Second day of our conference. For those of you who don't know me, I'm Mike DiFiore, one of the senior biotech analysts at Evercore who covers Kodiak. With me today is Victor Perlroth, CEO of Kodiak Sciences. Victor, welcome. Thank you so much for trekking down to Miami to be with us.
Thanks a lot, Mike.
Yeah, and glad to have you. And before we delve into Q&A, we'd love to give your overview of the business, what key data flow and catalysts we can look forward to in the next 12 months.
Yeah, well, thanks, Mike. It's an excellent time to take a look at Kodiak. We have three phase III molecules that we're exploring in four phase III studies. We have three phase IIIs that we expect to read out over the next 12 months. The first will be the GLOW2 study, which is with tarcocimab in the diabetic retinopathy indication. We expect to have last visit in January of 2026 for that and top-line data in March. We have tarcocimab and KSI-501 that are in the DAYBREAK study. We expect last visit in August of 2026 and top-line data, let's say, mid-September. That's in wet AMD. We hope to file the BLA for tarcocimab very shortly thereafter.
And then about 12 months from today, let's say, we expect to have the phase III top-line data for the KSI-101 molecule readout from the PEAK study. So that's a lot of work at Kodiak over the next 12 months. And it really is going to bring together, let's say, more than 10 years of work. And we'll all see the outcome from all three molecules across those three studies together.
Got it. Very helpful. A lot to dig into, a lot of significant news flow coming up in the year ahead. But let's just start with KSI- 101. That seems to get mostly investor attention lately. KSI- 101, for folks who don't know, that's your IL-6 VEGF bispecific, the naked protein, the unconjugated form that you're developing for MESI. First of all, how did you come upon MESI as the ideal indication to pursue these targets?
MESI is just a simple term that describes very clearly what the drug's goal is. We're looking at patients who have macular edema, which is fluid under the retina. It's macular edema secondary to inflammation. It's patients who have macular edema, and they got that macular edema due to some immune insult, which relates and drives inflammation. It's actually very simple. Most of the population inside of MESI is what's called a UME or uveitic macular edema population. That's a little bit of a complicated word. Another way to say it is just inflammatory macular edema or other drugs that may be approved. One of them, Xipere, is macular edema associated with uveitis. All of these are sort of circling around the general concept of having macular edema, which is really kind of like a physical finding.
You sort of see it on the OCT imaging. You just see this massive fluid under the retina. And it's related to autoimmune or inflammatory attack.
Sure. I want to focus on the phase I, 1b APEX study. You just reported 20-week results. Again, for folks not that familiar with the story, these patients received four consecutive monthly doses and then just followed along afterwards. And the latest data point is at 20 weeks, correct?
Correct.
Okay. And this was kind of comparable in the sense to Roche's phase III trial because they both entailed patients getting four monthly loading doses. Again, but Roche enrolled UME patients. You guys enrolled MESI patients, which is kind of a heterogeneous group of diseases that can include UME. But since both trials entailed four-monthly dosing, a lot of folks kind of compare. And clearly, if we just focus on how Roche did in terms of BCVA gains, the drugs seem more or less equal. I mean, they seem more or less equal. But where Kodiak really seems to shine is that when you look at patients who achieve at least 15 or more letter gains versus Roche's IL-6, why did that happen despite having slightly lower BCVA stats at baseline?
So given that the letter gains were equal, why did more patients on Kodiak's drug achieve greater than 15 or more letter gains, would you say?
Yeah, thanks, Mike. Well, that's a thoughtful question. And the focus on 15-letter gain is important because that's really the area that FDA focuses on. So while the mean change in the BCVA in the MEERKAT and the SANDCAT seems to be in the ballpark of the mean change in BCVA seen in the top dose levels in the APEX study, our study with KSI-101, we disagree that they're similar results. Remember, we, as you mentioned, recently presented the week 20 data in APEX. And the 10-mg dose achieved as a mean change as a group a 15.4-letter gain. And the five-mg dose was close behind at a 13.4-letter gain. So the closer you get to the 15-letter gain as a group, that means the chance of an individual patient to cross over the threshold greater than or equal to 15 increases kind of like exponentially in a way.
