Okay, I think we can get started. Hi, everyone. Thanks for attending the Wells Fargo Healthcare Conference. My name is Sadia Rahman , and I am honored to introduce our next presenting company, Relay Therapeutics. From Relay, we have the Chief Commercial Development Officer, Peter Rahmer. Thanks, Peter, for being here.
Yeah, no, thank you for having us. It's a pleasure to be here at the conference. I look forward to the chat.
Great. Let's start with RLY-2608 in breast cancer. At ASCO, Relay Therapeutics reported longer follow-up data with RLY-2608 from the rediscover trial, and that's looking really promising compared to benchmarks like capivasertib's 5.5 month PFS. Can you walk us through the most recent update at ASCO and how efficacy of RLY-2608 with fulvestrant is differentiated from approved PI3K and AKT inhibitors?
eah. Like you said, the main reason for the update at ASCO was really to show the further maturity of the data and the consistency of the data over time. We originally presented the doublet data at SABCS in December of last year. That had a median follow-up of nine months. We were seeing strong progression-free survival there with a median PFS of about 10 months. When that data matured further at ASCO to 12 months, that median PFS continues to hold up at about 10 months. In second-line patients, the median PFS continues to be around 11 months. We made the decision to move forward with the phase III trial in the middle of last year based off of the evolving rediscover data. I think these data continue to confirm our confidence in running that study and the POS of that study.
It compares, as you highlighted, quite favorably to what will be the control arm of that study. That study will be RLY-2608 plus fulvestrant versus capivasertib. AstraZeneca's AKT inhibitor is the current commercial leader in this setting, which has been demonstrated at about a five and a half month median PFS in kind of apples-to-apples patient population. We are feeling very confident as we move into the phase III study. The other non-selective inhibitors in the space tend to perform in and around the same. You have alpelisib that's approved in the second-line plus setting. It's also about seven months, six and a half months, depending on the data set you look at. We feel pretty confident in not only the near-term execution of the phase III trial, but then the eventual commercial opportunity that's in front of us given the landscape.
Digging into the ASCO data, were there any specific patient subpopulations within HR-positive, HER2-negative breast cancer where the combination performed better? How might that inform your positioning strategy?
Yeah. Maybe the most obvious statement to make is we are a PI3K alpha mutant selective molecule. That is the only population we are studying, HR- positive, HER2 -negative, PIK3CA-mutated patients. Within that population, the median PFS is 10.3 months. That's a near doubling in and of itself against the current standard of care. Within the different mutations, within PI3K alpha mutations, we do see a super performance in what is called the kinase population. It's about half of that population. There we have a median PFS in that subpopulation of about 18 months right now. We feel very excited and confident about the performance of RLY-2608 across all mutations and the potential not only in second line, but to think about moving forward into earlier line settings.
Talk about the safety profile. How do you expect the relatively clean safety of RLY-2608 to translate in the real-world setting, and maybe to impact real-world adherence and duration on treatment compared to the approved agents?
Yeah, it's actually a hallmark belief that we have at Relay that's really driven all of our drug discovery, both our target selection and drug discovery programs over the years, is that we go after genetically or clinically validated targets. Usually what we're trying to solve is, traditional means have not allowed the space to actually create, in most cases, a selective or potent enough inhibitor of the actual driver mutation or target you're going after. We've been able to be the first company to create a PI3K alpha mutant selective molecule. It kind of gives you two key advantages. One kind of begets the other. The non-selective molecules, the reason why they run into the inability to hit the target hard enough is because they're usually inhibited by dose intensity due to the tolerability profile.
By creating a selective molecule, we're able to increase that therapeutic index, hit the target harder, and that comes with a more tolerable safety profile. We think that that's going to be a hallmark characteristic of the molecule as it moves into the later stage development and commercial.
Can you provide some context on the hyperglycemia and GIAEs seen so far and how they compare?
