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

Thanks to the Relay Therapeutics' team for joining us as our next company presenter at the BofA Annual Health Care Conference. I'm Jason Gerberry, SMID cap biotech analyst. I'm joined by [ Chi Fong ], who is my colleague. And we've got on stage with us Don Bergstrom, President, R&D; and Pete Rahmer, Chief Corporate Development Officer. So gentlemen, thanks first off for joining us.

Yes. Thanks, Jason. Appreciate you having us.

So quiet first half, but it seems like you guys are going to have a pretty busy second half. Maybe if you just want to set the stage for some of the updates that you're going to have in second half because it seems like you're hitting a pivot point where there was limited updates, but now second half looks it's getting a little bit busy.

Yes. But we would agree with that assessment. So coming up in the second half, probably the one we will probably spend a decent amount of time talking about is on RLY-2608, our PI3K-alpha mutant selective inhibitor. We've -- that molecule has been in the clinic for a couple of years now. We now have it moving forward into kind of expansion stage of the study at 600 milligrams twice daily in HR-positive, HER2-negative PI3K-alpha mutated metastatic breast cancer. And there, the questions that we answered early on were around safety and efficacy and showing meaningful differentiation against the existing nonselective inhibitors. And I think the remaining question is more around on the efficacy side and understanding the comparability against some of the nonselective inhibitors. And especially when you start to think about ahead to the regulatory endpoint, which is progression-free survival, that is just a -- it's one of those endpoints that it just takes time to generate that data. And we're pretty excited to be able to point to the second half of this year as a point in time when we think we'll have some pretty robust data on -- to that end, against understanding the profile of 2608 when it comes to efficacy and ostensibly the second-line patient population in breast cancer. So that's a key data point that we're looking forward to, and we continue to be very excited about that opportunity. We are -- even in the face of that data coming in the back half of the year, we are starting to put some preparations in place for a potential pivotal study next year. So continue to have a lot of confidence in that program and its ability to drive meaningful change for patients with PIK3CA-mutated breast cancer and beyond. Additionally, with RLY-4008, our FGFR2 inhibitor, we had shown initial tumor-agnostic data last fall at the Triple Meeting. And at that time, we said we were pausing regulatory and commercial preparations in cholangiocarcinoma to allow the tumor-agnostic data sets to catch up. And so we can kind of understand the profile of the molecule in that setting. And we pointed towards minimizing spend in '24 against that program, but generating some more mature tumor-agnostic data, going to have a conversation with the regulatory authorities about a potential tumor-agnostic regulatory path forward. And then in the second half of this year, we'll come -- put those 2 things together, a data update, some regulatory feedback and really provide the view of the strategic direction of that program moving forward. And then just yesterday morning, we put out a press release announcing that will be hosting a new program and platform event next month. We had guided last year to showing at least one new program this year. And so that event will be the forum for us to show a couple of new programs. And that is really quite exciting for us. We continue to have a very productive platform, the Dynamo platform. We were one of the first in this space to really bring together cutting-edge computational and experimental techniques to solve previously unsolvable problems in target oncology. We then expanded that to genetic disease. And so this will be a forum for us to really remind people of how powerful and productive the Dynamo platform has been when it comes to solving very hard problems. And so these programs that we will talk about will be -- the intent is to be able to demonstrate potential either first-in-class or best-in-class profile. It will be against targets that are known, but have had liabilities that have not been solvable through to the traditional means, and we'll be able to identify and articulate a novel approach that the platform has afforded us.

When we think about the platform, I think with both FGFR2 and PI3K-alpha, you've demonstrated an ability to develop more selective molecules and solve for certain I guess, dose-limiting side effects and potentially be able to generate better efficacy. I guess the overlay of that has been things like Project Optimus, IRA, all these different things that have -- from a stock perspective, I think made the end market opportunities more challenging, I think, for investors to crystallize what the light at the end of the tunnel looks like. So as you think about advancing new programs into the clinics, what do you say the learnings from how the markets evolve and how you go about product selection?

Yes. So it's important to remind folks that the platform is not created to solve biology question. It is created to solve novel insights of how to go after a target once you've chosen a target of interest. And there, the platform has evolved better than we have anticipated and the ability -- in its ability to help us identify novel ways go after genetically or clinically validated targets because that's what we choose to point it at today. And therefore, we continue to have a large focus on that target selection part of the process. That is more of a strategic question that we ask at the very beginning of any drug discovery effort. And then once we set our sights on a target of interest, we can then set the platform off to do what it can do, which it has done repeatedly now, which is identifying novel ways to go after these protein targets of interest in helping us identify novel chemical matter to be differentiated for patients in those settings.

