Rhythm Pharmaceuticals, Inc. ($RYTM)

Good day, and thank you for standing by. Welcome to the Rhythm Pharmaceuticals Conference Call. [Operator Instructions] Please be advised that today's conference is being recorded. I would now like to hand the conference over to your speaker today, Dave Connolly, Investor Relations. Please go ahead.

Thank you, Daniel, and good afternoon, everyone, and welcome. This afternoon, we issued a press release with the Phase III top line results for our EMANATE trial. You can access the press release as well as the slides that we will be reviewing today by going to the Investors section of our website. Listed on Slide 3 are the speakers for today's call. David Meeker, Chair, President and Chief Executive Officer, will walk through these data. And Dr. Alastair Garfield, our Chief Scientific Officer; and Hunter Smith, our CFO, are available to answer questions on this call as well. Before we get started, I would like to remind everyone that these statements we make on this conference call will include forward-looking statements. Actual events or results could differ materially from those expressed or implied by any forward-looking statements as a result of various risks, uncertainties and other factors, including those set forth in the risk factors in our Risk Factors section of our SEC filings. In addition, any forward-looking statement made on this call represents our views only as of today and should not be relied upon as representing our views as of any subsequent date. Except as required by law, we specifically disclaim any obligation to update or revise any forward-looking statements. With that, I'll turn the call over to David.

