Good morning, everyone. I'm Andy Berens, Senior Biotech Analyst at Leerink Partners. Thank you for joining us on the final day of our Global Healthcare Conference in beautiful and sunny Miami. We're very excited today to have Brian Sullivan from Celcuity join us. Thank you, Brian.
You're welcome. It's my pleasure.
Why don't we start with a brief overview of the company, and then we'll dive right in. There's been a lot of interesting data in your area that we're going to talk about, but I think it would help to have an overview of the company for those who may not be familiar with Celcuity.
Sure. I started the company a number of years ago, initially to focus on development of a platform that could analyze and characterize the activity of signaling pathways in living tumor cells. We have essentially then evolved over time towards the development of our own drug. Our lead asset is gedatolisib. It is a pan-PI3K mTOR inhibitor. We have three current programs in process, a phase III program that will have two readouts this year, one in a wild-type population, PIK3CA wild-type, the next one in PIK3CA mutant patients. We just initiated a phase III study in first-line breast cancer patients. The other study I just alluded to was a second-line study. We have a phase I-B/II study in process in castration-resistant prostate cancer. We can go into more detail later.
OK. Why don't we talk about the rationale behind bringing in a drug that hits two nodes in the pathway? Other companies have chosen to go very selective or even super selective. What do you think the advantages are from gedatolisib versus that approach?
Sure. While we were developing our platform, one of the pathways of interest to us was the PI3K/AKT/mTOR pathway. We call it the PAM pathway. One of the questions that I wanted to answer internally was this: how could a pathway that's so important as an oncodriver, had the highest proportion of patients with mutations associated with the pathway, be so underdrugged? At that time, when we started looking into this, there was no approved PAM inhibitor, or there was one everolimus mTORC1 inhibitor. We started developing a way to characterize signaling activity in this pathway. Along the way, as part of our development process, we characterized how good these various inhibitors are. One thing popped out, and one drug in particular stood out to us. It was gedatolisib.
Gedatolisib is actually a pan-PI3K mTOR inhibitor, so it actually hits six nodes, the four class I isoforms, and then mTORC1 and mTORC2. It was clearly superior as an inhibitor of PAM pathway-associated activity. We thought, wow, this is a fantastic drug, clearly differentiated. We had a relationship with Pfizer. Pfizer held it at the time. We thought, let's collaborate on development of this drug. That is when we found out that they were outlicensing it. Our decision was really fundamentally based on this conundrum of underdeveloped pathway probably reflects the minimal targeting of this pathway. That reflects the nature of the pathway itself. It is very complex. Most pathways that are drugged in oncology have a single target. This pathway has six targets, these class I isoforms and mTORC1 and mTORC2.
When the pathway was discovered as an oncodriver 20 years ago, because of the understanding of this complex structure, every major pharma had a pan-PI3K mTOR program in place. None of them made it out of phase I. The reaction was, or response, was to narrow the focus to inhibition of only single nodes in this complex pathway, PI3K alpha, mTORC1, AKT. The problem with that approach is that you essentially are compromising efficacy, because the biological imperative requires you to inhibit, or to optimize efficacy at least, to inhibit the pathway comprehensively, which requires control of all those nodes. On the one hand, these drugs have gotten approved. They got through the therapeutic window because of the challenges with toxicity. In our view, they haven't optimized efficacy.
Internally, and we've published data that has shown that with a pan-PI3K mTOR approach that gedatolisib takes, the potency and our ability to inhibit cell proliferation is 300 x greater than it is with single-node inhibitors. It really provides a kind of a mechanistic assessment of how important multi-node inhibition is relative to single-node inhibition. The drug, we think, solves the riddle that has been challenging to optimize the potential impact targeting this pathway can offer patients, at least initially in breast cancer and hopefully prostate cancer as well.
OK. Maybe it would help to just broadly go through the data that you've demonstrated to date. The drug's in phase III. There's a couple of big readouts coming out. We'll talk about those shortly. What have you guys shown with this approach? I know that some of the narrow drugs that you've mentioned, the response rates are usually single digits, low double digits. Some of them have been able to alleviate the hyperglycemia you see with some of the other agents. What have you guys demonstrated that could support the hypothesis you just mentioned?