Yeah. The key off-target toxicity associated with non-selective PI3K inhibitors is when you hit PI3K alpha wild type, it ends up leading to drug-induced diabetes or hyperglycemia. Our selectivity window has allowed us to push very high PI3K alpha mutant target inhibition while keeping the grade 3 hyperglycemia, which is the most problematic, at bay. We currently have a 3% grade 3 rate of hyperglycemia, and that compares quite favorably across the class of therapies. On the GI front, that's been specific to capivasertib. Capivasertib has a diarrhea rate of about 80%, and it comes with a lot of grade 2, grade 3 diarrhea. Our diarrhea rate is meaningfully below that, and the diarrhea rate we do have is largely grade 1, which is much more tolerable.
Got it. Okay. The phase III ReDiscover-2 trial recently initiated, and it's comparing RLY-2608 versus capivasertib plus fulvestrant in CDK4/6-experienced PIK3CA-mutated breast cancer. Can you provide an overview of this trial and how the data thus far has informed the design of the study?
Yeah. It's a global randomized study of RLY-2608 plus fulvestrant versus capivasertib plus fulvestrant, randomized one to one. There'll be 540 patients. The inclusion and exclusion criteria will mirror the rediscover patients pretty closely. The one thing we'll change a little bit is making sure the inclusion criteria pushes us more towards a second-line patient population. In rediscover, given that it was the first in human study, we were a bit more flexible in the number of prior therapies that patients were able to have. Here we'll be a little bit more strict and make sure we get into the true second line and some third line patients. Otherwise, it'll look and feel a lot like the rediscover study, the HbA1c. The metabolic criteria will be the same. We'll stratify for mutation type, visceral disease, and geography. The primary endpoint of the study will be progression-free survival.
It'll be tested hierarchically in the kinase patients and the intent to treat population. We have reserved some alpha to make OS a key primary secondary.
Got it. Can you discuss the decision to pursue the 400 mg BID fed dosing strategy versus the 600 mg BID fasted? What were the reasons behind this? What's your confidence in the selected dosing strategy?
Yeah. We've taken a path that is very common amongst oncology drug development where you start your initial dose exploration in fasted patients, and then you independently study a food effect. Over time, if you have a food effect, you can determine how to introduce that into your dosing schema and regimen. We probably explored the food effect a little bit earlier than most, probably. That was largely driven because we knew we were moving into vascular malformation, which is a disease where you really want to be able to treat. It's a genetic disease that you have at birth. It happens in utero. You are going to want to be able to treat those patients in a fed state. We started to explore fed dosing, and we identified that we had a positive food effect that increased the exposure of RLY-2608.
We then ran 400 milligrams fed in some patients to get some data there. We took all of that data to the FDA and showed that the 400 milligram fed exposure is quite comparable to the 600 milligram fasted exposure. The FDA had no questions about us taking the 400 milligram fed dose.
What are your expectations on the control arm of capivasertib plus fulvestrant?
Yeah. I think it's, as you stated earlier, the CAPItello- 291 data is probably the best data set to get that estimate from. They've shown a couple of different ways to look at that, the population that we're studying in. We are running this study, our study, in a 100% CDK4/6 pretreated population. The CAPItello- 291 study had about 30% of their patients that were CDK4/6 naive. You can't really look at the overall PFS. You really have to look at that CDK4/6-experienced subset. In that pathway-altered CDK4/6-experienced subset, those patients did five and a half months of median progression-free survival. That's really been what we have been benchmarking the pivotal trial design on.
Got it. How should we think about the powering for the median progression-free survival, primary endpoint, and key secondary endpoints such as overall survival? How confident are you that PFS is enough for approval versus also requiring the OS data?
Yeah. We feel quite confident that we haven't gone into the specifics of our powering assumptions. What we have said is you have, you know, the clear data that we've used to benchmark as we went into taking consideration for the designs of the study. You know, we've talked about the capi data showing five and a half months as a median PFS that they have been able to demonstrate. Our ongoing data, you know, 10.3 months across all patients, 11 months in the second-line patient population. Those were obviously key data sets that informed our trial design parameters. On the OS side, it's pretty well precedented in the past 5 years or 10 years in this space that meeting progression-free survival is a sufficient regulatory endpoint without the requirement for OS at the time of approval.