Thank you, Dave, and thank you all for joining this afternoon. It's been a long road, ending not exactly where we had hoped, missing the primary endpoint in all 4 cohorts, but with some positive elements, which will position us well going forward. As expected, we saw positive signals in the POMC Het cohort. Somewhat unexpectedly, the SH2B1 cohort was negative. And surprisingly, we did see a clear signal in the SRC1 cohort. We will look at each of these in greater detail. We continue to look forward to our March 20 PDUFA date, and I will have no further comments about that on today's call. In addition to the signals in the POMC and SRC1 cohorts, we have learned more about each of these genes, which will position us to design a better study going forward. We know patients with impaired signaling through the pathway have a difficult time losing weight with diet and exercise, and we have not seen a clear placebo effect in any of our trials. The stronger placebo response in some of the cohorts is consistent with a larger number of patients having benign variant and a pathophysiology more consistent with general obesity. Finally, this trial suffered from an exceptionally high dropout rate, which becomes extremely challenging when you perform the conservative multiple imputation analysis on the primary endpoint. Slide 6 shows the top line data. The values are the least square mean difference between setmelanotide and placebo for the modified intent-to-treat analysis using multiple imputation for the missing data points. As you can see from the p-values, we missed the primary endpoint analysis for each of the 4 genes. Now before we dig more deeply into the data, let's briefly revisit how we got here, beginning on Slide 7 with a familiar cartoon of the MC4R pathway. Our approach to identifying genetic populations likely to respond to setmelanotide was always a holistic process based on evolving scientific and clinical data that supported MC4R pathway dysfunction as a contributor to obesity in these patients. The genes under investigation in our EMANATE study all impact POMC neuron function and are predicted to reduce alpha-MSH, suggesting that replacement with setmelanotide would be effective in reestablishing pathway function and reducing body weight. The success of this approach is predicated on understanding that the gene variants we enrolled do indeed compromise MC4R pathway function. As shown on Slide 8, this variant classification process is a significant challenge for both the medical genetics field and Rhythm in no small part due to the fact that scientific understanding of variants is always evolving, leading to a dynamic classification process. For our POMC, PCSK1 and LEPR heterozygous cohorts, we had a good understanding of our enrolled variants that enable more accurate patient enrollment, which as you will see, is borne out in our genetically confirmed cohort analyses. By comparison, much less was known about the SRC1 and SH2B1 cohorts where almost all variants are classified as VUS, variants of unknown significance. As a reminder, on average, only 20% of VUS variants end up being classified as P, pathogenic or likely pathogenic. The gray box in the upper right corner shows which variant classifications were enrolled in each cohort. For POMC, PCSK1 and LEPR Hets, we enrolled pathogenic, likely pathogenic and suspected pathogenic variants, the latter representing variants still classified as VUS, but the data is increasingly pointing towards that variant being pathogenic. For SRC1, all patients were VUS, whereas the SH2B1 gene had a mix with the 16P11 deletion patients by definition, having loss of function of the gene. Slide 9 shows the number of patients enrolled in each cohort, which is noteworthy for the small number of patients in the leptin receptor Het cohort. This is an extremely rare indication, and we were not able to enroll fully. The trial design was a standard double-blind, placebo-controlled trial randomized 1:1 with a primary endpoint of change in BMI at 52 weeks. Slide 10 shows the demographics, which were similar across cohorts. On average, patients in each group suffered with severe obesity with BMI values greater than 40. The adverse event profile as shown on Slide 11 is similar to other trials run with setmelanotide with injection site reactions, GI complaints and hyperpigmentation being most common. Slide 12 shows the dropout rate with an average 40% to 60% of patients discontinuing with the exception of the LEPR Het group with a very small number of patients, the dropout rate was similar between treated and placebo patients. Slide 13 shows the reasons for discontinuations, aggregating patients across all 4 studies. Patient decision was the most common reason for patients on placebo and adverse events were the most common reason for patients on setmelanotide. Now beginning with the POMC Hets cohort on Slide 15, let's dig into the clinical data. As you might expect, we have looked at the data several ways. We missed on the primary endpoint in the modified intent-to-treat population using the conservative multiple imputation method required by regulatory authorities where the imputed data is informed by the opposite arm. Specifically, a patient who discontinues in the setmelanotide arm at whatever time point will be considered as though their outcome at week 52 was the same as the placebo group. On Slide 15, we show a post-hoc analysis using the last observation carried forward methodology where the last value a patient has prior to discontinuing is used for the final analysis. This showed a highly statistically significant difference of 5.53% for the same modified intent-to-treat population. Moving to Slide 16. As noted in the intro, we have utilized an in-vitro assay to assess loss of function for each of the variants in the POMC and LEPR genes. This allowed us to classify variants as pathogenic, likely pathogenic and suspected pathogenic. The study protocol included reconfirming the genetic eligibility and variant classification for each patient in order to conduct a prespecified analysis for genetically confirmed patients. That reconfirmation resulted in a net 11 patients being removed for not having that genetic confirmation. On Slide 16, we show using the last observation carryforward methodology in the genetically confirmed cohort, a statistically significant difference of 6.8%. Slide 17. Finally, using the rationale that a response to setmelanotide, a highly specific peptide of the MC4R receptor provides further evidence that the MC4R pathway is central to that patient's disease, we looked at those genetically confirmed patients who completed the study. This additional post-hoc analysis showed a highly statistically significant lease square mean difference of 9.7%. The short story for the LEPR Het cohort is summarized on Slide 19. These patients were hard -- rare and hard to recruit. A small number of patients recruited had one variant, which on reanalysis was down classified by our central lab. We do not anticipate doing further work on this patient population. The SRC1 cohort, as we have highlighted multiple times, by virtue of the fact that all variants enrolled were classified as VUS had a low probability of success. Despite that, the primary endpoint modified intent-to-treat analysis with multiple imputation for the almost 60% of patients who discontinued from this cohort was a least square mean value of minus 4% and showed the strongest trend towards significance with a p-value of equal to 0.12. On Slide 21, in the same post-hoc analysis we performed for the POMC Het patients using the last observation carried forward methodology, we saw a highly statistically significant lease square mean difference of 6.24%. On Slide 22, we looked at the post-hoc analysis of genetically confirmed patients who completed the study where we saw a statistically significant difference of 8%. Finally, on Slide 23, to give you an example of the process by which we can work to better understand loss of function and the ability of patients with a true loss of function variant to respond to treatment, we examined those patients who had a variant in the critical binding domain for the protein for the SRC1 protein. 19 patients carried this variant with 12 of the 19 completing the trial. In those 12 patients, the placebo-adjusted difference was 12.7%. The most disappointing cohort was SH2B1/16p11. We had hypothesized that the 16p11 deletion patients who, by definition, would have loss of function of the SH2B1 gene, which is embedded in the deleted region, should respond. The modified intent to treat showed no effect on average. And when we look at the genetically confirmed completer analysis shown on Slide 25, the difference was only 3% and not statistically significant. A further analysis looking only at the known loss of function 16p11 patients on Slide 26 similarly showed only a modest effect, which did not reach statistical significance. In summary, we are disappointed by the top line results, but encouraged by the findings in the POMC Hets and SRC1 cohorts. We will continue to interrogate these and other DAYBREAK genes from our Phase II study to improve our ability to determine loss of function variants. Our priorities going forward for our next-generation therapies will be the HO studies, the Prader-Willi studies, BBS studies and patients with confirmed loss of function genetic variants. Patients with genetically impaired signaling through the MC4 pathway represent a significant unmet medical need and addressing that need is one of the 3 key pillars of Rhythm's strategy. The other 2, as you know, are anatomic abnormalities of the hypothalamus, which acquired HO is perhaps best known, and we look forward to the upcoming PDUFA date. The third pillar is Prader-Willi syndrome, a genetic disease but with unique challenges and representing a large unmet need. I want to close by thanking the patients, their families and our clinical investigators and their teams. This trial was not easy. It was statistically negative, but not negative in terms of the learnings, which will move us closer to developing a meaningful treatment for these diseases. And with that, we can open it up for questions.