Sure. In our early-phase study, we enrolled 138 patients, four different cohorts, based on their prior treatment history. In the treatment-naive cohort, patients received gedatolisib and then standard of care, which was palbo and letrozole, reported 48 months median PFS, objective response rate of 79%, which compares very favorably to published data for the palbo-letrozole regimen. With our phase III dose, in patients who had prior CDK, we reported 12.9 months median PFS and objective response rate of 63%. Those numbers compare very favorably to current standard of care. One of the other very relevant factors is that because of the approach the drug takes by inhibiting it comprehensively, it is able to, at least the preliminary data suggests, induce comparable efficacy independent of the status of the PIK3CA pathway.
PIK3CA wild-type patients and PIK3CA mutant patients get about the same response. Whereas the single-node inhibitors to date, at least, and we wouldn't expect it to change over time, have only demonstrated activity in patients who have a PIK3CA mutation. The drug has the potential, really, to offer an all-comer solution for patients and for doctors to prescribe, greatly simplify the treatment paradigm. More importantly, particularly for wild-type patients, because there really are very limited options that have an endocrine backbone right now that offer meaningful benefit.
Right. I just want to define wild-type, because we have some other drugs that have read out, one yesterday. Your wild-type includes ESR1 mutant patients, right?
Yes. Essentially, we're not factoring in ESR1 status. When we define wild-type, it's PIK3CA wild-type. That includes ESR1 mutant and wild-type. Similar for mutations of PIK3CA, we're enrolling both ESR1 mutant and wild-type. We haven't found in our early data—we haven't published this—but we did not see any difference in outcomes based on ESR1 status.
OK. That's irrespective of whatever endocrine therapy is potentially used?
Yes.
Interesting. OK. In terms of some of the concerns I've heard about investors, going broader, great, maybe more efficacy, but they're going to have more safety toxicity. What have you seen with the key events of interest?
Sure. One of the great characteristics of the drug is how well tolerated it's demonstrated, at least in the early studies. Over 500 patients have been studied with the drug and have published results. In breast cancer, we had cohorts of patients that received a weekly dosing. The discontinuation rate for those patients was around 9%. That's almost exactly the same level as you would see for palbociclib letrozole. The addition of geda to that regimen didn't add incremental toxicity that required patients to discontinue. Whereas with the phase III dose, for the second-line setting and beyond our current studies, it's a three-week on, one-week off schedule. The discontinuation rate was less than 4%. Again, below the published data for discontinuation rate for CDK and hormonal therapy.
Our conclusion is, and we'll see with our phase III data, is that geda's proven to be, at least so far, or demonstrated at least, that it's well tolerated, doesn't require people to get off the drug, and essentially doesn't incrementally add toxicity relative to the doublet that it's added to. As we think about the positioning of the drug, we'd say that the safety tolerability will actually, I mean, again, we'll see. Indications would be that it'd be better tolerated than, for instance, the doublets that have PAM inhibitors today, and certainly as well tolerated as CDK4/6 and hormonal therapy.
Right. What about adverse events of interest, like hyperglycemia?
Hyperglycemia has been somewhat the third rail on this pathway. That's because the PAM pathway regulates glycolic function. That function takes place in the liver. Particularly with oral drugs that are processed first pass through the liver, you disrupt that activity and can induce very high levels of hyperglycemia. Geda is an infused drug, IV administered. You avoid the liver on first pass. Other characteristics of the drug result in optimal exposure for tumor control, but not so great that you are inducing unacceptable levels of hyperglycemia. We reported to date about 25% overall hyperglycemia. That compares with the other inhibitors out there that are 60%-70%-80% hyperglycemia when you measure glucose levels. Some of the alpha inhibitors coming out, it looks like they've figured out how to optimize targeting for mutant versus wild-type, and that can create favorable hyperglycemia.
For us, it hasn't been a reason for either discontinuation or dose reductions in the course of the development of that drug. We think hyperglycemia is off the table. The one adverse event that does require management is stomatitis. Stomatitis is kind of typically induced by this class of drugs. Interesting feature of our drug, though, is that it responds, even if you don't have a prophylaxis for it, if you've basically given the drug, you experience a stomatitis, responds very effectively to steroidal mouth rinses. That's why we haven't really seen discontinuations due to stomatitis, because most patients, great majority, respond well and reverse stomatitis with this mouth rinse.
