Okay, great. Thanks everyone for continuing to join us here at Guggenheim's second annual healthcare innovation conference. Yeah, Brad Canino, Senior Analyst here. I'm really happy to be joined on the stage by Relay Therapeutics. We've got Sanjiv Patel, President and CEO, Don Bergstrom down there, President R&D, and Pete Rahmer, the Chief Corporate Development Officer. Relay crew, always great to see you. Thanks so much for joining us.
Thank you. Thanks for the invitation.
Maybe Sanjiv, if you'd just like to kick off by introducing the portfolio and development priorities at Relay today.
Yeah, I mean, I think RLY-2608, it's the first PI3Kα mutant selective inhibitor to enter the clinic. It's front and center for us. Three different trials ongoing. We have a post-CDK4/6 metastatic trial in hormone receptor positive, HER2 negative breast cancer. That pivotal trial started this summer. Behind that, we're getting ready for selection of what could be a frontline strategy in the metastatic setting with ongoing triplets with RIBO, PALBO, and abemaciclib. And then behind that, we have a vascular malformations trial that we're very much excited about. We have preclinical assets in KRAS and NRAS, and then we have ongoing research. And so we're set up with cash into 2029 with a range of catalysts ahead of us over this period. We are heads down execution mode at the moment.
Okay. I'll let you take this question wherever you'd like to go, but the question is, what about the clinical profile of RLY-2608 that you've shown gives you confidence that it could produce superior benefit versus capivasertib, which is the control arm of your Phase III trial in second line breast cancer?
I mean, the profile of the asset, you know, since we've been working on this all the way back to 2017, was all around if you could dial out some of the wild type toxicities, hyperglycemia, rash, dermatitis, diarrhea, you could keep these patients on greater dose intensity. No one's arguing that the PI3Kα is not a fundamental driver of the disease here. It's just, can you hit it hard enough and inhibit it over a sustained period to allow these patients to have the benefit that they need? The first generation inhibitors that have been approved, alpelisib, inavolisib, and then the non-selective inhibitors, capivasertib, have all proven out the fact that you can get the efficacy. But the safety profiles of these molecules are so challenging that unfortunately patients can't stay on them.
What we've shown in the data that we've shown over the last three years now is that it is definitely more selective and therefore dials out many of the toxicities that we talked about, allows these patients to stay on treatment. In this second line setting, we've seen PFSs now in the double digits. That's meaningfully greater than the five and a half months we see for capuzertib. Most importantly, patients are staying on treatment, and it is much more tolerable than some of the treatments that we've talked through. I think everything that we thought about the preclinical profile has translated into the clinic, and that's obviously now translating into PFS.
Great. Now, other competitor programs, when you look across the landscape, are using alpelisib in the control arm. The question is, why have you selected capivasertib, and do you expect the efficacy to be a higher hurdle or a lower hurdle for one or the other?
I mean, alpelisib on its face has a higher PFS than capivasertib, but the sales figures that we've seen over the last two years since the launch of capivasertib show that capivasertib has almost totally taken the market share from alpelisib and is at a run rate now of $800 million annually. That's with a PFS of 5.5 months. Even though it has nominally a lower PFS, physicians are using it because of their nominally better safety profile. Now what we see in real life is capivasertib is also a very challenging medication to take. The rates of hyperglycemia are much probably greater than we saw on the label. Even though alpelisib has a higher efficacy bar, we've chosen capivasertib because that's now the de facto standard of care.
We feel very confident that we'll clear the five and a half months' worth of PFS that's set out.
Okay. Now within the mutant selective class, how do you see 2608 as competitively positioned relative to the Scorpion Lilly molecule that recently had updated data at ESMO last month?
Yeah, I mean, I think those of you who follow the field, you know, have eagerly been anticipating the data disclosure that we saw in Berlin. We saw monotherapy data last year from the molecule in the kind of 20s for the response rate. That unfortunately deteriorated in this update. On the Dublin update, which is the first kind of like-for-like comparison in a population where we've shown 39% confirmed objective rate of response, what we saw from the Scorpion Lilly molecule was a 20% confirmed response rate. That's similar to what we saw with alpelisib and inavolisib. I think the questions really are is, is the molecule going to be more differentiated in any way than the current early generation molecules? I think over time that will hopefully improve, but it's got a long way to go to get to the 39% that we've set as the bar.
In the triplet, although the data was early, the confirmed response rate, as you rightly saw, was zero. Again, hopefully those responses, the unconfirmed responses, will confirm, but there is still a lot of daylight here between that and the current standard of care. I think net, I think we feel pretty confident of our position of being both the first and potentially the best in class here.
