Kodiak Sciences Inc. (KOD) Earnings Call Transcript & Summary

Good afternoon, and welcome to the Kodiak Sciences Phase IIb/III study in wet AMD top line results call. My name is Marion, and I will facilitate the audio portion of today's interactive broadcast. [Operator Instructions] Today's conference is being recorded. At this time, I would like to turn the conference over to John Borgeson, CFO of Kodiak. Please go ahead.

Thank you for joining Kodiak Sciences' conference call. I'm John Borgeson, Kodiak's Chief Financial Officer. Joining me today are Victor Perlroth, Chairman and CEO; and Jason Ehrlich, Chief Medical Officer and Chief Development Officer. After our prepared remarks, we will open the call for Q&A. The webcast portion of this call contains a slide presentation that we will refer to during the call. Those following along on the phone who wish to access the slide portion of this presentation may do so on the Investor Relations section of our website. An archive of this webcast will be available on our website. I would also like to remind you that remarks made on this call today include forward-looking statements about Kodiak that are subject to risks and uncertainties. All is described in more detail on this slide. A more complete description of these and other material risks can be found in Kodiak's filings with the Securities and Exchange Commission, including its 10-Q for the quarterly period ending September 30, 2021, which has been filed with the SEC. I will note Kodiak's next SEC filing, its Form 10-K for the annual period ending December 31, 2021, is expected to be filed in the near future. Kodiak does not undertake any obligation to update publicly any forward-looking statements, whether as a result of new information, future events or otherwise. Now I'm pleased to turn the call over to Victor Perlroth, our CEO. Victor?

Thank you, John. Good morning, everyone, and thank you for joining us. We are a public-focused retina company, and we're here today to discuss the results of Kodiak's Phase IIb/III study of KSI-301 in patients with neovascular wet age-related macular degeneration. I trust that many of you have read Kodiak's press release of this morning, announcing the top-line results from the study. For this webcast, and the accompanying slides, we will discuss the Phase IIb/III study data and the implications for KSI-301, our ABC Platform and for Kodiak. Of course, these data are very fresh, and there is a lot of data to be looked at, which we don't fully understand yet. Before we jump in, let's remind ourselves of what we are trying to accomplish here at Kodiak every day. We are a biotech focused only in retina. We've developed a technology for retinal medicines called our Antibody Biopolymer Conjugate, or ABC platform, which we used to design and develop new product candidates. These are biopharmaceuticals in structure and are designed, manufactured and administered in the same and/or very similar way as today's existing standard of care medicines in the retina. Today's available retinal medicines are clinically beneficial, but must be injected frequently into the eye to achieve this result. Many attempts over the past 20 years to achieve a biopharmaceutical-based approach to increase durability have failed and the opportunity for a more durable solution is substantial. Our ABC platform was designed from the ground-up to enhance durability, with the key elements to demonstrate being a clear biological effect and a strong duration of this effect, and doing these safely. There remains an opportunity to uniquely improve on the durability of currently-approved agents. If we can demonstrate these elements, we believe the potential for our ABC medicines and our ABC Platform to treat and prevent retinal diseases is tremendous. DAZZLE -- next slide -- is our randomized, double-masked, active comparator-controlled Phase IIb/III clinical trial, evaluating the efficacy, durability and safety of KSI-301, our novel Antibody Biopolymer Conjugate, in treatment-naive subjects with wet AMD. The trial randomized 559 participants, approximately 80% of whom were enrolled in the United States. The study had 2 treatment arms, KSI-301 5 mg on a flexible, long-interval regimen versus aflibercept 2 milligrams on a fixed, short-interval regimen. In this study, 3 monthly loading doses were administered to all subjects at 0, 4 and 8 weeks. Subjects on aflibercept were then treated at fixed, 2-month intervals, and subjects on KSI-301 were assessed starting 3 months after the completion of the loading phase. That is, beginning at 20 weeks. And based on predefined disease activity criteria, these KSI patients were treated every 3, 4 or 5 months. As a result, patients in the KSI-301 group did not receive dosing more frequently than every 3 months at any point in the study after the loading phase. The primary endpoint of the study was the average change in best-corrected visual acuity or BCVA score, which is a measure of the best vision that a person can achieve, when reading letters on an eye chart, including with corrections such as glasses -- the change in BCVA from baseline at year 1. For the assessment of the primary efficacy endpoint, the KSI-301 patients in all 3 groups, those dosed every 3, 4 and 5 months, were pooled together, and their BCVA was compared as a group to the aflibercept group, who were dosed every 2 months. The results show that although KSI-301 demonstrated strong durability and was safe and well tolerated, it did not meet the primary endpoint of showing non-inferior visual acuity gains for subjects dosed on extended regimens, compared to aflibercept given every 8 weeks. A prespecified secondary analysis at year 1, assessing durability, shows that 59% of patients in the KSI-301 arm, achieved 5-month dosing. And this larger group showed visual acuity gain and anatomic improvements in the eye comparable to the overall aflibercept group. KSI-301 was safe and well tolerated in the study, with no new safety signals identified. Allowing treatment with KSI-301 no more often than every 12 weeks after the loading phase for every patient has turned out to be insufficient for some patients. Nonetheless, we believe the results demonstrate a clear anti-VEGF effect, strong durability and a reassuring safety profile. We think that these data continue to support the potential of our platform, our ABC Platform, to significantly extend treatment intervals in retinal disorders in a safe and convenient manner and to do that broadly. Looking forward, our BEACON ongoing Phase III study in retinal vein occlusion will have the primary endpoint visit completed in all patients this June, with the top-line data anticipated to follow shortly thereafter. For our GLEAM and GLIMMER long-interval studies in diabetic macular edema as well as the DAYLIGHT short-interval study in wet AMD, for these Phase III studies, we expect top-line data in early 2023. As our understanding of KSI-301 and the different patient populations within retinal vascular diseases evolved, our study designs, importantly, have also evolved. One significant factor that likely contributed to this Phase IIb/III study missing its primary endpoint, the undertreatment of a strong minority of patients, is addressed in BEACON where we do proactive dosing every 8 weeks and is also addressed in GLEAM and GLIMMER, where we have tighter dynamic retreatment criteria and dose as frequently as every 8 weeks. And this element or challenge is also addressed in DAYLIGHT, in which all patients are proactively treated on an every 4-week regimen versus, in this case, aflibercept on its every 8-weeks regimen. But I don't want to get ahead of ourselves, so at this point, I will turn it over to Jason, who will walk us through in some detail the specifics of our Phase IIb/III study design, the data. And we will begin together to discuss their implications. Jason?