Yes. Okay. So maybe shifting to the PI3K-alpha program and Don, can you just remind us sort of the improvement relative to the existing standard of care on hyperglycemia and some of the early efficacy and maybe how you would define sort of therapeutic benchmarks, right? Because it's a fluid and evolving space from both second line and there's ultimately like looking ahead to frontline.

Yes. So I think when we look at the nonselective PI3K-alpha inhibitors, there are 3 toxicities that are shared across the class that have really been limited both the doses that patients can be started at as well as the ability for patients to be able to maintain dose intensity and those are hyperglycemia, diarrhea and rash. Those are the AEs that most frequently led to discontinuation of alpelisib and it's -- I mean, of alpelisib in its pivotal trial. So we hypothesized that with mutant selectivity, we should be able to be superior across these key toxicities. Now hyperglycemia is one that we see is really critical for physicians and paramount in them thinking about what drug they're going to reach for. They are Grade 1 hyperglycemia, which is a lab abnormality doesn't require any medical intervention. Grade 2 typically is met with starting an oral anti-hyperglycemic. I think oncologists have shown that they're largely comfortable with those. Grade 3 is when patients start requiring injectable insulin sometimes need hospitalization. And I think that's sort of the red line that we've seen that physicians aren't comfortable with and frequently are calling endocrinology consults, if they want to keep a patient on a drug or just stopping the drug and moving to something else. So our goal was really to dial out Grade 3 hyperglycemia. And in the early data we shared to date, we've been able to do that successfully. And I think we've also seen lower rates of diarrhea and lower rates of rash. So that gives us confidence that the profile that we are able to generate in the lab is translating into a superior safety profile in patients. Now when we look at these comparisons, it's important to take into consideration who are the patients who are being treated. And we have used enrollment criteria in our ongoing trial of 2608 that are relatively broad and relatively reflective of a Western patient population. We can enroll patients who have higher BMI. We can enroll patients who are most likely prediabetic. And the most recent data disclosure that we saw from Roche on their PI3K inhibitor inavolisib, which they disclosed in December at the San Antonio Breast Cancer Symposium. That was an incredibly highly selected patient population that really had strict limits of fasting plasma glucose and hemoglobin A1c that functionally excluded about 50% of a Western population, just based on the metabolic criteria alone. So I think one of the key questions is, as you're benchmarking us, who's our patient population and what's the hyperglycemia number in that patient population. And then when you get to the efficacy side in the post-CDK4/6 setting, at this point, the two most relevant comparators of the two agents that have full approval in the post-CDK4/6 setting, alpelisib and capivasertib, alpelisib in the real world in patients who have seen prior fulvestrant, prior CDK4/6, maybe prior chemotherapy. It's been reported by Novartis and others that you're seeing a real-world PFS in the neighborhood of 6 months for capivasertib, which is an AKT inhibitor from AstraZeneca that was approved in late 2023 in their Phase III trial in the post-CDK4/6 patient population PFS was 5.5 months. So I think as we look at the benchmark ultimately for what the approval of an agent that you might run head-to-head against either of those agents, it's going to be in the neighborhood of 5.5 months to 6 months that you're looking to beat for progression-free survival.

And do you think you'll have interpretable PFS? Because I recall, like last year, we met and focus was on CBR. And maybe just -- I guess I don't fully know what the CBR benchmark might be.

Yes. So I think at the point that the PFS data are not yet mature, CBR is the relevant end point to look at. We anticipate CBR will be most robust when we get into a disclosure in the second half of this year. So the CBR benchmark for alpelisib in the [ by lead ] population, which is a patient population that was less heavily pretreated than ours is likely to be, was 46%. And capivasertib did not report the CBR, but with a 5.5-month median PFS and post-CDK4/6 patients, you have put that in probably a low 40s for CBR. So I think somewhere in the 40s is what we're looking at for the benchmarks. And what we'll be looking to do is improve on that substantially.

Yes. Go ahead, Pete.

Although we would anticipate we won't have a mature enough data set to show a median PFS point estimate. We -- in addition to the CBR, we can do a 6-month landmark PFS analysis against this data set, which becomes very germane when you talk about these benchmarks that have a median PFS in and around 5.5 months to 6 months. And so if we can show a CBR rate that's meaningfully north of 50% and then median -- a landmark PFS 6-month analysis, again, meaningfully north of 50%. You're talking about CBR rates in the mid- to low 40s and PFS in 5.5-month to 6-month range, that starts to become -- that's why we continue to say that this will be a robust and interpretable data set against those benchmarks.

