Relay Therapeutics, Inc. (RLAY) Earnings Call Transcript & Summary

Great. Good morning, everyone. Thank you so much for joining us. Really pleased to have with us the Relay team. With us, we have Sanjiv Patel, President and CEO; Don Bergstrom, President and Head of [Technical Difficulty] and Pete Rahmer, Chief Corporate Development Officer. To start here, Sanjiv, last week, you disclosed 3 new pipeline programs. Could you take us through the rationale behind pursuing these opportunities and how you leverage your platform and their design and just maybe to level set just an overview of the Dynamo platform and what we'll see for your pipeline over the next 12 months?

Thanks to Goldman for the invitation. I think we've all enjoyed the 24 hours stuck in this building. And so to come back to your question, the Dynamo platform, there's a lot talked about the computational tools and their impact on broader society. We see the platform as just a constellation of tools. Those tools are both experimental and computational. They're combined together to solve very discrete problems slightly better than they're currently sold. And as you know, it takes hundreds and hundreds of steps to make a medicine. And so if you just stack it all together, we do things slightly more efficiently, slightly more effectively. And hopefully, in the end, it results in solving challenges and problems that others can't solve. And so we target the platform towards as low as risk as we can take in the clinic and on the biology side. And so it's a validated targets with genetically defined patient populations, where we think the market opportunity is significant. And it's amenable to our approach, and that has led us to traditionally precision oncology, which is why again last week what you saw us stacking on top of our PI3K-alpha franchise, a selective NRAS inhibitor because it fits exactly the phenotypes that we've just described. And then several years ago, we used the same criteria to identify genetic diseases as an area of interest for us. And so what you saw last week was our unveiling program with Fabry's disease, fits the criteria that we have described. And then finally, we unveiled a PI3K-alpha mutant selective program targeted against vascular malformations. And so we believe these are all very attractive markets, low biology risk in the clinic, significant opportunities commercially and amenable to our platform. And so maybe I'll hand it over to Pete and talk about just kind of the data that we're going to see over the coming 12 months.

Yes. So the -- probably the most anticipated data set we have upcoming is for our PI3K-alpha mutant selective inhibitor, RLY-2608 in our ongoing breast cancer study, called the ReDiscover study. There, just to remind folks, that molecule went into the clinic in early '22. And we showed pretty clear and definitive differentiation across safety, tolerability from the nonselective inhibitors and kind of the open question here is, given that the regulatory endpoint is, progression-free survival is generating enough data at an optimized dose with enough follow-up to be -- give us all a bit of a sense of what that durability and potential PFS difference looks like. And that's really what we're aiming towards doing in the back half of the year with the 2608 disclosure we have coming up. So it will be a pretty robust data set at an optimized dose for us, that's 600 milligrams twice daily in combination with fulvestrant. And what we are guiding towards is having at least 40 patients that will have had at least 6 months of potential follow-up. And that's relevant because the comparator -- the most likely comparator in a potential registrational study in this setting that's 5.5 months [indiscernible]. So it's working towards this data set of hopefully very clear and interpretable differentiation also on the efficacy side. The other guidance that we have, milestones in place for the year, upcoming is an update on our FGFR2 inhibitor, RLY-4008. And so you may recall when we showed our initial tumor agnostic data in the fall of last year, we said at that time, you're going to deprioritize the regulatory and commercial efforts for cholangiocarcinoma and let the tumor agnostic data mature and minimize the spend against that program until we get a clear picture of what the full profile of the molecule is across tumor types and then also have a regulatory interaction and really understand the regulatory path forward. Both those things are continuing to happen and guiding towards that update in the second half of this year also. So pretty rich kind of data strategic update coming across these -- there are 2 lead clinical programs. And then the new programs that Sanjiv mentioned, for vascular malformations, will be going into the clinic in Q1 of next year with RLY-2608 initially, but we do have an additional pan-mutant selective molecule that we can move into development there. It would likely take forward for any pivotal studies in that indication. And then the Fabry program will go into the clinic in the second half of next year and also the NRAS program.

