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

Good afternoon, ladies and gentlemen, and welcome to the Kodiak Sciences Third Quarter R&D Webinar. My name is Keith, and I will facilitate the audio portion of today's interactive broadcast. [Operator Instructions] Thank you. At this time, I'd like to turn the conference over to John Borgeson, CFO of Kodiak. Please go ahead.

Thank you for joining Kodiak Sciences Third Quarter R&D Webinar. 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 presentation discussing recent data from our ongoing Phase Ib study of KSI-301 and highlighting continued progress with our KSI-301 pivotal program, we will open the call for analyst Q&A. 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 Events and Presentations section of our website. An archive of this webcast will be available on our website soon after the conclusion of the call. I would also like to remind you that remarks made on this call today include forward-looking statements regarding our business, financial guidance, the initiation, enrollment, conduct and results of clinical trials, our regulatory strategies, our research and development activities, risks related to our business and certain other business matters. A more complete description of these and other material risks can be found in Kodiak's filings with the Securities and Exchange Commission. Kodiak will not undertake any obligation to update publicly any forward-looking statements whether as a result of new information, future events or otherwise. During the call today, we will review data from our ongoing Phase Ib study of KSI-301 and contextualize KSI-301's profile by benchmarking our data to current anti-VEGF agents on the market. We will also highlight how we view the probability of success of our pivotal studies of KSI-301 in light of the Phase Ib data. Finally, we will discuss our plans announced earlier today to expand our manufacturing partnership with Lonza. I will now turn it over to Victor, who will provide his introductory thoughts.

Thank you, John. Good afternoon, everyone, and thank you for joining us for our Third Quarter 2020 R&D Webinar. We're holding this call to discuss the recent data from our Phase Ib study of KSI-301 presented at the ASRS 2020 Annual Meeting in more depth and to assess what this data means to KSI-301's likelihood of success in our expanding pivotal study program. And our products' place in a competitive landscape that simply put is evolving in a manner that's very favorable for Kodiak. Ours was designed to be a product for everyone with retinal vascular disease, and I think the data are showing this to be the case. We will also discuss our ongoing pivotal study in wet AMD, DAZZLE, and reveal more detail on our planned pivotal studies in DME and RVO, which are scheduled to begin enrollment later this quarter. We will also talk briefly of recent advancements in our commercial manufacturing capabilities and plan. Today, KSI-301 is well characterized. It has been assessed in more than 300 patients, with over 150 patient years of exposure, with impressive safety, strong efficacy and incredible durability, with the majority of patients going 6 months or longer in wet AMD and DME. KSI-301's performance continues to be validated with maturing data, and it continues to surpass the high expectations we had for KSI-301 when we first initiated clinical studies. And with over 540-injection administered in the Ib and many more in our pivotal, we remain very pleased by the safety profile of KSI-301. KSI-301 is demonstrating a safety profile that is tracking with the gold standard anti-VEGF agents today, notably, LUCENTIS and EYLEA. By the time we file our single BLA in wet AMD, DME and RVO, we will have clinical data on KSI-301 in over 1,000 patients in concurrent pivotal studies. We've designed a high-margin of safety into our pivotal clinical studies. Now the exploratory Phase Ib study was designed to push KSI-301 in terms of dosing intervals and to provide us with a broad view of our product safety, efficacy and durability profile. The superb safety, efficacy and transformative durability we've seen so far in the Phase Ib study give us a high confidence of demonstrating a transformative profile in these registrational studies as well. We now take it one step further, and we have applied the learnings from the exploratory Phase Ib study to further optimize each one of our pivotal study protocols to build in a high-margin of safety and predictability. And for each of wet AMD, DME and RVO, Jason will walk us through those optimizations, those details later in this discussion. With all of these developments, we're excited for the journey that lies ahead. As is now very well accepted, there remains a truly great unmet need among patients receiving anti-VEGF therapies today. And despite the promise of today's medicines, as demonstrated in their registrational clinical trials, visual [gains ] are not maintained. In the real world, patients cannot be treated frequently enough and are overextended between doses under treatment leads to disease progression and permanent retinal damage. The retina community knows well that patient lives cannot be forced fit into high-intensity treatment regimens, which lead to patient burnout and/or abandoning of treatment. A new better medicine is required to shift the curve, to reset what has been appropriately called an epidemic of preventable blindness. We believe KSI-301 stands the best chance among treatments currently in development to directly address this great need. And we have made significant progress with our pivotal study in wet AMD, DAZZLE, and we are on track to initiate pivotal studies in DME and RVO this quarter. As to diabetic retinopathy, without DME, we're moving forward with the preparations, but the data of pivotal study initiation is not yet locked, as we assess the impact of COVID on this less acute yet high unmet need indication. As to manufacturing in line with our 2022 vision, which sees us going commercial in 2023, we don't want to be one of those companies that turns the cards with excellent top line data and then everyone waits 2, 3 or 4 years for the commercial supply. It is our intent to be able to supply millions of doses in year 1 from a facility designed from the get-go to flex up into double-digit millions of doses per year to supply a growing market demand. Kodiak remains focused on thoughtful execution of our KSI-301 pivotal program, the clinical studies, the manufacturing efforts, the regulatory strategy and the pre-commercial readiness. We continue to build our extraordinary team, both internal and external, and we're very thankful for the strong interest in KSI-301 in our ABC Platform by the retina community. We're looking forward to a poise commercial opportunity, and we're optimistic for the future care of patients. Adjacent competitive intravitreally injected anti-VEGF biologic agents in therapeutic candidates, notably Beovu and abicipar are being held to account and their safety challenges having been observed from early in their development. The port delivery system, beyond the limited accessibility and the challenged economics of surgically implanted port delivery systems, their top line data, it is an unfavorable safety profile that was unveiled, with 20% of patients experiencing significant safety events, including previews to a number of what we could call late events. And for example, all of the patients with conjunctival erosions or retractions that reflect an opening of the device from the back of the eye to the outside world. 