To avoid that and to really, hopefully, optimize the patient experience for the phase III studies and ongoing studies going forward, we're giving patients the mouth rinse as a prophylaxis upfront for the first couple of cycles. There's some data that suggests you can reduce by 90% grade II or greater stomatitis events if patients for the first two cycles use this dexamethasone mouth rinse. It can be very impactful. We think going forward, we'll see when we report our data. We would expect to report more favorable stomatitis events than was reported in our early-phase data.
OK. I guess, just to close the loop on mTOR, that drug, there have been a number of mTOR inhibitors for a while. What have they shown in this type of setting?
Sure. There's one that's approved. It's an mTORC1 inhibitor called everolimus. Novartis has that drug. That drug had challenges upfront. When it was launched, stomatitis wasn't very well managed by this dexamethasone mouth rinse. They had a 24% discontinuation rate when it was launched. It had trouble getting much penetration. That's likely a function of just daily dosing, high molar amount of drug given to the patient over the course of every month. They subsequently found, though, and they're the ones that ran this randomized study, found that in the case of their drug, if they give the dexamethasone as prophylaxis, patients take it while they're receiving treatment, it really greatly affects it. That's been very helpful for that drug. There's no data right now in the post-CDK setting for that drug's efficacy.
There have been a couple of retrospective studies that have looked at chart, done chart reviews, and found it's offering three to four months in the wild-type setting. Again, it's not randomized data. It's not clear, really, what that drug is doing. Again, kind of confirmed the importance of mTORC, at least mTORC1. We think mTORC2 is also important as a part of the feedback loops in this pathway. Geda inhibits both. We think we're able to optimize mTOR-related activity as well. Again, it's been demonstrated to be very relevant.
OK. By hitting the pathway in two locations, in two nodes, are you able to lower the amount of coverage you need for each individual pathway? Like, in other words, by adding mTOR, does that mean you don't have to hit the PI3K pathway as hard to get the same level of efficacy?
I think sometimes people conflate potency with toxicity. I think in this pathway, the toxicity is obviously related to potency and on-target effect, but more related, or as much related, to target organ exposure. For instance, the liver for hyperglycemia, GI for GI-related toxicity. Again, as an infused drug, you avoid those two organs in first pass. What is interesting about geda is that it is equally potent against all six of these nodes at low nanomolar concentrations or subnanomolar concentrations. When you compare the activity of the drug as an inhibitor of tumor cell proliferation, you need 12 nanomolar to get a half maximal effect. Whereas with the single-node inhibitors, to have half maximal effect on their proliferation potential, they need 300 x more drug. That just reflects the more limited biological potential of inhibiting a single node.
If you do inhibit this pathway comprehensively and, in effect, shut off these resistance mechanisms, you subsequently need a lot less drug. Against each individual target, for instance, our drug, when you assess it biochemically in a sulforhodamine assay, is 10-fold more potent against alpha than alpelisib, or the Relay compound, as an example. It's a get the best of all worlds. By being very potent comprehensively, you ultimately need less drug to induce the on-target effect that you're seeking.
OK. Before we talk about the trial you're running and the big readout you have coming imminently, why don't we just talk about a couple of the data sets that have come out recently? I know we've chatted about EMBER before. I guess we can start with that one. I also like to talk about SERENA-6 , which I thought was a brilliant study. I'd like to hear your perspective, because I don't think we've chatted about that since the top line came.
Sure. The EMBER- 3 study, I think it's a confusing study because of the mixed populations they enrolled. They enrolled first-line patients and second-line. They didn't break out that data. Very hard to, essentially, with what they published, to know exactly what that drug is doing. When we tease it out, it looks like nominal effect as a monotherapy, no effect on ESR1 mutant wild-type patients. We think nominal effect in that population, even with the abema addition. We don't think there'll be much overlap with that combo, which is a question of whether it'll get approved because of the trial design. Take that off the table, assume it does. We think there'd be very limited overlap with what we're doing. As far as this SERENA-6 study, no, great, very interesting switch strategy for patients to develop these mutations.
