Good morning, everyone. I hope everyone's enjoying day two of London and our healthcare conference. Luckily, it was not raining as much as it was yesterday, so that is an incremental positive as far as I'm concerned. My name is Akash Tewari. I'm a pharma and biotech analyst here at Jefferies. I have the pleasure of hosting the Relay Therapeutics team, Sanjiv, Don, and Pete. So glad to have you. Sanjiv, why don't I give it to you for some introductory remarks, and then we'll get started.
Thanks very much for the invitation. Those of you from the U.S., hopefully you saw the sun this morning, which was wonderful to see. Apparently, it does come out here in the U.K. Thank you very much to Jefferies for the invitation. We're excited to be here. It's an exciting time for the company. We showed some very exciting data a couple of months ago in hormone receptor positive, HER2- negative, second line plus metastatic breast cancer with a PI3K alpha mutant selective molecule, RLY-2608. I'm sure we'll talk a lot about that during the next 25 minutes. We're off to the races now, putting together a pivotal trial, Phase III , and trying to get this medicine to patients as rapidly as possible.
Behind that, we have an exciting NRAS selective molecule, as well as a foray outside of oncology into rare disease with a Fabry program. Recently financed, and now we have a balance sheet that is $850 million, and we have a team to execute against it, and so we're excited about the years ahead of us.
Understood. Sanjiv, before we get into San Antonio and more data, I kind of wanted you to take a step back and talk, because there's probably few fields that are as competitive as HR-positive. The way we kind of see the space, it's about partnering with the right assets that are going to be relevant over the next decade. To me, we cover Pfizer. That CDK4 is something that I think could end up being a multi-billion-dollar opportunity. Talk to me about why you and Pfizer chose to do that combination study. What have you seen with the CDK4 data early on that got you so excited about pursuing a partnership?
Look, I think we're very excited about CDK4, but you're right to say that the evolution of the landscape is probably as great as it's ever been. And so the arrival of CDK4/6 a couple of years ago was the last real kind of seismic shift in this field. And you see now the evolving landscape in that with the CDK4 selective molecules arriving, such as Pfizer's atirmociclib. And we were very excited to see the data on the tolerability and then the greater efficacy that it leads to. You also see the move now on the endocrine backbone with the oral SERDs, and there's a plethora of data coming our way over the next year. And so as these two worlds evolve, we think there's always going to be a home in PI3K alpha mutated patients for a selective PI3K alpha inhibitor.
And so we want to create the backbone of choice. And as these two worlds evolve, we'll combine with what's the best of breed in each of them. And so what you'll see us do, and the Pfizer move was the first one, is to try and generate data with what we believe to be the best in breed molecules so that when they need to combine, they will combine with the best in breed of a PI3K mutant selective inhibitor, RLY-2608.
And maybe to that point, because I almost feel like it's kind of irrelevant. You get asked, oh, when are you going to have more monotherapy data? I don't see any patient getting monotherapy PI3K alphas anyway. And also maybe Don, I love your color. Where do you think PI3K alpha will ultimately head out? Because we focus on late line populations, but ultimately, I look at that Roche study where you have that triplet. You're talking about 15-16 month benefit, which is fairly profound. Is that really what we should be thinking about with your molecule? It's not about late line setting. It's actually the first line opportunity as a triplet.
So maybe we'll break it down. Maybe I'll take the first bit around the monotherapy. You're absolutely right, which is we don't believe that if we stack the commercial opportunities for RLY-2608, breast cancer is going to be top of the list in combination. Combination first with fulvestrant, and then as you see that we're running trials with ribociclib and atirmociclib. And then as we start to think about other opportunities, the highest one for us would be vascular malformations, 170,000 patients that have a PI3K alpha alteration. And so for us, it has never been a priority on the monotherapy side. And so I'll hand it over to Don to talk about the opportunities on the combination side.