11 of these went back to the OR, either for a [ tempset ] conjunctival repair, revision or removal of the device. That with a tightly curated set of surgeons in the pivotal study and which we understand each surgical procedure in the pivotal study witnessed by company representatives and now shifting to virtual VR trainings towards a commercial rollout. And fundamentally, the early state of the gene therapy platforms and as to high-dose EYLEA and faricimab, both in the lexicon of the field, step edits in a way. With high-dose EYLEA, I don't think anyone feels there is any kind of disruption here. And with faricimab, we see more of a commercial play as a higher molar dose LUCENTIS, still with high treatment intensity in terms of both of loading doses and treatment intervals and no possibility for any kind of designed in durability, although they will try for durability, nonetheless in wet AMD. And in DME, it's hard to see a lot of white space for them with how strong DME -- in DME KSI's efficacy and durability data are, and in our case, with only 3 loading doses. KSI with its powerful combination of safety, efficacy and designed in durability is unveiling itself to be the single, what we call Generation 2.0 anti-VEGF, a product for everyone. A first-line agent that like EYLEA in its day, can rise to own significant markets. The remarkable market share growth of EYLEA of a safe, effective anti-VEGF biologic, which in its day, had a longer dosing interval that disrupted LUCENTIS in the wet AMD market. Even at its earliest days, one had had only U.S. approval for wet AMD, whereas our concurrent regulatory strategy in the 3 big indications is designed to pull that curve forward. We remain independent as a company, and this independence provides us with the flexibility to adapt to both R&D and commercial decision-making within the ever-changing domestic and global landscape. And we remain, of course, well capitalized with a high-quality group of long-term investors who understand what it takes to execute, commensurate with the opportunity. We remain on track with our 2022 vision. We may complete enrollment for DAZZLE this year, including U.S. and ex U.S., but we are not pushing hard on this and rather focusing our efforts towards first patient in for the 2 DME and 1 RVO pivotal scheduled for later this quarter. We remain focused. Stepping back for a second, a summary view of our KSI-301 clinical studies, in particular, the durability comparisons versus EYLEA. The objective on top of safety and efficacy to show unique and powerful durability versus EYLEA in each indication and then through our concurrent regulatory strategy across the indications at launch. The outcome a unique, what we call Generation 2.0 profile and a growing and high confidence of getting there with, as I've mentioned, the high margins of safety we're building into each pivotal study and which Jason will detail shortly. As we've previously discussed, we've deepened and accelerated the development program for KSI-301 on the basis of last fall's end of Phase II meeting with FDA. This was a very collaborative dialogue, and the net result is that we can achieve an approval for KSI-301 in the key anti-VEGF indications on the basis of the plan shown here, 2 pivotals in a first indication, and we've chosen DME and 1 pivotal each in wet AMD and RVO and 1 additional study for diabetic retinopathy without DME. The information shown on this slide reflects our latest view as to duration and size of these studies. Here on Slide 8, we see a view of company milestones. On the basis of prior clinical data, our plan was proposed. And on the basis of our maturing clinical data, as shared here with the retina community at ASRS, our plan is in focus. It's actually a simple plan, and we invite you to join us on this journey because 2022 is right around the corner. I will pause here and hand the call over to Jason, who will talk about our recent clinical data and the progress with KSI-301 in our pivotal study program. Thank you.

Thanks, Victor, and good afternoon, everyone. The Kodiak clinical team has continued to make great progress with our studies of KSI-301 and the maturing durability data that we've been seeing in the Phase Ib study really continue to surpass even our own expectations. So before getting into the updated data that were presented at this year's virtual ASRS meeting, I just wanted to thank everyone who's been contributing, our clinical investigators, the growing Kodiak team, our partners, and most importantly, the patients who are participating in our clinical studies and really helping us learn so much about KSI-301. So what we'll review now is the latest safety, efficacy and durability data from the Phase Ib study as they've continue to mature. We now have over 100-patient years of clinical experience with KSI-301 in the Phase Ib study, along with the data that are accumulating from the ongoing mass DAZZLE wet AMD pivotal. Along with the new data, we'll share some details in some case examples that provide additional color and richness beyond what could be presented in the short 6-minute talk that's online at the virtual ASRS. So we'll also put those data in the context of both the unmet need for longer-lasting medicines for retinal disease and in the context of the competitive landscape, which, as Victor just described as clearing in some important ways. We'll also review the designs of the pivotal studies and how each of those is informed and derisked by the Phase Ib data and then further optimized for success and a high-margin of safety to show noninferior efficacy and disruptive durability compared to the gold standard anti-VEGF medicine EYLEA. So with that, let's move to Slide 10. So here, we see the Phase Ib study design, and then we'll also talk about the retreatment criteria and the patient baseline characteristics in the next couple of slides. A few words about the objectives. This is an exploratory study. It's designed to provide a scientific and clinical proof-of-concept for the safety, the efficacy and the durability of KSI-301 and more broadly for our ABC Platform in patients with the most common retinal diseases treated with anti-VEGF therapy, wet AMD, DME and retinal vein occlusion. Now as we've mentioned before, we're studying treatment-naive patients because we believe this is the most representative population that lends the most confidence to the design of our pivotal studies. And we're studying 3 different disease indications because each has its own characteristic efficacy response to anti-VEGF therapy and a different durability profile. So by studying all 3 of them early on in development, we learned the most about KSI-301 and its potential for achieving a Generation 2.0 anti-VEGF durability profile. So again, as a quick reminder on the study design, treatment-naive patients with AMD, DME or RVO are randomized to receive the 2.5 or 5-milligram dose level in a 1 to 3 fashion open label. A 3 dose loading phase is followed by a durability assessment phase and an extension phase, yielding a total of 3 years of planned follow-up and protocol guided treatment. We've enrolled 51 patients with wet AMD and 35 each with DME and RVO, and this study is fully enrolled. The overall study duration