Okay. What do we know about both of these molecules in terms of their ability to treat both patient subgroups with the kinase mutations and the helical mutations? How are you incorporating that question into the pivotal trial you have ongoing in second line as well?
I mean, I think as we've been consistent, we have activity across all PI3Kα mutations, both the helical and the kinase. And then both molecules that we've talked about here, Scorpion, Lilly, and Relay, the, you know, the data is all in the public domain. They have relatively similar selectivity on the kinase. Obviously, we've set some very high numbers on the kinase. I think, you know, even if something was more selective, as we've seen from OnKure with its, you know, very highly selective 1047R molecule, we don't see there's much more efficacy that's there to take. On the helical, I think we believe that we are more selective than the Scorpion molecule. It's probably looking at their data from their ctDNA. It's where they're clearing the ctDNA least well.
We imagine it's probably why they didn't break out the efficacy between helical and kinase.
Okay. Now this is a question that comes up a lot now, especially after we've seen the Scorpion profile, which you've given comments on and where you think you compare. But Lilly is moving forward into a frontline study with that compound. How do you expect to maintain a competitive position with 2608 if the competitor is prioritizing an earlier line?
First of all, let's not discount the post-CDK4/6 market. It is a significantly sized market, especially given the fact that increasing use of CDK4/6 in the adjuvant setting will occur over the next decade. The post-CDK4/6 trial that we're running, we believe will still have a very significant commercial potential. Now, you know, if there was an approval in a pre-CDK4/6 pretreated population, again, what we've talked about around the Scorpion molecule and the profile of Loxo Lilly's trial, there's high risk there. Obviously that trial will read out over multiple years. We continue to do all the work around setting and launching our own pretreated pre-CDK4/6 trial. We run triplets with abemaciclib, ribociclib, and palbociclib. We want to have all bases covered here. I wouldn't discount the fact that a post-CDK4/6 treated population is a significant commercial opportunity.
Okay. Can you talk a bit more about the status of your own triplets in terms of how long they've been in development and, you know, what the timelines are for those?
Yeah, Don Bergstrom.
Yeah. Yeah, so as Sanjiv mentioned, we've been looking at 2608 in triplet combination with fulvestrant and then ribociclib, abemaciclib, palbociclib. All the work we're doing for dose finding is in patients who have been previously treated with a CDK4/6 inhibitor. It's relatively standard to do dose finding work in those populations before expanding into patients who are naive to a CDK4/6 inhibitor. I think we've been very encouraged by the data we've seen with the abemaciclib, as we heard, you know, from Alima, who presented before us. That is an option if we are considering that we want to, you know, go to skate to where the puck is moving. Because it is, you know, I think a differentiated profile from what's been shown in the public domain, a differentiated profile for CDK4 selective inhibition compared to what you see with CDK4/6.
There are a number of those selective CDK4 inhibitors in development. PALBO, you know, I think we're continuing to be encouraged by as well. That provides a different opportunity given that PALBO likely will be losing exclusivity in the near future. That now gives you the opportunity to go on the backbone of an agent that, while, you know, today is losing market share in the frontline metastatic setting, will be going generic in the near future and globally is still widely used in the first line metastatic setting. I think we are really, you know, we're generating the data right now to understand how to move forward. We have some key strategic decisions that we'll make as we design the plan.
Okay. Now the first mover PI3K in the space with the triplet was Roche's wild type inhibitor inavolisib. Very different profile and ended up only going after the endocrine-resistant population, which was the smaller subset. Do you see the 2608 triplets as also positioned just for the endocrine-resistant or expanding out to the endocrine-sensitive patients and why?
Yeah, I mean, I think there's no reason why we would just be in the endocrine resistant patient population. We think a 2608 approach could be equally beneficial in both the endocrine sensitive and endocrine resistant patient population. In fact, Roche did start with the endocrine resistant population, but they have an ongoing endocrine sensitive study as well. I think the challenge that we've seen with that molecule is, as many of you may know, the inclusion criteria that Roche set for the INAVO 120 Phase III trial. This was the positive trial that ultimately led to the label used very, very strict selection criteria for patients based on their metabolic fitness. They had to have low hemoglobin A1C. They had to have low fasting plasma glucose.
