So welcome to the next Fireside Chat with Relay Therapeutics. My name is Michael Schmidt, biotech analyst with Guggenheim. It's my great pleasure to welcome the Relay team. We have Sanjiv Patel, CEO, Don Bergstrom, Head of R&D, as well as Megan Goulart, who runs IR at Relay. Welcome, guys. Thanks for joining us.
Thank you.
Thanks for the invite.
So, Sanjiv, I think most people know the stories, so we'll just-
Yeah
jump right into into some of the pipeline products, and I'm very curious to to talk about those. Perhaps starting with RLY-2608, which is your PI3K inhibitor. And so, you know, this is a very interesting drug, I think, that's sort of going into the breast cancer space. Just to set the stage, remind us, you know, what, how the drug compares, perhaps pre-clinically, to to alpelisib (Piqray), which is approved from Novartis, and how that, how those features could translate into its clinical potential.
Well, the development of RLY-2608 parallels kind of the emergence of precision oncology. And if you think back, you know, some of the original efforts against PI3K alpha were kind of these dirty multi-kinase inhibitors, and obviously, their utility was very limited due to the toxicity. The next generation of efforts against PI3K alpha were much more selective, and that has yielded alpelisib, which is, you know, selective for PI3K alpha, but unfortunately is not mutant selective. And again, although it's approved, its actual use in the real world is limited by the off-target toxicities of grade 3 hyperglycemia, rash, diarrhea.
So our hypothesis, all the way back in 2017, was if we could create a mutant-selective inhibitor of PI3K alpha, we'd be able to make the off-target effects as minimal as we could and reduce these rates of grade 3 hyperglycemia, rash, and diarrhea and make it more tolerable, and then that, in turn, would lead to greater efficacy. And I think some of the early clinical data that we showed last year from our trial through dose escalation showed exactly that, that we have lower rates of grade 3 hyperglycemia, diarrhea, and rash, and hopefully, over time, that will translate into a greater efficacy profile.
Thanks. And yeah, we've seen some news this past December also from Roche, who is also has a, you know, next gen PI3K inhibitor in development. They had positive phase 3 data in breast cancer. What do we know about that drug? How does it compare to yours, and how does this news read through to your program?
Yes, that's exactly right. At San Antonio, Roche showed the results of its phase three trial INAVO120 for inavolisib. Inavolisib is exactly as we just talked about, it's a non-selective PI3K alpha inhibitor, and the data was remarkable. In a subset of first-line patients, in a triple combination, inavolisib, fulvestrant, and palbociclib, it showed a clear efficacy result, doubling the PFS versus the control to 15 months. So a very impressive outcome. As we've talked about before, the non-selective inhibitors have limitations due to their off-target effects. And so alpelisib, the approved agent from Novartis, was unfortunately not able to have a dose that they can use in a late line trial with a CDK4/6 inhibitor.
But Roche did find such a dose, and so inavolisib plus palbo plus fulvestrant was trialed, and it was done so in a narrow patient population, in that these patients were endocrine resistant, which is about 40% of the hormone receptor-positive, HER2-negative patient numbers, and in metabolically relatively healthy patients with relatively low hemoglobin A1C and fasting blood glucose. And so we believe that this kind of trial encompasses about 20% of the first-line patient population. And so a remarkable result and, you know, real proof of concept for the mechanism of if you could inhibit PI3K alpha in these patients, could you see an efficacy benefit? And I think INAVO120 clearly showed that.
Now, it gave us great confidence because if we could, with our selective PI3K alpha approach, dial down some of these off-target effects, then there is a hope that you could use this triple combination in a much broader patient population. And we're talking about large numbers here, so this could be a very significant commercial opportunity. And so we were very excited at the start of the year to announce that, you know, we're going as fast as we can now with RLY-2608 in our doublet combinations with fulvestrant. We've also initiated a triplet in combination with fulvestrant and now what is becoming the market leader, ribociclib.
Right. So we'll talk about those in a second. But, so we saw some phase 1 data from RLY-2608 last year, and, you know, I know you have since advanced a 600 mg twice a day expansion cohort, but I'm not sure if it's been declared as the official RP2D, but so what needs to-- what do you need to see and, you know, what, what else needs to be accomplished before you can determine an RP2D?