So let's start with 2608. You've spoken about the knowledge and experience that you've built up over time with PI3K alpha. Could you walk us through your strategy for the franchise, including how you plan to position yourself in the evolving HR-positive HER2-negative breast cancer paradigm?

I mean I think we've always seen PI3K-alpha as a significant precision medicine opportunity. And I think -- again, we've laid out 3 areas. There's obviously breast cancer. Inside of breast cancer, there are multiple lines of therapy and there are multiple combination partners. And so just having a suite of assets there, especially in the age of IRA is important to us. Second opportunity is distinct and sits outside of solid tumors in vascular malformations. And so obviously, that consists of a range of different diseases inside of that. So again, a different molecule that's distinct, maybe advantageous to us. And then the final pillar is solid tumors outside of breast cancer. So we think that all 3 of these represent significant commercial opportunities. And maybe we'll just walk through them 1 at a time. So maybe Don, you can just take the landscape around breast cancer.

Right. So in breast cancer and hormone receptor positive, HER2-negative breast cancer, current first-line standard of care in the metastatic setting, is usually an aromatase inhibitor combined with the CDK4/6 inhibitor. And then in the second line in patients who have PI3K pathway mutations, we've now seen 2 approved agents alpelisib, marked as Piqray by Novartis, is a nonselective PI3 kinase inhibitor, again, used in -- primarily in second-line patients post CDK4/6 inhibitor. Real-world PFS is about 6 months and comes with significant toxicity liabilities. And a real challenge for patients to stay on that drug with 1 of the hallmark toxicities being hyperglycemia and especially grade 3 hyperglycemia, which oncologists are very uncomfortable managing and frequently will discontinue use of the drug once they encounter that. The more recently approved agent that Pete mentioned would be our most likely comparator is capivasertib, which is an older, relatively nonselective AKT inhibitor, marketed as Truqap by AstraZeneca. I think we're seeing the market move towards that, largely driven by the fact that the rate of grade 3 hyperglycemia is reportedly lower than what you see with alpelisib. And again, that's a toxicity that treating physicians want to try to avoid. And that is in the setting of no better PFS than what you see with alpelisib. The clinical trial PFS was 5.5 months for capivasertib. So as we think about how you position 2608 in this landscape, what we've already shown with 2608 is proof of mechanism that we can achieve our target inhibition of mutated PI3K-alpha, while sparing hyperglycemia not seeing grade 3 hyperglycemia. And that's at our 600-milligram dose as we move forward. So I think that positions 2608 well to be a treatment paradigm for the PI3K mutant patient population. The first pivotal trial that we could likely start would be a doublet fulvestrant in patients who have previously been treated with a CDK4/6 inhibitor and most likely, the comparator could be capivasertib, again, with a benchmark of 5.5 months progression-free survival. There's also opportunities to move earlier in the treatment paradigm. We have seen positive data from Roche with inavolisib, their nonselective PI3K inhibitor in frontline in combination with palbociclib, a CDK4/6 inhibitor. But that was a place where their patient population was very, very heavily selected for patients with good metabolic fitness, using metabolic inclusion criteria that would exclude 50% or more of a typical western population. So we think there's opportunities to move earlier into the frontline setting in combination with the CDK4/6 or a CDK4 targeted therapy. So that comes back to the other news that we disclosed last week. That we've entered into a clinical trial agreement with Pfizer to combine RLY-2608 with their selective CDK4 inhibitor, which we view as an emerging standard of care in the treatment of ER positive, HER2 negative breast cancer.

There's clearly been a lot of interest in the pan-mutant isoform-selective PI3K-alpha inhibitor, the 2608, as you talked about. At the Phase I/II update last July, you reported clinical benefit rate -- response rate and safety data in combination with fulvestrant. At this point, what are your thoughts on the dose level and efficacy profile?

You want to take, Don?