[Operator Instructions] Our first question comes from Phil Nadeau with TD Cowen.

Just a question on next steps to dive in a little bit further. It seems like the imputation analyses that you did computing what placebo would have done are very conservative and maybe not appropriate for a study with such a high dropout rate. So I guess first part of the question is, why wasn't that the imputation method that was chosen? And then second, on next steps, it sounds like you're not going to take the data to the FDA to see if like the completer analysis or LOCF could support a filing. Is that fair?

Yes. Taking the second part of the question, that's correct. We're not going to file on any of these. So I think the way we're looking at this data, if it was robustly positive, we would, of course, as we've said, taken that to the regulators, FDA. Your question about whether they would accept a different approach, no. I think there -- across all of our programs, they've used the same, call it, conservative, they would call it standard multiple imputation analysis. We use it for HO. If we've done a less observation carried forward even in that trial with a much smaller dropout rate, we would have had a slightly better result. So it is a more favorable way of looking at that. The value in doing it that way is to help us internally. I think our goal was to try to use this data to understand does it work in that gene or not. I think we're exiting this saying, I still think it works on Leptin Receptor Hets, to be honest. That's so rare that's -- and it was challenging -- so challenging to recruit. I think it's unlikely, as I indicated in my comments, that we would do more work. SH2B1, that was pretty negative. It's not 100% negative, but I think we're not going to rush there to do additional work. Whereas for the POMC Hets and the SRC1, I think we will. It won't be the highest priority coming out of the list of things that I put forward there, but we will get to that. So does that answer your question, Phil?

Our next question comes from Tazeen Ahmad with Bank of America.

I just wanted to ask about the potential of looking at these same basket indications for your next-gen assets. Based on what you saw for setmelanotide, does that reduce your interest in looking at the ones that didn't work for the follow-on compound? And if not, can you give us a sense of how you're thinking about when you might try to look at the next-gen compounds in these indications?

Yes. Thanks for allowing me to clarify that. Yes, all additional work, future work will be done with the next gen. So we would not go back and run another trial with setmelanotide in these. I think this data is relevant in the regulatory conversation, and we know that, that it's providing with the subsequent -- some post-hoc, some not post-hoc, but the subsequent analysis, I think, give us reasonable confidence, and I think regulators would look at this as supportive evidence for the role of MC4R in these indications, and then we need to run a positive trial with one of the next gen, which we would do. As I mentioned in my response to Phil, we -- our initial priorities, not surprisingly, are going to be to run the HO studies, definitely getting Prader-Willi up and running as quickly as we can. We want to have the next-gen run in BBS so that we check the boxes on our approved indications and get those taken care of. And then we'll be looking at the additional genes, and we'll be weighing both what we've learned here and also what we know from what we learned in our DAYBREAK studies, and we'll begin to pick and there's probably 5 plus or minus genes across this group, which would be of the highest priority.

Our next question comes from Derek Archila with Wells Fargo.

Just wondering if the level of discontinuations you saw in the trial were expected. And maybe just remind us how these populations are different from Bardet-Biedl and HO in terms of sensitivity to adverse events seen with setmelanotide.