was originally 9 and then 18 months, and now we've extended the treatment and follow-up period to 36 months total, so we'll continue to generate long-term outcomes data in advance of in parallel to the pivotal studies. Outcomes include vision, measure the change in best corrected visual acuity or BCVA, using the standard ETDRS testing protocol and retinal anatomy, which is measured as change in retinal central subfield thickness or CST using optical coherent tomography imaging also called OCT. We also obtain other images, such fluorescein angiography, color fundus photos and OCT angiography. So a really rich data set that we look forward to learning more about the anatomy from these different kinds of imaging tests over time. Let's go to Slide 11, we'll review the retreatment criteria. So these are the retreatment criteria for the Phase Ib, which are specific to each disease state. We designed these in conjunction with the investigators and based on our evolving knowledge from past clinical trials as well as our objective to understand the clinical durability of KSI-301 in these different patient populations. The criteria really meant to individualize the dosing regimen to the patient based on disease activity, vision that changes in vision and retinal anatomy. And of course, we also want to understand chronic safety. So in the case of wet AMD, which is the most chronic of the disease is under study, the maximum retreatment interval is every 6 months, meaning that retreatment has given 6 months after the last treatment in wet AMD, even if the patient doesn't meet other disease activity criteria. In addition to the protocol specified objective criteria, because this is an early phase study, investigators can also retreat at their discretion if significant disease activity is present that does not meet the other criteria. On the next slide, Slide 12. We see the baseline characteristics of the 3 different groups: AMD, DME and RVO. Again, all these are treatment-naive patients. The baseline characteristics are what you would expect for a U.S.-based naive study population. And since we allow patients with vision as good as 20/25 Snellen to enroll in the study, the baseline vision in the wet AMD and DME groups, in particular, is very good, better even than in some of the recent Phase III studies. So patients are presenting to the retina specialists sooner and with a better vision, so it's important to allow these patients in the study. So these baseline characteristics are important to keep in mind when evaluating data within and across trials. Let's turn to Slide 13, start to look at the latest data. I'll start with wet AMD. Here on Slide 14, we see the efficacy data in wet AMD, measured as change from baseline in BCVA and OCT center subfield thickness. These are data from the 31 wet AMD patients who reached the week 44 visit prior to the ASRS data cutoff date of June 9, 2020. You can see the initial improvements in BCVA and OCT from the initial treatments as expected for a potent anti-VEGF. From baseline to week 12, the patients gained an average of 6.7 letters off of their good starting base of roughly 65 letters and an improvement in OCT center subfield thickness of 93 microns. In the period between week 12 and week 44, that is months 1 to 9 after the loading phase, the treatment effect is maintained, both in terms of best corrected vision and OCT, with an average of just 1.32 injections per patient. And notably, 44% of those treatments were the mandated every 6-month doses. And indeed, 58% of these 31 patients received their first retreatment at week 32, 6 months after the last loading dose. And then further supporting the extended durability, we see only a slow fluctuation in the OCT over time, which compares favorably to the OCT fluctuations observed with existing anti-VEGFs that are given on shorter dosing intervals. And the stability in best corrected vision over this interval, also consistent with the prolonged duration of KSI-301. Now turning to Slide 15. In the Phase Ib study, the average retinal thickness, or OCT CST data, as reported by our clinical investigators, includes the height of pigment epithelial detachments or PEDs. These are an anatomic feature in some patients with wet AMD. The treatment success in patients with PEDs doesn't necessarily imply flattening or complete flattening of the PED, but rather eliminating intraretinal and subretinal fluid, particularly when the PED is very high prior to anti-VEGF treatment. And additionally, comparison across studies of OCT mean center subfield thickness values is difficult because it's often not clear or not disclosed in presentations and publications, whether the data include or exclude the height of the PED, whether or how the data are corrected for different OCT machines, whether they're using the center subfield or center point thickness, among other reasons. So on this slide, the graph showed the BCVA and OCT change in the wet AMD subjects, either including or excluding the 4 patients from the 31, who had a very high PED at baseline, which we define here as 500 microns or more in total CST. The best corrected vision in OCT curves are similar in shape to those of the full cohort, but excluding the high PED patients, the CST values are lower at baseline and over time and the standard of the mean error bars are narrower. You can see how those 4 patients with high PEDs, thus pull the overall average CST value up on the prior slide. So excluding those 4 patients, you can see the average retinal center subfield thickness after treatment is recovering right around the 300-micron mark. So overall, we're really pleased with the data out to week 44. So let's now look at the latest data on durability of KSI-301 in wet AMD. So in wet AMD, remember that most patients treated with EYLEA or LUCENTIS get treated every 8 weeks or more often. Relatively few can go longer than 8 to 10 weeks and many patients get treated monthly or on occasion even more often. So our objective all along with KSI-301 has been to do a lot better than that, get the vast majority of wet AMD patients on an every 12-week dosing regimen or better, with the majority on an every 4 or 5-month regimen or even longer. So we present these data as swim lane plots, so you can see for how long each individual patient was followed and when they were retreated. Since the study is ongoing and the patients had variable follow-up time based on when they were recruited into the study, we think this is [indiscernible] apparent way to review and understand the data. As an upgrade to our prior presentations, we've indicated the different dosing regimen durations by [indiscernible] so the color of the bar represents a distinct treatment interval and the diamonds show each retreatment given. The pink bars denote the maximum duration of 6 months. You can see a lot of mandatory [ Q6-month ] retreatments on the plot. The bars in orange represent a retreatment interval of 3 months or less and the green and light blue 4 and 5 months. Purple means that the first retreatment hasn't yet been given. So overall, 92% of our wet AMD patients have achieved a time to first retreatment of 3 months or longer. 