triplet regimen we choose.

And one thing I haven't taken a look at is this sort of PI3-kinase expression levels prior to sort of a front line. So like an adjuvant setting, for example. I mean, obviously, this would be something much longer term. But is there an opportunity to have combinations in the adjuvant or early line settings of breast cancer?

Yes, absolutely. As Sanjiv mentioned upfront, these are truncal mutations, the foundational to the disease. And so that onset in the adjuvant setting, you're going to have 40% of the patients with a PIK3CA mutation. We do think it's a driver event in those patients. And so as you take a step back and look at the big picture of what the opportunity is for Zovega in breast cancer, certainly, we have near-term plans for the ongoing, the second-line metastatic disease in HR-positive, HER2 negative. We have frontline plans you can absolutely contemplate adjuvant opportunities in HR-positive HER2-negative disease.

And then I think over time, we'll think about even moving into other breast cancer settings such as HER2-positive disease or even triple-negative disease. There's a high frequency of these PIK3CA mutations, who require probably different combinations with different agents, but certainly an opportunity to help address those patients, too.

Great. And maybe switching to vascular mal formations, and you talked about it briefly earlier. Just curious where are the majority of these patients treated in the community setting? Do they need to travel to specialized treatment centers?

Yes. So I would say just to start at the broader picture with vascular anomalies, there's about 170,000 of these patients with PIK3CA-driven disease in the United States from a prevalence standpoint. The vast majority of these patients exist on a more mild spectrum of the disease. They probably are sit and being treated within their primary care physician or dermatologists because there's some -- in the more mild presentation of the disease, it's more of a cosmetic cutaneous presentation that doesn't require more aggressive intervention for the disease.

I think the context of the disease that we're thinking about is the more moderate to severe. And there, you have -- there's a subset, where alpelisib has accelerated approval right now. It's called PIK3C-related or spectrum. So it is a syndromic form of the disease that involves usually multiple limbs and/or tissues or critical organs. That's where you'll see the bulk of our initial data is in those patients.

Those patients tend to find their way to what are called vascular anomaly centers. There's probably 2 to 3 dozen of them in the United States, another 2 to 3 dozen in Europe, where the vast majority or at least our understanding today is vast majority of the moderate to severe patients make their way to those types of centers. They tend to overlap with children's hospitals because this is a disease that is caused through a mutation that happens in utero.

So these kids are born with this disease and these lesions start to kind of grow with the child as they grow. And so they tend to find their way early on to pediatric hematologists, oncologists in the more moderate to severe presentation.

Great. Yes, I was going to lead to my next question is sort of you recently started a pediatric cohort of patients, and that, I guess, makes sense from that standpoint because I guess the breast -- majority of diagnosis probably happens when these children are so young. Maybe for the data that you're going to be disclosing in the first half, what type of data disclosures that we could expect from that -- the cohorts of patients that you're evaluating?

Yes. So just to maybe level set the study we're running because of the breadth of experience we had in oncology, when we went into vascular anomalies, we were able to go into a dose randomization study as opposed to a traditional dose escalation. So we're randomizing against 3 different doses of Zovega. The highest dose is the RP3D in oncology, so 400 milligrams BID, then 300 milligrams BID and 100 milligrams BID.

So a good spectrum of target coverage. And it will be randomized. So it will be about an equal number of patients across each of those doses. We said that the disclosure we'll make in the first half we will have at least 20 patients that will be efficacy evaluable.

And what we mean there is they will have reached at least that first 12-week MRI time point. Safety database will be larger than that, but at least 20 patients of efficacy evaluable. It will be largely PROS patients, the PIK3CA related or spectrum patients with a spattering of lymphatic malformations and venous malformation patients.

Got it. And when we think about, I guess, potential pivotal design assuming success here in these early clinical trials, how should we think about that ultimately?

So the current precedents here from alpelisib, again, similar to the approach that we're taking where they took their nonselective molecule in oncology that they try to apply here. That drug has accelerated approval right now with -- it was based off of retrospective chart review of 37 patients treated under compassionate use kind of speaks to the level of unmet need in these patients.

And then the current -- they failed their first confirmatory study, EPIK-P2, and they have a new confirmatory study ongoing.

That is a single-arm study, 105 patients. So we believe the accelerated approval pathway, depending on our data and interactions with the agency would be available to us here. And the bookends, if you will, are the 37 patients currently that gave alpelisib their accelerated approval and the confirmatory study of 105 patients. It will be data and regulatory interaction dependent, but that's what's the press release today.

And I guess that first sort of miss study, is that -- is this sort of they're attributing that for alpelisib to design or maybe the types of patients that they -- are any learnings from that, that you can use as you sort of think about your pivotal design?

Yes. So their accelerated approval there's no dose optimization because there's just patients treated on compassionate use. So they just took the oncology dose and backed off a little bit, so they used 250 milligrams once daily. Because of that, when they went into the confirmatory study, the physicians on that study required them to go to half that dose because they were concerned about safety and tolerability. And so they went from 250 down to 125 in the confirmatory study, saw about half the efficacy.

Actually quite encouraging for us to see that because we can see a clear dose relationship in terms of target coverage and dose intensity. And so we know from our oncology experience and everything we've done with the PK profiling against alpelisib, we can sustain much higher mutant target coverage than alpelisib can throughout the day.

And so all of these pieces give us a lot of confidence as we move into our study that we should be able to -- be able to demonstrate far superior efficacy with a better tolerability profile.

And have you had conversations with physicians around their experience in alpelisib and then ultimately, what that could translate for you?

We've talked to physicians globally about their experience with alpelisib. We've seen -- there's been some publications, another -- a number of case studies coming out of these academic institutions. It's clear that there's a tolerability profile issue on the table for what is limiting the full potential of the target.

Great. And maybe taking a step back, some of the run rate conversations that you've had with cash. And ultimately, when you think about sort of the studies that you have ongoing in hepatic malformation as well as breast cancer. ultimately, where does that cash take you from sort of a readout across these 2 pillars?

I mean what we did over the coming last few years has taken us our cash into 2029. So that obviously fully contemplates executing the second-line metastatic breast cancer trial. It goes a long way towards -- obviously, we don't know the exact trial we will run yet in the vascular anomalies, but it should take us a long way towards getting that towards the top line data.

Obviously, we have money for moving our NRAS and Fabry assets forward too as well as running research. And so all of that is contemplated in the current cash that we have. The thing that's kind of sitting outside the cash window that we have is obviously running this frontline trial. But we believe we have a lot of catalysts ahead of us, the 3 that we've talked about and the data that we're about to show. We'll have ongoing updates from all 3 of these areas.

And so we do think we have a lot of levers as we start to think about how do we path from here towards having 3 very large commercial opportunities in 3 different Phase IIIs all executed. Hopefully, we'll have revenue in the back half of this decade. So we think we sit in a very nice position as we kind of execute all these opportunities.

Great. It looks like we're up on our time. Sanjiv, Peter, thank you so much for your participation. But thank you for our listeners, and we'll be back soon with our next session. Thank you.

Great. Thanks very much.

Welcome once again to the 46th Annual TD Cowen Healthcare Conference. I'm Yaron Werber from the TD Cowen Biotech team. It's a great pleasure to moderate the next fireside chat with Relay Therapeutics. With us today, we have Sanjiv Patel, President and CEO; to his left, we've got Peter Rahmer, Chief Corporate Development Officer; to his left, Don Bergstrom, President of R&D; and with us today, Jaena Han as well from our team. So I think we have a few slides to go through, and then we'll actually start the presentation thereafter in Q&A on vascular malformations, which is a new and important new area, and then we'll talk about breast cancer.

Excellent. Well, thank you, Yaron. Thank you to Cowen for the invitation. We're excited to talk about the exciting year ahead that we have. It's been set up because over the last few years, we've built the foundation. As you know, we've all lived through tumultuous times, and we made some very deliberate choices as a company to get through those so that we could focus all of our precious capital on advancing zovegalisib, which is our PI3K mutant selective molecule.

And so to do that, we obviously focused our research organization. We staggered the entry of our preclinical programs that focused on NRAS and Fabry. And we out-licensed what was at the time, our most advanced clinical program, lirafugratinib. All of that extended our cash runway into 2029 and allowed us to focus our effort in 3 very large clinical areas, second-line hormone receptor positive HER2-negative breast cancer, first-line metastatic hormone receptor HER2-negative breast cancer and then PI3Kn-alpha-driven vascular anomalies. All of these are very large commercial opportunities. And our goal with the capital and resources we have over the coming years is to deliver these medicines to patients and make sure that we can meet these unmet needs. Over the coming year, we're going to share 3 different data sets that will hopefully derisk meaningfully, these 3 opportunities in the minds of our stakeholders, including our investors.

And so let me go through those 3 disclosures that we hope to make. The first one will come in our second-line metastatic breast cancer area at ESMO TAT in Paris in a couple of weeks' time. Up to date, we've shared data from our 600-milligram BID fasted cohorts, which have shown a PFS over 10 months and response rate close to 40%. As you know, a couple of years ago, we did a food effect study, and we wanted to dose our pivotal trial using food.

And so we selected an equivalent PK dose, 400 milligrams BID fed. And so the data from this disclosure will come from 57 patients who have had 400 milligrams BID fed as their dose. So the same dose that we'll use in our Phase III pivotal trial, ReDiscover-2. So the goal would be that it would show consistency with the data that we've already shown and meaningfully derisk our Phase III trial. On that note, we're kind of heads down executing. We're enrolling across the world, and we're very happy with how things are going.

In our frontline approach, what we'll share this year is hopefully the design of our trial as we think about how to penetrate this very large market and the go-forward regimen that we hope to use. And then in the new area that we'll talk a lot about, I'm sure, today, around vascular anomalies, where there is an agent with accelerated approval where they have shown a 27% response rate in their accelerated approval. And then in the kind of failed confirmatory trial, they showed an 11% response rate at 12 or 14 or 16 weeks, I think 16 weeks.

What we'll show is 20 patients' worth of data at the 12-week efficacy endpoint. And what we hope to show is meaningful differentiation from that. And hopefully, that will provide proof of concept for this mechanism. And then hopefully then, we'll be able to share next steps. So the goal of the year is to meaningfully derisk all 3 of these very large commercial opportunities and then for us to continue our journey towards getting these medicines to patients.

So with that, maybe I'll turn it over to Q&A.

Yes. So I'll start us off. Clearly, vascular malformation and vascular anomalies are really coming into focus for you guys this year. Let's kind of zoom out a little bit. In Q1, you started your Phase II ReInspire study -- in Q1 of last year, you started your ReInspire study of zovegalisib in adults and children with vascular malformations. Could you provide an overview of kind of this indication as a whole and why vascular malformation should be amenable to treatment with a PI3K-alpha inhibitor?

Yes. Maybe I'll start and maybe, Pete, you could take it. But obviously, the condition is driven by PI3K-alpha mutations. And we've seen proof of concept with the nonselective PI3K-alpha inhibitor alpelisib and then the nonselective kinase inhibitor sirolimus in these patients. So we know that if you can inhibit this pathway, you do see responses.