I think that's what we're looking at with our drug versus vamikibart. For example, with a 15-letter mean gain as a group in a normal population, it would mean 50% of the population is above that 15-letter threshold. You can see that in the MEERKAT and SANDCAT, where a gain of 9 letters consistently results in about a 25% three-line gainer. That's supported by DOVETAIL, which was similar around 25%. With 11.9-letter mean gain, they got 34% three-line gain. With 12.8 letters, they got to 43% three-line.
Sorry to cut you off. I was actually comparing the 16-week data versus your 16-week data, in which the BCVA letter gains seemed about equal. But you're right. Your 20-week data is absolutely higher.
Yeah, even though we're not dosing anymore, the data continued to strengthen. So basically, every single letter of mean change matters. And achieving dryness in the highest number of patients possible is critical. So we know that inhibiting IL-6 is very useful. We saw that in our data. We saw it in the Roche phase I data, DOVETAIL. And we saw it in their PEAK and PINNACLE data, too. And it leads to vision improvement. But treatment success means achieving real dryness, which is like absence of fluid. And that's not accomplished in 50% of the patients on the DOVETAIL. So I think one last consideration regarding the vision, like the starting vision, you mentioned the baseline. We found it noteworthy that as a group, the top two KSI-101 dose levels are getting to 20/25 Snellen equivalent vision by week 20. So that's 79 letters.
That's a key differentiator because from the MEERKAT and SANDCAT, they're only getting to, say, 68-73 letters at week 16.
Got it. Got it. In terms of drying, that's where KSI- 101 really does seem to differentiate versus Roche's IL-6. In terms of retinal drying, a single dose of KSI- 101 showed superior retinal drying than four doses of Roche's IL-6. I guess, to what extent did the differences in baseline CST thickness, as well as in disease state MESI versus UME, maybe have factored into the magnitude of drying here?
Well, it's really important to remember what treatment success really means, which is to dry the retina. And the most direct way to measure that's by looking at the absence of the intra and subretinal fluids. So in APEX, we showed that with 101 by week 8, more than 90% of the patients achieved total absence of both intraretinal and subretinal fluids. Whereas in DOVETAIL, only 50% of the patients did that. Or alternatively, to say, half the patients had persistent fluid for the duration of the study. And that's a huge difference. And they haven't actually disclosed what those data look like for the MEERKAT and SANDCAT studies. So an indirect way to look at dryness is using the OCT, the central subfield thickness, the CST, the measure of the thickness of the retina.
There's not really like a magic CST number that translates to dry versus not dry. So we call it like a dryness corridor. And it's between like 250 and 325 microns. So I mean, to answer your question, irrespective of where you start, like the baseline, the goal is, as a group, to get deep into the dryness corridor. And we've shown that we clearly get into that corridor with a single dose, whereas it takes four doses of vamikibart to barely reach the shallow end of the corridor in MEERKAT and SANDCAT. So it is an interesting and important point that in the KSI-101, the 10-mg group, it had the thickest CST across APEX, DOVETAIL, MEERKAT, and SANDCAT. So we don't really believe that baseline CST plays a role at all. Yeah.
Great.
So I mean, to say, we think, if anything, that it's remarkable that even though we do, to your point, have a more heterogeneous kind of group of diseases, we're achieving a real dryness in basically every patient, whereas vamikibart has a 50% of treatment failure in UME by this metric.
Got it. Very helpful.
Victor, is it possible?
If I may.
Is it possible that part of what's happening is obviously there's a mechanistic basis, but is it also possible that perhaps as we think about the dosing frequency that's being employed, there was a week 12 in play in the Kodiak study? I don't think there was a week 12 in play in the DOVETAIL. Does that matter at all, in your opinion?
I mean. Whether you're looking at week 16 or week 20, I think we have a matched in the SANDCAT and the MEERKAT, I think they have the four-monthly doses, which we have. So I don't think so. I think it's with the combination of the VEGF and the IL-6 together, I look at it as like a 1 plus 1 equals 3 or a 1 plus 1 equals 10, plus our drug is at a higher dose, say, five or 10 mg.