82% went 4 months or longer. Actually, most patients have not received their first retreatment until 5 or even 6 months after the last loading dose. Essentially, half, 49%, reached the 6-month cap without retreatment after the initial loading dose. As the data mature with some patients followed for as long as 18 months on study, we also see the duration of the additional retreatments. And remarkably, 2/3 or 68% of these wet AMD patients have achieved a 6-month treatment interval at least once during follow-up. So remember that this is an anti-VEGF treatment naive population. So there's no preselection for VEGF responders or for patients who might require less frequent dosing, there's really no way to identify those patients in a previously -- in a treatment naive population. So this result is really beyond our expectations. And I think these data really underscore the potential of KSI-301 and the ABC Platform to achieve truly long interval dosing with an intravitreally administered therapy. And so as Victor mentioned, we see the product developing to be the product for everyone with retinal vascular and exudative diseases. So let's flip to the next slide and look at a case example. On this slide, we see OCT images from a patient who has had 2 long interval 6-month dosing cycles with KSI-301. You can see the pretreatment OCT at the top at day 1 showing distortion of the retinal architecture and fluid in multiple compartments. A month after the 3 monthly loading doses at day 1, week 4 and week 8, you can see that then at week 12, the patient has gained 8 letters on the eye chart. And their disease is nicely controlled as shown on the OCT with restoration of the retinal anatomy. That patient was then followed monthly without disease recurrence and reached week 32 where treatment was mandated since it had been 6 months since the last dose. The vision had continued to improve with a gain of 12 letters from baseline. That mandatory treatment was given and no further treatment was indicated in the next 6 months either. So at week 56, their vision and OCT were both stable with 11 letters gained from baseline. So 4 total injections over year 1 and a really great outcome. And then, of course, that patient was again retreated under the mandatory every 6-month dosing at week 56. So then overall what have we learned here so far? So the data collectively demonstrate that KSI-301 has a potent anti-VEGF effect on both best corrected vision improvement and retinal drying in wet AMD patients. The clinical benefit appears in line with existing anti-VEGFs, and we're observing substantially longer durability of a clinical effect with KSI-301. The very high proportions of Phase Ib wet AMD patients who have been extended to 4, 5 or even 6 months without receiving retreatment, also further supports the design of our ongoing pivotal study DAZZLE, in which KSI-301 is administered to naive wet AMD patients on an every 3, 4 or 5-month regimen as compared to EYLEA on an every other month regimen each after 3 monthly loading doses. So we'll talk more about DAZZLE in a bit. And I'll turn it back over to Victor for a moment to give us some context around these data.

Thanks, Jason. As we see KSI-301's durability profile come into focus, we thought it would be useful to step back from the clinical data for a moment and to think about how to benchmark KSI-301's emerging durability profile against the state of the art. To do that, let's take a quick walk down memory lane. Last October 2019 at our R&D webcast, we spent time exploring these questions in detail with members of the retina specialist community. How have things evolved? Let's go to the next slide. Last year, we spoke about how the initial core goal of a next-generation biologic in speaking with retina doctors should be to bring nearly all patients to a 12-week interval. Remember that LUCENTIS' best efficacy per label is every 4 weeks, monthly. EYLEA in wet AMD is labeled for every 8 weeks for best efficacy. And real-world data suggests maybe 25% of EYLEA patients on monthly, 50% at every 8 weeks and a quarter every 12 weeks. So our major advance would be to bring essentially all patients to every 12 weeks. In speaking with retina specialists and KOLs, they often say, if you give me more durability than EYLEA with the same safety and efficacy, why wouldn't I shift all of my patients? This is because even small differences, small improvements, such as 6 weeks to 8 weeks or 8 weeks to 12 weeks, these make such the difference in the ability of the system of patients of retina specialists of health systems to keep the patients on drug over the long term, such that they don't get treatment fatigue in patients don't abandon therapy. And by the way, what was the impact of that clinical benefit? Those additional weeks of durability that EYLEA brought on the marketplace and the usage of anti-VEGF. Let's look at the next slide as a reminder. On Slide 20, we look at the EYLEA versus LUCENTIS market share and revenues worldwide over time. In the year after EYLEA was first approved, notably only in wet AMD and only in the United States initially, those 4 additional weeks of durability on its label led it to capturing nearly 20% of the branded anti-VEGF market in the United States. By 2015, in its fourth full year on the market, EYLEA had overtaken LUCENTIS as the market leader, despite not having all the indications approved until 2014. Since then, the gap has continued to widen because of this difference in durability, combined with similar safety and efficacy as LUCENTIS is important for physicians, patients and health systems. And now we are approaching our time with Kodiak and KSI. And of course, it will be a more complex market that we will be dropping into, but we will have the same core attributes. As we have called it from a technology standpoint, same where it matters, different word counts, an anti-VEGF biologic with the same safety, same efficacy, a unique and disruptive durability. And as an intravitreal biologic, the same mode of administration and reimbursement. And we have a development plan that brings the 3 major indications to the market concurrently rather than serially, as was EYLEA's development approach. Turning to the next slide. So last October, we did the thought experiment together. We asked what would be the impact of a next-generation therapy that didn't just bring almost all patients to 3 months, but brought, say, 80% to 4 months and 60% to 5 months. Well, we think it would be disruptive because these durability improvements are really important for physicians and patients. It's important to remember that as a community, we are in the same clinical situation with respect to durability of a safe anti-VEGF that we have been in since EYLEA was first approved back in 2011. Let's move to the next slide. So how is KSI-301 doing against our hypothetical next-generation medicine? Let's move to the next slide. Here, we see from our Phase Ib data in wet AMD that KSI is exceeding the bar for a hypothetical next-generation anti-VEGF. Essentially, all of the EYLEA treatment interval is contained in the left margin of the KSI-301 durability profile to date. It is a remarkable conclusion. And as we have mentioned, KSI is exceeding our expectations. We have indeed come a long way as a molecule and as a company in 9 months. And with the progress and execution of our KSI-301 development plan, I look forward to sharing the next 12, 18, 24 months with all of you. It's a bit of a put on your seatbelts moment that we are in. I will turn it back over to Jason as he looks from a bit of a different lens to benchmark KSI-301's durability against EYLEA.