Maybe I'll hand that one over to Don Bergstrom, Head of R&D.
Yeah. So, we disclosed some data last August, on our dose escalation that was triggered by actually opening a 600 milligram expansion cohort. It was the first cohort that we opened. So we had dosed 17 patients at 600 milligrams in dose escalation, and I think what we were seeing in the initial data at 600 milligrams was entirely consistent with the profile we were looking for, for a recommended Phase II dose. So of the 17 patients that were treated at the time we did the data cut, 15 of the 17 remained on study and continued to be treated. Of the 5 patients who had measurable disease and had at least one post-treatment scan, we saw a tumor regression in four of the five patients with a confirmed response in 1 of the 5 patients.
In the seven patients who we had at least six months follow-up on, six of the seven achieved clinical benefit for an interim clinical benefit rate of 86%, which compares very favorably to what's been reported for agents like alpelisib and inavolisib in a similar population, where there's either 46% or 48% clinical benefit rate, respectively. So I think all of the data were pointing towards 600 milligrams looking like a candidate RP2D. We opened an expansion cohort of 20 patients. These are all post-CDK4/6, ER-positive breast cancer patients being treated in combination with fulvestrant. We enrolled that 20-patient cohort over the remainder of 2023 and disclosed in last month that that 20-patient cohort was fully enrolled, and we're expanding it to enroll an additional 20 patients at the 600 milligram dose.
We also, at the same time, are opening a 400 mg expansion cohort, not because we think it's actually a preferable dose, but to use as a contrasting data set to support our discussions with FDA around Project Optimus guidance, to be able to support the 600 mg dose as being the optimal dose to take forward into eventual pivotal development in the post-CDK4/6 setting. I think, you know, today, the data we're seeing are pointing us towards the 600 mg dose. I think what we need to do is to continue to generate the data that will allow us to have the productive discussion with the regulators around 600 mg being the right dose to take forward into a pivotal trial.
So it sounds like the 20 additional patients that you added on top of the prior 20 and this at the 600 is purely to serve as another comparator to the 400, or what drove the decision to add more patients?
Yeah. Yeah, so I think, when we look back at Roche's—Genentech and Roche's development of inavolisib and ask what data they had at the time they decided to trigger their phase III program, they had about 60 patients' worth of doublet data in the post-CDK4/6 setting, and they had dose escalation data in combination with palbociclib in the triplet setting. You know, so I think that is one of the external benchmarking databases, and that's essentially the database that we'll have over the course of this year, both for internal decision making as well as for discussions with regulators.
Then what are you hoping to see for efficacy in that large cohort now at 600 to advance?
Maybe-
Confident, yeah.
Megan, you can take this one.
Yeah, I think that you can look at it a few different ways. Don mentioned the existing treatments where you're seeing about 46%-48% CBR or PFS. You know, alpelisib and BYLieve had PFS of about 8 months, though that's a little lower in the real world. Inavolisib is about seven, and capivasertib is about 5.5 post-CDK4/6. So that's kind of one level to look at it. The other thing that we're thinking about is really just what we're hearing from the clinical community, which is that they really want to see double-digit PFS because no one's been able to achieve that yet. So that's really what we're focused on going forward.
Right. And, you know, is, you know, response rate per se has been, you know, difficult in these late-stage breast cancer patients. Is that a factor as well, or do you mainly look at PFS and CBR?
So, you know, response rate is, at best, a secondary endpoint in this patient population. I think seeing responses gives you, you know, as we're seeing right now, gives you some comfort that you're at an active dose. But from the point of view of thinking about further development and what the actual registration path is, it's supportive at best.
Okay. And then, as you noted, you did recently open this triplet cohort with ribociclib and, I believe, guided to sharing initial data later this year as well. And so perhaps talk about what your plans are for the triplet, perhaps relative to the doublet strategy in terms of next steps.
Yeah. So I think obviously combining with ribociclib is the path we think that, you know, per current standard of care, is the best path to getting into eventual frontline metastatic development. Ribociclib, as Sanjiv mentioned, I think is, is emerging as the market leader in this patient segment based on being the only CDK4/6 inhibitor that has shown an overall survival advantage in this patient population. So I think physicians who are evidence-based are leaning towards using ribociclib. So as we think about the options of how to eventually move into an earlier line setting, there are either segmented options you could take. So again, Sanjiv mentioned the way that Roche designed the INAVO 120 trial. It was in a highly segmented patient population in the front line that only addresses about 20% of the patients.