You start bringing all of these things together.
Right. Remind me, how many patients were in APEX trial?
I think we had like 41 across the three dose groups.
OK. 41. Because I remember when I saw some of the data sets that you guys posted, it said preliminary analysis. So I couldn't tell what was going to be more updated about it.
We're going to have the full 24-week data. It'll be presented at Angiogenesis, I think, in early February.
I see.
Yeah, and what that's going to be nice to see is because we stopped dosing right after the four doses, and then we go out to week 24, so it'll be nice, and like I said, we saw continued improvement of the vision from week 16 to 20, even though we weren't dosing, which is cool, and then to kind of see what happens as we go to week 24. There's not really a bad outcome because if the patients continue to improve, you say, wow, that's like really cool. If some of the patients rebound and then they need more therapy, well, I mean, that's also a good outcome because then we can give them more therapy.
In phase three, what is the dosing frequency?
It's going to be monthly, day zero, then at month one, two, three, four, and five, and then week 24 is kind of the endpoint.
OK, got it. So, you're not interval pushing? Is not the goal here?
No, not at all. In fact, we have a very high dose strength throughout 100 mg/ mL. We're dosing five mg and 10 mg , and the physicians really asked us to try to push the dose high because these are very, very sick patients, and the goal isn't to go for durability. The goal is to drive the patients into disease remission, and then basically, we go out from 24-48, and we have these PRN criteria where if the disease reactivates in the pivotal, they can be retreated.
Right. Is it unreasonable to, when we look at your 5 mg, which is, again, bispecific rates, you're hitting VEGF and IL-6, so is it unreasonable, and I realize it's very crude, to think of the 5 mg you have as being the equivalent of 2.5 mg on Roche IL-6? Because you're delivering at least the same amount of IL-6, and yet you're seeing double the efficacy on the patients with above a 15-letter. So your 5 mg, which is effectively the equivalent of the IL-6 in the Roche study, if it's delivering something double, clearly VEGF is very active here. And it validates it further. And then we can decide how we interpret the 10 mg from there.
Yeah, I mean, it's a little bit of math that you want to do carefully.
Because of the proficiency on each side, et cetera.
Yeah, we have a trap antibody fusion structure. So it's about 190,000 molecular weight. So the antibody that's 150,000, and the trap's around 40,000. And so clearly at 5 mg, we have a good amount of both anti-VEGF binding and VEGF trap. And then we have a good amount of anti-IL-6 binding.
Got it. Were you expecting a dose response into 10 mg?
I mean, part of it is like with the small n that we have in APEX, people say, oh, you have a clear dose response of 2.5, five, and 10 mg.
2.5 to 5 is very clear.
Yeah, that may or may not be.
Maybe you hit a peak at five.
Right. That could be. But I think we'll find out when we have more patients. Right.
I'm sorry, just sorry, Mike. The trial design in terms of dosing versus the dosing frequency and the actual doses for Roche, what do you expect that to be?
Say that one more time.
For Roche, where do you expect them to be shaking out on the dosing frequency and the actual dose when they go to a bispecific?
When would Roche go to a bispecific in, say, a MESI equivalent indication?
Right.
I think what we've done by having dosing monthly through a 24-week endpoint. We've shown that the vision can continue to improve. We showed that in APEX. And I think if they were to run a new study, they may copy our design of doing monthly doses through the endpoint. And then after that, through 48 weeks, that's more of a safety database. So I don't think you would want to continue dosing monthly in the second six months. You want to decrease that tenor a little bit.
And remind us readout timing for phase three?
We believe that we can have full last patient in for PEAK in the first half of 2026, and that we're on track for top-line data for PEAK in around 12 months from today.
Yeah. It's interesting about the design, the PEAK and PINNACLE trials, the phase three trials, Victor. They're identically designed in terms of their dosing scheme. But they're enrolling kind of two distinct MESI subpopulations, where PEAK is kind of the sicker trial and more severe patients. And PINNACLE, it can include severe patients, but also it could include mild patients. Wait, why did you design the trials that way?