Yeah, and I think you're spot on, Akash. So PI3K alpha mutations in breast cancer are truncal . These happen early in the development of the disease. They're driving the disease. It's not dissimilar really from HER2- amplification in HER2-positive breast cancer, where it's a clear driver. It defines a subtype. And when you look at the evolution of HER2-directed therapies in that space, HER2-directed therapy is used in essentially every treatment setting, starting from early breast cancer with neoadjuvant adjuvant treatment all the way through the metastatic treatment landscape. And I think with PI3K, again, given that it's a dominant driver, given that it's a truncal mutation, you can see a similar evolution in treatment in this space.
Obviously, the first place we're starting, the most tangible place that we can start, is in patients who have seen a CDK4/6 inhibitor, and it's where we can generate our first registration, but it certainly wouldn't be the place where we generate our last registration, and as Sanjiv already mentioned, there's the opportunity to move into the earlier line metastatic settings, but over time, as I think we look at this population and look at the biology of this mutation, ultimately, there will be opportunities to move into early breast cancer as well. Understood. Now, Sanjiv, you mentioned something about we're not just thinking about CDK4, but we're also thinking about SERDs. Obviously, it's a very competitive field there as well.
But I think when I look at our team, and I think probably why we missed your early data, was there was synergy between the PI3K alpha and fulvestrant that wasn't showing up from a monotherapy perspective. And that's, I think, the failure we had on a research perspective. Do you feel like that might even get better with the SERD? What are you seeing with the SERD data sets that get you excited about a combo? And do you feel like there might be a differential response in wild type overlaying PI3K alpha with the SERD?
Yeah, so we have seen throughout our preclinical models synergy between estrogen receptor pathway signaling and PI3K alpha pathway receptor signaling, and we do see synergy there. And I think when we look at our data that we've generated in the clinic to date, we've generated very limited monotherapy data. Again, it's not really a development path we can follow. We have reported monotherapy responses. I can't tell you precisely what the monotherapy response rate is. I think in the data we disclosed last year, we had two positive breast cancer patients who we treated. One of them was a confirmed response. I'm not going to sit here and say we have a 50% response rate.
Yeah.
But I think there is.
I'm happy to see this data.
I think clearly there's monotherapy activity. The patients we're in right now with fulvestrant are a patient population where we wouldn't expect to see much contribution from fulvestrant on its own. 52% of our patients have seen prior fulvestrant or an oral SERD, 42% are ESR1 mutants. Yet when we put our two drugs together, we're sitting at a 33% response rate across the PI3K alpha mutated patient population. We're sitting at 9.2 months median PFS, which is markedly more than you'd expect to see from certainly fulvestrant alone. So I think there's certainly some data that suggests there could be synergy there. Now, as I mentioned, 42% of our patients are ESR1 mutants. We've not seen a strong signal of differential activity of the fulvestrant plus 2608 combo and ESR1 mutant versus ESR1 wild type.
We seem to have a dominant effect that is relatively independent of ESR1 mutation status. I think the trial that we'll propose for our phase three trial will be independent of ESR1 mutation status. We saw capivasertib just granted its label about a year ago, regardless of ESR1 mutation status. So I think right now there's a clear path forward with fulvestrant. I think as we start seeing data from the oral SERDs, it will be interesting to see if there's any more oomph, if you will, that's brought to the table by having an oral SERD that has some more activity in the ESR1 mutant populations. But at this point, there's nothing in the data that we're seeing that suggests that we're not getting a reasonable amount of synergy from fulvestrant, even in ESR1 mutant patients.
Understood. Now, this is a discussion that I think the investor community and I have a lot when we talk about our PI3K drug-drug interactions. I think it becomes critical, especially if PI3K alpha goes into a triplet setting. Can you talk to me about what work you've done with the metabolism of RLY-2608? Because even with your backup compound, that was stuff you were talking about. And the DDIs, or lack thereof, with both CDK4 and next gen SERDs, but also with CDK4/6s, which seem to be CYP3A4 metabolism.