Thank you, Victor, and good morning, everyone. Before getting into the study and the results, I wanted to take a moment to thank everyone who contributed, our clinical investigators, the whole Kodiak team, our partners and most importantly, the patients and family members, who are participating in the study as well as the broader portfolio of clinical studies of KSI-301. We just unmasked the initial data from the study, so I'm sure there will be further insights to come. And we are sharing the important top-line results of the study with you today, including some insightful details that go beyond what is often presented for these kind of studies. Victor already provided you with an overview of the study design. As a reminder, the study enrolled patients with treatment-naive neovascular or wet AMD. The inclusion criteria for typical studies in the field have included patients with baseline best-corrected visual acuity, or BCVA, of 25 to 80 letters on the standard ETDRS eye chart for a vision of approximately [ 20/25 to 20/20 ] on the Snellen eye chart you may be familiar with from your own visits to the eye doctor. Patients were randomized 1:1 to KSI-301 or aflibercept. Patients who received KSI-301, received 3 additional months in loading doses and then were assessed for disease activity beginning 12 weeks after the last loading dose after week 20 visit. By default, KSI-301 subjects are assigned to every 20-weeks or every 5-month dosing. If the disease activity criteria are met sooner, then those patients drop down to receive treatment every 12 or every 16 weeks. As you can see from the orange boxes in the slide, there are several visits throughout the study at first year, at which the dosing interval can be reduced from every 20 weeks to more often, every 12 or every 16 weeks. But dosing more often than every 12 weeks was not allowed in the study by protocol. Aflibercept patients, the active comparator, were dosed every 8 weeks after the 3 initial monthly loading doses. To match the study, all patients visit the clinic and are assessed monthly, and they received sham or pretend injections on visits where active injections are not required. The physicians in the clinic team who evaluate safety and efficacy are all masked, and they did not know what treatment a patient is randomized to receive. Only a separate physician, who performs the eye injection is unmasked. That's standard for studies in the field. The primary efficacy endpoint of the study is the main change in vision from baseline in year 1, defined as the average of the weeks 48 and 52 visits, as you can see on the chart. The primary efficacy endpoint is evaluated by standard MMRM statistical model, which is designed to impute missing data and adjust for baseline differences. The non-inferiority margin for the study is set at 4 eye chart letters. This means that the lower bound of the 95% confidence interval of the difference between the treatments cannot cross minus 4. Let's turn to the next slide, Slide 5, subject dispositions. From 785 patients screened, 559 were randomized and 557 treated. So the analysis -- that included 277 patients in the KSI-301 arm and 280 in the aflibercept arm. In this study, more patients in the KSI-301 group, 14%, discontinued from the study treatment, compared to the aflibercept group, 9%. Most common reasons for early study treatment discontinuation in the KSI-301 arm were related to disease progression, 4%, or ocular adverse events. Most of those were [ signs of ] disease progression, 4.7%. In the aflibercept group, the most common reasons for early discontinuation were withdrawal of consent, 3.6%; and death, at 2.9%. As Victor mentioned earlier, we believe that undertreatment resulting from not allowing dosing more often than every 12 weeks had a negative impact on the study. And you can see this coming through in the number of KSI-301 patients who discontinued treatment early, relative to the aflibercept group and the reasons for continuation. Let's turn to Slide 6. The baseline demographics of the enrolled patients in the study were well balanced. Our study enrolled patients from both the United States and Europe, and over 80% of the patients who enrolled were from the United States. Looking to Slide 7. The baseline ocular demographics. These were also well balanced and typical of patients with newly-diagnosed wet AMD. The baseline BCVA was on average 63.6 eye chart letters, which is higher than in other recent studies. Nearly 90% of the patients had vision of better than 50 eye chart letters at Phase I. The optical coherence tomography, or OCT, central subfield thickness, measured in microns from the internal limiting membrane to the retinal pigment epithelium, was 350 to 360 microns between the groups, on average. Other OCT characteristics and baseline IOP were also well balanced, as you can see here. Turning to the next slide, Slide 8, here is the average change from baseline in BCVA and CST as adjusted by the MMRM statistical model. As you can see, there's an initial improvement in BCVA in the KSI-301 arms for 12 weeks and then a diminishment of effect over time. A tough graph to see and not what we had wanted or anticipated. The aflibercept group did exceedingly well, with a 7 letter improvement at year 1, averaged over weeks 48 and 52, actually the highest final BCVA for aflibercept group in a large, randomized study that I'm aware of. The study, unfortunately, did not meet its primary efficacy objective of demonstrating non-inferiority to BCVA of KSI-301 from aflibercept, even though the majority of KSI-301-treated patients did achieve durable vision gains, as we'll see in the next slides. We believe this is in large part due to the impact of undertreatment in some of the patients, because under the protocol, they were not able to receive KSI-301 treatment more often than once every 12 weeks. You can see on the right side, the change in OCT central subfield thickness over time. We'll look at this in more detail in subsequent slides. Now you can see that there's a separation of the graphs in the BCVA and OCT curves early on in the study. This difference could be in part due to random chance or differences in patients that didn't randomize out as [ 1.5 to 2 better ] differences have been seen among groups treated identically in randomized studies with aflibercept before. Or another possibility, in part, this could be due to differences in the biological effects of the drugs through the first 3 doses in some of these wet AMD patients. Our ongoing study of monthly KSI-301 in patients with wet AMD may help address this question. On the next slide, Slide 9, you can see that over 59% of the patients completing year 1 in the KSI-301 arm achieved 5-month dosage, with robust and durable visual acuity gains and anatomic improvements, comparable to the aflibercept group. This chart looks at the best-corrected vision. So the study dosing interval selection criteria achieved the important goal of identifying the majority of the patients who could do well with every 5-month dosing in KSI-301, which is, I believe, an important accomplishment. These patients had an average vision of nearly 70 letters or 20/20 vision at year 1. Turning to the next slide here, focusing on the 30% of patients who are assigned to Q12-week dosing. I would note that most of these patients were assigned to the Q12-week dosing regimen at the first available disease activity assessment opportunity at week 20 in the study. Here, you can see how these patients, on average, were the worst performers, and they dragged down the results for the KSI-301 population as a whole. We believe that these patients could have benefited from more frequent dosing than what was allowed under the protocol, and this is, I think, also supported by the OCT data, which we'll review next. So turning to the next slide here, focusing again on the majority of patients who were on the Q20-week dosing interval and looking at the OCT/CST values. These patients on 20-week dosing with KSI-301 had OCT outcomes comparable to the aflibercept patients, with both groups getting patients on average to normal retinal thickness levels. For the most part, you can see the 95% confidence intervals overlap here. You can also see the nice stability in the OCT curve for the KSI-301 group, with little evidence of rebound for OCT [ flammer ]. On the next slide, let's again focus on the Q12-week patients. You can see how these patients had an initial improvement in OCT and by 12 weeks were at a similar level to the Q20-week group. But the Q12-week group had fairly quick recurrence of disease activity, which then got worse between weeks 16 and 20. The additional doses given at week 20 and only every 12 weeks thereafter were not able to control these patients' disease effectively at that point. We believe that if this group of patients were given more frequent therapy, their vision and anatomic results may have been better. So that's our overall vision and anatomic results, with an important look at some of the subsets to help us understand a bit more where our study design and treatment regimens performed well and where they fell short. Let's take a look at our safety, turning to the next slide, Slide 13, which is a summary of treatment-emergent adverse events. Overall, treatment of KSI-301 was safe and well tolerated. The main study eye treatment-emergent AEs were somewhat higher in the KSI-301 arm, although we would say much of that difference could be attributed to the higher rate of AEs related to disease progression for the KSI-301 group, as I had mentioned before with respect to discontinuations. There were 6 SAEs, or serious adverse events, in the KSI-301 arm. One was a procedure-related endophthalmitis; one was a retinal detachment that was not procedure or drug related; one was a cataract; and 3 of which were related to worsening of wet AMD. Systemic safety was comparable between the groups, and there were no new safety signals observed in the KSI-301 arm. On the next slide, Slide 14, with respect to intraocular inflammation, the rate of KSI-301 was 3.2%. This is within the range that has been reported for other aflibercept studies in wet AMD lately, which has ranged from 1% to 4.5%. In our study, the rate of intraocular inflammation in the aflibercept group was 0. Of the cases reported for KSI-301, in all cases, the clinical findings resolved, and no cases of intraocular inflammation with retinal RVO, or retinal vein occlusion, were observed. We believe that these safety data, coupled with the safety observations across the ongoing masked Phase III studies, continue to suggest the favorable ocular safety profile for KSI-301. Now turning to the final data slide, just a summary of what we discussed so far. I would also note that we expect the full data from the study to be presented at an upcoming ophthalmology conference, and there's more to learn from the study, as we analyze it in more detail. With respect to the study's primary endpoint, unfortunately, non-inferiority of mean change from baseline BCVA for treatment-naive wet AMD patients treated with every 12- and 20-week KSI-301 versus patients treated with aflibercept Q8-weeks was not demonstrated. We believe, this is, in large part, due to the impact of undertreatment in some patients because treatment with KSI-301 more often than every 12 weeks was not allowed in the protocol. We did observe robust durability at year 1 with nearly 60% of patients on Q20-week dosing. These patients on Q20-week dosing achieved meaningful reductions in CST and improvements in vision comparable to aflibercept Q8-weeks. KSI-301 was safe and well tolerated, intraocular inflammation event rates were 3.2% of KSI-301 patients. No cases of IOI with retinal artery or retinal vein occlusion were observed. There was a higher overall rate of AEs and discontinuations due to AEs in the KSI-301 arm, at least in part due to undertreatment in patients who may have needed treatment more often than 12 weeks, allowed in the protocol. With that, I'll turn the call back to Victor.

