So I think for all those reasons, it’s a legitimate ask. I think there’s a strong argument to be made there. But I also want to reiterate that our guidance right now is that the full ENCORE data set will be needed for approval. And if we get an opportunity to go earlier and the FDA is clear on that, then we will take advantage of it.
Andrea Tan: Got it. And then just on your peak sales estimates for brenso of the $5 billion plus. Just curious if you could walk us through the assumptions that get you to that $5 billion and how much of that is driven by bronchiectasis versus CRS without nasal polyps.
Will Lewis: Yes. So we haven’t gone into greater detail on that front, but I think the key drivers here, and I know a lot of you who are thinking about modeling, want to look at each of these. So price. When we think about the price here, we’ve talked about specialty asthma products like FASENRA, which is about $40,000 a year. We’ve referred to that historically as very likely a floor in price, assuming that the target product profile is consistent with what we saw in WILLOW. When we think about the addressable market, we’re talking about 1 million diagnosed patients at the time of launch. These are patients that would meet the criteria that we think would be suitable for effective treatment. Based on the results of the WILLOW and subsequently assuming success the ASPEN trial.
So that’s a very interesting addressable population out of the gate. There’s an opportunity to explore building beyond that, but we’re going to start with that as a point of guidance. And then penetration rate, I think for a first in disease drug with a novel mechanism and a very low treatment burden, like a once a day pill, it doesn’t really get much better than that. And given the safety we’ve seen to date, which is very compelling in WILLOW and is at least that good in ASPEN to date, as it was in WILLOW, we think that the penetration rate being very healthy is something you can rely on in your modeling efforts. I know some people have modeled it quite conservatively, but we think we’re going to have a very effective penetration rate.
We’ll get more specific about that as we move forward. And I want to take this moment to highlight that shortly after the data, we’re going to put out the top line data within a week, we intend to have a commercial day, if you will, a virtual commercial day, where we will walk you through each of the assumption sets for each of our three late stage programs or pillars. That would be ARIKAYCE frontline and Brensocatib and bronchiectasis, and how we’re thinking about it in CRS and HS and TPIP for both PAH and PH-ILD. So that people can update their financial modelings consistent with the data that we will have released by then, which will include new data on PH-ILD and PAH, and obviously the data from ASPEN. Where I’m trying to go with all of this is, I think the vast majority of the investment community right now is sort of in a holding period as they wait for the ASPEN data.
But very shortly after that data comes out, and presuming it’s positive, there’s going to need to be another rerating of this company, in my opinion, because the financial modeling that has been undertaken to date, I would put in the very conservative category. There are some people that do not give us any credit for TPIP in terms of revenue generation. There are folks who are modeling ASPEN and bronchiectasis incredibly conservatively, in my opinion. And on the other side of successful data, I think it’s going to be important to rapidly readjust that lens and see these three pillars for what their potential really represents. And to put numbers behind that, we think ARIKAYCE, all MAC NTM, that’s $1 billion plus product. We think TPIP between PAH and PH-ILD is a $2 billion peak sales product.
And we think brenso and bronchiectasis and CRS without nasal polyps alone is north of $5 billion in peak sales. So there’s a lot to model here, a lot to understand, and we’re going to walk you through it within days of the top line ASPEN results.
Andrea Tan: Great. Thanks, Will.
Operator: Your next question comes from the line of Jason Schwartz – sorry, Joseph Schwartz of Leerink Partners. Your line is open.
Joseph Schwartz: Great. Thanks for all the updates. So I have a couple questions on bronchiectasis. The first is, since the lower blended blinded rate of pulmonary exacerbations in ASPEN could reflect both treatment effect as well as some placebo effect, which we’ve seen in other bronchiectasis trials, including WILLOW. I was wondering what factors you think could drive placebo response the most in this setting, and whether you’re able to do anything to control for this. We’ve heard that things like better adherence to airway clearance and just reversion to the mean can occur in these trials. I’m wondering if you think, are these the biggest factors to consider or are there other things that could drive placebo effect more. And are you able to do anything to minimize that impact?
Will Lewis: Yes. So when we look at the particular aspect of exacerbation rates in the study, there are several important things to remember. The first is, how do you define an exacerbation, and how do you let patients into the trial using that definition? I want to remind everyone, our definition of an exacerbation is very strict. There not only has to be an exacerbation declared by the patient and the physician, but the physician then has to document a change in treatment, including either prescribing an antibiotic or admitting the patient to the hospital. So that treatment response makes that a very substantive event that they are trying to address. That exacerbation or event is then adjudicated by a third party. So there are several layers of protection to ensure that what we’re seeing are real events.