Essentially, the patient population that was enrolled was younger than a typical breast cancer patient population, had lower rates of prediabetes and lower rates of obesity than a typical breast cancer patient population in the U.S. When you now have seen that drug be introduced into widespread clinical practice, there have been significant challenges with hyperglycemia leading to a dear doctor letter sent from the FDA to physicians in the U.S. about the need for being able to monitor very thoroughly for glucose to avoid cases of diabetic ketoacidosis, of which there have been several reported. I think we've seen a more tepid launch for that drug compared to what we've talked about for the capuzertib numbers.
You know, I think honestly, especially in the U.S., physicians are just trying to figure out how to take the data from that very elegant clinical trial and actually extrapolate it to the patient population who we're treating here in the U.S.
Okay. Don, you touched on this a little bit, but maybe just to double click around the aspects of the triplets, how are you collaborating with Pfizer between both the PALBO and abemaciclib triplet arm? How do you think about the choice for a potential CDK partner as you move into pivotal studies?
On the case of how we work with Pfizer, for a TERMO cyclid, we have a clinical trial collaboration ongoing where Pfizer supplies us a TERMO cyclid and we execute the triplet dose finding and expansion at Relay. With PALBO, the supply agreement does not cover palbociclib, so we are just purchasing commercial PALBO for that study.
Okay.
I think in the long run, what we're trying to do is figure out the best possible triplet regimen for the eventual frontline study that we would decide to go run. I think we kind of keep all options open as we continue to collect the data and think about what the best path forward is there.
Okay. Now for additional opportunities for the molecule, what is your thesis to support bringing 2608 development into vascular malformations?
Yeah, fortunately, it's a very simple one that is quite similar to our thesis bringing it into oncology. You know, targeted therapies have held kind of the same truth for a while, which is there's a driver of the disease, in this case, picked to be PI3Kα mutations. And what you would like to do is hit just the driver of the disease as hard as you can for as long as you can to maximum benefit for patients. That's proved out to be true for us in oncology, where we are going to be the first in class and likely first to market for this class of agents. In vascular malformations, that's the same goal. You currently have alpelisib with accelerated approval in one subset of vascular malformations, PROS.
That in and of itself is a sizable patient population, 5,000-15,000 of those patients in the United States. A lot of the same liabilities hold true there in vascular malformations as it does for oncology for alpelisib. We think that a more selective molecule that can not only spare PI3Kα wild type, but also the other PI3K family members will allow us to get to better target coverage, better dose intensity, lead to a better tolerability profile, and ultimately better efficacy for these patients.
Now this has moved into a phase one through trial in Q1 of this year. It'd be helpful if you describe some of the key components of that study and what you're looking to learn from it.
Yeah, so the first part of the trial is a dose optimization part of the trial. I think this is important because when alpelisib was developed in vascular malformations and PI3K-related overgrowth spectrum, there was not really any dose optimization done. The dose that was used was a lower dose than the oncology dose, but, you know, with no real understanding of what the optimal risk benefit profile was for the dose. You have seen some uncertainty on their part about what dose to take forward. The accelerated approval label for alpelisib for Vijoice, which is the vascular malformations label, has a 250 mg dose in it for adults. The confirmatory trial they ran, they used a 125 mg dose because the investigators were concerned about safety at 250 mgs. That trial failed. Now they are back running a new confirmatory trial back at a 250 mg dose.
You need to know what the right dose is to use. We have designed a trial where as patients come on, they are randomized across one of three doses, the top dose of which is the 400 mg dose, which is our oncology recommended Phase III dose, and then two active doses below that. We were able to do this based on the extensive experience we have had with 2608 in oncology patients, where we did not need to step through higher dose levels sequentially as we changed diseases. We were just able to take these patients and randomize them from the outset. The goal there is to give us the idea of what the right dose is to give us the optimal balance of efficacy and long-term tolerability.
Because remember, this is a patient population who for the right drug should be able to be dosed chronically for the rest of their lives. It's a congenital disease. You know, very long-term treatment would be anticipated here. We'll do the dose optimization, pick the right dose to take forward, and then we'll go into an expansion at that dose or doses to be able to further characterize the safety and efficacy of that dose. You know, I think we anticipate that our initial patient population in this dose randomization, we're focusing across the spectrum of PI3Kα driven vascular malformations. We anticipate it will be largely the PI3K-related overgrowth spectrum or PROS and then lymphatic malformations, just given the frequency of PI3Kα mutations in these two subsets of the disease. It's 100% in PROS and 80% in lymphatic malformations or PI3Kα mutant.
We expect those are the patients we'll see.
Okay. You have not given timelines on data yet, so I will not ask you on that. I think the question that comes up is, is it hard to find these patients? Is it hard to get a good sample size enrolled into the clinical trial? What are you finding?
