Good morning, everyone. I'm Tara Bancroft. I'm one of the senior biotech analysts here at TD Cowen. I want to thank you so much for joining our Sixth Annual Oncology Innovation Summit. For our next session, we have a fireside chat with Celcuity, and it's my pleasure to introduce Brian Sullivan, who is the CEO and co-founder. It's a privilege to have you here, Brian, and thank you so much for joining us.
It's my pleasure.
Before I get started, yes, of course. I just want to remind anyone listening that you can email me questions at any time, tara.bancroft@tdsecurities.com, and I'll make sure that I incorporate it into the questions. Brian, maybe you could start with just a general update overview before we get into specific questions on VIKTORIA-1 and more.
Sure. As a company, we started, and I started the company, to develop a platform to quantify signaling pathway activity in live tumor cells. We have now since evolved to focusing on treating cancers involving the PI3K/AKT/mTOR pathway, which we refer to as the PAM pathway. This pathway is probably one of the most important oncogenic pathways. It is the most highly altered pathway, which correlates with role as a cancer driver. We think it represents one of the largest untapped drug development opportunities in solid tumors. Our lead assets are gedatolisib, PAM, PI3K/mTOR inhibitor. We have three studies ongoing: two in breast cancer, one in prostate cancer. We have an ongoing phase III study evaluating second-line patients with HR-positive/HER2-negative advanced breast cancer. That one is one where we expect to report data in the third quarter for a wild-type cohort.
For a PIK3CA mutant cohort, data we expect to report in the fourth quarter of 2025. We also have a phase III study that we're just getting going that's evaluating first-line patients, HR positive, HER2 negative advanced breast cancer. We expect to begin enrollment this quarter. We have an ongoing phase one B study in prostate cancer, a second-line study for men that have metastatic castration-resistant prostate cancer. We expect to report top-line data for that study actually later this month or later in June.
Okay, great. I guess before we get into details on the data readouts that are expected, I think it would be helpful to kind of set the stage for the ways in which gedatolisib is actually differentiated in this space. There is so much competition now going on. I think just also mechanistically.
Sure.
One of the things that KOLs are doing that we're seeing is kind of bucketing it in with this PI3K class and the safety concerns that are associated with that. I want to get your thoughts on that and how it's different from PI3K only class and also some rationale for how it could work in wild-type patients versus mutants as well.
A lot of ground to cover there. Unlike most pathways, the PAM pathway has multiple components or targets that have to be addressed. There are feedback and feed-forward loops between the PI3K isoforms like PI3Kα or AKT or MTORC1 that can cross-activate the ones that are uninhibited. This essentially results in adaptive resistance when one target is inhibited in the absence of the others. When the pathway's role as an oncodriver was first discovered, nearly every major pharma company initiated development of a PAM PI3K/MTOR inhibitor because of the understanding of this adaptive resistance and that that was key to optimizing efficacy. The critical role the pathway plays in these metabolic processes makes it difficult to optimize pathway inhibition.
Safety issues also led to the evolution of drugs that essentially just focused on single targets, let's say PI3Kα , AKT, or mTORC1 as a way to get through what was believed to be a very narrow therapeutic window. There are these approved inhibitors that hit alpha, PI3K, mTORC1, AKT. These drugs have not been shown to be effective in patients lacking PIK3CA mutations, particularly in patients who have had prior CDK4/6. Unlike these other inhibitors that really just address single targets, gedatolisib inhibits all these targets at sub-nanomolar or low-nanomolar potencies, concentrations. We have done analyses comparing how these approved single-target PAM inhibitors perform and found that gedatolisib is 300x more potent than the other drugs in vitro and the only one that is cytotoxic.
is also important to point out that gedatolisib is the only drug that has comparable potency and efficacy in our in vitro studies in tumor cells with or without PIK3CA mutations. Again, that contrasts with the approved single-node inhibitors that are three to six times less active in tumor cells when they lack PIK3CA mutations. The mechanism is highly differentiated. I think it is just because the most recent drugs that have been developed and approved happen to be single-target inhibitors. We think we will reestablish the paradigm as requiring comprehensive PAM inhibition in order to optimize treatment of patients with dysregulated PAM activity.
Okay. Definitely very great.
I guess the question was, okay, why do we think gedatolisib will work? What's the rationale for that in patients with these mutations? It reflects two factors. The first factor relates to the PAM pathway's role as a regulator of critical cell functions, including regulation of the glycolytic process. That's particularly relevant in tumor cells because they require a tremendous amount of glucose relative to normal cells to survive. That's their energy source. That makes the PAM pathway a fundamental tumor driver independent of the presence of an activating mutation like PIK3CA, particularly in tumor types like HR-positive breast cancer. The second factor just relates to this complex structure of the PAM pathway.