Thank you, Jason. In conclusion, while we are highly disappointed that DAZZLE did not reach its primary endpoint, we do feel there are valuable insights into the potential for KSI-301 and the value of the ABC platform more broadly in the treatment of retinal disorders. For KSI-301, the ability to take the majority, the true majority of the patients with wet AMD to a far more convenient every 20 weeks -- that's 5-month dosing interval -- without sacrificing benefits in visual acuity, when compared to aflibercept dosed on its every short week -- a short interval every 8-week regimen. We believe this is a compelling proposition to patients and the medical community with KSI-301. Even more so, when considered in the context of real-world outcomes for these patients, where there is extensive evidence demonstrating that patients struggle to achieve frequent dosing, and as a result, often, do not achieve the same level of benefit seen in clinical trials. Furthermore, KSI-301 is continuing to demonstrate a well-tolerated safety profile, as clinical experience with the drug is expanding on a daily and monthly basis. So what is the regulatory path from here? Actually, our regulatory plan is still largely intact, with 4 additional Phase III studies reading out over the next year. GLEAM and GLIMMER are our 2 Phase III studies in diabetic macular edema. They are long interval studies, which also allow for up to every 8-week dosing for patients in these more intensive anti-VEGF therapy. And importantly, these 2 studies, GLEAM and GLIMMER, have always been and are designed to serve as our primary approval set. The BEACON study in patients with retinal vein occlusion also uses a more intensive dosing than KSI-301 -- every 8 weeks -- and is intended to serve as the basis for approval of KSI-301 in that indication, on top of GLEAM and GLIMMER. In wet AMD, we believe that the DAYLIGHT study will clarify the ability of KSI-301 when dosed aggressively to effectively manage a group of patients with wet AMD, similar to those who had poorer outcomes in the DAZZLE study. If successful, DAYLIGHT can form the basis of approval in wet AMD, albeit for monthly dosing. We believe, over time, we will be able to generate data to expand the label to include longer-interval dosing in wet AMD, which we believe physicians may be interested in exploring given the DAZZLE results. Over the coming weeks, as we continue to analyze DAZZLE further, we will review and consider if there are any changes to these ongoing study protocols that we should make in light of the results reviewed today. On our ABC platform more broadly, we're excited to be bringing our second product, KSI-501 into the clinic and plan to file its IND later this year. We are also continuing to make progress on KSI-601 in our triplet molecules, which we believe have the potential to add additional value to our ABC platform. And we look forward, of course, to updating you on these programs, as they progress. We are fortunate to be in a strong financial position to date, with over $731.5 million in the bank as of December 31, 2021. This can be expected to fund the company for more than 2 years, based on our current plans, and provides ample resources to weather any near-term volatility. In closing, despite today's setback, we remain excited about the potential for KSI-301 and our ABC platform to deliver significant value for patients, physicians and payers, who are still struggling to meet the challenges of retinal vascular disorders. We appreciate the support of our shareholders and are sorry for the disappointing top-line results we have announced today. We remain committed to doing our best to deliver value to our shareholders over time. On behalf of all of us at Kodiak, we'd like to thank the patients and their caregivers as well as the many physicians and staff in the United States and abroad, who participated in this important study. And we will now open the floor to analysts for questions. Operator?