We have, you know, a number of these therapies, alpelisib, capivasertib, the alpelisib in a frontline setting, and others that have gotten approval on progression-free survival. We are, again, preserving some alpha to be able to have OS as a key secondary endpoint, but we don't believe that'll be required at the time of finalizing. You know, these are all data-dependent, regulatory-dependent.
Got it. Can you discuss the commercial opportunity in the CDK4/6-experienced patients, what the unmet need is there, and how could durability of RLY-2608 play a role in its commercial potential?
Yeah. We think this kind of second-line setting is a massive commercial opportunity. Our TAM calculations put it at about $2 billion- $3 billion in the U.S. and major global geographies. There's probably about 30,000 or so patients in those major geographies. You start talking about a therapy that can treat these patients for hopefully something around a year. The durability is really what makes that TAM bigger than what some of the precedented therapies base are, because so far, our data is in the neighborhood of doubling the duration of therapy in which these patients can be treated. The durability is a meaningful contribution to the commercial opportunity. The unmet medical need, for the past 10 years, unfortunately, we haven't seen much of an improvement in second-line progression-free survival.
Whether it's CDK4/6 retreatment, the non-selective inhibitors, NextGen oral endocrine therapies, and different permutations thereof, you end up consistently coming out with somewhere between five to seven months of progression-free survival. I think the physicians in the space are really excited and really eager to have meaningful improvements in therapy for the patient.
Great. Beyond the doublet, you have ongoing triplet studies where you're looking to move into earlier lines. Can you discuss the triplet combinations that you're evaluating?
Yeah. We currently have a protocol open that allows us to explore triplet combinations with ribociclib, temociclib, so Pfizer CDK4/6 inhibitor, and palbociclib. We think that, you know, one permutation, one of these regimens will be what we would anticipate taking forward into a frontline study. We're in the midst of generating that data. I do think it's data that we'll have in hand within this cash runway window and should be able to provide clarity around that frontline strategy with that data in hand. We're encouraged with what we're seeing thus far, especially in the temociclib arm of the study. We have a lot of investigator enthusiasm for that regimen. We do think that there's a medium-term path to where we can likely pursue a frontline registration study as our next phase III study in breast cancer.
Got it. For these triplet dose finding studies, what do you need to see to move these programs forward? What are you focused on on the tolerability side? What are your expectations there?
Yeah. First and foremost, just establishing combinability because that's been a barrier for some of the other non-selective inhibitors. I think establishing a dose and combinability with these molecules, optimizing the dose with these regimens. I think what we're going to look for is that all the toxicities you're seeing are just the contribution of the parts. You don't want to see additive or, sorry, you don't want to see synergistic toxicities. I feel like we're pretty encouraged thus far with what we're seeing on that end. I think we would make the decision based off of safety, tolerability, and probably early efficacy measurements such as response rates, clinical benefit rates. The current standard of care, CDK4/6 plus AI in these patients gets about 16-1 9 months of median PFS in PIK3CA-mutated patients in the frontline endocrine sensitive setting.
We would probably make a decision on how to move forward before we would have mature median PFS data from this regimen.
In relation to ribociclib versus temociclib, how do you view the trade-offs there? How do you decide between which one you would move forward?
Yeah. I think the easiest answer there is we'll make a data-driven decision. Ribociclib is the only molecule of the three CDK4/6 inhibitors with an overall survival benefit, a clinically significant overall survival benefit in the endocrine-sensitive patient population. Therefore, it has been gaining a decent market share in the frontline metastatic setting. Palbociclib still maintains about a 40% - 60% market share in the frontline metastatic setting, so still a very significant combination partner on a global basis. Palbociclib goes off patent late 2027. That's also a consideration as you think about what regimen to move forward with. Ribociclib is probably one of the more toxic of the three. The key driver is going to be in which of the regimens do we think can bring the greatest benefit to patients in the frontline and the sensitive setting. We're in the midst of generating that data to inform that.