With the various PIK3CA mutations and mTOR complexes that kind of cross-activate each other, the pathway can continue to function if only a single node, such as PI3K alpha or AKT, is inhibited. As a result, in patients that lack mutations, in particular, complete blockade is required of the pathway to induce an antitumor effect. That is required to be done at low-nanomolar concentrations. This is another observation I think is important. If the only relevant feature of this pathway was the presence of PIK3CA mutations, you would not expect to see any activity on wild-type tumor cells in In vitro studies. Our studies with breast cancer cell line models and prostate cancer models and endometrial cancer models demonstrate that potency and cytotoxicity of gedatolisib, which again blockades this pathway completely, is nearly identical in tumor cells regardless of whether they have a PAM pathway alteration.
Our preliminary phase one B data provides a demonstration of this where the objective response rates, PFS rate at 12 months was comparable in the patient cohorts that had PIK3CA mutations versus those that lacked PIK3CA mutations.
Okay. With that setting the stage, I think now the most helpful thing would be to get into expectations for the phase three wild-type data that you said are now coming in Q3. I guess what would be helpful is hearing what you believe would be kind of the minimum HR that you would consider to be good data in this wild-type update. Not yes for hitting stats. We know that you have good powering and everything, but really in relation to other studies in the space. We recently got Serena 6, other marketed products even. I know there's a range of HRs for other different studies that we've seen from CAPItello at 0.59 and EMERALD 3 at 0.68. Some context there would be great.
Sure. Certainly, hazard ratio of 0.5 or below, I think, would be considered very compelling. As far as I know, no study in patients that have HR-positive, HER2-negative advanced breast cancer who are receiving endocrine therapies that include a portion of patients who've had prior treatment with CDK4/6 inhibitor has reported a hazard ratio below 0.5 as its primary endpoint. In terms of specific references, I guess there's been recent data, EMBER-3 data, HR for Imlunestrant, the oral SERD, was 0.62 in patients with ESR1 mutations, was not active in patients lacking those mutations. When Imlunestrant was combined with abemaciclib, CDK4/6 inhibitor, the HR was 0.57 for its primary endpoint. The EMERALD study evaluated another oral SERD, elacestrant, which was not active in patients lacking ESR1 mutations, but in the patients that had ESR1 mutations, the hazard ratio was about 4.55.
As far as the Serena 6, I'm not sure that's a relevant comparison because that really is a modified first-line indication. The patients in that study were switched to an alternative endocrine therapy and oral steroid before their tumors progressed. They are really just trying to optimize the endocrine backbone of that before progression sets in.
Okay. What have you heard from KOLs of what HR that they would want to see in order to use it? I know you said 0.5 would be highly compelling, but what do you think you're hearing from KOLs specifically?
I think when KOLs are evaluating results from a study and then trying to assess how those results compare to other studies, they're looking at a variety of factors. Certainly, hazard ratio is a critical one because it provides the relative reduction of risk for one drug to another that allows you to do some comparison if you've got similar patient populations. I think the delta in the PFS between the median PFS between the control arm and the study arm is very relevant. I also think the ratio, how much better was the study arm as a multiple of the control arm? They'll look at those three factors and draw a conclusion from that and from there make a decision. Obviously, safety will be an important component of their assessment as well.
They'll factor in those various efficacy metrics, the overall tolerability of the regimens, and make a decision based on that.
Yeah. Okay. Thank you. I guess the next thing I'd love to get your thoughts on is the idea of absolute median PFS. I know all of this of what you just said makes total sense, but the street has kind of settled on the importance of what the absolute number is. I'm curious what you would say to that, to us doing that.
Sure. No, I think obviously it's very easy just to evaluate top-line numbers. To be frank, the problem with using absolute median PFS is that it doesn't take into account the performance of the control arm. For all clinical trials, it's the relative difference between the study arm and the control arm that matters when assessing the clinical activity of these study drugs, not the absolute performance. Most of the studies in this setting use a very similar control, fulvestrant or endocrine therapy. The relative comparison really is very important. I think it might be helpful. Let's take results from two hypothetical trials just to illustrate this. Imagine hypothetical trial one, where you have a study arm that reports nine months median PFS and the control arm reports five and a half months median PFS.
In that hypothetical trial, then you have a study arm and median PFS that was 1.6x the five and a half month median PFS reported for its control. The incremental PFS was three and a half months. A lot of numbers there. Sorry. Okay. Let's consider a second hypothetical trial where the study arm reported seven and a half months and the control arm reported two and a half months median PFS. In this second hypothetical trial, then study arm's median PFS was 3x two and a half month control, and the incremental PFS was five months.