We knew we wanted to evaluate the full spectrum of MESI as a disease severity and the pivotals. We can demonstrate the benefit of 101 in patients with very severe disease and patients with milder disease, kind of as you mentioned. The simplest way to do that would be to include the full disease severity spectrum in each trial and make them identical. The thing is, then you end up having double the number of sites.
I see.
So this is a nice way where we can have one set of sites, and we can be enrolling both pivotals kind of at the same time.
Very smart. Very smart to do that.
So you know it took SANDCAT and MEERKAT quite a long time to enroll their pivotals. So we.
Just given the deep level of drying seen in the phase one APEX study, how should we think about durability after six loading doses in phase three? I know we'll find out once APEX reads out in 24 months. I assume that there's lots of heterogeneity here, just due to the nature of the disease. Any thoughts on how durable it could be?
Yeah, it's a little bit like what I said before, that our objective isn't to drive durability, but to drive the patients into remission, and then it's really a good outcome either way. If patients can stay in remission for a good amount of time, that's a great outcome. If some patients are a little bit like within the wet AMD framework, a lot of patients can do well on infrequent dosing, but actually, there's a certain proportion that need monthly dosing, so I think we'll end up with like a spectrum, and I guess we win either way, really, and we're going to find out.
Have you ever disclosed that the powering of the trials and the effect size assumptions? Is that public?
Both studies are going to be more than 90% powered using a conservative approach of treatment effect of 40% versus 15%.
15%. OK.
Yeah. But I think the point is that I guess the lesson from the Roche studies is you don't want to be underpowered. So we are enrolling more quickly than we had anticipated. So we also need a safety database for filing. So our objective really is going to be to bring more patients, both into PEAK and in particular into PINNACLE, and then to basically bring that safety database inside the primary endpoint. So that will be highly powered in both studies. But PEAK is, as you mentioned, a bit of a more severe group. So although we may and will bring more patients into PEAK, we may end up bringing substantively more patients into PINNACLE. And that's our plan.
Got it. I do want to get into some commercial questions. Just given the patient journey in MESI, it could be like two years before they're eligible for third-line treatment, where 101 is expected to be positioned. So is the 150K addressable patients a conservative number, just given that 300K patients meet the 101 trial inclusion criteria?
I wouldn't say that 101 is expected to be positioned as a third-line treatment. Interestingly, at the Roche investor relations day after their SANDCAT and MEERKAT readout, I was pleased to hear the ophthalmologists on the panel say they thought that even with the vamikibart clinical profile, despite one study hitting the stat sig and the other not, even so, she was saying if a safe and effective biologic were available, she wouldn't even use intraocular steroids anymore.
I just assumed that it would just be naturally for patients. The patients would mandate just step through off of steroids first before even considering one.
I think what they were saying is that these biologics would be first line after topical steroids. So I do think that the 150,000 number is very conservative as to addressable patients.
Got it. Got it. Any initial thoughts on pricing? I know it's too early yet. Is Tepezza a reasonable analog here if we think about pricing?
I do think these are very sick patients, and they don't have good therapy today. I think the concept of more orphan pricing is certainly a good direction. Maybe we see how we do before we talk about the pricing.
Got it. And the last maybe 30 seconds ago, KSI-501 now, this is the polymerized version bispecific. It seems like it may be an improved tarcocimab because of the IL-6, which is a good problem to have. So if that's the case, how should we think about cannibalization or erosion to tarcocimab sales if both are approved?
Yeah, well, we're very excited to see the results of the DAYBREAK study. It's a nice design where we have tarcocimab and 501 against aflibercept. If we end up showing something really special with 501 in wet AMD, as you mentioned, that's a great problem to have. And I'll look forward to solving it because we solve it ourselves.
Unfortunately, we're out of time. This has been very helpful. Thank you so much for spending time with us, Victor.
Thanks, Mike. Good to be here. Appreciate it.