Yes. So when we made that disclosure last August, it was based on the data that we were seeing across our dose escalation cohorts where we had identified 600 milligrams is being a potential optimal dose. Based on the fact that we're achieving our target inhibition, we were clearing the mutant ctDNA. We were seeing early efficacy data that was promising. We're starting to see responses. We're seeing good duration of benefit. And importantly, on the tolerability side, we weren't seeing grade 3 hyperglycemia. Overall, we were seeing a good overall tolerability. And importantly, in this disease, patients were able to stay on drug with their scheduled dose intensity. And that's really what you need to be able to win in a PFS-driven environment, especially as you move early where you're looking at patients potentially having to take your drug for 2 or even 3 years, having that chronic tolerability is key. So this 600-milligram dose was very promising to us. We had announced last August that we would open 600 milligrams as our first expansion cohort. We opened that by the end of last year. We had dosed about 40 patients at the 600-milligram dose. We also did simultaneously open a 400-milligram cohort. 400 milligrams is really what we consider the minimal biologically effective dose. And we did open that cohort, not because we feel it's an optimal cohort, but to provide us with the data set that would make us well positioned for interactions with regulators, especially in the Project Optimus era.

And you next plan to present data from the doublet in the fourth quarter where you guided that at least 40 patients treated at 600 BID will have at least 6 months of follow-up. What are the relevant clinical benchmarks here? And how are you thinking about the interpretability of the upcoming efficacy data?

Pete?

Yes. Beauty of this one is it's quite simple. It's 5.5 months. So the clear burgeoning commercial standard of care in this setting is capivasertib, as Don alluded to. And they had a very robust clinical study CAPItello-291. And in the true second-line patients that are pathway altered. They had a median PFS of 5.5 months. And so we do believe that the data that we intend to present, as you outlined it, will be quite interpretable against that benchmark. And again, our goal is unequivocally to show differentiation against that. We can do that in a couple of ways in this data set. First and foremost, 1 metric we can report on is 6-month PFS landmark analysis. And that is extremely relevant here given that the benchmark meeting PFS is 5.5 months. So if we can see a 6-month PFS landmark analysis rate decently north of 50%, especially give us all confidence that in the head-to-head study, well-controlled randomized trial that we could beat that 5.5 months mark. And so I think that's going to be probably the key metric that folks will be looking for. We'll also report on clinical benefit rate. Again, that's a 6-month endpoint. So it's the stable disease or better at 6 months, which is helpful in this comparative -- across our comparison context. And then we will report on the typical metrics of response rate and other supportive endpoints, safety and tolerability full update. We should have over 100 patients in a safety database at that point in time, over 60 of them at 600 milligrams twice daily. So it should be a pretty robust and interpretable data set.

You'll also have triplet data with CDK4/6. Maybe help us understand what you hope to learn from this data set and whether it can enable you to move to earlier lines.

Yes. I mean as we've seen over time, we don't think a mutant selective PI3K inhibitor -- I'm sorry, nonselective PI3K inhibitor has been able to combine [Technical Difficulty] which is first becoming the established market leader in the CDK4/6 world. Obviously, for us, we feel very confident that we can demonstrate that we could get a combinable dose of RLY-2608 with ribociclib. And so that would be the goal in any disclosure to show safety and tolerability. Obviously, as we announced last week, we've also now have a clinical trial collaboration with Pfizer to combine RLY-2608 with their first-in-class selective CDK4. And so our goal is obviously to show that if you want to go earlier and earlier, the kind of safety and tolerability is important as you stack on each of these various AEs on top of each other and so again, the selective approach that we have with RLY-2608. And we believe the early data we've shown -- clearly shows that there could be a potential safety and tolerability advantage of the nonselective inhibitors makes it better for patients that are going to be on therapy for long periods of time in early and earlier lines. So I think that's the goal of the disclosure would be to show combined ability and obviously, safety and tolerability.

And once you do have these 2 data sets in hand, how are you thinking about the registrational pathway from there?

Maybe, Pete, do you want to take that one?