We're very happy with our experience to date. Now it's true, this is a much more nascent disease setting than breast cancer, for instance. In the context of vascular malformations, the disease has only been fully characterized for about the past 15 years. The EPIC P2 study, so alpelisib's failed confirmatory study, was the first clinical, first and only thus far clinical study that's reported out for use of a systemic therapy in these patients and a subset of these patients. We are definitely paving some new ground here to learn about the patient journey and the treatment decisions that both these physicians and patients and families go through. All that said, we are very happy with the pace of enrollment into the study, the reception of the molecule by investigators on a global basis.
We are very optimistic to our ability to execute and enroll patients and eventually be able to present some data in the not so distant future.
Okay. Pete, you touched on this a little bit by suggesting the number of PROS patients that might be out there, but can we expand a bit more on what the commercial opportunity might be in this setting? I think my team has done work and kind of estimate that alpelisib is maybe $300 million and still growing in this segment with a very particular drug profile as well. As you think about the opportunity for 2608, what has your initial work suggested that could end up being?
It's an extremely large market opportunity. If you take the comments that Don made, this is a congenital disease where ideally you're treating patients as early as you can in childhood and you're treating them hopefully chronically to keep the disease at bay. There's 170,000 PI3Kα mutant vascular malformations from a prevalence standpoint in the United States today. The question there is amongst those 170,000, where on the severity spectrum do they sit and how many of those would be seeking chronic systemic therapy to control their disease? As you can imagine, it doesn't need to be a very large portion of that 170,000 to get to a multi-billion dollar opportunity for 2608 in this setting.
I think the first two areas of focus in PROS and lymphatic malformations, where you have the disease is almost completely characterized by PI3Kα mutations. I think we feel very optimistic that we'll be able to demonstrate the utility and effectiveness of 2608 in those patients and then branch out even further to the other vascular malformation settings. We've done some market research to suggest that anywhere from 20%-40% of patients in those two subsets would seek chronic systemic therapy that is well- tolerated to control their disease. That would be an extremely meaningful market opportunity for 2608 over time.
Yeah. I guess kind of two follow-ups on that that are related. What do you know about the current use of the available systemic therapies today? So alpelisib, sirolimus, et cetera, that's used today and how much of their profile actually leads to that number?
Yeah, so we don't have great numbers on this. You've done some very good work on it probably on the alpelisib front there where you can track Scripps data and you can get to a rough estimate of the number of patients that are likely that are seeing alpelisib today. It's probably in the high hundreds to maybe cresting 1,000 patients or so that have seen alpelisib. The open question is how long are they staying on that treatment? There we can rely on our experiences with talking to investigators. The interesting thing is in this setting, when you go to medical meetings, you're not seeing data presented from large global trials. What you're seeing are case series from specific institutions. It's a different way of understanding a disease space than we're used to from breast cancer.
With serolimus, serolimus has been used in these patients probably for about the past five to eight years. I think across in PROS, again, where there's 5,000-15,000 of these patients with both alpalasib and serolimus, some of the estimates that we've seen in the market research we've done suggest maybe a quarter of that patient population has had experience on one or both of those drugs at some point in time. Very few of them are able to tolerate the drug to stay on it chronically, which would be the desired treatment pattern that you'd like to put them on.
Okay. What do you foresee as the potential regulatory path for VM given what's happened with alpelisib over the past few years?
Yeah, so you know I think we've got the regulatory precedent from alpelisib, which is a rather unusual regulatory precedent in that the original accelerated approval was not based on a prospectively conducted clinical trial, but on retrospective analysis of a compassionate use series of patients. And there I think, you know, while the trial design was somewhat atypical, what was established was what the registration endpoint would be, which was a radiographic endpoint using MRI of the vascular malformation lesions and analyzing a volumetric sense and looking for a 20% reduction in the volume of the lesions with therapy. That was the endpoint on which the approval was granted. That's been the endpoint both in EPIC P2, which was the failed confirmatory trial, as well as the upcoming EPIC P4, which is the new confirmatory trial.
I think the precedent would be that that would be the primary efficacy endpoint. That is the endpoint we're looking at in our ongoing re-inspire trial. With regard to how many patients we would need, single arm, double arm, I think that's going to require further interactions with health authorities, which are interactions that we'll have when we have data in hand.
Okay. Great. Thank you so much for joining us. We'll have to stop there and look forward to the continued progress. Thank you.
Thank you.
Thanks.