In the past, in other studies where the definition hasn’t been as strict, there’s been a little bit of shift around the number of events and where they’ve fallen. So we use this definition in WILLOW. It’s why we saw a clear, statistical significant impact on both doses, on the measure of that study. And I think it’s why we have such confidence going into this one. As we think about assumptions, the observed rates in most trials are right around 1.35 or so. We saw 1.37 in WILLOW, we had assumed 1.2 more conservatively, and for ASPEN, we kept that assumption at 1.2. So the powering of this study is very healthy relative to observed rates. Where the placebo rates have been lower in other studies, it is almost entirely driven by the fact that patients were recruited in areas that the placebo rates drop, because they are receiving better medical care.
There’s one study in particular where up to 30% of the patients that were recruited came from Russia or Eastern Europe, where it is commonly the case that by having patients from those areas, simply including them in a clinical trial where they get better medical care, results in a drop of the actual placebo rate. I’ll remind you that WILLOW had about a 13% recruitment from Eastern Europe. No patients from Russia and ASPEN has less than 10% from Eastern Europe, and no patients from Russia. So the probability of having those kinds of influences or aberrations in our study is extremely low. And for that reason, we feel like the study is well powered. We have seen the right level of blended blinded events that we want to. If this drug has a treatment effect as it did in Phase 2, we will be able to capture it in Phase 3.
Joseph Schwartz: Thanks, Will. That’s very helpful. And it kind of leads nicely into my next question, which was actually on the antibiotic prescribing behavior in particular, given I think there’s some different propensity to use antibiotics in different regions around the world and even within countries such as ours, I’m just wondering if that could be a potentially confounding issue at all. Have you looked at WILLOW in that sense, and are you able to do anything to control for that kind of behavior?
Will Lewis: Yes. So, once again, what we want to do with these studies is have enough patience and enough variety that it controls for any idiosyncratic aberration that may come from small patient numbers or responses. Happily in WILLOW, I’ll give you an example, probably the most important antibiotics that we kept a close eye on is the use of macrolides, because they have some anti inflammatory effect associated with them. Interestingly, in that study, we did not see a difference between those who were given that or were not given that. And so in Phase 3, we kept that just as it was. And those patients are going to end up being equally distributed across the different arms of the study, and we didn’t see an influence from it.
But if it were to be there, the larger numbers and the distribution across the study should protect us from any aberrations that might hypothetically flow from that. I want to be clear, though, we didn’t see any distinction in Phase 2, but because of exactly what you’re talking about, we looked at it very carefully. So we feel good about the design of Phase 3, because it has tried to compensate for all of these different influences that in some cases have been in evidence in prior Phase 3 studies that have failed.
Joseph Schwartz: Great. Very helpful. Thanks, again.
Operator: Your next question comes from the line of Ritu Baral from TD Cowen. Your line is open.
Ritu Baral: Good morning. Thanks for taking the question. Well, I just had some questions on the ASPEN statistical analysis plan. And then my second question was just some clarification on ARISE versus ENCORE. For stats, it sounds like you guys now are splitting the alpha between the 10 milligram and the 25 milligram dose versus any sort of hierarchical analysis. Is that correct? And then could you describe further the type of analysis that you’re going to be using and how it plays into what you said about. I think you had previously said that the study was powered down to show an effect size of low 20s reduction. I’m just wondering if that – when you said that that was alluding to a p value of 0.01 or that 0.01 to 0.05 range that you mentioned today.
Will Lewis: Yes. So the whole thing to understand is, better than most in the Arcana of statistics, is how do you control for multiplicity? And what we’ve done in our statistical analysis plan, at least as we conceive of it right now, and we’ll finalize the details with FDA as we approach the final data set reveal is to use what’s called the truncated Hochberg analysis, where you basically preserve some alpha to look at each of the different dosages independent of one another. So if 10 wins, then we can go to 25 and we can look at secondary endpoints in both. If 10 fails, we can still look at 25 and still preserve alpha to look at secondary endpoints underneath 25. Not all statistical analysis approaches permit that, this is one that has been used commonly, it’s with precedent.
So we don’t anticipate it being controversial, but that’s how we’re thinking about the control for multiplicity and the ability to look at both the primary endpoint for each of the doses and preserve some residual alpha for the secondary endpoints. And so when we talk about being able to detect an effect into the low 20s, it contemplates that approach. So hopefully that’s responsive to the question.
Ritu Baral: Got it. And you are testing that 10 milligram first before the 25?
Will Lewis: Yes. I think, however, you want to think about it, the 10 or the 25, they get equal amounts of support for whether or not you clear that initial hurdle. And then some residual alpha is preserved for secondary endpoints under each of those doses.