Yes. I think the -- we have a couple of regulatory interactions ahead of us. One, kind of as you -- part of your question earlier on dose, we have to go have a conversation with the FDA on dose. That's why we're accumulating these data at both 600 and 400 just to enable ourselves to have that conversation in the back half of the year. And then additionally, we are currently thinking about and designing a potential registrational study in the second-line setting. And so once we have that -- once we've had the dose conversation, we also need to have kind of an end of Phase II -- conversation around the Phase III trial design. And so that -- those will probably be sequenced. And so that would enable us to be ready to do -- to enter into a Phase III study -- start a Phase III study in the second line setting likely in the middle of the back half of next year. That trial, we believe, is the -- some of the considerations of that trial that we need to really continue to think through is in such an evolving landscape. The level of pretreatment with CDK4/6, level of pretreatment with chemotherapy, prior fulvestrant, the studies that have come before us that we will be compared to, particularly CAPItello-291, had very minimal CDK4/6 pretreatment, 30% had no prior CDK4/6 and the rest had 1%. On the fulvestrant side, they had 0% of patients had prior fulvestrant treatment. Chemo was below 20%. In our ongoing ReDiscover study, we do have patients that have seen multiple CDK4/6 inhibitors, have been -- have seen 50% or so of prior fulvestrant. The chemotherapy pretreatment range in the metastatic setting is upwards of 30%. So it's going to be pretty important for us to think about running a Phase III trial and really a contemporary patient population. And our goal here is to design a study that the outcome is -- can be interpretable and have utility for the broadest patient population possible. In the second-line setting, in HR-positive, HER2-negative PI3K-alpha mutated patients, there's 14,000 patients a year in that setting. So if you can provide a differentiated treatment for those patients and one that they can stay on for a good amount of time, that's an extremely meaningful market opportunity for a company of our stage and size. Let alone starting to look into the future, and that -- this is what some of the supportive triplet studies are doing, is really enabling us to think about that next pivotal study, which would be in the frontline setting. That would start to unlock 18,000 to 20,000 patients per year and then you're talking about years of potential treatment in that setting. So this kind of the building blocks we're putting in place now to enable us to do over the next couple of years.

And I know it's early days, but can you speak to the time lines of the combination program with Pfizer CDK4?

Pete, do you want to answer?

Yes. So obviously, that deal was just signed. We're guiding towards initial -- initiating that study by the end of this year. So I think -- let us get a little bit closer to treating a patient in that setting and then we can start to think about guiding to potential data there.

Don, was there anything to add with regard to the trial?

No.

You plan to start the clinical development program in vascular malformations, driven by PI3K-alpha in the first quarter, as you mentioned. How are you thinking about which indications to initially pursue and the path to proof of concept?

I mean, so it's a big gain opportunity for us, over 170,000 patients with PI3K-alpha altered vascular malformation. And it's a spectrum of different conditions. So maybe Don, you can just kind of help us educate around the different diseases involved in.

Yes. So I think we're really -- when you're looking across the spectrum, the first that we'll talk about is PROS or the PI3K-related overgrowth spectrum, which is really a collection of distinct diseases, but the 1 commonality as all of them are driven by mutated PI3K alpha. So that's a population that's 100% mutated. It's a population where today there's existing proof of concept from alpelisib, where there's some accelerated approval in this patient population. And that was based on a retrospect cohort review of 37 patients that were treated in a compassionate use program using a primary response endpoint of radiographic response in the lesions, 20% volumetric reduction. They saw a 27% response rate. So 1 of the first questions we can ask is we actually have a benchmark in that population. We know PI3K is a driver in that population. And 1 of the first questions is in patients with severe PROS coming in with a mutant selective inhibitor, can you take advantages of having, again, that chronic tolerability that comes from sparing wild-type PI3K-alpha and ultimately drive the response to a higher than 27% response rate using the same endpoint. As you then go across the other subtypes of vascular malformation is the 1 within the most frequent PI3K mutation is lymphatic malformations where about 80% of patients have a PI3K mutation. There's not an existing benchmark today. There is no alpelisib trial running in that population. But again, this is a place where we think early on, we'd want to try to include patients with PI3K-mutated lymphatic malformations to establish the breadth of signal 2608. As you get into venous malformations and central cavernous malformations, the PI3K mutation rate gets lower. I think you'd have to do some prospective selection of those patients to identify the PI3K-mutated patients. Our trial, we anticipate would be open to those patients, but we anticipate they would not be represented at the same rate as PROS and lymphatic malformation.

Delve a little further into that, maybe speak to the commercial aspect here, the ability to identify these patients where they are now and how you think of this -- the ability to kind of initially go and get, then target that large market that you mentioned?