Yes, it's a good question. I think there's more for us to learn here about exact reasons why. But just to highlight a few things. Let's take Bardet-Biedl, where we had a 20-plus percent discontinuation rate. Those patients, many of them had caregivers who are working with them. So there were other support structures around them, which made it easier for them to stay in a trial. We also had a shorter placebo period. HO in contrast was a full year study. Most probably also had supportive structures around them to a higher degree than this population, but they also were seeing a larger benefit. And so motivation to stay on there might have been greater. I think this trial, we started running back in 2022. It's been going for a long time. We are running against the emergence of GLP-1s, for example. And so I think there's a lot of patients in this study who might have been looking over the fence at other things they could be taking, if they weren't seeing a dramatic response and being in a trial like this has a pretty significant burden in terms of commitments and the number of tests that need to be done. So there's no one answer, but the aggregation of things, timing, level of burden, not the most dramatic response on average, I think, left us where we are with a high discount rate.

Our next question comes from Seamus Fernandez with Guggenheim Securities.

This is Ed Lang on for Seamus. Just following up on some of the discontinuation rates. Any more details you can provide on specific AEs or subject decision that comprise the majority of dropouts versus some of the expectations? And really just curious potential to mitigate with the next-gen or [indiscernible] assets. I'm wondering if this is more Sentinel related or whether there's other optimizations that can be done. Second, just curious in terms of prioritization of next-gen assets here, which ones you may think could be more appropriate? And then third, is there an opportunity to further exploit dose for patients reaching max dose of Sentinel?

Okay. You're breaking the one-question rule there. Let's see what we can do here. So I think in terms of discounts, I don't -- the split, as we said, the placebo group was basically patient choice, and that's everything, mostly the list that I ran through. I think those in the set group had adverse events. The adverse event profile, again, as highlighted, is exactly what we've seen in the other trials where we didn't have that kind of discount rate. So I think you couple an adverse event with maybe not getting the result that they had expected, whatever, it's invariably a combination of factors for them. Have you run a better trial, I think the focus on running a better trial will be taking what we've learned from this, and we have learned and there's stuff we're in the process of learning today, to be honest with you, about how to better identify patients with true loss of function. So you take that into the next trial, which says we want to minimize the number of patients who don't have true loss of function. They're not going to benefit. And if you can truly be more specific about the true loss of function, again, they'll run a better trial. And then it's just execution on the trial. And there are things that we just didn't execute on well, either as a company or as a CRO, we're working with. And so those are learnings. Again, Rhythm, we've matured a bit as a company. And so I think we can just run a better trial. But the bigger factor will be getting the right patients in, of course. Your question about higher doses, I don't think -- that's an interesting question. When we come to the next-gen, I mean, they're not apples-to-apples in terms of dosing. So what is the dose for the weekly and what is the dose for the daily oral, we'll figure that out. But I think for any population we're in, we're always asking the question of, are we in the right dosing range and for Set, I think somewhat luckily, for the most part, we have been. We'll continue to ask that question going forward.

Our next question comes from Corinne Johnson with Goldman Sachs.

Maybe you could talk a little bit to your last point about the trade-offs between getting more specific on identifying the patients that would respond best to the therapy versus the pace of enrollment and finding those patients? And related to that, how has the genetic testing progressed, if at all, since you started EMANATE that you could kind of take advantage of in a future iteration of this program?

Yes. So that -- I would take the trade-offs part of it. As you know, Corinne, my background in rare diseases have really been shaped by I think this is true across all medicines. It certainly works or all diseases certainly works in rare diseases. You want to get responders. And the more diluted you are by patients who don't respond, even though you can feel good about a "higher faster rate of enrollment, you're not serving anybody and you end up in a place where we are today probably with a nonsignificant result, and we probably could have done better if we've been able to better identify the patients. So that's going to be the highest priority. I think we have insights into that, certainly for the POMC Hets and for the SRC1 that will allow us to be better at that. The genetic testing is a key part of the genetic strategy. And I think many of you have heard us talk about this, which is Rhythm's goal is not to be a testing company, but we recognize in a rare disease space, the company often has to provide the testing. We do that today. We've learned a lot from that. Our goal ultimately, as we get more genetically driven diseases approved, then the value of testing goes up. And so you're running a panel. And if you're looking for one, your hit rate off the panel will be low. If you have 5 potential genes on that panel, which might translate to responding to therapy, then by definition, you'll have a higher hit rate and you're likely to test more. So there's a virtuous circle that we'll be entering into as we begin to develop this further. And then ultimately, back to us not being a testing company, as obesity evolves here, 2 things -- multiple things will happen. But one of the things that's going to happen is as GLP-1 use becomes broader, it's not everything, it's a hammer and everything is a nail. That's no longer true. And increasingly, people will recognize GLP-1 failures, how do they work them up. One of the obvious things you can do is test those patients for their genetics. And so these things are going to come together and Rhythm and the MC4R pathway diseases are going to surface, I think, in that kind of dynamic. So did I answer your question?