I think there's going to be an operational component associated with deployment of that study. We'll certainly optimize first-line experience for these patients, switching hormonal therapy to optimize for patients that develop an ESR1 mutation. I don't think it will necessarily impact us downstream. I think the PAM pathway is involved intrinsically in this disease. Our first-line data really provides a pretty interesting demonstration. If you were to look at the impact of hormonal therapy in the first-line setting with letrozole, you'd see about 14 months median PFS. If you bolt on palbo, the CDK4/6 inhibitor, you get a near doubling with that drug. Clearly, CDK4/6 is important. And our data, which again, you have to be wary of cross-title comparisons, it's a single-arm study.
Nonetheless, we reported 48 months median PFS, which we think suggests and kind of provides demonstration of the relevance of this pathway in this disease overall, not essentially as a consequence of resistance to the initial treatments they've received. That is why we made the decision to go into the first line as early as we could, because we think that the hypothesis, the rationale for that, is very strong scientifically.
Yeah. I almost had like to hear your thoughts on this. I mean, I don't know if there are data out there yet. I feel like if you can actually bring in a drug that's active at ESR1 mutation to the front line, or whatever you want to call it, front line 1.5, it could, obviously, it should block the ESR1 mutation from being an escape pathway. It could put more emphasis on PAM, which.
Exactly. The PAM pathway is involved. Really, when people think of breast cancer, they should think of it as involving three cooperative pathways: your pathway, CDK4/6 pathway, and the PAM pathway. Optimal tumor control, at least our data suggests, certainly would require comprehensive inhibition of all three of those pathways. Certainly, you can optimize, in the case of ESR1 mutation patients, a control of the pathway. That does not make the PAM pathway less relevant. As you said, it suggested it could make it more relevant. I think the PAM pathway, in our view, clearly needs to be inhibited to optimize outcomes for patients. We will have data, hopefully, that confirms that in the phase III setting. Certainly, the preliminary data suggests that.
OK. Why don't we talk a little bit about the phase III pivotal program, the readouts that you have? I know you've been asked probably 10 x as many times as I have about the bar that you're looking for, and the first readout, and then the second readout. Why don't we go through the two binary events you have that are very big catalysts?
Sure. We expect to report out next end of Q1, sometime in Q2, data for our wild-type. We have three arms. We're evaluating geda, palbociclib, and fulvestrant, the triplet, and comparing that and one primary endpoint versus fulvestrant as the control. We have another arm that's geda, fulvestrant, Arm B, that we're also comparing to fulvestrant. We're testing those hierarchically. We're sparing alpha. We're optimizing that analysis. As far as expectations, if you were to look at the data that's been generated the past few years with fulvestrant from in patients who've had prior CDK, you'd see fairly consistent results, essentially in this two- to three-month range. We think it's most likely that the control would, don't want to be too precise, but in that range.
We have heard pretty consistently from regulators, as well as KOLs, that what really will determine or establish a new standard of care is a delta relative to, in this case, this control of three months. That is not necessarily what we hope for. However, three months delta PFS will essentially, according to the regulators and KOLs and certainly the community docs we are speaking with, motivate usage of a regimen that offers that and would allow you to get very, very meaningful penetration. The market itself is very, very significant, $5 billion serve market potential with proprietary drugs. There is a there there. Even I am not suggesting and projecting anything from a penetration standpoint. However, it does point out that even with just a 30%-40% penetration, you can build a blockbuster indication. Certainly, more than three months is better.
We would hope, but again, the data will be reported, and we'll see soon, to have as maximum a delta as we could. The delta to be in the game and to be relevant, to be a potential standard of care is about three months.
OK. How important is the doublet arm in the readout, do you think?
Essentially, we added that primary analysis really to give us the potential to submit that for an approval as well. We wanted to make it a registrationable endpoint. If it is registered, it was at our option, our discretion. It is not required for success. If A/C is successful, that will carry the day. The rationale for it is to just provide doctors, if the data bears this out, with an option. This is not a homogeneous population. Some patients become not only refractory to CDK, but they cannot tolerate them for whatever reason. If we have an option that allows doctors to prescribe their patients geda, fulvestrant, as an example, it may not provide optimal tumor control, but for that patient, it may be the best option available to them.
It would provide a way to, we think, maximize overall penetration of the drug in this setting.
OK. And then the fulvestrant arm will have patients with ESR1 mutation in it.
Yes.
What percentage do you think you will get based on how it is?