Yeah, so based on our profiling of 2608, we wouldn't anticipate that we'd have a high risk of CYP mediated DDIs. And I think that continues to be our position as we look at potential combination partners. We have seen what we've characterized as a favorable interaction with ribociclib. It's not a CYP mediated interaction. We have some hypotheses about what the interaction may be. It's likely a pathway that's not as commonly characterized as you're looking at molecules. But it is a specific interaction with ribociclib where ribociclib is actually increasing the blood concentration of 2608. So this is very different from the Arvinas situation where Arvinas is changing the concentration of the combination agent, which is the approved agent. Now you need to use a non-label dose. In this case, we anticipate we'll be able to use label dose of ribociclib.
What it means is we'll just have to use a lower dose of 2608. We are moving forward with the triplet dose escalation to be able to understand ultimately what that dose of 2608 will be. We anticipate it will be lower than the 600mg dose we're using in the fulvestrant combo.
Understood. And can you just expand a bit on what is that about the metabolism pathway that you're seeing with ribo in RLY-2608? And then I think number two, you think about ribo, it is kind of a CDK4 biased, CDK4/6. Can you talk about the synergy specifically with CDK4 and a PI3K alpha? Because obviously the Pfizer combo CDK4 selective.
Yeah, so I'll start with the second part first. So we've done preclinical modeling in vitro and in vivo xenografts looking at 2608 with atirmociclib. We see strong synergy between the two. It's largely comparable to what you see with the CDK4/6 inhibitor. I think one of the things we're most excited about with the CDK4 combination is the ability to mitigate some of the cytopenias that you see with the CDK4/6 class that ultimately limit the dose intensity of CDK4. And we think with a combination with atirmociclib, you should be able to overcome that. And I think it's supported by some of the early data for atirmociclib that you're able to achieve a reasonable degree of efficacy, especially in CDK4/6 pretreated patients while mitigating a lot of the cytopenias.
Understood. Obviously, kind of one of the hypotheses of this class is, hey, we want to limit the rash, the hyperglycemia. I think you had very modest with fulvestrant Grade 3 plus hyperglycemia in combination. Talk to me about with the CDK4/6 profile. Because to me, it seemed like fulvestrant was really the component that was adding maybe some of the Grade 3 plus there. Talk to me about what you anticipate the combination CDK4/6 toxicity profile to look like. What's the bar for you to move forward here?
Yeah. I mean, I think here, obviously, you'd anticipate that you'd see probably most of your Grade 3 AEs are going to come from cytopenias that are associated with the CDK4/6, especially with ribociclib. You're going to expect to be in a 40% plus rate of Grade 3 cytopenias in a CDK4/6 pretreated population. We wouldn't anticipate a priori to see any type of synergistic toxicity between a ribociclib or an atirmociclib with 2608. We'd expect at worst to see a sum of the parts. And I think that's exactly the hypothesis that we're testing right now in the triplet development.
Understood. And then maybe just for the CDK4, the early data we've seen there, which I think is pretty interesting, is in combination with fulvestrant, you're seeing about a 15% Grade 3/4 neutropenia. So a pretty marked reduction even from ribo. Where do you think that CDK4 combo Grade 3 neutropenia rate would be? Should it be similar? I mean, let's put it this way. Am I crazy to think a triplet combination between fulvestrant, CDK4, and your molecule would likely have a Grade 3/4 neutropenia rate under 20%?
There's no reason right now, and I'm going to caveat this by saying we're just starting the trial, but we wouldn't expect to see any additional contribution to neutropenia from the addition of 2608.
Understood. Now, maybe to that point, I think when we look at, especially as we go into that earlier line setting, it's not just about potency. It's also about effective coverage over a period of time. I mean, certainly that's the lesson Pfizer probably learned the hard way with the PALLAS studies. When you look at the Roche data in first line with the triplet, where do you see room for differentiation in a triplet combo in terms of overall exposure and target coverage over a period of time?
You want to start with just the bigger picture and then we'll get into the coverage.
Yeah, so I think one thing to keep in mind about that INAVO120 study, we learned a number of things. Obviously, the control arm was palbo plus fulvestrant. That control arm probably underperformed from historical average of what you would see in that regimen in an endocrine resistant patient population. But you saw by adding the mechanism of PI3K alpha mutant inhibition, a doubling of the PFS. That is probably the most informative piece of that data set. Now, they had to do it by going into a bit of a very metabolically fit patient population, a patient population that doesn't really represent the majority of the Western breast cancer patient population. And that will yet to be seen how that affects uptake in use. The other limitation of that data and that trial was that they ran it against palbo and fulvestrant.