Our next question comes from Mike Ulz with Morgan Stanley.

This is [indiscernible] on for Mike. Due to the rare nature of these genes, is there anything that you can do to accelerate enrollment for future next-gen trials? And then how much did hyperpigmentation contribute to discontinuations?

Yes. And the second part, I don't know -- that's probably in the data set. I don't know it today. I'll just answer that by saying, overall of our experience, hyperpigmentation is about 5% of patients who discontinue about 5% of it is due to the hyperpigmentation. I would guess it's probably similar here in terms of overall percentage. What can we do to accelerate enrollment? That's just increasing testing and increasing awareness. And like I said, I really think this dynamic of patients who fail to respond to GLP-1s, that world is exponentially growing. And so the pool of patients who have a history of early onset obesity, who failed to respond to a GLP-1, they're going to be prime candidates to be tested. And one of the things about the VUS challenges that we talked about, which is, okay, you got tested, but your results show that you have a VUS. I don't know what to do with that. If you are an individual who's failed a GLP-1, had a history of early onset obesity and certainly, if you had hyperphagia, the probability that your VUS is pathogenic is higher than if you're just a random individual with a variant in that gene. So I think these are things that will increasingly make the next trial easier to run and to recruit into.

Our next question comes from John Wolleben with Citizens.

Just wondering if you could talk a little bit more when you say additional work with the next-gen, what that could look like? Would they be smaller Phase II studies or a chance for a Phase II/III pivotal program?

Yes. So here's one. So the first thing I want you to take away is the priorities for Rhythm right now are these more important from a company strategic standpoint, trials, the Prader-Willi [indiscernible] and BBS, as I indicated. On the genetic side of the equation, some of -- there's some investigator-initiated work going on now. You could imagine, and we are looking at it, that some of the syndromes where, by definition, if you have a syndrome, we know you have loss of function to the extent that we have some insight into that through early proof of concept, then we might run a study there, and we would run study, call it a 2, 3, whatever. But I think -- and the dosing for the most part, when we say 2, 3, you're sort of -- the 2 part is trying to figure out your dosing and the like. We'd run 1 trial and however, we would have to design it. But we -- for each of these genetic, it would be only one trial. And to the extent we go back to POMC Het, for example, or SRC1, this data, as I said earlier, would be supportive in that regulatory filing in terms of the role of MC4R pathway, and we would be enrolling patients with a slightly narrower scope, but with a higher probability of having loss of function in their gene.

Our next question comes from Lisa Walter with RBC.

It sounds like the plan is to explore development path forward with the next-gen MC4R agonist. But I'm wondering what the strategy might be to ensure fewer placebo arm discontinuations going forward even with these next-gen assets. Any color here would be helpful.

Yes. I mean that's a challenge. I think -- so aside from just supporting patients in any trial, which is what a well-run trial does, you just try to get patients to hang in there. I think we did a really good job in our HO study and those patients hung in there. One of the things that helps placebo patients stay in is, a, a belief that they don't have other good options. So I'm a GLP-1 failure, nothing else has worked. This is my hope. And I know that when I finish the trial, then I have a shot, I will get the drug. So that's the biggest thing. And then secondly, back to expectations of having a reasonable response. So you manage placebo groups by having the reward at the end be meaningful to them in addition to the just general support.

I'm showing no further questions at this time. This concludes the question-and-answer session. I would now like to turn it back to David Meeker for closing remarks.

Okay. Well, thanks, everybody, for tuning in on short notice here. back to where I started. I'm disappointed in this. But actually, this is a solid building block going forward. And I think you've heard me speak before, and you'll hear me speak again. I think I'm pretty excited about where Rhythm is and all the things we have coming down the pike here. So we look forward to our next conversation. Thanks.

This concludes today's conference call. Thank you for participating. You may now disconnect you lines.