And so what we've been able to show in breast cancer is if you can [ mutantly ] selectively inhibit the pathway, try to reduce some of the off-target toxicities, you can achieve [ greater target ] coverage, and we see meaningfully increased PFS. We hope to see the same thing here in vascular anomalies to see meaningfully differentiated response rates. And so that's why we feel very confident that this is an area that we can be successful in.

Yes. So among vascular anomalies, we have 3 kind of key subtypes. We have the PI3K-related overgrowth spectrum, lymphatic malformations and venous malformations as well as cerebral cavernous malformations. Are there differences between the subtypes and how sensitive they are to PI3K-alpha inhibition?

We don't think so. I mean, obviously, we have to run the experiment. It hasn't been a robust data set produced with the molecules that we have. And so we'll run the experiment. We think a priority, there is not.

And then how do you expect enrollment of between these kinds of subsets in your Phase II study?

Yes. In the ongoing Phase I/II study ReInspire and in the initial data disclosure, we imagine the bulk of the patients will be PROS patients. They tend to be the most severe patients. And because that's where you have alpelisib with accelerated approval, those patients are actively seeking treatment and are discussing systemic treatment options with their physicians. But over time, we think that we will be -- we will -- well, one, we'll have some lymphatic malformations and venous malformations in the study. And over time, the distribution will probably grow.

Right. And if we zoom out a bit, you say that PROS is the most kind of common. But in general, overall, what the prevalence of VM and what percent of that would be these PROS patients?

Yes. It's a good question. So at the highest level, there's about 170,000 PIK3CA-driven vascular anomaly patients in the United States on a prevalence basis. And then if you narrow down to the 3 subtypes that are most likely to be in our study over time, which is PROS, LMs and VMs, that's about 100,000 patients. 5,000 to 10,000 of them are PROS, 60,000 to 65,000 are lymphatic malformations and then about 20,000 to 25,000 in the venous malformations. And we think of that 100,000 of the cumulative -- those 3 subtypes, we think about 25,000 of them would seek chronic systemic therapy to treat their disease. So you can think of that to correlate with the severity spectrum of these patients.

We believe that today, based off of the understanding the -- what's driving uptake of both alpelisib and PROS and then off-label and then sirolimus use, you're sitting at about 25% of those 3 populations seeking systemic treatment. The goal would be to present a therapeutic intervention option like zovega that could expand that patient population over time.

Great. And if we can just double-click on kind of standard of care. You've mentioned alpelisib and sirolimus. Kind of, what options do patients have? And what does the typical patient journey look like?

You want to take it, Don?

Yes. So as Pete mentioned, there's really a continuum of severity in these patients. So this is largely -- these are largely congenital disorders. Patients are born with them, frequently diagnosed in childhood. Patients with quite severe manifestations, so who have either large lesions or multi-tissue lesions. Those are the more severe cases. They tend to be picked up a little bit earlier. These are patients who were referred into these centers of excellence, multidisciplinary centers of excellence, of which there are roughly 30 or so in the U.S., usually children's hospitals.

And they will undergo a number of either procedures. So they could have surgery, they can have sclerotherapy. So essentially, when you have toxins injected into the lesion, other types of interventional radiology procedures. But eventually, a large proportion of these more severely impacted patients will look for systemic therapy because the local therapies, whether it be surgery or sclerotherapy, both -- eventually, a multifocal disease, you can't control the disease anymore. And invariably, after either an excision or after sclerotherapy, the lesions will grow back, right? So these are patients frequently who will end up looking for systemic therapy.

There are then patients with less severe disease who especially may present with skin disease. Frequently, it will be seen by dermatology. Frequently may go years with misdiagnosis, until eventually, they'll see somebody who will put the pieces together. And so really, the disease is quite heterogeneous in terms of the way it presents and quite heterogeneous in terms of the experience patients have through their diagnosis and treatment.

Right. And then can you remind us what you've disclosed so far on the trial design for Phase I/II ReInspire? Kind of in terms of target enrollment, primary endpoint, what doses you're evaluating, et cetera?

Yes. So we were able to leverage when we started the ReInspire trial. We're able to leverage the fact that we had treated at that point, about 200 to 300 solid tumor oncology patients with zovega. So consequently, we had a good idea of what the safety profile looked like as a function of dose. So as we started ReInspire, we didn't have to go through sequential dose escalation in the new patient population. We're able to open randomized dose optimization in patients 12 years and older, where we are randomizing patients across 3 different dose levels of zovega with the top dose level being 400 milligrams BID fed, which is our oncology recommended Phase III dose. And then we're looking at 2 doses below that, 300 milligrams BID and 100 milligrams BID.

So these are all biologically active doses that sort of span the range of target coverage. And we'll have 45 patients who are randomized 1:1:1 across those 3 doses, so 15 total. We then, once we finish that part of the study, can move into cohort expansion. So in the same patient population, but now with what we learned from the dose randomization, we can move forward with either a single recommended Phase II dose or a couple of recommended Phase II doses. We'll continue enrolling in a single-arm fashion there with an ORR endpoint. We feel that there's still the potential opportunity, given the fact that there are no approved therapies for these patients, to have an accelerated approval path forward. And we'll obviously need to talk to health authorities, but it's possible that the data we generate in the single-arm cohort expansions could be used to support an accelerated approval application.

At the same time, as I mentioned, this is a congenital disease. So everything I mentioned is for patients 12 and older. We have recently opened dose escalation in patients between 6 years old and 11 years old. That will be weight-based dosing. We will initially dose patients at a dose that we think should give us exposure comparable to what you see at the 100-milligram BID in the adolescent and adult patients. And then with time, we have the option of opening as well, a 2- to 5-year-old cohort.

Right. And so even though ORR is kind of the primary endpoint for potential accelerated approval, when you think about vascular anomalies as a whole, what do you think is the most clinically relevant thing? Are we looking just at ORR, at DOR, PFS? And then kind of what is the bar that you want to meet here? Do you want to just be better than [ Vijoys ] was in their pivotal trial? Do you want to be better than their failed confirmatory? What are you looking for here?

Yes. So the precedented regulatory endpoint here is ORR. So it's measured by volumetric MRI. And what the FDA has looked for a few different times here is these patients generally are scanned about every 12 weeks. And so they're looking for your confirmed objective response rate, so 20% reduction or greater by volumetric MRI in a confirmed response. And the bar in EPIK-P2 was to exclude 15% in the lower bound of the 97.5% confidence interval. So that is at least the currently defined regulatory hurdle.

And I think our -- and like you said, the bookends of the best that Apellis has done to date is 17% in their failed confirmatory study, 27% in their accelerated approval study, which, to remind everyone, was retrospective chart review of 37 patients treated on compassionate use. So by no means, a traditional clinical study. And then in their failed confirmatory study, the response rate at the most comparable time point that we'll be able to show in our initial data, it was 11%. So ultimately, we want to show a profile that's an approvable drug. And it's clearly, we want to beat the 11%. And ultimately, over time, we want to make sure that we have a data set that could clear the regulatory hurdle.

Right. And then on the safety side, what kind of profile do you think would be acceptable given that kind of these vascular anomalies are benign?

I mean, I'll start, and then Don. I think you characterize -- we all characterize these types of lesions as noncancers or benign. But the symptom involvement and the burden of the disease on the patients is far from benign. And so the safety and tolerability will be extremely important because the goal is to treat these patients chronically, starting in childhood. So it will be a critical component of how we assess what dose to ultimately bring forward. But we're also going to make sure that we are choosing a dose that has the best risk-benefit profile to allow patients and physicians to make their best treatment decisions.

As folks that are understand rare diseases, there's commonly a lot of decisions being made by physicians and patients on -- around dose. And like if you look at any of the old sirolimus studies, there's constant moving around of the dose being used in those patients. So we'll kind of put all those factors together. We definitely want a safe and tolerable drug that can be dosed chronically, but also one that offers the maximum benefit for patients over time.

So do you see a similar profile to kind of what you've seen in breast cancer with single-digit rates of Grade 3 hyperglycemia and stomatitis, do you think that's good enough for vascular anomalies?

I think it's really just going to come down to the risk-benefit profile, so like what kind of response rate you're going to get for that kind of tolerability profile. And the totality of the data will tell us what the best trade-off is going to be.

Makes sense. And then also, how might zovegalisib be differentiated from Palvella's QTORIN rapamycin, which is also in development for lymphatic malformations and venous malformations?

Yes. That whole team has done an amazing job of bringing forth another option for patients that have no options today. So we definitely applaud all their efforts and are rooting for them to get that drug registered and bring another option forward for patients.

We don't think that these patient populations, the one that they're pursuing and the one that we are pursuing, are going to have that much overlap. And so their drug is obviously -- it's a topical, so it can only treat topical cutaneous lesions. And as their KOL has described it, it's called going after the microcystic lymphatic malformations. So those are lesions that are cutaneous and 2 centimeters or less.

And so I think there's a lot of these patients. There's a lot of unmet medical need. We believe that the bulk of the patients that we'll be seeing are going to be either mixed or macro cystic presentation. And so yes, we don't think that the patient populations are going to overlap that much.

Great. I'll pass it on to Yaron to cover breast cancer.

Great. So as you mentioned, at TAF -- sorry, at TAT, we're going to see the data from the 400 mg BID, 57 patients. And if I recall correctly, previously, we've also seen sort of a 50s patient number at the 600 once a day, non-fed. The -- is the data going to be long enough to look at PFS?

Yes.

And the -- what about -- and I assume, on response rates as well?

Absolutely, yes. We should be both presenting both ORR and PFS.

And it sounds like there -- we can read between the lines that you're comfortable with the data and they're probably going to be fairly comparable.

Our goal is to show consistency between the 2 different cohorts. And I think we've been public to say that both cohorts we shared with the FDA are -- and that totality of the data was part of the granting of the Breakthrough Therapy Designation that we got in January.

In terms of maybe more data that we can expect this year, it sounds like you're saying this year, we'll get -- I imagine both the CDK4 combo with Atirmo and the CDK4/6 combo and the Phase III trial design. Is that sort of fair?

I think our goal is to share data specifically for the go-forward regimen. What more data we share at the time, we'll decide. Because obviously, it's going to be a lot of information to get over. We're going to need to share all the relevant benchmarks in the second line, the data of the regimen that we go forward, the frontline trial design, the next steps. So we just want to make sure that whenever we put out a disclosure, it is easy to digest. And so we'll see when we digest -- when we put it all together, we'll see what else we'll share, but we'll definitely share that.

Okay. As you think about what would make sense and would be innovative in frontline, I imagine there's an opportunity potentially to think more on the CDK4 axis as opposed to the CDK4/6. You could do a combination, I presume, with either one depending on tolerability and things like that. But would it make more sense to advance a CDK4 forward? And that could be actually a good -- in combination with the PI3, it sort of even derisks the CDK4 path to market depending on how their own data shakes out?

Look, I think the premise of a frontline regimen has been kind of -- the path has been blazed by the INAVO120 trial. It showed that if you add a PI3K inhibitor to a CDK4/6 and AI in the frontline, you add significant efficacy. The challenge in that regimen is the tolerability.