So one option could be to go into a similarly segmented population, potentially an endocrine-resistant population, but potentially with more liberal inclusion criteria around metabolic fitness, given the lower hyperglycemia liability that 2608 has. There's also the potential to consider going into a broader frontline patient population, which would require a larger and longer trial, but would result in a TTP that likely would include about three years of therapy with 2608, which in a large population, with that long duration of therapy, could be a very large commercial opportunity.
Right. So we're looking forward then to seeing data from both the triplet and the doublet cohorts in the second half. Is that correct?
Right. So the triplet will be dose escalation data. We anticipate that most of the patients, if not all of the patients that are enrolled in dose escalation, likely will have seen prior treatment with a CDK4/6 inhibitor, so will not reflect a pure frontline patient population. But we do anticipate in the second half that we will have that initial dose escalation data.
Great. And then I know you underwent a sort of pipeline prioritization, you know, exercise last year, but you do have a more selective H1047R inhibitor, I believe, in development still. So what are your plans for that, and how does that fit into the scheme at this point?
I think based on what we've seen so far, we're not sure we need it. And so obviously, the 2608 approach is to have a pan-mutant approach. And obviously, we're seeing very encouraging early signs, both on the safety side and on the efficacy side. I think we'll see how it progresses. I think the hypotheses are that, you know, could you use a selective 1047R either in combination with a pan approach or following? But I think we'll see how the 2608 data pans out, and should we need it, the 1047R approach, we'll have it.
Right. Okay. Then perhaps switching gears over to RLY-4008, which is your FGFR2 inhibitor that is already in a pivotal registration study in cholangiocarcinoma. But we've also seen data last year at the Triple Meeting in, you know, in a basket study of other tumor types with FGFR2 alterations. So, you know, I know you've said the cholangio trial is sort of in, you know, fully enrolled now and in follow-up stage. But, you know, I think perhaps on the tumor-agnostic setting, so what are your plans at this point in time based on the data we've seen last year?
I mean, I think you've kinda summarized our thinking well, which is the cholangiocarcinoma fusion, pivotal cohort, is fully enrolled now, and obviously, the early cuts that we showed, of that data are encouraging. And obviously, our molecule, RLY-4008, is exquisitely selective, and will allow that data to mature. The basket trial, data that, that you referenced from last year's Triple Meeting, again, showed encouraging signs of activity in other alteration types and other tumor types. And I think our approach in 2024 is to allow all this data to mature, and then in the second half of the year, take a look at the totality of the data and then plot which is the most rational pathway forwards, and at that time, engage the regulator.
What are some of the aspects you need to align on with the FDA, specifically in that case?
Right. So, you know, I think when we've talked about the pivotal cohort so far, that has been focusing exclusively on cholangiocarcinoma. I think the data that we showed at Triple Meeting shows that the potential for RLY-4008 could be broader and could be across histologies, as defined by FGFR2 alterations. So I think one of the key questions is: how do we take these databases that we have? One, which is obviously exclusively focused on cholangiocarcinoma, the other, which has a very diverse set of tumor types, where we've seen responses, I believe, across five different tumor types with FGFR2 fusions, supporting that there's the potential for broad activity across various tumor types.
How you harmonize those two databases into a filing, either that could support, you know, potential tumor-agnostic labels for 4008, and it's sort of operationally how you would think about bringing those together.
Okay. And will we, will we see additional data? I know, I think you've been adding more patients to those cohorts in the tumor-agnostic setting.
Megan?
Yeah, we've guided to data and regulatory updates for tumor-agnostic in the second half of the year.
Great. And, you know, typically, you know, investors in many cases think about about 30% response rate, you know, as a, you know, bar in many types of cancers. Do you think this could apply in this case, and/or are there other endpoints that you think regulators may consider when thinking about a approval decision?