The data to date shows that there's no difference in outcomes between patients with ESR1 mutations or wild-type with fulvestrant. Essentially, the control arm is going to be what it is, independent of the mix of patients with that mutation. If the data is consistent, if the patient population is consistent with what's been reported with some of the SERDs studies, you'd expect between 30% and 40% of the patients to have an ESR1 mutation.
OK. Subsequent to this readout, you have another phase III readout in mutations.
Yes. Exactly. There we're comparing our triplet, gedatolisib, palbociclib, fulvestrant, to alpelisib and fulvestrant. Alpelisib was reported median PFS in the 5.5-7.5 month range. Again, not a huge amount of data. We have our internal assumptions about what we expect there. Again, similar delta, I think, is required to be considered clinically meaningful, about three months, and if we're able to do that. What's interesting is that another PAM inhibitor was recently approved called capivasertib, an AKT inhibitor. It did not directly compare to alpelisib, but their data, if you just do a cross-trial comparison, was comparable, or you could even argue maybe a little worse, lower efficacy, but much better safety profile. There has been a rapid shift away from alpelisib. They have also, because of the better tolerability, been able to expand the penetration of that drug.
In less than a year, they're at a $600 million run rate, primarily just because of a safety advantage, which I think highlights how desperate oncologists are to find better solutions for these patients. Obviously, what they're hoping for is better efficacy and good tolerability. If they're not even getting better efficacy, they'll switch very quickly to a drug that offers their patients better quality of life.
When we see the first data release in the wild-type patients, what would you expect to see as a step up to the mutant population?
Early phase data suggest that it's comparable, but there's a numerical advantage relative to the wild-type data. Again, this is preliminary data. Sixty percent of patients in the mutants cohort were progression-free at 12 months. With the data published for alpelisib as an example, I think less than, it was about 25%, were progression-free at 12 months. We think, again, if you just look at that preliminary data, it suggests we should see, we hope, a meaningful delta. That same number for the wild-type patients was around 49%. That, because of the way the KM curves are shaped, can create somewhat of a delta. It's hard to predict because you don't quite know the slope of the curve. We would expect there to be a differential. I don't want to speculate what that would be.
I would expect it to be higher in the mutant population than in the wild-type population, the PFS.
OK. Just maybe with the last minute we have, one of the concerns of some investors, and I know we've chatted about this before, is the implications commercially of an IV drug in a setting that seems, I mean, fulvestrant obviously doesn't seem to be going anywhere. A lot of the physicians are hoping to have oral options. What do you see as a?
The coin of the realm in oncology, as is the case in most diseases, is efficacy. The guidelines basically, or for the most part, are oriented around recommending regimens that offer superior efficacy relative to the alternatives. The second criteria that they emphasize is tolerability. They want to maintain as long as possible the patient's quality of life. That is the benefit-risk calculation. The route of administration really isn't a consideration that's even taken into account with these guidelines and typical recommendations. Certainly, all things being equal, if you have two drugs that offer similar safety and efficacy, if one has a better route of administration, that's going to be an advantage.
If you have a regimen that is offering better efficacy and is well tolerated, I think there's very few, and again, based on our internal research, we think it's a very small number of docs that wouldn't want to make that an option for their patients. The other, I think, consideration to understand is that breast cancer, the largest drugs are all infused drugs. Herceptin was launched as a weekly drug 20 years ago, 25 years ago now, offering a three-month delta in efficacy. Obviously, they built that into a $6 billion peak revenue drug, added Perjeta, another infused drug. Certainly, in HER2, the new ADC has not found any limitations with infusion. There is an infrastructure in place. These docs know how to essentially deploy and manage infusion. There are other advantages in this setting from a managed access standpoint.
These docs get to recover costs when they infuse patients. They capture some of the cost of the prescription for these patients. Because it's a medical benefit as an infused drug, patients aren't exposed in the same way they are with a prescription to copay risk. There are a lot of actual advantages from a doc's perspective in prescribing their patients infused therapy. There is the compliance question. They know their patients are taking the drugs because they're showing up, and they'll know if they don't show up, which is a consideration with some oral therapies. Overall, we just don't think it's a relevant factor, given what we think the safety and efficacy that we hope to report will be.
Great. We are waiting for the data, as I am sure you are, and a lot of the investors. We look forward to when you turn over the cards.
Yes. Great. Thank you.
Thank you, Brian. Thanks, everyone, for joining us.