The current commercial standard of care in Western countries for endocrine resistant patients is ribo plus fulvestrant. And there's a couple of different studies that suggest ribo plus fulvestrant gets you about 15-16 months of median PFS. So I think the combination of the limitation of those patients to being very metabolically fit and the fact that they didn't go against the current commercial standard of care, and you really in a cross trial comparison don't see much of any differentiation from median PFS between the triplet and the doublet, you're probably going to those two things may impact the overall uptake of that triplet. But the biggest takeaway remains that this mechanism matters in these patients. And that's really what continues to give us confidence and excitement about having an agent like ours that really allows you to maximize that dose intensity.
The other thing you saw from that study is even though it was a median PFS of 15.3 months, the median duration of those patients on INAVO120 was only nine months. Again, back to your point, Akash, the improvements we've made in selectivity, you can see it already in our Phase I-B study with even more beat up patients that were mitigating a lot of the side effects and toxicities that INAVO120 sees. It's resulting in greater dose intensity. When we eventually move into these triplet settings in both second line and front line and maybe even earlier, we do think we're going to be able to even advance further the ultimate efficacy profile for patients in those settings. Understood.
Now, maybe just stepping back and looking at the class of molecules these selective PI3K alphas, obviously, I think Scorpion had data that they showed earlier this year. It's funny because I actually see a lot of similarities in terms of the principle of going after multiple isoforms, selective reversible between your product and Scorpion. We'll get more data from Lilly and Loxo later on. Obviously, that's going for just one isoform, but it's also an incredibly selective molecule. But Lilly's not moving forward with their drug. They're moving with a backup. Scorpion has seemingly higher monotherapy response rates right now, but obviously N is low. But then they also have some Grade 3/4 liver toxicity. I'd love to get just kind of a level set. After you've seen the Scorpion data, you've gotten the Lilly announcement. Where do you think your molecule stacks up from a competitive perspective?
I think we feel great. So if you kind of unpack what you've said. On the pan mutant side, obviously, the only one in the clinic is Scorpion. I think the monotherapy data they showed was encouraging. But I don't think on the efficacy side, it was anything that we haven't seen with the non-selective inhibitors. As Don pointed out, we had very few patients dosed at therapeutically active doses. And I think we have no endometrial patients and only the two breast cancer patients, of which one was a response. So I think if we went off and put effort behind it, which we won't, given the fact that it's not a commercially viable area, I think we would stack up very favorably on the monotherapy. And so really the bar on the combination is, what does it look like versus standard of care capivasertib?
And we showed 9.2 months and 33% response rate. So efficacy numbers that we've not seen before. On the tolerability front, as you rightly pointed out, it only had 1.9 months' worth of follow-ups. So it's very hard to see how you could really get a good picture. But as you rightly pointed out, they hit a DLT with paresthesia and myalgia, and there was Grade 3/4 liver tox. So definitely something to keep an eye out. So I think on the pan-mutant front, we feel very solid in our position as being ahead. On this kind of isoform-selective front, as you saw, Lilly Loxo entered the clinic very similar time to 2608 and has not shown data yet. We'll see it at San Antonio, but they've announced a discontinuation, which obviously is never a good sign.
And so on the kinase specific element of that, so the like for like, I think we showed 53% response rate and 10.3 months' worth of PFS in that subgroup. Those are numbers that we've just never seen, like the standard of care on the response rates for alpelisib are in the high teens. So we're putting two to three X against it. So for someone to come with a molecule that's going to beat that, you're going to have to put massive numbers up. And so again, we feel very confident. And just so that we can level set, there again, we have seven and a half months' worth of follow-up on our patients. And as that goes, we imagine those numbers could get better.