Thanks, Victor. So what is the durability of EYLEA versus KSI-301 based on randomized trial data and real-world data? So let's take a look at the next slide, Slide 24. On this slide, we're looking at EYLEA durability from the rival study. This was a randomized study of the treatment extend approach to wet AMD treatment comparing EYLEA and LUCENTIS. In this study, treatment can be extended in 2-week intervals. And it's an important study because many patients are treated in real-world practice using a treat-and-extend approach. And what we see is that, perhaps surprisingly, nearly half of the EYLEA patients actually can't go beyond monthly EYLEA. And the rest of the EYLEA patients were evenly distributed across the other dosing intervals, roughly 10% to 15% each who are on intervals of every 6, 8, 10 or 12 weeks in this study. On the right side of the graph, in the darker bars, are the proportion of KSI-301 treated patients on each interval from our Phase Ib data. So this is what we believe to be a Generation 2.0 durability profile, with nearly half of the patients on every 6-month dosing and 92% on 12-week or longer dosing. Let's go to the next slide and look at this in a different way. On this slide, Slide 25, we're looking at some really interesting real-world data that were just presented at the virtual ASRS that took place this weekend, looking at treat-and-extend intervals from a collection of U.S. retina practices. Here, we see that over a 2-year period, the average EYLEA interval was every 8 weeks, and the average of the maximum interval ever achieved by EYLEA treated patients was 9.6 weeks in this cohort of 428 patients. By contrast in our Phase Ib data, the average first interval was 19.3 weeks and the average maximum interval achieved by wet AMD patients considering all available follow-up in the Phase Ib is 20.6 weeks. So this would not just be an incremental improvement for patients or physicians. Let's move to the next slide and return to the KSI-301 data. Here on Slide 26, one of the things we've been eager to see is data consistency in the Phase Ib study over time. So it's been reassuring to see that as the data set continues to mature, the vision, OCT and durability outcomes have indeed remained consistent over time. Here, we can see that consistency. So comparing the outcomes we presented at the Angiogenesis Meeting back in January, February earlier of this year versus the data that were just presented at ASRS. The durability proportions remain stable, as do the vision and OCT outcomes, both with more patients and over time. So let's then turn to the DAZZLE study, our ongoing pivotal study of KSI-301 versus EYLEA in wet AMD. Let's move to the next slide, please. So this is a simplified schematic of the DAZZLE pivotal study design in its first year. In this study, patients with treatment naive wet AMD are randomized 1:1 to receive KSI-301 every 12 to 20 weeks or EYLEA every 8 weeks, each after 3 monthly loading doses. The determination of the treatment interval for patients assigned to KSI-301 is based on disease activity assessments who are both OCT and vision are measured and compared against prior data, similar to other recent and ongoing Phase III studies in the field. So by default, patients assigned to KSI-301 are on every 20 week regimen. And then if the disease activity criteria are met before 20 weeks has elapsed, that is 12 or 16 weeks from the last dose, then the treatment interval is correspondingly shortened. So if the patient was on a Q 20 week interval at that visit when the criteria are met, they could go to 16 or 12 weeks. Likewise, if they're on a Q 16-week interval, they could go down to Q 12 week. The primary end point is at 1 year, and it's a noninferiority comparison to EYLEA with a 4 letter noninferiority margin. The 1-year end point is measured as the average of the best corrected visual acuity change from baseline to weeks 48 and 52. And all of the KSI-301 patients are analyzed together as a single group with respect to the primary comparison to EYLEA. And then, of course, we'll also report out the proportion of KSI-301 patients on each of the different dosing intervals. Now not shown here in the second year of the study, patients whose disease is stable can have their KSI-301 treatment interval extended. And then also for patients originally randomized to EYLEA, we're rerandomizing that group 1:1 to either continued EYLEA or switching to every 8-week of KSI-301. So those will be useful and important outcomes, too, that we'll have for the second year of the study. On the next slide, we look at the disease activity criteria for DAZZLE that are used for determining the KSI-301 treatment regimen for its duration. So here on Slide 28, in the Phase Ib study, remember that the disease activity criteria are used to determine when an additional treatment is given. But in DAZZLE, the criteria are used to determine whether the interval is either 12, 16 or 20 weeks. So remember then that the objectives are a little bit different in Phase Ib, it's an exploratory study, had a small number of sites. And we want to learn as much about durability as we can. And in DAZZLE, we're letting a large global noninferiority study head-to-head against the gold standard of EYLEA. So we tightened up the criteria, removed some of the subjectivity and then these criteria and the overall approach to the dosing regimen determination in DAZZLE are, like I said, very similar to other recent and ongoing Phase III programs in wet AMD. Essentially, what you want to try to do is customize the dosing interval per patient in a way that is feasible in a large multicenter double mass trial. So in terms of the specific changes compared to the Phase Ib. one example here on the table is that the dosing interval is shortened in DAZZLE, if a patient has an increase in CST of 50 microns or more, along with a decrease in vision of 5 letters or more versus with 75 microns in Phase Ib. So we have a little bit tighter disease control as an example. Looking then to the next slide, we've also made other optimizations to the design of DAZZLE, and these are also baked into our upcoming pivotal studies in DME and RVO to further optimize their likelihood of