I think most oncologists would select a regimen with seven and a half months median PFS that was three times longer relative to its control and offer five months incremental benefit over one that reported nine months median PFS that was only roughly 1.6x relative to its control and only offered three and a half months incremental PFS. That is why, again, it is just not the way clinical decision-making takes place. It is just looking at the one number out of context, as in most things in life, context matters. In this case, I think it is obvious that the absolute PFS really just is not the right metric to use to establish some way of comparing one study regimen's results to another regimen's results in a clinical trial because this example, I think, provides a pretty good illustration of that.
Yeah. Definitely. Thank you for that context. I feel like that's one of the most illustrative ways you could have said that. I appreciate that. I guess then not to belabor this point, but I guess one of the things that I still think is tougher to get our minds around is the time that it's taken to reach the study event. Totally understandable. This is an event-driven trial, of course, but to what extent can you have confidence that it's not because the fulvestrant arm is outperforming? Like you said, with the five months, 5.6 or whatever it was in postMONARCH versus your expectation for three months, what underlies that and your confidence?
No, it's a great question. Oftentimes, the differences in controls have nothing to do with the drug or just randomness. It really is more a function of which patients were enrolled, what their tumor burden was, what the enrollment criteria were. That's where you can get the variation. That's why it's important that you assess the relative performance of the study drug to the control. There have been four studies randomized that have used fulvestrant as a control that have very similar enrollment criteria to ours. The average median PFS for fulvestrant in those four studies was about 2.8 months, 2.7 months. I think it's pretty reasonable for us to expect between two and a half to three months. That's just based on that data. I don't have any insight into anything. I just think, okay, those studies had very similar criteria.
Statistically, four different dips into that pool yield a very similar result. The studies that used fulvestrant as a control, like EMBER-3 and postMONARCH, did not have patient populations that were comparable to what we expect to enroll in VIKTORIA-1. EMBER-3 is a very complicated study because it enrolled four different subgroups of patients, first line, second line with CDK prior treatment, not CDK prior treatment. It is very hard to get a beat on what the relevant subgroup that compares to one we are enrolling actually reported for fulvestrant. postMONARCH study included a significant percentage of patients with non-measurable disease, and that results essentially in a significant upward skewing of progression-free survival for those patients. It is very hard to compare that patient population to ours. The relative difference between the control and the study arm is, again, as I said earlier, what is most relevant. We will see.
I mean, we'll soon know how good our reliance on this historical data is.
Yeah, definitely. Okay. I guess that all makes sense. I think one of the other things that I have a lot of conversations on with people is trying to understand it in context of the phase I data that you've gathered so far. I think it's clear from previous conversations why you would compare this dataset with one of the arms versus another in the phase I. Maybe just remind us again how we can use the phase one data to kind of increase our confidence in the phase III or make connections between the two.
In our phase I- B study, the results we viewed as very favorable. In the wild-type cohort, 60% of patients or so had an objective response. At 12 months, 49% of the patients were progression-free, which, again, we thought was very good relative to the data from currently available therapies. The patient population we enrolled in the phase I- B study actually had more stringent requirements. To be frank, patients with more tumor burden than what we expect to enroll in the VIKTORIA-1 study. In particular, all 100% of the patients in the phase I- B study had visceral mets, whereas we do not expect that in VIKTORIA-1. In fact, there were no bone-only patients in the phase one B study.
Bone-only patients tend to be much more responsive or rather have much longer PFS periods because of just the nature of how bone mets progress relative to mets in solid organs. In the VIKTORIA-1 study, we are enrolling patients who have bone-only mets as long as they have a measurable lytic or lytic blastic lesion. I guess another factor that differentiates the phase-I from the VIKTORIA-1 that I think kind of gives a sense that phase I-B study had kind of a tougher hurdle to demonstrate results. In the phase 1b study, patients were allowed to have prior chemo. Roughly a third of patients had prior chemo in the advanced setting. We're not enrolling patients who've had prior chemo, but those patients were excluded in the VIKTORIA-1.
That's relevant because patients who've had prior chemo in the advanced setting tend to be less responsive to subsequent targeted therapy. I think the phase I- B study actually is a good illustration. I think sometimes these phase I- B studies are optimized to get quick enrollment, which means essentially allowing everybody that they can get just to get the data, which is a fair approach. What that can do is create a skewing of patients that may not be representative of the population you'd enroll in a phase III study, and therefore the results may not necessarily translate. That's why you see kind of a lot of phase III studies not replicate. We've somewhat taken the opposite approach, but we'll see soon enough how that translates.
Yeah, of course. I guess while we're on this topic, can we think about how either the wild-type data can read through to the mutant data and/or how we should think about the phase one context that we have? How should we start now thinking about the expectations of the mutant data?