We have a significant number of patients that on the kinase domain specifically that sit in response or in stable disease that are still ongoing. So we imagine our 53% and 10.3 months worth of PFS on the kinase will only improve. And so I think we feel pretty confident on the isoform selective piece as well. So as we sit today, we have definitely a lead. We know that that will translate in the commercial world as it has in all kind of precision oncology classes. So the goal for us is to try and get this to the market as rapidly as possible and maintain that lead. And then for people to put significantly bigger numbers up to try and beat us, which I think is going to get harder and harder.
Understood. Now, your team had pointed this out rightfully. When you look at the CDK4/6, it's not just about response rate, but it's about clinical benefit rate and how that really evolves over time. You've started to show that as well with your data. You're going to have more updated combination data at San Antonio. I will be there. It's going to be fun. So what have you seen in terms of the translation of clinical benefit rate to then actual responses? And teach us how to look at the updated data you're going to be having at San Antonio. How much more follow-up are you going to have? Could we expect that 33% response rate to improve with increased durability? Because it does seem like, which is quite interesting actually, a lot of the responses you've gotten seem to happen quite quickly.
So should we expect a deepening of responses at San Antonio?
Pete, you want to take that one?
Yeah. So just to be clear, in our September conference call when we initially disclosed the doublet data, we articulated there that we're not going to be updating the doublet data at San Antonio. So this is just an opportunity for us to take data that we've only presented in a corporate form to date and put it into the kind of more normal course medical meeting venue. So we can obviously, one of the key goals here is to launch a Phase III study next year. And so continuing to grow awareness for these data among the medical community is a huge priority. And so that'll be the key focus at the meeting itself is to continue to drive awareness among the medical community. But as Sanjiv alluded to earlier, we do think that these data can grow stronger.
I think to your question of how do these different endpoints correlate with each other, the reason why we've been telling folks not to get overly indexed to response rate is because if you look across a lot of different pathway inhibitors in HR-positive or HER2-negative breast cancer, you don't see great correlation between response rate and the ultimate regulatory endpoint of progression-free survival. Now, you do see better correlation between CBR and progression-free survival, which is logical because CBR is a measurement of stable disease or better at six-month time point. But we're encouraged to see the response rate continue to grow. It's the most objective measurement you can have of the impact on a tumor. So it is definitely a good fact pattern that we're seeing.
But we're going to continue to focus on this confluence of data maturity and ultimate progression-free survival because I think that's going to be the ultimate underpinning of our confidence in the ability to be successful as we continue to move forward.
Understood. Maybe to that point, do you think we will get updated combination data maybe sometime in 2025? Is there an appropriate forum for that?
Yeah, there's definitely a good chance that we will, once we get to full maturity of this Phase Ib. We're not enrolling any more patients into it, and so once we get to full maturity of the Phase Ib experience, either through manuscript or at a medical meeting sometime next year would be a logical point in time to do another update.
Understood.
If we ever find it to be informative against the probability success of the molecule, we can update these data really at any point.
Understood. Now, Sanjiv, you alluded to, okay, you're looking at CDK4, but it sounds like your team is looking across the space in HR positive to see other kind of synergistic targets that make sense. When you look at the milieu of assets that are out there, where do you think there's biological rationale to combine with your molecule?
Yeah, I mean, we've already touched on the oral SERDs. I think we're going to learn a lot in the next year about where that space is going to shake out. That will help provide a guide map for us about where to think about combining with oral SERDs. But again, I think here the constant would be 2608. We think you'd be able to generate data with a multitude of different oral SERDs to support usage in the real world. I think the CDK2 possibility as we think about, again, what are some of the resistance mechanisms for CDK4. We've got development happening in the CDK2 space. We've actually brought a CDK2 inhibitor of our own to the IND ready space, but haven't brought it into the clinic yet. That's a mechanism, again, where we've shown there's preclinical synergy between PI3K inhibition and CDK2.
So I think as you see these mechanisms emerging, as you see growing validation in breast cancer, there'll be opportunities to explore them. But I think the constant we'll always see is that PI3K is the backbone in PI3K alpha mutated patients.
Understood. Thank you so much. I really do appreciate it.
Thank you very much.