Yes. I mean, we believe it's a very attractive commercial model for a company of our stage. And maybe Pete can talk through the details.

Yes. I think 1 thing to make very clear is this is a pretty nascent setting -- disease setting. It's -- for those that tuned in last week to our R&D event, we had a KOL there, Dr. Adrienne Hammill, who's 1 of the first KOLs in this space, was trained under the pioneer in the space [indiscernible] the very first study that actually looking at vascular malformations was only conducted in 2009. So I think -- and I preface my answer with that because there's still going to be some learning to do here, especially in terms of commercial setting how to find these patients. But today, it sounds like in talking to some of these KOLs, there's about a dozen or so centers of excellence that exist in the United States. That are seeing a large bulk of these patients, especially those that have reached that severity level of seeking the very [indiscernible] systemic therapies that are available today. And so I do think within PROS, there's about 15,000 of these patients in the United States, completely driven by PIK3CA-alpha mutations. In lymphatic malformations, as Don mentioned, 80% of them are driven by PIK3CA alpha mutations. And so they -- in both of those settings, they're clinically diagnosed with their presentation, not necessarily needing a genetic test to confirm that diagnosis. And there, if you add those 2 together, you get into the neighborhood of 70,000 to 80,000 patients in total, just in those 2 disease settings. And today, it seems like about 25% to 40% of them are seeking systemic therapy and those that are seeking the systemic therapies seem to be concentrated across these centers of excellence, where you have a vascular anomaly specialist team. So there's definitely a good large proportion of the -- at least in those 2 disease settings where the commercial model is really amenable for a company of our stage and size. It truly is by a rare disease like commercial model. And we have some precedence on the -- it's always too early to talk about pricing, would you haven't even treated a patient in that disease setting, but there is some precedence on the pricing side there too. The Vijoice, which is the brand name of alpelisib in the setting, charges about $350,000 a year.

Switching over to your selective FGFR2 inhibitor, you will have data as well as commentary on the regulatory side in the second half. I guess where will you stand with regard to alignment with the FDA? What can you disclose at that time point?

Do you want to take that one?

Yes. So I think what we disclosed at the end of last year was that we are active in patients outside of cholangiocarcinoma that have tumors that are driven by FGFR2 alterations. With the response rate that we reported at the time last year in patients with FGFR2 fusions or rearrangements, that was consistent with some of the recent approvals that we've seen in a tumor-agnostic setting. So as Pete mentioned, we made the decision at that time to continue to build the database outside of cholangiocarcinoma to enroll more patients, to get more follow-up, to get a more precise and robust estimate of duration of response, speaking about the ability to then take these data at FDA and really ask a few key questions. Ask about the agreement that there's unmet need in this setting, which would be the benchmark for being able to pursue accelerated approval; ask about the data that we've generated is it sufficient in terms of the number of patients and importantly, the distribution of tumor types to be able to be considered for an accelerated approval path. So those are really the key questions that we need to come to agreement with the FDA. And we anticipate we'll have a regulatory interaction this year to be able to ask those questions.

And just to add on top of that. And I think with that information in hand, with the data we have in hand is then incumbent upon us to take a step back, look strategically at the entirety of the portfolio and the opportunity sets we have in front of us. And just really ask the capital allocation question and where can we produce the most value? And so I think that is the third piece of this that will be, one, discussing internally and then part of when we make -- when we report out data and kind of an outcome of regulatory interaction is really kind of try to give some guidance on the strategic path forward and how this fits into our very large growing opportunity of medicines to put our capital behind.

On the Fabry and NRAS programs, they're clinical bars, but they're also competitive bars in some cases. So help us understand how you're going to demonstrate differentiation through the studies initially so we can understand the proof of concept here.

Let me take the 1 at time, maybe start with Fabry, Don.