Yes. Yeah. I mean, certainly, response rate has generally been the primary efficacy endpoint that's been used in these labels, along with durability of response. So it's not just getting a response, but also understanding the quality of those responses and what the durability is. It's one of the drivers for us to continue to allow the data to mature so that we know what both of those parameters are. And, you know, I think the 30% number, again, is one that people have arrived at largely based on some, you know, precedents and then just the underlying statistics for how the FDA has been looking at these types of approvals. But all of these, again, are gonna need to be validated with a regulatory interaction.
Okay.
I think the other part that regulators have made clear on the tumor-agnostic is just wanting to see good diversity of tumor type as well, and diversity of effect across tumor types. It can't just be primarily one tumor type driving the response. They wanna see, you know, at least several in the majority.
Right. Great. Okay, thanks. And then as we think about the sort of commercial opportunity for 4008, we've obviously seen extremely high response rates in cholangio, and there's obviously products on the market. You know, how do you think about the commercial potential in cholangio, based on perhaps factoring in duration differences to sort of approved drugs?
You wanna take that one, Megan?
Yeah. So from a cholangio perspective, I think we're looking. It's a fairly small commercial opportunity, about 1,000 patients a year in the US. And so that's why you heard us at the Triple Meeting say that we're not gonna be filing that as our first filing application. And so we're aligning the timing of that with the tumor-agnostic opportunity, which is the larger opportunity. And so if you look at by alteration type, for the FGFR2 fusions, you have about 13,000 patients a year in the US. And for amplifications, you have about 7,000 patients a year in the US.
Or if you were to look at kind of across tumor types, for HR-positive breast cancer, where we've seen a strong signal across alterations, you have about 7,000-8,000 patients a year in the U.S., and then about 2,000 for gastric across alterations. So, those are definitely the bigger opportunities. I think, all of the different pieces come into play, though, in terms of our commercialization strategy, because it's still a matter of finding those patients, especially given that they're not existing approved therapies. And typically, patient identification follows once you have an approved therapy. And so, while they are bigger opportunities, they're not as well identified as well.
Okay.
So something that we'll take into account as the data mature and as we make our decisions on go-forward plans.
Great. And just one more on the tumor-agnostic. There's two baskets, the fusions and then the mutations or amplifications or maybe three baskets, I suppose. What's your understanding or your best guess assumption for the size of these cohorts that could potentially support regulatory filings? They can be small, but sometimes not. So what are your thoughts there?
Yeah, I mean, I think that's, that's again, where we can look at precedent, and you can see that there have been cohorts anywhere from 40 patients to about 150 patients. But I think to be able to say definitively where we would fall in that range would require us to be on the other side of a regulatory interaction.
Great. Sounds good. And then, you've again, as part of sort of some pipeline prioritization, last year, you paused development of a CDK2 program that you had announced in the past. But you're investing into new molecules, and you're planning to announce two or more clinical or preclinical programs this year. Can you just talk about how your pipeline strategy has evolved, perhaps, in the recent year?
Yeah, I mean, I think last year, as you rightly point out, on our CDK2, which was an IND-ready asset, and our ER alpha degrader, which is a development candidate, we paused them primarily because of our kind of clinical execution focus on RLY-2608. And so I think both programs we remain very excited about, and at some point in the future, they stand ready to be unpaused. And I think, well, you know, we've seen obviously a lot of excitement around both programs over the last year, looking at data that we've seen from programs that are in the clinic. I think behind that, we have a preclinical pipeline of seven plus, and our commitment is this year to reveal at least one new program.
And the profile of that will be very similar to the kind of previous, kind of innovative programs that we've shared, such as PI3K and our FGFR2 selective program, in that we believe that it'll be against validated targets, with a defined patient population, and that we can get into the clinic and show very rapidly proof of concept data. So we look forward at some point this year to sharing those. But I think it all speaks to the productivity of the approach.
You know, we have a unique approach using a combination of experimental and computational approaches, and there's a lot talked about the productivity and promise of some of these approaches. I think all we can say is now we have had four clinical programs all created by our platform, our CDK2 program that sits at IND, and then this pipeline behind it. So we look forward to sharing this over the course of the year.
Great. Well, with that, I think it's time to wrap up. Thank you to the Relay team. It's been a pleasure of seeing you today.
Great. Thanks very much for the invitation. Thanks.