Ritu Baral: Got it. Which of the secondary endpoints have you decided to reserve alpha for?
Will Lewis: Well, you reserve alpha for effectively all of them, and then you go hierarchically through the secondary endpoints to see whether or not you clear them.
Ritu Baral: Got it. Okay. And then the meeting that you will be having on frontline ARIKAYCE, will that solely be on ARISE or are there any questions on ENCORE QLB or statistical analysis plan? Anything around ENCORE left to nail down at that meeting? And it sounds like you’re waiting to do an in-person meeting versus the usual Zoom or questions they like to do over paper.
Will Lewis: Yes. So to be clear, there are two groups that we are dealing with at FDA. The first is the PRO group, that’s a specialized group at FDA that looks at the creation and elements that are necessary for an appropriate patient reported outcome tool to be used. That’s where the first dialogue has been happening. We’ve had written exchange with them. It’s been very encouraging. We’ve sent them the information they want. There’s always an exchange and request for additional analysis, when that stuff goes through, they then review it. We then have an in-person meeting to run to ground any final remaining questions or points they have. And at the conclusion of that in-person meeting, we should have their blessing that this PRO is suitable for use in ENCORE.
At that moment, once we have that PRO sort of finalized for ENCORE, then we can turn around and go to the review division and say, based on the ARISE data and the fact that this PRO is suitable in ENCORE and therefore in ARISE, is it reasonable to pursue a Subpart H pathway for earlier approval for all MAC NTM. And so that’s the cascade that’s going to unfold here. So the specific question is, are we going to be talking about ENCORE or ARISE with the PRO division? We’re really going to be talking about the PRO and whether it is where it needs to be from their point of view. Once that’s constructed, then we will use that PRO as they have blessed it, to frame out the final elements of the statistical analysis plan for ENCORE. And then in parallel, we would go to FDA’s review division and ask about whether ARISE is adequate under Subpart H for an earlier approval.
But to be clear, things that are discussed with the PRO group are more like, what is the minimally important difference in treatment effect that they want to see used in the PRO? We’ve proposed a level based on what we shared with you last fall. The FDA could go above that. They could go below it. For us, it doesn’t matter, because the drug works in all settings, at all levels. That was the real breakthrough moment of the PRO last fall. So whatever the FDA sets it at, we’re going to be fine. It’s just – it really is up to them, what they feel is the appropriate threshold.
Ritu Baral: Understood. Thanks for taking the questions.
Operator: Your next question comes from the line of Vamil Divan from Guggenheim Securities. Your line is open.
Vamil Divan: Great. Thanks for taking my questions. Maybe one, just follow up on brenso and then one ARIKAYCE on the guidance. So on brenso, obviously there’s a longer trial with ASPEN than it was in WILLOW. Question we’ve gotten a few times from investors, just sort of the safety side of things and how long have patients, at this point, any patient has been treated with brenso? And is there anything you’re particularly concerned about as you look at a 12-month trial as opposed to what you saw in six months. And then in terms of the guidance, I know you give it more on a global level, but I’m just curious, especially on the Japan side, where we have a little bit less visibility, if you can maybe comment on the trends you’re seeing there or maybe just broadly speaking to your growth expectations in Japan relative to what you’re expecting in the U.S. for 2024. Thanks.
Will Lewis: Sure. So when we think about the time of exposure in the ASPEN study versus the WILLOW study, I would tell you that should help us dramatically, because if you think about the Kaplan-Meier curve, it was separating and continued to trend apart at the different doses versus placebo, and that was within a six month timeframe. And actually, quite strikingly, remember that this drug takes about two to four weeks to get to full pharmacodynamic effect. So we really were looking at about five months of treatment effect in the WILLOW study, and to see that dramatic of an impact was unbelievably encouraging. So going out a year should help as we look for the events to take place and patients to improve. I will just share that we have had some patients in open-label extension and through a special dispensation that have been on the drug, I can think of one that’s been on the drug for more than three years at this point.
And the Data Safety Monitoring Board has met repeatedly. And as we’ve said publicly, the safety profile we see in ASPEN is at least as good as what we’ve seen in WILLOW. We’re obviously blinded, but in terms of number of events and what we’re seeing, it looks really good. And the WILLOW study was striking in that we had a higher dropout rate in placebo than we did in the treatment arms. So I think the drug is looking quite safe. We have to obviously examine the data in detail, but that is a really positive point. And it goes to the heart of why we have regulatory confidence here. Because if we see what we saw in WILLOW, we know that the FDA and other regulatory authorities start with safety, is the drug safe? And then they look at efficacy.