Got it. Any enrollment update, or can you talk about how enrollment is progressing for these studies and when we might see an update from them?
Yeah. We're happy with how they're progressing. We're not in a position to guide to data just yet on these, from the triplets, but you know, very happy with how they're progressing. We are eager to understand the data and be able to put a plan in place to move forward in the frontline setting. We want to make sure we have clear, comprehensive, interpretable data in hand and want to be able to articulate exactly what that frontline strategy is when we do make that decision.
I think you've said that the triplet data will be disclosed once there is robust data. What would constitute robust data in terms of, you know, patient numbers or duration of follow-up?
Yeah. I think you want to have an established dose and have a good end within at that established dose. You want to make sure you have a clear understanding of the safety and tolerability of that regimen. I think a robust enough of an end where the response rate point estimate, which will probably, it's the first efficacy point estimate you can get, is interpretable against the current benchmarks. In the frontline setting, ribociclib plus AI does about 50% from a response rate standpoint in the frontline endocrine-sensitive patient population. That'll probably be something that we'll be measuring against. That being said, the bulk of the patients that we are running these triplets in is in CDK4/6-experienced, so they won't be frontline patients.
You'll have to, we'll be weighing our response rate against what you would expect to see in that CDK4/6-experienced population, which from the CDK4/6 retreatment studies, you know, tends to be somewhere between 5% and 30% the other CDK4/6.
Mm-hmm. Got it. So then the NextGen PI3K space is becoming more crowded, but it seems like 2608 has a good lead on the competitor programs. Can you discuss how you view the competitive landscape for these NextGen molecules and where do you think you net out?
Yeah. In short, they're no different. They have not shown really any reasonable data-driven differentiation from RLY-2608. As you said, we are clearly first. We've established a pretty high best-in-class benchmark to be beaten with clinical data. We've seen some conjectures of different differentiated selectivity profiles, but I think we really, you know, if you start to think about the data sets that we put forth within the ends and the follow-up, you know, about 60 patients' worth of data in the doublet setting with fulvestrant over 12 months of median follow-up, the highest ORR, CVR, and progression-free survival that's ever been seen. In most cases, doubling the numbers we've seen with the non-selective inhibitors. I think any of the fast followers behind us, there's a high bar to beat clinically. Honestly, as we look at the preclinical profile of these molecules, don't see any clear difference.
kay. Moving on to vascular malformation, can you discuss the rationale for PIK3 and vascular malformations? Do you believe that the data available for alpelisib de-risks your approach in PROS?
Yeah. So, you know, just a vascular malformation is an umbrella term that captures a few different subsets of diseases, all driven by malformations, originating from different vessels within the body. PROS stands for PIK3CA-related overgrowth spectrum. That is a, it's an instance in where 100% of those patients have PIK3CA-driven disease, and it tends to be a multifocal disease, meaning they have manifestation of the disease at different organs or different parts of their body. That is probably the third, and I should say there's about 170,000 PIK3CA-driven vascular malformation patients in the U.S. from a prevalence standpoint. Within PROS, there's about 5,000 - 15,000 of that 170,000. It's the smallest subset, but it's 100% characterized by PIK3CA mutation. That's where, as you referenced, alpelisib originally pursued development. I should say it wasn't even Novartis that actually pursued that development.
That was actually driven by an investigator in France who was using the drug in a compassionate use setting. That initial accelerated approval label was a retrospective chart review of 37 patients at a center in France. It wasn't even data generated by Novartis. That was encouraging data, though. It's clearly clinical proof of concept data because you could see across a couple of other studies that placebo in these patients does nothing. That was further well validated in Novartis' full approval study or confirmatory study, EPIC P2, which failed, but also confirmed that the placebo does nothing in these patients.
Given the failure of EPIC P2, how do you think RLY-2608 can differentiate?