[Operator Instructions] We will now take the first question from Umer Raffat from Evercore.

I'm still processing the data and trying to internalize your observations on more intensified dosing on the subset of patients. And I guess, where I'm a little confused is the green line, the every 16-week arm -- and presumably, if the blue line, the every 20 weeks in patients that are stable are tracking just like aflibercept, and that makes sense, but then why is the green line where it is? I mean, presumably, they could have gone through a more intensified dosing to take a step down, and their BCVAs should have tracked more or less akin to aflibercept, had they taken a step down? So I guess, how should we interpret that? I think, that confuses me, one. And also on efficacy, I don't know if there's a clear explanation for it. But in the every 20 weeks subgroup on BCVA, there's a dip from week 44 to 52. It looks like a couple of points or so. And just trying to understand what that is. And then on safety, I'm trying to think about how the profile relative to EYLEA would evolve, as we go to a more intensified dosing. And also, was there any vision loss in this existing trial?

I'm sorry, Umer, on the last part, just to make sure we got the question properly.

Yes. Was there any vision loss in the patients that did have the 3.1% of inflammation?

Well, thanks, Umer, for your questions. We understand that it's important to process the data, and we went through that same process and also found a lot of, let's say, comfort in the strength of the Q20-week group, the 60% of the true majority of the patients and how well they did. And their top level vision doing really as well in [ what's a computer's ] analysis, both for the aflibercept group and also the KSI-301 group in Slide 9. As to your question about the green line, which is really the Q16W patients, I think the most important element there, Umer, is that it's really a very small number of patients. And as a result, I think, what's important to remember is that wet AMD, it really is highly variable disease. And so you really have a fairly small number of patients in the group. And that's really, I think, the most important element. I think, we really think at a high level, Umer, when we want to understand and process the visual acuity data, it's more useful to take a look at the larger numbers. So that's why when you look at the Q20-week group and you look at the strong data, you can get a lot of comfort. And in some sense the ABC platform and KSI-301 takes us to what I sort of call like the promised land of long-interval dosing, true durability with the biologic and that the orange line helps us to really understand what happened in the study. And together with the orange line, one can supplement that in our own mind in pulling, maybe half or slightly more, in terms of the adverse events relating to progression of disease, right, which from an MMRM or sort of the last observation carried forward equivalent, right, is another force that's pulling only the visual acuity down. So on the high end, you have the real strength of the Q20-week patients and the durability of that effect and the strong vision gains that we see. The Q16-week group really just represents a small number of patients, okay? And it's difficult to interpret, okay? The orange line shows us what happened in the study. And then when you supplement that with the discontinued patients, right, some of which related also to lack of efficacy. In some sense, the patients, when they're overextended, right, basically fall down with prodigious [ launches ]. And those forces really destroyed DAZZLE. Okay. So on the efficacy, perhaps -- Jason, I think Umer is asking, when you look at the last 2 data points, right, for the Q20 week and you look at the 48- and 52-week data points, I think there's an impression, let's say, that the visual acuity of that group is maybe decreasing. And/or I think the higher-level question is, well, is the effect right in this Q5-month dosing, right? Is it sort of falling off? Or based on the fatality of the data that we see, that we have, right, how strong do we think the benefit is and how durable and how robust is KSI-301 in the strong majority of the KSI-301 arm patients?