success. So first, we're recruiting a very similar patient population, treatment naive wet AMD, mainly from the United States, with best corrected vision up to 20/25 at baseline, and we expect 80% of the patients to come from the United States roughly. And recruitment in the U.S. and Europe is ongoing, and it's almost done in the United States. We have a tighter dosing interval in DAZZLE, ranging from early 12 to 20 weeks as opposed to 24 weeks in the Phase Ib. It's clear that many patients can go every 24 weeks based on the Phase Ib data, but we're mindful that we're going up against a good comparator in a noninferiority study. So likewise, DAZZLE is designed for tighter disease control and less subjectivity in determining the treatment interval compared to the Phase Ib for that reason. And then finally, we have high statistical power for noninferiority should be well over 90% based on the study design assumptions. So we're confident that this study is designed for a high probability of success. So with that then, let's turn to the latest data in diabetic macular edema. So personally, as an ophthalmologist, having worked in the DR and DME field for some time now, going back to the original LUCENTIS Phase IIIs before there was any intravitreal medicine approved in the United States for diabetic eye disease, I'm really very excited by what we continue to see in the KSI-301 data. Remember that in DME, EYLEA is labeled for every other month dosing after 5 monthly loading doses, and LUCENTIS is labeled for monthly dosing. And this need for frequent intensive treatment was confirmed in DRCR.net collaborative group study of DME patients called Protocol T, which was published a few years ago in The New England Journal. Almost all of the patients required 6 initial monthly loading doses under the DRCR treatment algorithm and a median of 9 to 10 doses was administered in the first year of therapy for all of the agents tested, EYLEA, Avastin and LUCENTIS. They all needed about the same number of injections per year. So we have a twofold objective with KSI-301 in DME: first, to reduce the number of initiating or loading doses; and second, to extend the treatment interval in the maintenance phase to 3 months and beyond. So let's go to the next slide, and we'll look at the data. Here, we see the KSI-301 efficacy data in treatment naive DME, again measured as change from baseline in best corrected vision and OCT center subfield thickness. These are data from 18 DME patients who reached the week 44 visit prior to the data cutoff date. These patients had good starting vision approximately 70 letters. They experienced the visual acuity increase after the 3 loading doses and maintained a gain of 6.6 letters, with a mean of just 0.61 retreatments. So notably, 2/3 of these patients never required retreatment during this 10-month follow-up period. Consistent with the extended durability effect of KSI-301, we again see only slight fluctuations in the OCT over time, comparing favorably to the OCT fluctuations observed with existing anti-VEGFs given on shorter dosing intervals. Let's turn to Slide then 32. The swim lane plot for DME durability. Colored bars in pink and orange continue to denote retreatment intervals of 6 months or longer and 3 months or shorter, respectively. The diamonds indicate each retreatment and purple means that the first retreatment has not yet been needed. So far, 97% of the patients have gone 3 months or longer before their first retreatment or put another way, only 1 patient of the 33 required retreatment at 2 months after the loading doses. 76% have gone 4 months or longer and 70% have gone 5 months or longer before their first retreatment. 2/3 have gone 6 months or longer. And remember, there's no cap to the treatment interval in this DME cohort. Almost half of the patients, 45% have not required retreatment to date, denoted by the purple bars. Some of these patients have gone for a year without needing additional treatment, which is really quite remarkable when you think about the treatment intensity in the first year with LUCENTIS or EYLEA in Protocol T. Let's move to Slide 33. Here, you can see a case example of this with a DME patient who had really pretty severe and diffuse DME at baseline. You can see the large cysts and fluid in the retina in the pretreatment OCT at day 1. This patient has received 3 loading doses of KSI-301 and has received no additional doses in the subsequent year. They've gained 8 letters from baseline [ running ] of Snellen equivalent of 20/20, which is a really nice outcome for 3 injections. Moving to the next slide, Slide 34. Again, looking at the efficacy and durability data, as they stood earlier in the year to now, we see consistency in the maturing Phase Ib data, both with longer follow-up and with more patients. Looking then to Slide 35. So how do the KSI-301 durability data in DME compare against the best treatment regimen for the best medicine in DME EYLEA. In the NEI funded DRCR Protocol T study, essentially, as I said before, all the patients treated with EYLEA received 6 monthly loading doses and then 3.2 additional doses on average in the rest of the year for a total of 9.2 injections on average in the first year of that study. These were U.S. patients who, in many ways, are similar to the ones in our Phase Ib study. And in our study so far, the DME patients have received 3 monthly loading doses and an annualized 1 additional dose for a total of 4. So considering that EYLEA is labeled for 5 monthly loading doses and in every 8 weeks, we think KSI-301 has a Generation 2.0 anti-VEGF durability profile in DME as well. So then how will we be studying KSI-301 in the DME pivotal studies? Let's move to the next slide, Slide 36. This slide shows the schematic of our pivotal study design in DME. As Victor mentioned, we're going to be running 2 of these Phase III studies in parallel with identical design. In each of the 2 studies called GLEAM and GLIMMER, we will randomize 450 treatment naive DME patients to either KSI-301 every 8 to 24 weeks after 3 loading doses or EYLEA every 8 weeks after 5 loading doses per its label. Approximately 450 patients will be randomized per study, and the primary endpoint is the