There are some clear learnings between the accelerated approval and EPIC P2. In accelerated approval, the dose used was 250 mgg. That resulted in a response rate. The endpoint here is a, the clinical endpoint is radiographic response by volumetric MRI. The hurdle is 20% reduction in volumetric MRI at six months. The response rate in the accelerated approval study was 27% for alpelisib at that 250 mgg dose. When they went to the confirmatory study, keep in mind, there was no true dose exploration done in the accelerated approval. Physicians were a little reticent to use the 250 mg dose for safety concerns. The dose used in the confirmatory study was half that dose, 125 mg. What you saw was about half the efficacy. The response rate ended up being about 17%.
What that tells us is that our hypothesis that because of the lack of selectivity that alpelisib has, they're likely leaving efficacy on the table. It gives us even more confidence that they're, given that they have a clear dose response from 125- 250, that as we can continue to push dose intensity in these patients and likely drive to greater efficacy. We have to prove that out in our own clinical study, but I would say our confidence is growing the more we see alpelisib data.
Got it. Makes sense. How should we be thinking about the TAM for RLY-2608 and PROS and the expansion opportunities to other vascular malformations?
Sure. Our first in human study in vascular malformations will be run in both PROS and lymphatic malformations and be open to some of the other malformations. We'll be focusing in PROS and lymphatic malformations. There's about 5,000- 15,000 PROS patients in the U.S. There's probably about 60,000- 65,000 lymphatic malformation patients with PIK3CA mutation. The key commercial question, which we can't answer today, is, you know, these patients exist on a severity spectrum. Not all of these patients will need or want to immediately move to chronic systemic therapy to manage the disease. Today, we don't know exactly what % of that 60,000 will ultimately choose chronic systemic therapy to manage the disease. From a total addressable market (TAM) perspective, it doesn't need to be a lot.
A lot of patients would be a very large opportunity, given that we're talking about this would be taken chronically. I think it's early days to estimate a TAM, given we don't have a lot of precedence in this space, but we're certainly talking about in the multiple of billions of dollars in the U.S. alone.
What are your plans for expanding into other tumor types with PI3K mutations?
Yeah. As most folks may know, PI3K alpha is the most commonly mutated kinase in cancer. It's about 13% of all solid tumors. It's enriched in HR-positive, HER2-negative breast cancer. There, it's 40%. There will be opportunities over time. We sit in a capital-constrained environment, so we have to be thoughtful about how we deploy capital to those other opportunities. We will prioritize breast and then vascular malformations for now. As we generate data that hopefully allows us to continue to access the capital markets, there's other breast cancer indications such as triple negative breast cancer, which we initially want to think about. You can move outside of breast into some other tumor types where probably, again, more hormonally driven tumor types where it seems that the PIK3CA mutation appears to be more of a driver event.
I guess speaking of capital, how are you thinking about balancing your OpEx for RLY-2608 in breast cancer versus vascular malformations as well as advancing your preclinical program?
Yeah. We have to think about it very carefully. I think the way we look at the world is we can't be confident that there's easy ways to go access the equity capital markets anytime soon. What you saw us do is some pretty dramatic reorganization over the past tail end of last year and into this year, extends our cash runway into 2029. That will clearly get us through three what we think are going to be clear value-creating inflection points. It'll get us to top-line data of the phase III trial ReDiscover-2. We'll generate clear triplet data to inform the opportunity in the frontline breast cancer setting. It will get us to clear and interpretable vascular malformation proof of concept data. I think from there, as we kind of, we haven't guided to those three milestones specifically, but they all can happen within this cash runway.
As far as your earlier pipeline, are you considering partnering some of those programs? What would you be looking for in partnerships?
Yeah. I think philosophically, partnering for any of our programs is always on the table. Our obligation as a publicly traded company is to maximize the value of each of these assets. If that comes through partnership at any point in time, they will always be considered. We don't have the capital to fully prosecute Fabs and then RAS into the clinic. We will, like we said, progress those to a high IND, and then we'll make business decisions at that point in time. Partnering is certainly an option for those programs. That requires counterparties that are not in our control to me.
Great. I think we're just about out of time. If anyone has any questions in the audience? No? Okay, thank you so much for joining us.
Thank you again to Wells Fargo for having us. It's been a great day, so thank you.