And so the point you make is exactly the right one, which is this is about showing a tolerable profile. This isn't about proving the efficacy. That's already been proven. And so between the 3 triplets that we run, our goal is to try and figure out which is the one that is going to be tolerable for patients to take for 3-plus years. And the premise of having a CDK4 selected and a PI3K-alpha mutant selected together, the premise would be that would be the most tolerable. And so obviously, we have to let the experiment play out.

When you're thinking about a -- what would be the challenges to that kind of a regimen? I guess the one thing that comes to mind is the CDK4 head-to-head against CDK4/6 fails, how does that impact your ability then to get that regimen approved?

Yes. So I think, obviously, there's an ongoing trial of atirmociclib in the frontline, doublet versus doublet, the FourLight-3 trial. I think if we were to move forward with atirmociclib, we'd be looking at a slightly different hypothesis, right? The CDK on the efficacy side would not necessarily need to be superior to the existing CDK4/6 regimens for efficacy. We'd be looking at our triplet being superior. So the success of that approach does not necessarily hinge on the success of the FourLight-3 trial.

Now there are questions about what the contribution of [ parts ] would be. We think that recent FDA guidance that came out last year for the development of these types of novel agents and looking at contribution of [ parts ] would be consistent with potentially using FourLight-3 data to support that. In a 1,000-patient trial, you would anticipate that you'd see on the order of about 200 patients who would be PIK3CA mutated that would receive atirmociclib that could potentially be used to support the contribution of [indiscernible]. But all of this, if we were to go that route, would require us to work through with the [ health authorities ].

Right. So you can potentially use the data that Pfizer is generating and get access to that data?

That would be consistent with recent guidance. Yes.

Any questions from the audience? Maybe an update to your ReDiscover-2 study going head-to-head against Truqap, both in combination with fulvestrant. Any sense how far along are you, enrollment-wise?

Yes. As Sanjiv hit at the top of his remarks, we're very happy with the progress being made thus far. We are open globally now and enrolling patients. We're not currently guiding to full enrollment or top line data. I think once we have all sites open globally and have a good handle on what the enrollment rate is with all sites open, we will probably endeavor to try to give some kind of target date to full enrollment. Not quite there yet, but very happy with how the trial is rolling out globally.

Okay. There's been a lot of questions over the years on the Scorpion compound, 478. And now that we've seen its profile a little bit more, can you maybe remind us of your relative efficacy in all mutations and in the kinase mutation and tolerability relative to 478?

Yes. I mean, I think, obviously, we've done this one to death over the years. The monotherapy data, I think, started out strong and then deteriorated. Obviously, we've never really focused on that. We don't see that as the focus. We saw immediately for us to focus on the doublet with fulvestrant, where we showed 39% response rate. I think their latest data update was in the 20s. And then in the triplet, I think they showed some very early data where, again, they were below what they saw in the doublet. So I think on the efficacy front, I think we put out a bar which no one has beaten yet and no one has really even come close to.

I think on the tolerability profile, there's some questions on the compound. I'm sure that they'll work through them. But obviously, you still see a very short follow-up, significant Grade 3, Grade 4 liver function abnormalities. And so that -- really, the question to answer on that compound is what is the dose going to be, and then what's the tolerability profile going to be. And I think we look forward to seeing the long-term follow-up data where we can really do head-to-head comparisons and really come to a conclusion on what really is this compound.

In Phase III, they're still doing some dose finding in combination with CDK4/6s, right?

That's right, yes.

And what was the reason for that?

I mean, I think they obviously moved very, very fast as is their way. And this is their ability, the size of a company that Lilly can do. But I do think there is a significant question on what is the dose. And so I do think that's the fundamental is what's the tolerability profile going to look like and what's the balance between tolerability and efficacy.

And as you know, these patients are going to be on the frontline on this regimen for 3-plus years. And as we've said right at the beginning of this, tolerability is what matters here, and that's critical. And so we'll spend time making sure that we've got the right frontline regimen, and I'm sure they'll do the same.

There's -- we've been getting questions from -- Celcuity's geda is expected to have the Phase III VIKTORIA-1 data. It sounds like it's going to be late this quarter or early next quarter. So they've shown 9.3 months as a triplet in the PIK3A wild-type, and now we're waiting for the PIK3A mutant data. Can you remind us what they've said in the past on their PFS? And -- we think it was a very small sample. And what do you think about that triple overall, just given the lack of convenience?

They -- so they've not shown any robust data from their Phase I/II study at that dose and regimen. I think they had a bullet in a presentation that said they had 11 patients that were CDK4/6 experienced treated at that dose and regimen, and they said that, that was a 19-month median PFS. That would obviously be a very compelling PFS if that held up in a randomized controlled Phase III trial.

As you highlight, this is a 3-drug regimen where geda is administered 3 weeks on, 1 week off intravenously. And then it's combined with palbociclib, so CDK4/6 pretreatment. Then it's combined with endocrine therapy, fulvestrant. So...

And steroid [indiscernible].

And then there's prophylaxis steroids and prophylaxis antihistamines. So it's -- quickly accumulates to a 5-plus drug regimen. In the second-line setting, these metastatic patients are living very active lives, and it's going to be a very challenging regimen regardless of the PFS in true second-line patients.

The mechanism of the molecule is a multi-targeted drug hitting multiple pathways. The safety profile from the VIKTORIA-1 study in the wild-type patients looks a lot like an mTOR inhibitor. And all we know is that we have a lot of data in mTOR inhibitor, in breast cancer that with everolimus -- and everolimus does worse in mutant patients than it does in wild-type patients. So at the end of the day, you need to see the experiment play out. But we don't believe in any outcome that it is truly a long-term competitive threat to our regimen in second-line patients.

Great. Any maybe final question from the audience? Well, maybe a final question for you. What about doing a PI3 and SERD combo?

Can you get it, Don?

Yes. I mean, I think it completely makes sense. And one of the things that we've been doing over the recent past is looking for the SERD space to settle out. I think we're starting to see that happen. Last piece we'll see fall -- or one of the last pieces we'll see fall coming up imminently is going to be the persevERA trial. But I think this is an avenue we definitely, I think, could go into in the future.

Yaron, you should have asked the question, should it be a combo with SERD.

Lasofoxifene is a good drug.

Well, terrific, everybody. Thank you so much for coming. We appreciate it.

Thank you.

All right. Good morning, everyone. Thank you so much for attending. My name is Zaki Molvi. I'm an associate on Akash Tewari's team. And I'm here with the Relay Therapeutics team. We've got Sanjiv and Pete. I'll let them give some opening remarks, and then we'll get into Q&A.

Thanks, Zaki. Thanks to Jefferies for the invite, and thanks to all of you for attending this morning. I think as we've been doing all year, we're in execution mode. We are heads down now executing Phase III with our asset RLY-2608, which is a PI3K-alpha mutant selective inhibitor. We believe to be the first in the clinic. We're excited about the progress that we're making there. Behind that, we're doing all the work to look at the various triplet regimens that we could potentially use in earlier lines of treatment in breast cancer. And then we've commenced dosing patients with PI3K-alpha-driven vascular malformations. And so behind that, we have NRAS-selective asset as well as the Fabry asset. We have close to $600 million on our balance sheet, and we're excited to generate value over the coming years.

Awesome. So I guess, I think at ASCO, we saw your updated data. We've got a 39% ORR that continues to mature from the last cut, 10.3 month PFS, 11 months in pure second line. But we're also seeing the other PI3K alphas, the SERDs, the KAT6s looks like 5 to 7.5 months. I think just to level set, I think what people want to understand is what's going to be the right competitive bar in second-line plus setting? And how much -- how confident you are that, that 10 months will -- understanding that it may degrade, may not in a larger Phase III, will that still land in a place that's competitive relative to those?

Yes. It's a question that's really been circulating over the last year, 1.5 years because there's so many different mechanisms of action that have read out. And obviously, the hope of all these different mechanisms, KAT6, oral SERDs, CDK4/6 retreatment, some of the nonselective pathway inhibitors was that you could improve that 5.5 to 7 months' worth of PFS. And unfortunately, what we've seen is nobody has been able to move that bar. And even more interestingly, the bar could have actually been lowered. And so capivasertib was approved a couple of years ago now with actually a lower PFS than alpelisib, the current that was approved prior to that. And capivasertib has taken significant market share and is on a run rate now of $800 million annually just because it has a perceived better safety profile than alpelisib. And so although we haven't seen the bar move from 5.5 to 7 months, we actually may have seen it actually decline. And so we feel pretty confident sitting on a 10-month PFS, even with some degradation that we will be successful in the second line.

Got it. And I mean, for that second-line study for ReDiscover-2, you're going head-to-head versus Capi, which I think is great, makes a lot of sense. And I think just looking at the way you've powered the study, I feel like you are kind of powered for at least maybe 2, 2.5 months delta over Capi. I mean, is that kind of a fair assumption that you're making internally?

Look, I think we want to make sure that we have a successful trial against capivasertib, but we want to make sure that it's also going to be commercially viable. And so yes, we think somewhere between 2 and 3 months is the minimum that you'd want. As we sit today, it's 10 months PFS across all lines. It's a pretty robust data set now, over 60 patients, over 12 months' worth of follow-up. And so in second line only patients were at 11 months. So we believe that as many people point out, there could be some degradation in the real-world trial that we run across multiple geographies. But we still feel pretty confident that we'll be able to clear the hurdle.

Got it. And in terms of, I think, the dose, you announced earlier that you'd be going with a 400 mg BID fed dose, whereas the previous data that we saw was 600 mg fasted. And I think you've shown pretty nicely that it looks like the PK coverage is relatively similar, if not even maybe a little bit better just with the fed dose. But one thing is when I looked at the CAPItello study, right, we know Truqap on the label, it doesn't have a particular restriction. But when I looked at the CAPItello study, I think at that time, it was a recommendation of fasting before the dose. So I mean, how do you think this kind of compares? How will this data set be comparable if you're now going to fed?

Yes. So you're correct. We did notice a food effect with 2608, moved to 400-milligram BID fed, very similar exposures. These are exposures that get us on average about IC90 target coverage throughout the dosing interval. And then interestingly with Capi because of the tolerability profile, they had to move to intermittent dosing, so dosed 4 days on, 3 days off. And the label is agnostic to food. So I don't think -- we don't anticipate that being a challenge at all in the study that we will be asking patients to eat. It can be anything, low fat diet, high fat. It doesn't really matter what it is, just getting some food into the system. And it's very convenient because it's a BID dosing. So take it with breakfast, take it with dinner, and it shouldn't be an issue.

Got it. And actually, on kind of the diet part, in terms of the hyperglycemia, right, you guys have shown, I think, lower -- generally lower hyperglycemia rates than the other PI3K alphas, but definitely, it's still a class effect with hyperglycemia. And what can you tell us about for ReDiscover-2, there will be monitoring of fasting glucose. I mean, how will patients be monitored? And with the Fed regimen, will that have any kind of effect or artifact on hyperglycemia?