Yes. So in Fabry's disease, I think you're looking at a setting today where there is an approved small molecule chaperone. It is an inhibitory chaperone of the alpha-gal protein, which is the mutated protein that's disease causing in Fabry's disease. And that's important because it comes with some limitations of that therapy. It's got a very strict every other day dosing schedule. It's only active in a subset of mutations, representing about 40% of patients or so. And our hypothesis and the data we showed last week is if you could identify a more potent and robust non-inhibitory activator of alpha-gal, you should be able to do 3 things. One is you should be able to increase relative to what you can do with migalastat increase the amount of alpha-gal, that's active and stabilized the increase activity and showed a greater effect on substrate reduction. We've shown preclinical data supporting that for our molecules. So that would be 1 of the questions as we got into patients in early development. As you look at the biomarkers for alpha-gal activity, you see superior levels to what's been shown previously. We've also shown that we can address a broader range of mutations so that obviously would be another question as you get into some patients who today are considered nonamenable to migalastat therapy. And then finally, and this is more of a longer-term question from a development perspective, we've shown the ability to be able to combine with enzyme replacement therapy. So in the minority of patients that aren't appropriate for a small molecule activator or chaperone, then you combine with enzyme replacement therapy and increase the activity that you see there. So I think as we get closer to the clinic, we'll come forward with more specific clinical designs. But I think those 3 pillars of what we're trying to do with this program will be a goal to establish the differentiation early in the development of the asset.

And then on the NRAS, it's more of a classic precision medicine player, which is we think there are a large amount over 25,000 patients in the U.S. that have NRAS-driven solid tumors. If you could exclusively selectively inhibit NRAS, we believe that should translate into better clinical efficacy than the current kind of drugging other parts of the pathway. But you can do it without the associated tolerability challenges associated with this kind of nonselective and other nodes in the pathway. So it's a classic play, deeper inhibition, less tolerability challenges, should translate into a better outcome for patients. So Don, if you want to just talk about how we see it.

Yes. I think as we look across the spectrum of NRAS mutated tumors, you see concentration in melanoma, colorectal cancer, non-small cell lung cancer. It's actually a distribution that's very similar to what you see for BRAF mutations, the V600E mutation. And we would anticipate that as you are starting to look towards development of the asset, you'd focus on -- especially on those 3 diseases, with the goal, as Sanjiv mentioned, of being able to show that you can hit NRAS hard without seeing the associated toxicity associated with either pan-RAS inhibition or downstream node inhibition. And skin toxicity is 1 of the major categories of AE that you see in that space with the pan-RAS inhibitor from Rev Med showing about 80% rate of skin toxicity. We would believe based on everything we know about NRAS that if you hit NRAS selectively, you should be able to avoid that class of toxicity. So that will be a key question. We'll look to answer early on. And then I think we anticipate that you'd be looking for monotherapy responses across those NRAS-mutated patient populations to establish proof of concept.

Speak to the business development side here, where you have acquired and partnered to acquire technologies, how are they being integrated at this time point? And are you looking for more on the forward?

I think, as you know, the platform consists of this kind of constellation of different tools and approaches. And so we're probably 1 of the most experienced kind of exponents of using the tools in combination with each other to make INDs and ultimately, clinical compounds. So I think we see pretty much everything that's out there. We acquired almost 3 years ago now a company called ZebiAI that helped us build an internal capability around using big data in DNA-encoded library screens, much more kind of insightful way using machine learning to try and get the starting point much more rapidly. And so we've been able to successfully integrate that. Over time, our platform continues to evolve. We continue to incorporate new approaches, double down on things. I think, for example, the molecular dynamics that we do now internally, we pretty much know the capabilities that we need in-house. And so we can build things. There are plenty of tools that we've used over time that we found to be not productive. And so it's a constant evolution. I think on the business development side, probably our bigger focus is what to do now with this kind of very broad set of assets that we have in the clinic. Obviously, we talked about the 3 pillars of PI3K-alpha vascular malformations, breast cancer, other solid tumors. We have our FGFR2 inhibitor. Obviously, now we have Fabry's and NRAS. So that's a lot of assets to prosecute ourselves. So probably our focus is much more on how do we bring all these assets to patients as rapidly as possible. And we'll continue to look platform assets, but the threshold for us to bring 1 of them inside is high.

Great. Well, with that, Sanjiv, Don and Pete, thank you so much. Really appreciate the time today.

Great. Thank you.

Thank you.

Thank you.