Yes. So Umer, I mean, I think, overall, those results, through the end of the first year of the study, I'd say, are pretty stable. I don't think the small differences that you're seeing at the -- say, between week 48 and 52, are representative of some declination in their trajectory, right? It's within like a letter. So there's always like some wobble month-to-month of visual acuity. What's important, these patients are reaching on average of 20/40 vision. And when we look a little bit further into it, right, like the OCT is stable in that patient population. And yes, so I mean, obviously, there's -- ultimately, we would need more data to come from -- for the second year of the study, so we can see what happens with those patients. But I think, overall, it, basically, can be interpreted as stable data.

I think his last question, Jason, was a little bit of a question around the IOI or intraocular inflammation. So first of all, we were looking at the masked data at a high level, in terms of IOI and felt, over time, very comfortable. We were very surprised to see what a clean study, we have run from the EYLEA aflibercept standpoint, with 0 cases for EYLEA. So congratulations there. I think we also felt very comfortable with the types of IOI that we did have in DAZZLE. I think Jason can provide a little bit more context.

Yes, sure. I'd say with respect to the intraocular inflammation cases, you can see by the -- [ in terms of the -- ] the terms that were reported -- they're really, for the most part, these were mild or moderate. For instance, like a [ trace to half plus ] [indiscernible] [ cells ] with several of them. And mostly, these resolved without treatment or with just a topical treatment. The IOI patients did not lose vision in a way -- in anything [ what ] was associated with the IOI events. But obviously, all these patients also have wet AMD and dry AMD underneath. But the inflammation, again, that we saw, didn't result in vision loss. And all of the IOI cases, those clinical-findings resolved, with no vascular occlusions or anything associated with those.

So hopefully, that's a full set of answers to your very good questions.

We will now take the next question from Michael Yee from Jefferies.

Thanks, Victor and Jason, for the data. Two-part question, and I think they're pretty simple. I think what you're seeing here is that a portion of patients need dosing more than 3 months. Would you plan to rerun another study? What would be the next step at this point? Could you redesign another study to fulfill an AMD-type indication? Talk about that. And then second question is following up on IOI. Can you tell us if the 3-month dosing arm had higher inflammation than 3%? Or give us some comfort that all of the dosing arms for KSI were essentially balanced on inflammation.

Thanks, Michael. That's a really good question. What are our plans in wet AMD as a result of these data? And what's the regulatory path for wet AMD? I think, as I mentioned at the end of my comments, we actually are pleased in retrospect with our regulatory plan in the portfolio of studies that we're running. Now we didn't have visibility into the forces in DAZZLE. However, in some way, we were already looking around the corner in the portfolio or suite in the designs of these other studies. I will say that DAZZLE was designed when we didn't have the full suite of the Phase Ib data, whereas the other studies here benefited. And as we mentioned, in all of these studies, DAYLIGHT, BEACON, GLEAM, GLIMMER and even GLOW, we realized that when we want to bring all the patients to the long interval, that's really not credible. And in fact, it's really not necessary or needed in the community, right? As we look at this portfolio of studies, the objective is to get the drug approved. I think, now, with the powerful durability that we've shown, it's really a signal. So we've had the opportunity very quickly to have some important conversations already in the community, also with our IDMC, okay, the Safety Monitoring Committee as well, who recommends continuing with all of our studies in the same manner that they're currently structured to protocol. So that's very powerful. We also had the opportunity to have a preliminary discussion with the FDA around our regulatory plan [ until they review ] the data, which I think is also important for your question, Michael. So as I mentioned, what is the regulatory strategy? And important in your question, well, what's the future in wet AMD? Well, an important element is to run DAYLIGHT, which we're doing. BEACON, GLEAM and GLIMMER are fully enrolled, right? As we mentioned, BEACON's last visit will probably be in June, and we'll be able to turn that more quickly than we did through DAZZLE, which was around 3 months. BEACON, we'll be able to turn the crank on that more quickly, I think, from an operational standpoint. So we're actually looking right in the face of BEACON's data. GLEAM and GLIMMER, as we've mentioned, as we've announced, right, the enrollment is complete, and we're pleased, overall, with the protocol -- I think both BEACON and GLEAM and GLIMMER and in terms of pulling the dosing more frequently. And in DAYLIGHT, we've essentially maximized and created a fairly aggressive study design, right? It's KSI-301 every 4 weeks actually versus aflibercept, actually every 8 weeks. I think, from a standpoint of really trying to understand and to maximize the probability of success or approval of KSI-301 in wet AMD, I think DAYLIGHT is going to provide a very important data and our regulatory strategy of the 2 primary studies in a major indication i.e., GLEAM and GLIMMER in DME, provide the primary path for approval for the medicine and singular studies, let's say, in RVO with BEACON and let's say, DAYLIGHT for wet AMD as a package to provide that path. So then your question, Michael, is, well, how useful is a label that provides for a monthly dose? And I think the physician community, they know and their job is once the drug is approved, it's how to explore it, right, and how to use it in the patient population more broadly. So I guess I would say that going to the next level with your important question, right, is KSI-301 a first-line agent in wet AMD or -- which means all patients are started on it, right? And then the objective is through a monthly loading phase and a stabilization, then a shift of a true majority of those patients to that promised land of long-interval dosing. That's still, as a dream, is real and is supported in the data from DAZZLE and will be further cemented right in DAYLIGHT, right? That early phase in stabilization, maximally, with the knowledge, right, of what we've learned from DAZZLE. Do we need to run another different type of long-interval dosing study for KSI-301 in wet AMD? We do think that would be helpful. We actually don't think at this point that, that's a requirement. And especially following on other long-interval dosing data that we'll be able to be generating in other studies and other indications. So another possibility is -- and we've had the chance to speak with some physicians in the field, even in the context of the DAZZLE data and, let's say, an approval on the basis of DAYLIGHT, right? The idea that well, they would still want to put their patients on KSI-301 actually in wet AMD in a proactive basis, right, through a loading phase to try to see whether, in fact, those patients can get to the promised land of that long-interval dosing, which is important for patients. So that's a bit of a long answer to your very important question. I think in some ways, we're already beginning and are in the process of generating the data that will help answer these questions. And as we think a little bit more about our regulatory strategy about well, what other studies could be designed and run in wet AMD that will provide -- and let's say, how do we get a bit closer to what we were able to demonstrate in the Phase Ib study in wet AMD, right, where we allow monthly dosing all the way through Q6-month dosing and also allowed for physicians to dose at any visit, at any time, when they felt it was important for the patient. And that type of stabilization of long-interval dosing in some patients or that sort of induction in the early power, presumably, was also important. So that's a long answer to your important question, Michael.