change from baseline and best corrected vision at 1 year, again the average of the week 48 and 52 visits. The studies are noninferiority studies with a margin of 4.5 letters. At each monthly study visit, patients randomized to KSI-301 will undergo a disease activity assessment using data from best corrected vision and the OCT measurements. And then depending on their disease activity status, the dosing interval can be shortened, lengthened or maintained at the same interval. So this is different than in DAZZLE where the interval can only be maintained or shortened in the first year. And that's because in DME, you can get disease modification that curtails the need for therapy over time. And we've already seen that in many DME patients treated with KSI-301 that the treatment interval can be prolonged after the first retreatment. So we want to take advantage of that observation in the study. And then again, the minimum interval is every 8 weeks, the maximum interval every 24 weeks or 6 months. On Slide 37, the next slide, you can see the design in the second year essentially continues the same approach. EYLEA will stay on its every 8-week regimen and patients on KSI-301 will continue to be on the 8- to 24-week regimen based on the disease activity assessments. Moving then to the next slide, Slide 38. Here are the disease activity assessment criteria for the DME studies. Like in DAZZLE, all patients randomized to KSI-301 are on the longest interval by default. And then if they meet any of these disease activity criteria, the interval is shortened. So for instance, if they first meet the criteria of 20 weeks or 16 weeks after the last dose, they would go to a 20-week or a 16-week interval and so on. The criteria for shortening the treatment interval are tighter than what they were in the Phase Ib and the subjectivity is reduced. So for instance, the OCT criteria are pegged against the lowest previous measurement and the BCVA criteria are pegged against the best previous BCVA measurement. We also have some additional anatomic criteria. Again, the idea is to take what we've learned from the Phase Ib program and make some refinements on the margin to optimize the probability of success. Let's move to the next slide, Slide 39. So we've made other optimizations to the design of GLEAM and GLIMMER similar to what I described for DAZZLE. We're again recruiting a similar patient population to the Phase Ib, naïve wet AMD, mainly from the U.S. We have a tighter dosing interval from every 8 to 24 weeks, whereas there was no cap in the DME treatment interval in Phase Ib, that every 6-month cap is here in GLEAM and GLIMMER to ensure that we have sufficient treatment exposures for the safety database. Likewise, GLEAM and GLIMMER are designed for tighter disease control and less subjectivity, as I mentioned before. And finally, we have high statistical power for noninferiority in each study well over 90%. So we're confident that these studies are also designed for a high probability of success. With that, let's move to the next slide. Turning to the latest KSI-301 data on retinal vein occlusion, another important retinal disease and the third most common use of intravitreal anti-VEGF. Next slide, please. Here, we see KSI-301's efficacy data in treatment-naïve RVO, again measured as change from baseline in BCVA and OCT central subfield thickness. The 33 RVO patients that completed their week 44 visit by the data cutoff began with a lower visual acuity baseline of approximately 55 letters, which is what you would expect with treatment-naïve RVO, typical for this disease. After 3 loading doses, their visual acuity substantially improved with a 22.4-letter improvement at week 44 over 4 lines of vision gained. This vision gain was maintained with an average of just 1.33 injections, with only 1/3 of patients requiring more than 1 retreatment in this period. A sustained OCT response with an improvement from baseline of 370 microns was also noted, and you can see on the OCT graph how the macular edema was very well controlled. On the next slide, in the swim-lane plot of durability of individual patients with RVO, this is the disease with arguably the highest VEGF load. So the bars in pink and orange now denote retreatment intervals of 4 months or longer and 2 months or shorter, respectively. 94% of the RVO patients, 31 of the 33, have gone 2 months or longer before their first retreatment. Only 6% of patients had their first retreatment at one month after the loading doses. 66% had their first retreatment 3 months or later since the loading dose and 56% achieved a first interval of 4 months or longer. Remarkably, given that many RVO patients require monthly therapy for the best results with existing medicines, 71% of patients have achieved a 4-month or longer interval at least once during follow-up. Moving to the next slide, Slide 43, again, looking at the efficacy and durability data as they stood earlier this year to now, we again see the consistency in the maturing Phase Ib data with longer follow-up. We can see, in particular, the stability in OCT and vision outcomes from week 24 presented at Angiogenesis to week 44 presented at the Virtual ASRS recently. Looking now on to Slide 44, how do our durability data on RVO stack up? We are again seeing Generation 2.0 durability here. Looking at studies of EYLEA, patients all received monthly therapy for 6 months, which is the duration to the primary end point in the registrational studies, and monthly treatment is how EYLEA is labeled for RVO. In the second 6 months of the year in those EYLEA studies, patients received an average of 2.6 additional doses, essentially every other month for a total of 8.6 injections over the year. With our data for KSI-301, which should lend a lot of confidence to the results at a 6-month primary end point, we gave 3 monthly loading doses and then an annualized average of only 1.7 additional doses. So this would be only 4.7 doses over the year or roughly 4 fewer. So how will we be studying KSI-301 in Phase III for rental vein occlusion? Look at the next slide. This slide, Slide 45, shows the schematic of our BEACON study, the Phase III study in patients with RVO. This is a year-long study, in