We don't anticipate the fed dosing having any impact on the amount of hyperglycemia we'll see with 2608. In the study itself, we don't require at-home glucose monitoring for 2608. And Capi would be administered per label in the study. And they recently had -- I guess about 9 months ago now, they've updated their label to require at-home glucose monitoring. So those patients will need to -- per protocol will be asked to monitor their glucose because that is what the label requirement is for Capi. And the reason why that is -- in the clinical study in CAPItello, you saw a numerically lower hyperglycemia Grade 3 rate. But like I said, Capi is dosed 4 days on, 3 days off. They were measuring glucose at the end of the 3-day drug holiday. And we believe that, that artificially lowered the perceived hyperglycemia rate. And I think you're seeing that come to bear in the real world where they're now having a label update to require glucose monitoring.

Right. Got it. Makes sense. And in terms of -- we have -- there's another PI3K-alpha in the room, right, with the Scorpion Lilly molecule. And it's funny at ESMO, we saw -- I think even before ESMO, last time we talked, you mentioned that you've actually looked at that Scorpion molecule yourself and done kind of the in vitro coverage tests. And it seemed like their 100-milligram dose was just starting to touch the coverage that you were getting with your 400 mg go-forward dose. And now we have some 100-milligram data from ESMO, where it actually looks like their doublet data numerically very similar 40% ORR, the triplet, it looks like there's like 4 patients at that dose, like 3 out of 4, 75%. Who knows really until the data matures. But I mean, do you feel like that thesis is -- I mean, how should we look at the Scorpion data now that we have some idea of them at kind of comparable doses?

I mean I think it's the thing that we all look forward to. I mean there's a lot of speculation around the profile and how it could be significantly better than the current standard of care, including 2608. I think the data that came out was still very difficult to interpret. It was early. And I think that's the key thing, which is, yes, I mean, the numbers you stated were true. But if you really look at the confirmed responses, they were 0 in the triplet. And I think most of the doublet responses are also unconfirmed. So I think the only thing that we could really go away with on the efficacy side is still early, maybe it could get close to 2608, but we've set a pretty high bar at 39% ORR and 10 months of PFS.

I think the one thing that you can tell for certain from the data from ESMO is there is some unknowns around the toxicity profile. And so in their monotherapy that they showed back in 2024, there was an LFT signal. Clearly, they showed Grade 3 LFTs. And at that point, it was really only kind of less than 2 months' worth of follow-up. And I think the narrative then it was only at high doses. Now in this data set, they ticked out the high doses above 100 milligrams and that LFT signal doubled. And so we're seeing in the doublet 37% all grade LFTs increasing and then over 10% Grade 3. And that's close to now at 4 months' worth of follow-up. And they didn't break it out by dose. So you imagine that if you follow the narrative from the year before that those were higher at the 100 milligrams. So I think the real key question is on the efficacy is they could potentially get close to us at the 100 milligrams. The real question is, is it tolerable? And we'll find out over time. But the signals that we saw on the LFTs are definitely concerning.

Right. And to be clear, for your molecule so far in LFTs, we're out of the weeds there. And so I want to talk a little bit about triplet setting as well, right? Because you partnered with Pfizer to kind of run the triplet study with Atirmo. But we've already gotten a little taste of kind of PI3K-alpha mutant selective triplet data from Celcuity, right? They have gedatolisib triplet data. It looks like 14.6 month PFS in mutant. I mean, are you kind of looking at that as a potential preview of what this mechanism can do in the triplet setting? And how are you viewing that other than the efficacy, the other profile otherwise in terms of the IV dosing compared to you?

I think we were all thrilled to see the Celcuity data readout in the wild type. It was a positive trial. There's just so few options for wild-type patients that I think it is definitely a win. I think the question is unknown on the mutant side. It's a small number of patients that they showed the 14 months in. And so again, back to the question of how will that read out in a large multicenter Phase III. We'll find out, I think, in 2026. The real question is just going to be, again, how tolerable is this profile going to be. I think in the data that we've seen, you need multiple drugs here in the triplet to support it. You need dexamethasone mouthwash. You may need medications for the rash. So it's just potentially a 4- to 5- drug regimen, and it's an IV administration. And obviously, these are chronic therapies in a [indiscernible] oral market. So I do think there are some questions even if it does hit on the mutant to the commercial potential of this product versus an oral and the profile of 2608.

Got it. That makes sense. And also, we also have the Roche triplet data as well. It looks like coming in at a very similar PFS with a 15-month PFS. But I think you guys also have the ability to differentiate on safety. I mean, should we be thinking about triplet as like the bar for efficacy is going to be a 15-month PFS with generally cleaner hyperglycemia rates? I mean help us frame that.

Yes. So I mean you hit the nail on the head, and that is very mechanism validating data. They ran a very elegant study in frontline endocrine-resistant patients. It was a bit of a contrived patient population and that they had to go into very metabolically fit patients to mediate the hyperglycemia risk. But nonetheless, they proved that if you can thread the therapeutic needle and actually get patients to stay on that study and hit the target, you can have a pretty profound benefit. And we now see that they have an OS benefit. So that was very confirming to us on the commercial potential that we should think about for ourselves in the frontline setting. And yes, our goal is to move forward with a triplet in a frontline study. There's 2 open questions there is which triplet to move forward with and which study to run. But we feel very confident that once we get there, there's probably very high POS on that trial.

Right. And I think one of the things that -- as we talk about hyperglycemia, one of the things that you've done and you're doing it in ReDiscover-2 is the A1c criteria, under 7%, the fasting glucose, also upper threshold 140. And I think the discussions at SABCS had kind of talked about this as one way that you had kind of potentially limited the risk of seeing events. But how should we think about if the study is going to be run in that population, how is it going to look on the label? And then how might the risk be managed for a broader population that does include people with higher A1Cs?

Yes. I think over time, you'll see us test the hypothesis of going into higher A1c patients. Because we know that the data we generate, both in our current ongoing second-line trial and eventually the frontline trial we run, we know it's likely going to be compared against the different trials that Roche is running with inavolisib, which has very restricted metabolic criteria. The place to test that hypothesis is probably not inside these Phase III studies. We still by the entry criteria we have, these are prediabetic patients, anything over 6 is going to be considered prediabetic. In our Phase I/II experience, we had 1/3 of our patients either with a BMI greater than 30 or in the prediabetic range as it concerns HbA1c and fasting plasma glucose. So I think we could eventually run that experiment over time. I don't think it's the proper one to run inside of these studies given the cross-trial comparisons that folks will make in the future.

Right. That makes a lot of sense, right? The idea is you want to keep things consistent. We've seen these encouraging signals so far. And I think the idea is you want to be as best positioned as possible to kind of replicate these early signals that we've seen. To that point, I think one of the things that we were talking about as we discussed this 10-month PFS signal in the second line plus and then kind of replicating higher than that in the first line. Is there anything that you're doing kind of in recruiting for the Phase IIIs in terms of using repeat centers? Anything to kind of ensure reasonable limitations on the -- how much PFS might degrade?

So the key thing is the -- in the ReDiscover study, the Phase I/II, we had about 50% of our patients were third line plus. Over half of them had seen prior either oral SERD or fulvestrant. 1/3 of the patients had prior chemotherapy. Just by definition, as we move into this Phase III where we're putting a bit more restrictions on the number of prior therapies, we're likely going to see healthier patients, more predominantly second-line patients. And so I think completely agree with you that historically going from Phase I/II single-arm studies into large Phase IIIs, you should anticipate a bit of a degradation in the PFS you've seen. We probably get a little bit of balance in the other direction in that regard because we're moving -- we will control a bit more the amount of pretreatment that these patients have seen.

Right. Got it. Makes sense. And in terms of -- I think in your data, I wonder if there's -- we've talked about these kind of other experiments to kind of run. You've seen in your kinase mutation cohorts like an 18-month PFS in the second-line setting, which I think is kind of an intriguing signal. Do you think there's room to kind of say like potentially running a trial to confirm that signal? And I mean, is there an opportunity there?

I think we believe that we have a pan-mutant inhibitor on our hands in 2608. So sure, the data that we showed was -- at 18 months was, we think a high watermark. We think it's going to be over 12 months. But we also think that we have a molecule that covers the helical and so we're running the trial to try and get the approval across all mutations.

Got it. And now I want to talk a little bit about the vascular malformations opportunity. You're running an interesting trial there. We've already had the EPIK-P1, EPIK-P2 data from Novartis. It's funny, the EPIK-P1 data, right, they showed kind of like a 38% ORR, which kind of in the EPIK-P2, they did half that dose and they saw pretty much half the ORR, but also half the Grade 3 AEs. So I mean -- and also some KOL work that we did suggest EPIK-P2 is a little bit less severe patients, maybe not the ideal patient population. I know you are running a study where you're doing kind of an interesting 3 plus 3 dose escalation. So it looks like you can kind of get to more effective doses more quickly. But I mean, tell me a little bit more about how you're taking the learnings from those alpelisib trials that Novartis ran and incorporating them into your vascular malformations study.

Maybe I'll start and Pete, you can finish off. So those of you who are not familiar, vascular malformations is a significant opportunity driven by PI3K-alpha mutations. The only approved agent at the moment has accelerated approval is Novartis' alpelisib. And the studies that you cited here, one was the accelerated approval that got the approval and then the confirmatory trial unfortunately failed. And so there's plenty of precedent here for us to kind of work through as we figure out what the best pathway for us is. Maybe, Pete, you can just talk about the design.

Yes. So because of the experience we have of 2608 in oncology, you're right, we can -- instead of going through the traditional dose escalation, we're actually doing dose randomization across 3 doses simultaneously. And those -- each of those doses -- each of those arms could be up to 15 patients each. So it is a bit more efficient development. We'll -- the highest dose in that randomization will be the oncology Phase III dose. And the open question here is what level of target coverage do we need to really affect maximum efficacy in these patients while maintaining a good tolerability profile. We get a little bit of an understanding of that relationship between target coverage and efficacy, as you point out from alpelisib, they used 250 milligrams in the label dose, which is about the strangest approval you'll ever see.

It was based off of retrospective chart review of patients treated under compassionate use across 7 centers. There is no dose finding or justification done there. And that's why investigators in the confirmatory study required them to go to half the dose. And so I think it's difficult to really understand the safety profile of alpelisib at that 250 dose because that safety data was collected through retrospective chart review of compassionate use patients. It wasn't in the rigor of a clinical study.

And then the confounding factor of the EPIK-P2, like you said, it's done at half the dose. And so it's barely -- probably barely touching the PI3K-alpha pathway, the mutants from a target coverage standpoint. And so I think it will be interesting to see the profile over time. But yes, the clear goal here is better efficacy, better tolerability, get to these patients as early as we can in their disease progression. This is a somatic mutation that happens in utero, and the goal would be to treat patients as early in development as we can chronically. And I think that's the profile we'll be looking for as we continue to test this clinically.

Right. And in terms of -- I think one of the concerns investors want to know is how quickly will we eventually get data on this because these are smaller trials, take longer to enroll than the oncology trials. So when are we really going to be able to see some proof-of-concept efficacy there?