Right. On the margin, the patients who had PIK3CA mutations did better than the patients who lacked PIK3CA mutations in our early phase study. I would say there's a numerical superiority, not a statistical superiority. In the mutant population, 73% had an objective response versus 60% in the wild-type. Over 60% were progression-free at 12 months versus 49%. From a regulatory standpoint, from a biological standpoint, it's harder to induce a treatment effect on patients that lack an activating mutation, which has been the history in this setting. None of the PAM inhibitors that have been evaluated in a phase III study in patients that have had, or at least some of them have had, a CDK4/6 inhibitor, have demonstrated activity in patients lacking mutations.
If we report favorable results that offer a clinically meaningful benefit, we think that it would be reasonable to assign a higher probability of success in our mutant cohort. We'll have demonstrated ability to get over the, we think, the hardest hurdle to get over and that patients with mutations tend to be more sensitive to this drug class. We think by inducing comprehensive inhibition of that pathway will certainly offer the potential for superior outcomes.
Okay. How about in terms of safety? I think the fact that KOLs kind of unfairly bring up the PI3K class on its own and kind of hold you to that bar, can you give us some context in both the wild-type and the mutants? Because those comparisons to CAPI and ANOVA will inevitably happen.
Sure. The biggest challenge with drugging this pathway is that it controls these key metabolic processes. For instance, because this pathway regulates the glycolic system, you can induce high levels of hypoglycemia depending on how the drug is delivered, if it's orally delivered, processed to deliver where this glycolic regulation takes place. You can induce very high levels of hypoglycemia. The early drugs developed in this indication, most of them didn't get through the gate. The first one that did, alpelisib, induced very high levels of hypoglycemia, a very tough drug for patients. Still very high levels of hypoglycemia reported for drugs in this class. Gedatolisib is able to have reported much lower in its preliminary studies and also in a variety of other studies in different settings. About a fraction, 2/3's less hypoglycemia than has been reported with these other drugs.
No patients have been discontinued. Over 500 have published data for them. None have discontinued due to adverse events related to hypoglycemia. The discontinuation rate is another metric that we think is very important. 4% of patients discontinued in the early phase study who received the phase III regimen compared to, let's say, alpelisib, which was over 26%. Capivasertib, the AKT inhibitors, around 13%. The preliminary data suggests patients tolerate the drug well. Again, we'll know soon enough whether those results translate to the phase III, but we're optimistic just because of the number of patients that have been treated with this drug and that the profile has been consistent from study to study, safety profile.
Okay. I know we're quickly approaching time here, but I do want to give you a chance to talk about how you're thinking about the evolving market opportunity. I know that for the third space, I feel like especially on the street amongst investors, I feel like a big old wet blanket has been just placed on the entire class. We get comments from KOL sometimes about the utility of IV versus oral. In the context of how it's been evolving, especially in the last six months, one year, how are you thinking about how GETA fits in and will be used?
I think ultimately the coin of the realm in oncology or really any drug treatment is the efficacy. And then secondarily, and very importantly, would be safety. I think the route of administration, convenience is a factor. Certainly all things equal, if you had similar efficacy and safety, then the tiebreaker becomes convenience. I think if we are able to report superior efficacy and repeat the safety results from the earlier studies, I think that will weigh most heavily in decisions that physicians make as they're considering what drugs to treat. I think in the community setting, again, these doctors are used to infused drugs. I mean, the largest drugs in oncology are infused drugs. And breast cancer infused drugs are the largest drugs. HER2-positive patients receive Herceptin and Perjeta. As a class, those two drugs reach peak revenues of $9 billion.
It is not really a foreign concept. There are advantages for these doctors. They get to see the patients more frequently. These are sick patients, seeing them more frequently. Infused drugs ensure that the patients are compliant. They are getting their treatment. One of the challenges with oral drugs, especially if they have induced some side effects that are unpleasant for the patient, that can lead to inconsistent compliance taking the medication. Because of the nature of this process, infused drug is a health medical benefit. Different reimbursement path, to be frank, much less cumbersome. With these oncology drugs, it can be quite arduous for patients to get reimbursed. It can expose patients to fairly high copays. There is a certain financial toxicity. All these factors net-net, I think, again, it all hinges on efficacy and safety.
If we have that, I think we will be in a position to garner a significant share of the patient population, treat a significant share of those patients.
Yeah, definitely. I mean, I know I, for one, am really looking forward to the data in Q3 and Q4 and the prostate data, which we did not even get to. That is coming very soon. Brian, I just want to thank you again, as always, for your time, for joining us and educating us. Thank you, everyone, for listening.
You're welcome. Thanks for having me.