And I guess as you think about different therapies that have been sort of limited by the side effects, is the thinking that when you look at data cuts with those reductions or without that being able to avoid the side effects you can actually generate a step change in efficacy? Or is it more about having something that works comparable, but for everyone?

It's about changing efficacy. And I think the strongest proof we have for that is actually in the development of alpelisib itself in the SOLAR-1 trial, which was in patients who were naive to prior CDK4/6. So it's not purely interpretable relative to our data, but still in that trial. They actually did a retrospective analysis based on doing a cut at median dose intensity. So the label dose for alpelisib is 300 milligrams once daily. The median dose intensity was about 250 milligrams. So half the patients actually are getting less than 250 milligrams a day, half more than 250 milligrams. And when you looked at the differential PFS based on dose intensity, the patients who had higher dose intensity did significantly better than the patients with lower dose intensity. So there's clearly efficacy left on the table here. And I think with 2608 we're dialing out some of these AEs being able to hit the target harder and being able to maintain dose intensity, we feel we can pick up that efficacy.

And stepping back, is the -- how do you think about what's the market opportunity that you're truly chasing if Roche moves into frontline is PI3K-alpha? Is there still going to be a relevant second line plus subsegment worth further developing that all the way to goal line? Or is this more about validating and then knowing you going into pivotal Phase III in frontline say that that's the worthwhile investment?

I mean, in short, we think that even in the face of a potential approval for inavolisib [ study ] and INAVO120 study that there's still a meaningful size opportunity within the second line. If you play that out and the broadest label they would likely be able to get is in that endocrine-resistant patient population in the front line. And so that's probably 40% or so of the patient population, they then have the hurdle that they didn't actually combine it with the current commercial standard of care, which is ribo and fulvestrant. So you have to convince physicians to use like an inferior combination partner in palbo and fulvestrant. And then you're further restricted in the fact that the data set that they have was highly metabolically selected. And so even if they get a broader label in the endocrine-resistant patient population, physicians full well know that the way they got to that level of PFS was through choosing extremely metabolically fit patients. And that's not representative of the Western patient population. So if you really take all the cuts that they provided to get to the extremely positive Phase III study that they ran, it really only represents about 20% of the frontline patient population. And that's, again, not with the current commercial standard of care combination. So I think it's somewhere in between 0% to 40% of the patients that they would address in the frontline patient population. If you presume all of them saw inavolisib in that setting, you still have 60% of the 18,000 PI3K-alpha mutated patients in the front line. That would be available for subsequent PI3K-alpha inhibitor in the second line setting.

So I want to get that number right. Because it sounded like you were saying more than half of the patients second line will be PI3K naive based on the cuts that you just outlined.

At least, yes. At least 60% upwards to 80% if their label were going to be strictly restricted to what the patient population they ran the Phase III study.

Yes. That's something that we can only speculate on right now. Label will reflect that.

[ Chi Fong ]?

Yes, sure. Speaking of safety angle with your molecule and perhaps better combinability with a triplet regimen in the first-line setting, you do have a triplet in evaluation right now branded as Phase I and [ S&D ] second-line plus setting, but you -- I think you're going to get some safety, PK data, dose-escalation data. So can you talk a little bit about that data set in second half? And how might that [ inform ] safety differentiation, more specifically, I guess, combinability with CDK4/6 and fulvestrant?

Yes. So as you pointed out, [ Chi ], that will be -- that triplet started earlier this year. So it's going to be very early dose escalation data, really looking for acute safety and tolerability. And the ultimate goal there is to find a dose of 2608 that can be combined with the full dose of ribociclib. And the one thing that's a little bit easier to evaluate in this context is the other nonselective inhibitors have not been able to establish any dose in combination with ribociclib, which is becoming the first-line standard of care CDK4/6 regimen. And so if we can get to a -- if we can establish a dose in combination with the full dose of ribociclib, I think that would be a significant win and a significant step towards being able to demonstrate that 2608 truly has the opportunity to become the backbone PIK3CA-mutated agent of record for all patients across breast cancer.

Do you think you would get to the sort of a preliminary answer in second half? Or do you think that dose escalation might keep going?

I think we can get to the beginnings of an understanding by the second half. We may still have more work to do on the dose escalation side, but I think we'll have enough to tell us if we're on our way there.