It's a great question. It's one of the things that frustrates folks is that we are not giving guidance on any of our programs at the moment. And it's really out of interest for shareholders to make sure that when we do give guidance, it's with a really solid understanding that we'll have robust and interpretable data in hand. We're very happy with how the conduct of the study is going, the excitement amongst investigators, the pace of enrollment. We're just not quite there yet to feel comfortable giving guidance to when we could share that data. But rest assured, when we do, it will be the data set that we view to be interpretable and really informative for the ultimate next steps that we would pursue here.

The cash window that we have of close to $600 million in cash runway into 2029 allows us just to go out and execute. And so as you see in this window, we'll have the Phase III that we're running ReDiscover-2. We'll have the triplet data that we've talked about, and then we'll obviously have this VM, vascular malformations data as well as bringing the Fabry and NRAS program into the clinic. So there's plenty of catalysts ahead. I think the key at the moment is we're just not guiding to when.

Right. And I think having that data eventually is something that investors should wait for because it's 170,000 patients across all of the types of vascular malformations. I think one thing I want to hit on is, are there going to be subsets that you think might be the optimal responders? Because you look in the literature, you see, okay, there's different mutational burdens, patients who -- some patients whose disease relies more so on PI3K-alpha mutations than others, genetic mosaicism. I mean, do you think the trial program, when you do have that data, will be able to elucidate how big the opportunity is going to be?

I think the initial data will start to answer that question. It won't be the complete answer. Our view is that any of these patients with PIK3CA mutated disease should be responsive to a PI3K-alpha mutant selective inhibitor. You're correct. There's other -- biologically, there's other things happening inside of these lesions. But we do think that in the mutated patients, that is happening in an otherwise very quiet genome. And it's not like cancer where you have polyclonal disease that we do believe the PIK3CA mutation is driving this disease. There are some other mutations that we believe do happen in these patients. TIE2, for instance, is one of them, but we think that's mutually exclusive to the PIK3CA mutant patients.

Right. And one thing we noticed is there's only a few centers where there are specialists for vascular malformations. So when you think about this 170,000 number, I guess, are there -- is it commonplace to get PIK3CA testing? I mean, will those patients know that they are eligible for an inhibitor?

I mean I think in the PROS and the lymphatic malformations is a clinical diagnosis because 100% of patients with PROS have a PI3K-alpha mutation and close to 80% in lymphatic malformations. And so I do think that, that allows these patients to be treated much more rapidly. But there is a significant amount of testing that goes on here in both populations.

Got it. And with that, we are out of time. Thank you all so much for attending.

Good morning, everyone, and welcome to Morgan Stanley Global Healthcare Conference. I'm Sean Laaman, Head of U.S. SMID Cap Biotech Equity Research here at the firm.

Before we commence, for important disclosures, please see the Morgan Stanley research disclosure website at www.morganstanley.com/researchdisclosures. And if you have any questions, please reach out to your Morgan Stanley sales representative.

For this session, we have Relay Therapeutics. We welcome Sanjiv Patel, President and CEO; Don Bergstrom, President, R&D; Peter Rahmer, Chief Corporate Development Officer. Welcome, and thank you for your time today, gentlemen.

Maybe just to kick things off, Sanjiv, could you give a bit of an overview of Relay, and that will contextualize the questions that we have. But thank you for your time.

No. Thank you for the invitation. We're thrilled to be here. So Relay Therapeutics, for those of you that are not familiar, we're a company that's based in Boston, Massachusetts. We are very well capitalized, and we have cash now into 2029, and we are heads down executing across a variety of key initiatives.

First and foremost of that is executing with our first-in-class PI3K mutant selective inhibitor, RLY-2608. We've initiated a Phase III trial in post-CDK4/6 treated metastatic breast cancer patients. That trial is up and running, and we're excited with that going because we think this is a very large patient population that's currently unserved.

The second of our initiatives is that we are continuing to dose triplet combinations with RLY-2608 and fulvestrant, both through the ribociclib and the atirmociclib. As we look to generate what could be a triplet combination that we can use in CDK4/6 naive metastatic breast cancer patients.

And then our final initiative that we push forward in the clinic at the moment is using RLY-2608 to dose a condition called vascular malformations that are driven by PI3K-alpha mutations. This is a very large unserved population, and we are very excited about generating clinical data in this area.

Behind that, we have 2 preclinical programs that we've named an NRAS-specific inhibitor as well as the Fabry program, and we push those towards IND. And then we have a small research footprint now working on very hard to drug targets.

And so we have a very nice portfolio in front of us, but really, at the moment, it's all about heads down execution as we start to generate clinical data that will really meaningfully generate value for shareholders.

Wonderful, Sanjiv. Thank you for those remarks. We've got a couple of sort of macro type questions we're asking all the company and hopefully, we'll have a nice, consolidated industry view at the end. But with China's rise in biotech innovation, how do you think about -- how are you thinking about your competitive position here? And will this influence your R&D and business development strategy?

It absolutely does. Obviously, we see the very impressive rise of innovative work done in China, very specifically in the area that we're in small molecules. We see now a whole host of licensing deals from the large cap pharma with such Chinese-based innovative biotech. So I think it does change the way we think about things. And we really only focus on first-in-class very hard to drug targets. And so that's our entire portfolio. We think the kind of days of doing kind of second, third, me-toos are over because, obviously, you just cannot compete with the speed and efficiency of some of these companies. So yes, very impressive what's going on in China and does actually very much change how we think about things.

Wonderful. As an AI tech -- biotech company, can you describe the keyways your platform is leveraging AI and think about AI's future disruptive potential?

Yes, absolutely. Happy to talk about it at a great length. But maybe just -- I'll just give you kind of a very concise answer, which is the rise of AI, again, like we're seeing it in all aspects of our lives, from driverless cars to how Netflix provides recommendations for us as we turn our apps on. But I think in drug discovery, it's just a very hard process from beginning to end. It's thousands of very tough incremental steps as you go right through the process of choosing a target to filing an IND and then going towards the NDA.

So for us, we see it as a tool that can help you increase efficiency and help you design the right experiments. But we don't see it as a panacea for all. And we don't think any time soon will a drug be designed entirely by AI. We see it as an enabling tool and we think that people will be involved in drug discovery for many years to come.

And last question before we dig into some of your programs. So what has been the most impactful on Relay from the regulatory side? Has it been FDA? MFN is probably not quite relevant yet, but -- or tariffs?

Again, just given the stage that we're at, most of our regulatory interactions were done last year as we set up our Phase III. And so as we come into this kind of new regime, we haven't really had that much interaction. So we can't really comment too much on whether it's better or worse. And then in terms of being a preclinical company, again, we haven't really had that much impact on the tariffs and the various pricing changes that we've seen. We imagine that will impact us over time. So I guess on that one, we really haven't seen any change.

Wonderful. Now, to ask about some of your programs. So RLY-2608, allosteric mechanism of action, how does it differ from traditional PI3K-alpha inhibitors? And why is this significant for cancer treatment?

So PI3K is the most commonly mutated kinase in cancer. And over the last 3 or 4 decades has been at the center of drug discovery efforts across most of the large pharmaceutical companies. So we've seen various iterations of this. The final iteration that we've seen over the last decade is nonselective PI3K-alpha inhibitors. And so we know that this is a mutation that's implicated in a broad range of cancers. The challenge is if you inhibit both the wild-type and the mutant, you get a range of side effects including hyperglycemia, rash, diarrhea, stomatitis. All of these lead to patients just not being able to stay on drug and leads to lower dose intensity and then lower efficacy.

And so unfortunately, we've just not been able to produce a drug that has been efficacious in this field. And so when we set out back in 2017 to try and drug this target, what we were looking to do was try and create an exquisitely mutant selective inhibitor. And so the premise was, if you could dial out these wild-type toxicities and only really inhibit the mutant, could you generate a better safety profile. And then in turn, could you keep patients on drug at greater dose intensity and then could you generate greater efficacy.

And so that's exactly the premise of RLY-2608. We were able to discover a novel allosteric binding site that was preferentially seen in the mutant, and that gives us a handle on mutant and selectivity. And obviously, now we've shown a range of clinical data that starts to prove out the exact hypothesis, which is the data we've shown in RLY-2608 shows the potential for a better safety profile, and that has now translated in the data that we've seen in the kind of the early trial into what we believe to be better efficacy.

Wonderful. And what clinical efficacy data has been reported for 2608 in combination with fulvestrant?

I'll hand it over to Don Bergstrom.

Yes. So we've reported data out of our first-in-human trial, the ReDiscover trial. And our most recent disclosure was in June of this year at the ASCO meeting in Chicago. And what we've shown in the most recent update is we now have our efficacy evaluable subset of 52 patients. These have all been treated at a 600-milligram dose of 2608 fasted, which gives us the exposure level that we'll take into our Phase III trial. We see a confirmed objective response rate in these 52 patients with 39%. We see a median progression-free survival of 10.3 months. And these numbers are significant because they're almost double what we see for other PI3K pathway targeting agents in analogous patient populations.

So our comparator in our Phase III trial will be capivasertib. That was tested in the CAPItello-291 trial that led to its approval in PI3K pathway mutated patients. The reported PFS in PI3K-mutated patients in that trial was 5.5 months. So we're sitting at 10.3 months.

If you take into account the capivasertib population was largely a second-line population. They've never seen any prior fulvestrant. Our patient population includes a number of third line and later patients where half the patients have seen prior fulvestrant. So if we go a little bit further and just look at our true second-line patients in the ReDiscover patient population, we're sitting in 11 months PFS. So really almost double what we're seeing for our comparator.

Amazing. And could you please throw a little bit of meat on the bones around the safety profile, particularly regarding hyperglycemia, rash and diarrhea.

Yes. So we see very low rates of high-grade hyperglycemia. So where physicians really start being concerned with the management of hyperglycemia is when they make it to Grade 3 or more. And that's really when physicians need to start using injectable agents like insulin to be able to control hyperglycemia. This is where they're calling an endocrinologist to help manage the patients. And in many cases, especially in busy community practices where there may be limited access to endocrinology consults, that's a point where the therapy is just abandoned unless the patient is really benefiting significantly from it.

So what we've seen for other agents in this space in typical Western patient populations is that Grade 3 or higher hyperglycemia is typically in 20% or more -- 20% or higher of patients, and we are in the low single digits. And then for the other key AEs, which are very taxing for patients, diarrhea, rash, stomatitis, those can range for some of these agents, capivasertib, almost 80% of patients have diarrhea. You've got high rates of stomatitis. Again, for us, for high rates of any of these AEs, we're either in single digits or 0.

And what is the current status of the ReDiscover Phase II -- ReDiscover-2 Phase III trial? And what are its primary end points?

Yes. So we're currently enrolling that trial. We opened it earlier this year. And it's a 540-patient trial, randomizing fulvestrant plus RLY-2608 versus fulvestrant plus capivasertib in PI3K-mutated patients who have seen prior CDK4/6 inhibitor therapy. Primary endpoint is progression-free survival. And the way we've designed this trial statistically is to look at this in a hierarchical analysis.