That's very good. Can you comment on the 3-month IOI? That would give us some comfort as well, that the more frequent dosing does not necessarily lead to more IOI.

Yes, Mike. So with respect to the IOI, I think just a couple of points, right? So by design in the study, right, patients in the 20-week dosing will be scheduled to get 5 injections and the 12-week is like, I believe, 6 at year 1. So I wouldn't really expect any differences, in terms of IOI rate from the different dosing regimens. Don't have like a specific breakdown by dosing group, but looking quickly at the sort of the pattern and the timing of the data, I wouldn't say that there's any -- no obvious correlation there, in terms of the -- the timing of those events' occurrence were very scattered throughout the study. But we'll look at that more closely. I wouldn't expect, like I said, really any fundamental differences. I think what's really important, right, is the overall rate was low, the cases resolved and there wasn't any vision loss due to the IOIs. But of course, we'll look further into it.

We will now take the next question from Anupam Rama from JPMorgan.

Sorry about the disappointing outcome here. I know you've put maybe running another study in wet AMD on the table here. So does that mean that you won't be continuing with the second year of the DAZZLE study? I think in the second year of the DAZZLE study, you could move up or down in treatment interval. I think there was a portion of EYLEA patients that could go to KSI-301 every 2 months in the second year. Are there changes to the protocol you might be making in the second year? How do we think about that?

Thank you, Anupam. That's a great question and important also for our clinical investigator community. We haven't had a chance to do the [ broadening ] reach out. And I guess we need to speak with them and engage them, kind of, in a collaborative set of decisions around the second year of DAZZLE, given these results. I think the data that we do have from DAZZLE, really important data already. And we also recognize that the Q12-week group isn't doing as well as we would have liked. And so I think the simple answer is that we don't know yet, Anupam, what's the future there. The key is really to be pivoting to the set of ongoing studies, where there's an important future and that there's a regulatory aspect and an understanding that are critical. And from the standpoint of DAZZLE, I think we've learned quite a lot there and, I think, provide a very strong validation, right, through the long-interval dosing element and to the safety of the medicine, and to the clear anti-VEGF effect can be durable, and we see that in the data. So I think, that's the answer, is, we'll be circling the wagons together with the clinical community for DAZZLE, and we'll have to make some choices.

Yes. I mean, I would just offer maybe note, but of course, since the study took whatever period of time to recruit, right? There's actually a reasonable amount of post-year-1 data that's accumulated in many of the patients already, right? So regardless of what the ultimate set of decisions are for DAZZLE, there will be additional results from that second year, even if they're partial results that will be available potentially in the future.

We will now take the next question from Matthew Harrison from Morgan Stanley.

Great. I guess 2 for me. So first one, I don't know if you've looked at this, but can you just talk about IOI in terms of whether or not there was more concentration in some of the sham injections as opposed to the 301 injection? And then separately, I think there was about a 4% rate of discontinuations due to ocular AEs for KSI, compared to none for EYLEA. Could you just comment on what those were? And then finally, I guess -- sorry, there are 3. And then finally, just -- it sounds like you've addressed this, but is there a path to get the 5-month data on a label? Or you think it would only be through sort of [ med affairs ] that you would be able to discuss that data again, if you were successful with DAYLIGHT?

I'll take the third one,[indiscernible] to your questions on [indiscernible]. I don't think there is a path from DAZZLE for the 5 months. I think, there's a strong confidence from the community, right, and there'll be more discussion in the community, as people dig in and better understand the power of that durability. We were very pleased, as Jason mentioned, that the criteria that we used to identify this group by [ adjacent at 20-week, ] which was a 60% per the algorithm or a true majority of the KSI group. So from there -- and not a lot of variability in that group and very strong performance from that group, actually, from the beginning through the end of the study there. So there is a path for us to think about what kind of study we could run in wet AMD, more broadly. And we've initiated some discussions and some thinking there. And so we do believe that there will be a utility, but we're going to need some time to figure that out. In either case, I think, it's not unattractive, even with DAYLIGHT. But we certainly would be committed to exploring study designs that don't kind of bias against our ability to be approved only in the context of like long-interval dose, right, against EYLEA, which is a powerful competitor. Of course, it's dosed right on a shorter-interval label. So in some sense, there's complexity there, but we would be committed to spending the capital, once we have a little bit more information out of the Phase III program for KSI-301. So that's the short answer there. Maybe that's the long answer.