So in PI3K mutations, there are 2 main classes of mutations depending on the specific part of the protein that's mutated. The kinase domain gives rise to what we call kinase domain mutations. Typically, these are H1047, a specific amino acid that's mutated. And then nonkinase domain mutations, most frequently are in the so-called helical domain, E542, E545 or the amino acids that are mutated. So we'll look first at PFS in kinase domain patients. We anticipate that will represent about half of our clinical trial population. And then we'll look at PFS across the entire patient population. So all PI3K-mutated patients.

And we've really done this primarily because when we were designing this trial, there were, and there still are, a number of competitors out there who are actually designing drugs that just target these kinase domain mutant patients. So Lilly had a drug that just targeted these kinase domain mutant patients that they just discontinued at the very end of last year.

What we've seen in these kinase domain patients is a 67% response rate and an 18-month PFS in our heavily pretreated patients. So we're seeing very strong efficacy in these patients. And in the anticipation of there being potential competitors out there at some point that just target that patient population, we want to make sure that we're able to have claims in our label with appropriately -- appropriate statistical design, be able to call out these very high response rates and high PFS. So we've designed this hierarchical analysis for PFS. We also have OS as a key secondary endpoint, but the precedent in this space is for PFS to get the registration endpoint.

Wonderful. And could you map out the market opportunity for us for 2608 and how many patients could potentially benefit?

Pete, maybe you can take that.

Yes, sure. As you think about the post-CDK4/6 metastatic breast cancer setting in the United States alone, there's probably about 13,000 PI3K-alpha mutated patients there, probably goes to about 30,000 in the major global markets. As you then expand that into the front line and also inclusive of the CDK4/6 naive setting, that starts to about double those numbers on a global basis. And so these are -- even the second and the post-CDK4/6 setting alone, that's a multibillion dollar TAM that it would be an opportunity for us to start to penetrate over time.

Wonderful. And how was 2608 discovered using Relay Therapeutics platform? And what structural insights led to its development?

Yes. So our research team, when we started working on PI3K, there actually had never been a full-length structure of the protein that had been solved. And that is largely -- this is now largely a technical issue. It's the art of being able to actually make the protein and image it. We have very persistent and talented protein scientists who are able to make the full-length PI3K-alpha protein, use emerging structural biology techniques like cryo-electron microscopy to actually be able to generate the first known full-length structure of PI3K-alpha.

And then we're able to take that information that we get from our structural experiments and then use our computational methods to be able to start simulating the dynamics of how mutant and wild-type PI3K-alpha move and to try to identify differences between the mutated state of the protein and the non-mutated state of the protein. And what we started to appreciate was there were potential druggable pockets that were opening up in the mutated protein that weren't readily seen in the wild-type protein. And these pockets have never been seen in any of the publicly available structures up to this point.

And that informed our screening strategy to actually start identifying small molecules that bound to PI3K in those pockets that appear to just be present in the mutated protein. So we found a hit that eventually led to what became 2608. But as we progress this through drug discovery, we're able to show the mutant selectivity. We're able to show that it was potent against both helical -- or both kinase and nonkinase mutations and ultimately able to optimize the profile of 2608 to give us a compound that we thought would give us superior efficacy in patients while avoiding the wild-type toxicities, which is exactly the profile that we've generated in the clinic.

Would it be fair to say that AI-derived molecule, this would be probably one of the furthest ones in terms of clinical development that there are globally?

Yes. I mean I think we definitely, as Sanjiv mentioned, we definitely -- this was an AI-enabled project. It is a tool that we use across the various steps in drug discovery to make predictions that we can then go to the clinic or go into the lab and validate experimentally and then go back and iterate and update our models.

Coming back to Sanjiv's previous point, this is not, we pushed the computer button and it spit out the structure of 2608. There's a lot of, I think, human insight that went into discovery of 2608 that then was supplemented by using AI and ML to be able to broaden the chemical space that we could look at, make new predictions about what might work and then ultimately be able to go and validate those predictions experimentally.

And are you exploring or what triplet combinations are you exploring with 2608?

Yes. So we've initiated triplet combinations with ribociclib, so Novartis' CDK4/6 inhibitor. And then with atirmociclib, which is an investigational agent that's currently being developed by Pfizer, that is a selective inhibitor of CDK4. The idea there is selective CDK4 inhibition will avoid some of the hematopoietic toxicity associated with CDK6, give you the ability to get more effective CDK4 inhibition and ultimately, be more combinable. We need to remember that as we go into frontline patients, these are patients who are likely going to be on our regimen for 2 years or longer. So tolerability matters. So we entered in last year into a clinical trial agreement with Pfizer to get access to atirmociclib, and earlier this year, initiated that triplet development as well.

Sure. And beyond breast, where do you see 2608 applicable?

Yes, there are mutations in PI3K-alpha really across all tumor types. You see it in gynecologic malignancies, so cervical cancer, ovarian cancer, colorectal cancer frequently as PI3K mutations. Although in colorectal cancer, the PI3K mutations are frequently seen in conjunction with other oncogene mutations, including KRAS mutations.

So we anticipate, in some of the tumor types outside of breast cancer, there likely would need to be combination therapy that would be used to be able to target both RAS pathway signaling as well as PI3K pathway signaling. And that's why breast cancer for us is really the first indication that we're going after heavily because you do have existing proof of concept here. And in PI3K-mutated breast tumors, the PI3K mutation happens very early in the genesis of the disease and really is the sole oncogene driver that you see in these patients.

Sure, Don. And how was 2608 administered to patients? And what's the current dosing regimen that you've got that you're using in clinical trials?

So it's an oral drug. We currently are administering 2608 on a twice a day schedule continuously. So there's no planned breaks in treatment. In contrast to capivasertib, which can't be administered continuously, patients take it for 4 days and have to have a 3-day drug holiday for resolution of toxicity.

All of the dosing that we've shown so far has been using fasted dosing, which is very typical as you take a molecule into the clinic for the first time. You instruct patients to not eat before taking the drug. And the data we've presented has been at a 600-milligram fasted dose. Last year, we did run a food effect study, saw a small positive effect of taking 2608 with food on the blood concentrations that were achieved, and we're able to show that a 400-milligram dose of 2608 taken with food gives us the same blood levels of 2608 when it's administered 600 milligrams fasted.

We took those data to FDA. FDA agreed based on what we showed them that a 400-milligram dose with food was the appropriate dose to take into our Phase III trial. And that's the dose that we're currently looking at.

Wonderful. Last question on 2608. I don't know how much focus you put into this, but looking at Street estimates for what the current market opportunity is and flavor of probability of success. I don't know if you have any commentary on this.

There's -- we have about 13 covering analysts at the moment. I would say the variability in those models is quite broad. Of the subset that I think have spent a good amount of time putting some precision to those estimates and thought into it, the range is, again, in a peak sales range just in the second-line metastatic setting of about $1 billion and then probably a success range is somewhere between 40% and 65%.

Wonderful.

I think that's for breast cancer. I don't know if we're going to cover the vascular malformations.

I'm about to cover the vascular, yes, and breast, and we can get on that.

And that's just the post-CDK4/6 setting. When you get into the front line CDK4/6 naive setting, that grows by an order of magnitude.

Okay. So I'm moving on to vascular malformation. So I guess sort of how does 2608's mechanism as a mutant selective PI3K-alpha inhibitor specifically target vascular malformations compared to nonselective inhibitors?

Yes. So vascular malformation are -- it's a non-oncology indication. These are overgrowth, tissue overgrowth that are -- and form lesions, but they're not metastatic lesions. It's not a solid tumor. And in a large proportion of patients with vascular malformations, they have a PI3 kinase mutation that's actually driving the malformation.

So it's the same PI3K mutations that you see in solid tumors, but this is now in the context of a normal genome, so you don't have loss of tumor suppressors or other oncogenes that give you a metastatic tumor. You just have a PI3K mutation that's causing overgrowth of tissue. And the way the disease presents is a function of what tissue the mutation arises in during embryogenesis.

So you have some mutations that could occur early in embryogenesis where you may have a whole limb that's affected, where the whole limb overgrows. And that is one presentation of a disease subtype called PI3K-related overgrowth spectrum, 100% of those patients have PI3K mutation is on the order of about 5,000 of those prevalence in the U.S.

And then other manifestations include lymphatic malformations and the mutation happens in the lymphatic precursor. About 80% of lymphatic malformations patients, that 80,000 patients have PI3K driven lymphatic malformations. And again, depending on where and when during embryogenesis happens you get that mutation. Some patients may have several malformations, patients may have malformations that are impinging on critical vessels or critical organs and really require systemic therapy to be able to address those disease.

And then there's venous malformation that has a lower percentage of PI3K mutations and cerebral cavernous malformations where about half of patients have PI3K mutations.

So PI3K clearly is the driver in these patients. And the patients who need systemic therapy need to start it early in life, this is a congenital condition, and needs to take it chronically, most likely over the course of their lives. So you're treating children and you're treating children chronically.

Alpelisib, marketed in cancer as PIQRAY by Novartis, has been approved with an accelerated approval in this PI3K-related overgrowth spectrum or PROS. It's marketed there as Vijoice. So same API but marketed under a different brand with different pricing. And that has really established proof of concept for how you can actually control the growth of these lesions if you're targeting these patients with the PI3K-alpha targeted therapy.

The challenge for the drugs such as alpelisib, it's not selective, it's also hitting wild-type PI3K-alpha. So you see a lot of those toxicities that you see in cancer. In addition, you're in kids and you see slowing of growth as you're inhibiting wild-type PI3K-alpha.

So there's some real concern in this population that needs to have chronic therapy that you actually can't or don't want to keep these kids on therapy chronically. And this is where we think being able to come in with a mutant selective inhibitor such as 2608 could be able to leverage what we already know about these lesions in these diseases that hitting the mutant PI3K-alpha driving the tumor can give you control of lesions while avoiding some of the toxicities that are associated with hitting wild-type PI3K-alpha. So it's really the same hypothesis that we've already proven out in cancer. And now we've initiated the trial to prove that out in vascular malformations as well.

Wonderful, Don. What's the current status of your Phase I/II trial for 2608 in vascular malformations?

Yes. So we initiated the study late Q1 of this year. One of the benefits of having profiled 2608 so extensively in cancer patients is that has allowed us to instead of going to the traditional dose escalation part of the dose confirmation in the study, we are able to go directly into randomization of 3 biologically active doses in parallel. And so that will include the recommended Phase III dose in oncology, a 400 milligrams BID fed and then 2 doses step down out from that.

And so we are actively enrolling patients, happy with how the pace of enrollment and excitement amongst physicians is currently going. Not yet in position to guide specifically to when we'll have data yet. But yes, like I said, very happy with how the enrollment is going.

Wonderful, Pete. Can you subdivide the population for us in this indication. And maybe help us map out again what the patient opportunity or population opportunity is for you?

Yes. At the highest level, the opportunity, as Don pointed out to be at the top, about 170,000 of these patients have PIK3CA mutant driven disease in the United States alone from a prevalent standpoint. As you get into some of the subsets of the disease, about 5,000 to 15,000 of them exist in PROS. So the PIK3CA-related Overgrowth Spectrum, 100% of those patients obviously have PIK3CA driven disease.