So Matthew, with respect to your second question on the AEs related to discontinuation, I'd say, most of those in the KSI-301 arm were related to undertreatment, right, or adverse events that you expect with undertreatment. The -- your first point around the IOI and the sham injections, I'm not sure if I totally understood the question, can you clarify?

Sorry, I was just asking if you knew the proportion of IOIs that may have occurred post the sham injection versus post the KSI-301 injection? And if there was some more concentration in sham injections for some reason? I didn't -- I was just trying to get some detail on that. That's all.

Right. Yes. So I guess we [indiscernible] specifically looked. But just the sham injections, remember, it's not an actual intravitreal injection. So the eye is anesthetized with some little injection of subconjunctival lidocaine around the ocular surface, but there's no actual intravitreal injection. I would suspect not.

[Operator Instructions] We will now take our next question from Gena Wang from Barclays.

Sorry. I have 3 questions. So maybe I will cut down in one question with 2 parts. So the question is regarding treatment interval. For the 14% patients who discontinued, how would they categorized as a treatment interval? And also for the 59.4% 5-month dosing, what is the denominator?

Yes. So Gena, I think we analyzed that -- the durability at 1 year similar or consistent with how that's been done recently in the field for other studies. So the interval that we described, say, the 59% of the people who completed 1 year. So the denominator is the randomized set minus the number of people who discontinued, if that makes sense. And that's the same, as far as I understand it from the publications, for how this has been done with other recent studies that evaluate different durability groups.

We will now take the next question from Ellie Merle from UBS.

Just before we jump into the question. I think, Jason, Gena, also is asking a little bit about the discontinuations and how they feed from the different treatment intervals. And I think that the data show actually that the Q20-week has very few discontinuations. So I think, one of the things that people asked about from the Phase Ib studies was well, when we take people to, say, 5 or 6 months, are we overextending the patients and then, therefore, they may have a bad outcome more than [indiscernible]. And actually, in the data that we see, that's not the case. We have robust and strong and durable visual acuity and OCT situation in these Q20-week patients. So that's a high-level answer to your question. Of course, it's a complete analysis in the context of the durability when you get to 59%. As we mentioned earlier, it doesn't really matter how you want to slice and dice the numerator or the denominator for the durability. It's a true majority of the patients and for the algorithms in the field it's 60%. And in any case, whether you want to do the [ start ] of the population, it's a true majority. And I think what's important is that the group does extremely well and through that group, there are patients that at any important level, who are discontinuing because they're doing so well and essentially hitting that top level of vision and having an important and powerful resolution of the fluid as well, getting to a normal dry retina in that case. Sorry to interrupt, Ellie.

No worries. Just to follow up on the -- I think in the prepared remarks, you mentioned potential protocol changes for the ongoing trials. I guess, if you could just elaborate on some of the aspects that could be modified at this point. Or what you'd be considering? And then just -- I think you said you had a preliminary meeting already with the FDA. Just anything discussed there, in terms of potential protocol changes? And then just a clarification, just on your comments on the IOI rate. I just want to make sure I'm understanding your comments. So are you saying there was no difference seen between -- in IOI across the different dosing intervals? Or we just don't have the data yet, and you're still running the analysis? I just want to make sure I understand your comments correctly.

We couldn't hear all the questions, I think your -- [ this ] -- very well because of the connection. I think you were asking, was there a difference between the different dose regimens, in terms of the IOI. I think the important thing, right, I mean, with a number of the cases, really, you can trace levels of inflammation in the overall number of the cases being low, that is -- it's not that helpful to try and allocate them or to say when there's more injections there's going to be more IOI. That doesn't need to be supported by the data. And that's also not supported by the broader visibility, right, that we have across all the other studies that are ongoing, where we've given, I believe, many thousands of injections. So that's one answer. Your other question, I think, was, are there any changes to the protocols for the other studies that we're thinking about today or based on discussions with, say, [ IBM, ] [ CD ] or FDA or [indiscernible]? And at this point, all this feedback has been to continue running the studies as we have them planned. At the same time, our commitment is to take the information from DAZZLE, right, which we have announced today and also our deeper analysis into the data to then pause and then to quickly get -- if there's a utility to make any and all changes in the remaining portfolio of ongoing studies, to maximize the success of achieving the primary endpoints in all of those studies and getting the drug, therefore, up onto the market. So you can [ write ] the study is very close to the end. So we're -- always have felt and continue to feel the important power of the medicine in RVO and so for BEACON, that's really how much stuff we want to do there. For GLEAM and GLIMMER, we're going to think. We certainly have time there, if we wanted to make any changes, and we're going to take a detailed look and make some choices there. And for DAYLIGHT, there's no changes that we would think about because it actually turned out to be a well-designed study to address some of the questions that have emerged from DAZZLE and by adding it into the portfolio of our regulatory strategy on top of the 2 primary studies, right, with DME turns out to be nice. So we don't have any expectations to make changes, but we're going to explore all of our studies, right, so that we don't have to have this conversation with anybody ever again.

We will now take the next question from Andrea Tan from Goldman Sachs.

I'm curious, when you look at the baseline characteristics of the patients. For those who failed to achieve benefit at the 3-month dosing interval, was there anything that could have indicated a priori that they would need more frequent dosing?