Within lymphatic malformations, there's about 80,000 patients in the U.S. that have lymphatic malformations, and 80% of those have PIK3CA mutant driven disease. So that's about another 60,000.

That's the core focus of our initial clinical development. As you can imagine, as a chronic therapy to be used over the life of these patients, even just a few thousand of those patients being on disease -- on therapy for their lifetime is a multibillion dollar opportunity for us. And so the opening question, given we have only one molecule approved in just the PROS subtype and only with accelerated approval and not a lot of clinical experience and clinical data and therefore, commercial experience to reference, the open question here is just how many of those 170,000 patients would seek or be eligible, if you will, for chronic systemic therapy. But if it's even just a fraction of that number, it's an extremely large opportunity for us.

And I guess we've kind of covered this on the breast discussion. But just to sign off. So how might 2608 safety profile in breast cancer trials inform its potential use in vascular malformations, particularly regarding hyperglycemia and the other side effects?

I mean I think as Don covered earlier, these patients are going to be on therapy for the whole of their lives. And so these are not end stage metastatic breast cancer patients. And so the tolerability profile really does matter much more here. And so having a mutant selective inhibitor that can dial out some of the toxicities that we've talked about, hyperglycemia, diarrhea, rash, is much more important. And that's why we're so excited about the use of 2608 in vascular malformation.

Wonderful. Maybe to ask a few questions about the platform. So I guess, how does Relay Therapeutics' Dynamo platform integrate computational and experimental approaches to address previously intractable, inadequately addressed protein targets?

So we've been at this really since 2016. So one of the first companies really to enter this space of using computational tools and combining them with experimentation to try and make the process of discovery of new drugs both more efficient and more effective. And so I think as we've talked about, like each of the many hundreds of steps inside of this process can be broken down and each of those steps can be enabled in some way by computational tools.

And so we kind of break it down into 3 big groups. How do you identify a modulation hypothesis of where to try and bind a small molecule to a protein. That's kind of one area that we can use computational tools. And there, we've used long-time scale molecular dynamic simulation to stimulate how proteins move and how they can interact with a small molecule to identify novel binding sites, exactly as we did 2608 by identifying a novel allosteric binding site.

The second kind of step in how do you identify novel binders. And there, we use traditional experimental approaches such as hit finding techniques that everyone has access to. But we also have, over the last 10 years, used a range of virtual screening techniques, doing simulations and a variety of virtual libraries to define novel binders.

And then the final piece of our process is how do you optimize these binders into essentially drug-like molecules that will eventually become the development candidate. And there, we've used a range of algorithms, generative AI to identify novel modifications that we can make with molecules that eventually lead us to the development candidates that we've been able to deliver.

And so it's a lot of steps, each one slightly better given the computational tools that we have, string it together and you get development candidates.

Wonderful. I just realized I didn't quite finish off on the vascular malformation piece. But we talked about the estimates for breast, around about $1 billion peak sales. But in vascular malformations, maybe the same question.

Yes. I think today, it's a bit early to try to estimate that just given that we -- again, there's not well-precedented molecules in the market today. But depending upon the number of patients that ultimately see chronic systemic therapy over time, it could easily be a market opportunity that is in the single-digit billions of dollars.

Wonderful Okay. Maybe back to the platform. Sorry for jumping around. But what role does the Anton 2 supercomputer play in Relay's Dynamo platform? And how might the upcoming expiration of DESRES agreement August '25 impact their drug discovery capabilities?

I think it was very helpful to us over the first iteration of Relay. The power that Anton had was unsurpassed at that time. And so it helped us with a lot of the discoveries that we made in the first years of Relay.

Now obviously, computational power 10 years later has moved on a lot. You can do a lot on GPUs. And obviously, the cloud has moved so you can split up problems and spread it across the cloud, which is much more commoditized now and allows us to do a lot of things we used to do on Anton now using commodity hardware.

Wonderful. And final question from me. What -- I mean, there's lots to talk about, but what didn't I ask in our discussion that I, perhaps, I should have.

I mean I think the question that we get asked a lot is, what does it all mean, right? At the end of the day, does this whole use of computational technology make the process more efficient. And the answer to that is it's less about efficiency, it's more about output and productivity. In the end, we've been at this now almost a decade. We have put multiple programs into the clinic. We have a Phase III. We have another program that will hopefully file an NDA that we outlicensed.

And so for us, it's less about efficiency. It's more about output of high-quality molecules, and we feel that we've been able to demonstrate that. And that's why we're so excited about the potential for RLY-2608.

Wonderful. That might be a nice place to park the discussion. But thank you, Sanjiv. Thank you, Don. Thank Pete.

Excellent. Thank you to Morgan Stanley again for the invitation.

Wonderful. Thank you, gentlemen.

Good afternoon, everyone. Thank you so much for joining us. I'm Salveen Richter, biotechnology analyst at Goldman Sachs. I am really pleased to have with me Sanjiv Patel, President and CEO of Relay. Sanjiv, just to start here, could we just level set here with an overview of the Dynamo platform and where you stand today with the optimization efforts that have played out and the acquisitions of technologies as well as pipeline strategy and upcoming milestones that we should be looking at over the next 12 to 18 months?

Well, first, thanks for the invitation. In terms of the Dynamo platform, I think kind of close to 10 years ago now, we were one of the first wave of companies that put together computational approaches with experimental techniques to try and make this drug discovery process a little more easier than it has been. And so all intents and purposes over the years, we've done a great job of this in that we've been very productive. We've had now several programs enter the clinic, all kind of wholly created using our approach. But given the capital markets that we're in at the moment, we've made some tough choices over the course of the year to streamline our research portfolio down and focus on really kind of generating clinical data and value for our stakeholders.

And so in terms of the pipeline in front of us, our FGFR2 program, we out-licensed to Elevar and hopefully, they'll file an NDA imminently. Our PI3K alpha mutant selective program, RLY-2608 is about to start a pivotal Phase III trial, which we're very excited about in a large patient population. Behind that, we have a vascular malformation program. And then behind that, we have 2 assets that are nearing IND. That's an NRAS selective program and a [ Fabry ] program. So I think we have cash into 2029 now. And so I think we find ourselves in a very strong position with meaningful runway ahead of us and significant clinical catalysts.

So starting with your PI3K-alpha inhibitor [indiscernible] could you walk us through the design features that position it well within this evolving HR+/HER2- breast cancer paradigm? And how is it different from competitors such as Lilly?

Yes. So to start with and to level set, hormone receptor positive, HER2- breast cancer is a very large patient population, and 40% of these patients have a PI3K-alpha mutation. And traditionally, they've been underserved. And all of the kind of emerging approaches have unfortunately led to in the post-CDK4/6 world to a PFS somewhere between 5 and 7 months. And so we've seen a lot of promise around the oral SERDs, KAT6, CDK4/6 retreatment. But unfortunately, all of the data that we've seen over the last few years has still led us to this situation of PFS in the 5- to 7-month range. And the challenge for PI3K inhibitors over the last few decades has been a lack of selectivity. And so unfortunately, the inhibitors are nonselective, and that leads to a range of toxicities, including hypoglycemia, diarrhea, rash, stomatitis, which means these patients can't stay on therapies.

And so when we set out almost 10 years ago to create a selective inhibitor, we find a novel allosteric pocket and created an inhibitor that was mutant selective. And the hope was it would dial out these toxicities, allow greater dose intensity and that would translate into greater efficacy. And so that's exactly what we did. We entered the clinic as the first mutant selective inhibitor a few years ago. In terms of other mutant selective inhibitors, we know that Lilly has acquired STX-478, which has a similar [indiscernible] of at least kind of 12 to 24 months behind. And then there are a range of 1047R specific mutant selective inhibitors, which only target half of the population. And so I think we find ourselves to be ahead and differentiated from the current standard of care.

And last week, you presented updated dose escalation and expansion data in combination with fulvestrant at ASCO. What did you see as the key findings with regard to the updates? And how does it inform your view on the probability of success of the Phase III study?

Yes. The data we showed at ASCO was a continuation of a longer follow-on now, so 12.5 months worth of median follow-on from data that we had presented before at SABCS and in corporate updates. So the key was we'd already shown back in December of last year numbers that the class has not seen before. So we showed a PFS response rate and AEs that were just better than anything we've seen in the class. The key thing for us was just with a longer follow-up, could that data hold and be consistent. And that's exactly what we saw at ASCO last week now. So we showed in heavily pretreated patient population in the second line and the second line plus, a PFS of over 10 months. And in true second-line patients, we showed a PFS of 11 months.

We showed a confirmed objective response rate across all PI3K-alpha mutations of 39% and we showed in the subset of kinase domain mutations, a response rate of 67%. So these are numbers that are several fold greater than anything we've seen in the class before. And I think also very encouragingly, we show a very clean safety profile, very low rates of Grade 3 hyperglycemia, diarrhea, rash, stomatitis that have obviously plagued the nonselective inhibitors. And so I think what we were most excited about was the consistency of the data that held up with a much longer follow-up now over 12.5 months. That gives us great confidence to now start and run the Phase III.

And remind us why you see 5.5 months for PFS as the competitor benchmark?

The -- we're going head-to-head in the pivotal trial against capivasertib, AstraZeneca's AKT inhibitor, which is becoming the standard of care and it's off to a great start in its sales trajectory. And so that has become the standard of care in the field because of its perceived better safety profile. Although in the real world, we're starting to see now that, that may not be holding up. But in their pivotal trial, CAPItello-291, the subgroup analysis of patients that had already been treated with a CDK4/6 inhibitor that were PI3K alpha pathway altered, the PFS in that group was 5.5 months.

Now that included PI3K-alpha mutations but also AKT and P10 mutations. We know actually the AKT and P10 mutations do actually better. So the 5.5 months may actually be an overestimate for true comparator line for us. But that's the benchmark. The numbers that we showed, obviously, 11 months, there's a significant benefit for RLY-2608 versus CAPI. And so we're encouraged that we feel that we'll have a good chance of being successful.

And for the Phase III trial, could you speak to the population that you will evaluate and how you anticipate an initial label may look like?

Yes. So the population that we're going into is patients that have already been pretreated with a CDK4/6 inhibitor. So traditionally, that would be second line. But now obviously, we're seeing greater use of the CDK4/6 inhibitors in the adjuvant setting. And so we might see earlier and earlier use of RLY-2608. Although given the kind of uses has just begun in the Western countries, we imagine there'll be only a small number of patients that are adjuvant treated in our trial. But obviously, we want the label to reflect this because as you go out into the future, we imagine a much greater use of CDK4/6 inhibitors in the adjuvant setting. In terms of the kind of metabolic inclusion criteria, they're much more inclusive than some of the trials such as the INAVO120 trial at the approval of inavolisib. And so we're trying to offer this to a broader range of patients as we possibly can. So in summary, post-CDK4/6 patient population is what we will try and include in this.

And there's been some scrutiny around the need for a statistical -- statistically significant overall survival for oncology approvals. Just what gives you confidence here in your PFS primary endpoint for Phase III?

I mean there's a long track record of PFS being the approvable endpoint. All of the agents that we've seen in this class over the last few years have had PFS as their approvable endpoint. We know that there has been some speculation around how the FDA will approach this. It's probably not for us to speculate. All we can say is that we had a very successful engagement with the FDA. We will use PFS as our endpoint, but we'll also reserve alpha for overall survival. And so we think we feel confident around the design of having both endpoints.

And how are you thinking of the cadence that you will report additional data cuts from the Phase II study?

So I think we've just done that, exactly that, which is we reported data last September in a corporate update, and we showed data again from the same cut in December of last year at SABCS. And then I think we've just provided another update on the same data set back here at ASCO in June. So I think we've done what we can. This last update was really focused on the medical community. As we're about to start a global Phase III trial with a range of centers across the world, we want to make sure that this data is in the hands of medical professionals around the world.

So it's unlikely that we'll now update this data. 12.5 months, it's relatively now well baked, and we don't see it really changing too much more. And so I think we'll probably provide no further guidance on if and when we'll update this data. I think what we may do in the future is provide a kind of full Phase I/II manuscript describes it all. But I think in terms of this data set, I think this is probably the last data set that we'll show and probably not guide to anything else.

And maybe speak to your strategy for potential frontline expansion through triplet combinations and when we might see that data?

Yes, I think we are very excited about trying to provide 2608 to patients along their kind of treatment journey. And so obviously, this first pivotal trial that we'll run will be in post CDK4/6 treated patients. And so as we start to think about how we can move this in earlier and earlier lines of therapy, it does make sense for us to explore a range of combinations. And so I think over the last year, we've been exploring a triplet combination with 2608 plus fulvestrant plus ribociclib, which is the current standard of care. And then 2608 plus fulvestrant and then Pfizer's selective CDK 4 atirmociclib. And so we're not guiding to when we'll show data, we'll wait until we have a kind of robust and interpretable data set, but this kind of dose exploration and combinability is the first step for us in terms of defining what the next regimen would look like as we think about going early.

And will you, I guess, when you look at these triplet combinations here, will you ultimately choose only 1 triplet combination to advance either the CDK4 or the 4/6 inhibitor?

Again, we're not guiding to exactly how and when we'll make this decision, but it does make sense that obviously, as we start to look forward to the next pivotal that we run that we will choose 1 of these regimens. In terms of how we'll make that decision, it's really just going to be based on the data, the combinability, the AE profile. I think we have great confidence looking at the inavolasib triplet that they did with palbociclib that this is a valid mechanism in early alliance. Obviously, at ASCO last week, we saw overall survival data for that triplet to be positive, and there's great excitement around it. I think with RLY-2608's profile of having a very clean safety profile and that obviously translating into efficacy, we think there is a position for it in early and earlier lines.

In terms of what's the ideal regimen, these patients are going to be on treatment for a significant amount of time. And so the tolerability of this triplet is going to be important. And so obviously, the selectivity of 2608 is very helpful to that. And obviously, there is this theory that the CD -- selective CDK4s will also be more tolerable than the current CDK4/6s, taking away some of the kind of cytopenias that you see with the CDK4/6s and so obviously, we have to see the data play out. But I think there is a real emphasis on the cleaner profile as possible as we start to stack these tolerabilities on each other.

What could the design of a potential frontline study look like in that context?

I think I still -- we're just going to follow the data. So I think on that one, it's probably too early to define exactly what that looks like. I think we are very focused on trying to make 2608 available to a broader patient population as is possible. And so before we take that into account as we think about what the design will be.

And what about tumor types outside of breast here?

Again, PI3K alpha mutations exist in a range of solid tumors. And obviously, we focused and other developers are focused on hormone receptor positive/HER2- breast cancer, given it's such a large group of patients that are very well defined and identified. We know that other developers have started like Roche to look at triple-negative breast cancer. And so obviously, that could be an area of focus for us as well. And obviously, tumor types outside of breast cancer in general, colon, lung, I think all could benefit from this treatment, but probably not going to be a kind of short-term focus for us. I think given the focus on execution we have, it will remain focused on the pivotal trials that will run in hormone receptor positive/HER2- breast cancer.

You start with clinical development in vascular malformations, driven by PI3K alpha. And this is clearly a group that's very heterogeneous. You have varied manifestations. How are you thinking about which indications to initially pursue and the path to proof of concept?

So vascular malformations is a kind of very large unmet need. There's about 300,000 patients in the U.S. that have this umbrella condition named vascular malformations. Just over half of them, about 170,000 have a PI3K alpha mutation, and there are about 4 different phenotypes that can be grouped and I think for us, the kind of 2 phenotypes that we'll focus on now are PROS initially, PI3K-related overgrowth spectrum, and then a second much larger sector lymphatic malformations. There are about 80,000 of those and about 80% of them have PI3K alpha mutation. So still very significant numbers, but those are the 2 areas that we'll focus on. The PROS group is where Novartis' alpelisib has the accelerated approval, and that's where they're generating commercial sales at the moment.

And is an accelerated approval pathway here feasible in the context of these diseases?

I mean, over the years, our dealings with the FDA have always taught us never to speak on behalf of the FDA. And so all we can say is alpelisib today only has an accelerated approval. And their recent confirmatory trial, EPIC-P2, unfortunately, didn't meet the endpoints. And so we don't believe that, that will be converted anytime soon into a full approval. So I think all we can do is go full speed ahead at generating proof-of-concept data and then check in with the FDA, both the start of the pivotal and the submission of the NDA. But nothing today would say that the pathway of accelerated approval is off the table.

What about diagnosis of these patients? Are they currently being diagnosed? And so you have this pool of patients you go into? Or is there going to need to be a significant amount of identification and education on your side?

I mean I think it's an evolving kind of nascent field. And so unfortunately, the manifestation of these patients is very broad in terms of just the phenotype. And obviously, they present their primary care physician. And from there, depending on the severity of their condition, they end in various different referral centers. But there are only a small number of kind of vascular malformation centers in the U.S. and in kind of the major geographies across the world. And so I think, yes, there will be some education required. But in terms of for a company like us to deal with commercialization, given the kind of small number of centers that these patients do end up at, it is actually going to be quite manageable. And given the large number of patients here, 300,000, we do think it is actually a very attractive place for a company for us to operate.

Touching on your out-licensed selective FGFR2 inhibitor, can you provide an update on how the development is progressing here with your partner?

Yes. As you may remember, we out-licensed this program, our FGFR2, lirafugratinib to Elevar late last year. And they've been a great partner to work with. They are fully focused on getting this in the patient's hands across the world as rapidly as possible. And so I think they've been publicly stating that they will file an IND, hopefully, late this year or early next year. And so they've been a great partner to work with, and I think we'll continue to do everything we can to help them file that NDA.

You also have these Fabry and NRAS programs that you laid out in an R&D Day, but you're clearly -- you put it on the sidelines for now based on your resource allocation strategy. Walk us through, a, kind of the -- your interest in these programs, but b, what you think will end up playing out and could partnership be an opportunity for 1 or more of them?

Yes. I mean we are very excited about both of these programs. The NRAS selective program, I know that the whole industry has been working on trying to provide a NRAS selective inhibitor, and we've been able to create one. And obviously, it's a good scientific discovery. The utility of the NRAS inhibitors could be in a broad range of solid tumors, probably focusing initially on melanoma. So we feel very good about the potential patient population for these and the same for Fabry. Galafold from Amicus has shown great promise over the years and obviously generating significant revenues. But as many of you know, it only serves a part of the population. And it has a range of efficacy within that patient population.

So if you could just get to a broader range of the mutations and make them more amenable, we think that we could benefit a broader range of these Fabry patients that are currently underserved or not served at all. So I think we feel great about both programs. Now obviously, we sit here in 2025, in a very tough capital environment. And so I think we've done the prudent thing here and taken what we have is a very extensive balance sheet. We have $710 million at the moment and make sure that we can generate as much value in the medium term as we can. And so that's why you see us extend cash into 2029, which we believe will be well beyond the top line readout from the Phase III we're about start, allow us to generate proof-of-concept data for RLY-2608 in vascular malformation, allow us to generate the dose exploration work that we'll do with triplets and define an earlier line pivotal trial for RLY-2608 in breast cancer.

And so we'll bring both of these programs that we've talked about here, the NRAS selective and the Fabry program to IND and then look for options. And those options could be that things are progressing much faster than we imagined in the programs that we've just talked about and allow us to bring them in the clinic ourselves or as you rightly point out, look for partners. But we are very excited about the programs that we have. And obviously, we have this still research footprint that's focused on trying to solve some very hard problems.

And how about your manufacturing footprint at this point? Just as you look at these pivotal trials, your ability to then supply commercially?

Again, the great thing about all the molecules that we work on small molecules. And so the manufacturing is something that we've thought a lot about, it's all entirely outsourced. And so we can flex up and we've already started to think about the commercial supply for RLY-2608, have to think about launch later this decade now. As you know, it's not much of this decade left. And so we have to start thinking about it. But it's a much clearer problem because it's a small molecule. And so I think we feel good about our ability to manufacture at scale.

And how do you think about the field at this point, this motion-based design? Or the use of various AI/ML tools to create drugs. Just speak to how you stay at the forefront of what's playing out from a technological standpoint.

Yes. I mean those of you who have followed us over the years, we're the biggest skeptics of this. And that's because every day, our scientists are kind of at the cold face working with thousands of different problems. And so we've always found over the years that you have to really find individual discrete granular problems and then use these approaches to make those problems just slightly easier than they were prior. There's no magic button here where you can push and suddenly it provides you with an inhibitor. And so for us, some of the kind of hype and hubris that has been around over the years, has not unfortunately played out. We see this as a very complex task, making a medicine. And so the problems that we face is slightly different in every time that we approach a program. And so we stay ahead by just continuing to work on hard problems and see what works and what doesn't work.

And over the years, we've worked on many new computational approaches that have kind of fallen by the wayside. And I think over the last year or so, we've been able to hone that down onto the techniques and programs where we think it will going to have the highest impact. And I think we have to hold ourselves accountable. These companies should only be judged on their output. And over the kind of 10 years that we've been around, we've generated now many INDs, multiple clinical [indiscernible] and then multiple proof-of-concept data sets. And hopefully now, we'll be closing to filing an NDA for lirafugratinib [indiscernible] trial for RLY-2608.

So we think that's really the only measure. And any company that's talking about data, like how much data they have and how many preclinical programs they've generated. We don't think that's the right measure. And the right measure is how many drugs have you got approved?

As a last question here, what are you focused on from an execution standpoint in the near term? And is there anything you want to touch on with regard to strategy or the portfolio that we didn't speak to?

No, I think we've kind of covered the portfolio pretty well. As you rightly pointed out, been able to hone down our focus to very tangible clinical data that we want to generate over the coming years. And in terms of execution, that's all we're focused on there is standing up this pivotal trial, opening sites across the world for our ReDiscover-2 trial, generating awareness amongst the medical community and then generating this data and then making sure we're ready to file an NDA and then commercialize. So the entire company is entirely focused on that.

Well, with that, thank you so much.

Thank you. Thank you, Salveen, and thank you to Goldman. Thank you.