Yes, it's a good question, Andrea. And we started to look at that question, and I would say the short answer is, no. There aren't any obvious baseline characteristics. We'll continue to explore that, as we look into the data further. But I'd say that, sort of, historically in the field, there's not really been very good correlations between the need for treatment with any of the anti-VEGFs and any particular set of baseline characteristics. So often, it's that empiric determination in how the patients are responding to the initial courses of treatment and your ability to extend the different treatment regimens, that is the best way of figuring it out.

I mean, in the past, over the past few years, with some of the different study designs that allowed patients to be shifted to a slightly longer intervals, HAWK, HARRIER, for example, there were large efforts to do an AI to try and explore this question. And initially, I think people felt that, that effort would be promising. And then digging deeper and trying to go to the flatter of the curve of excellence, all of that really fell apart. So I think that's really to be determined. I think, what's -- and that's a useful question. I think, then flipping it into what's more important is, well, the patients that didn't follow the algorithm, got to the Q20 weeks, and they stayed there, and they did extremely well, and that's a majority of patients. In my view, there's not an unmet need for patients to be dosed every month or every other month. There are a number of agents that can be used. But the one thing that hasn't been shown, is what we've now shown in DAZZLE, I think, definitively, in a disappointing manner, right for the success of the study. But nonetheless, we have a set of criteria that took patients to that special place of long-interval dosing and that's special, and we need, as a company, to really build from that. And then I think we'll continue to showcase the safety with the imbalance that we have here on the IOI, compared to 0, they should [ have seen ] bigger. But in reality, we all know that our biggest IOI is not 0. So I think we're also in very good shape there.

We will now take the next question from ROTH Capital Partners.

Thanks for the really helpful update. I think the first one for me is just that I know you mentioned that more frequent dosing would have helped with the poor responders, and I can agree with that, since you could probably get closer to the efficacy seen with EYLEA. But what are your thoughts that these patients may need something more than an anti-VEGF? And then can you comment on more detail, at least, on where you are with KSI-501? And any comments on how you kind of plan to plug that into the treatment paradigm? And then the last 2 are just clarifications. I think the first clarification that I wanted to kind of figure out here, is that DAZZLE is going to be included in the BLA package? And if it's not, how do you think the FDA will view that? Because I thought all of these studies were being looked at as a whole, since typically 2 studies will be needed, right, for each indication? And then the last bit here, again, is just a clarification on whether the DAYLIGHT study, more frequent dosing is going to be only for wet AMD? Or do you plan to leverage that again across the other indications?

Great. Well, thanks a lot for the good questions, actually, and really the lead in to the KSI-501. So actually, we are always excited about the molecule. We did shift the manufacturing on a little bit, as we expanded the portfolio of KSI-301 studies. With COVID and everything, we had to make some choices on manufacturing slots. So we slotted in some additional KSI-301. But KSI-501, [ save from the ] [indiscernible] [ manufactured ] successful release of the antibody batch and moving into the drug substance bioconjugation very shortly. And we're hopeful that we'll be able to file the IND, let's say, in the September time frame. So 501 is a bispecific-type conjugate. And actually, the anti-VEGF portion is a trap, binding the IL-6 portion as an antibody. So it will be quite interesting, even from that standpoint. And to your point about, well, are there other mechanisms, right? And why is it that some patients need more intensive therapy? I think what's important about KSI-301 is, there are patients that through that initial phase, let's say, have 700-plus micron improvements in the OCT, KSI-301 can be very powerful in patients. But let's say, in other patients, to your point, additional mechanisms are needed, the KSI-501, right, with its potent anti-IL-6 as a conjugate will really help to answer that question. And I think, one of the challenges that we had, as we thought about KSI-501 as well, where should it live in our pipeline. When we think about KSI-301, whose job would be the perfect thing for everybody. I think in the context of KSI-501, a nice place for us to think about it is to think about the power, right, a special new type of power, maybe, in very difficult wet AMD patients and then also to be able to bring other conjugate elements. So I think, we do begin to see the vision of KSI-501 in that program, is something that we're excited about. Your next question is, well, what's the place for DAZZLE in the regulatory package. I think our regulatory strategy always has been to have 2 studies in the primary indication, okay? And then to have supporting [ single-phased ] studies beyond that. So we don't speak for FDA, we'll have to explore the full data package right at the time of the BLA submission. There's some good wood to chop between now and then. Then your third question is to, well, are we planning to leverage DAYLIGHT in different ways. And if you remember, the key aspect of DAYLIGHT was really to be able to provide monthly dosing and safety and exposure data done monthly, and to leverage that across the portfolio of indications. And so we believe that the ability to do that is still alive, and we believe it will be relevant for all of the retinal vascular disease indications. So thanks for those questions. And with that, I think we'll wrap it up. One more question.

We will take the last question from Neena Bitritto-Garg from Citi.

Just a question about the BCVA gains. On the slide where you show the gains for each of the dosing intervals that -- it shows that the Q12-week arm did have a gain in the first -- after the loading doses of about 6.5 letters or so. But when you look at the total BCVA figure on Slide 8, it looks like the max gain after the loading doses for the full group was really closer to 4. So I guess, can you talk a little bit about that? Was that due to significant underperformance in the patients who discontinued?

Jason, do you want to address that? I think it's an earlier question on the total mean [ of those 2 ] individual cohorts.

Yes, right, I mean, I think the -- of course, the total mean -- you see the initial improvement from baseline there, right? That includes -- the MMRM model, that includes like all of the patients, right? So that's maybe part of it and how the model adjusts differently for patients with different baseline visions and so on.

Okay. Well, thanks, everybody, for participating, and I hope that we were able to provide a lot of context. And hopefully, the answers to the question offered some additional context. We appreciate your support.

Thank you. That will conclude today's conference call. Thank you for your participation, ladies and gentlemen. You may now disconnect.