which we're going to randomize 550 patients with treatment-naïve RVO, either branch or central vein type, to either every 8-week KSI-301 after 2 loading doses or to monthly EYLEA for the first 6 months. The primary end point is at 6 months, week 24, and it's again noninferiority to EYLEA with a margin of 4.5 letters. In the second 6 months, patients in both groups will receive treatment on an individualized regimen, again, using typical disease activity assessments. This part will also be interesting as a direct head-to-head comparison of EYLEA and KSI-301 on the same criteria-driven regimen. The minimum interval is monthly, and there's no upper limit to the retreatment interval in this 6-month period. On Slide 46, the next slide, here again, we're looking at the disease activity assessment criteria. Remember that these, in this case, are only used to determine dosing in the second 6 months. The first 6 months of the study are fixed-interval dosing. The criteria are similar to the DME studies and again, a bit tighter than they were in Phase Ib, with the OCT and vision measured against the best previous measurements and reduced subjectivity. As in the wet AMD and DME studies, the idea is to take what we've learned from the Phase Ib program, make some further refinements and optimizations. On Slide 47, the next slide, one such improvement that we've made in the RVO study is to reduce the number of loading doses from 3 to 2 in the KSI-301 arm. This case example shows both the power of KSI-301 in treating the worst RVO cases and also the magnitude of effect achieved with 2 loading doses. You can see at day 1 that this retina is really under water at the beginning of this patient's journey with an OCT thickness of over 1,200 microns. That thinned out substantially after just a week, and the vision improved by 14 letters nearly 3 lines in that short time. Things continued to improve, and after the second dose, the retinal thickness was normal, and the patient had gained 23 letters to a vision of 2025. Our view of the data is that nearly all of the initial improvements in RVO are achieved with those first 2 treatments. So this allows us to move into a q 8-week, fixed-interval period from there through the primary end point at month 6. Turning then to the next slide. We've made other optimizations through the BEACON study similar to what we've described for the others, again, recruiting a very similar patient population to the Phase Ib, treatment-naïve patients, mainly from the United States, tighter dosing interval with q 8-week, fixed-interval dosing through the primary endpoint, and then in the second 6 months, tighter disease control and less subjectivity in retreatments compared to Phase I, and finally, high statistical power for noninferiority, again, well over 90% based on the study's design assumptions. With that, let's move to Slide 49, and we'll review the safety data to date with KSI-301. On the next slide, Slide 50, here, you can see the latest update. And we're really happy that while safety is a key objective of the study, we only need 1 slide to review it. I'm very pleased that the safety profile of KSI-301 continues to be excellent and very encouraging. Now with nearly 550 injections given in the Phase I program, we are continuing to track with a high-bar safety profile of the gold standard intravitreal medicines for safety, LUCENTIS and EYLEA. None of the serious adverse events that had occurred were drug-related, and they're typical of the systemic SAEs expected in these patient populations. There were only 2 events of intraocular inflammation resulting at a rate of 0.37% per injection or 1.5% on a per-patient basis. These events were mild in nature, both of them were traced to 1-plus grade cells in the vitreous on a standardized scale, where 0 is none and 4-plus is severe. These resolved completely, and there was no evidence of vasculitis or retinitis in either patient. Both patients have done very well with substantial improvements in vision from baseline. Actually, each of them has gained 30 letters or 6 lines of vision as of their last, most recent visit. Now of course, there's no guarantees in life nor any crystal ball, but we're very pleased with such a clean safety profile. And with the patient number and exposure length and the number of doses and of course, the repeat doses given so far, we remain very optimistic that we're tracking to a LUCENTIS or EYLEA safety profile with KSI-301. Let's move to the next slide, and we'll wrap up the clinical data and pivotal program design review. Here on Slide 52, we note the conclusions from Dr. Berrocal's presentation at the virtual ASRS, and we feel that it really hits all the high notes. As to my own conclusion, as we've all heard, the current gold standard medicines really require frequent eye injections in order to achieve the best clinical results. The state of the art in the field in terms of a safe, effective medicine has really stayed the same since EYLEA launched in 2011. So personally, as an ophthalmologist and as a drug developer, I'm very excited about the potential of Kodiak's ABC platform and KSI-301 to address this challenge of ocular durability and to put a therapy into the hands of physicians that shows the same safety and efficacy with really disruptive next-generation durability compared to EYLEA so that patients don't need to lose vision unnecessarily due to insufficient durability. I'm really pleased with the continued evolution of the longer-term durability, safety and efficacy data for KSI-301 that we've seen to date, and I'm excited and confident in our pivotal study designs that result from our Phase Ib data and our discussions with the FDA. So with that, I'll turn it back over to Victor and I appreciate your time and attention.

Thank you, Jason. Our development program, of course, is a combination of clinical development and manufacturing development, which is the job of our technical operations group and which integrates across manufacturing operations and quality operations with teams and facilities in both the United States and Switzerland. Moving to Slide 54. Earlier today, we announced in a joint press release with Lonza, a preeminent contract manufacturing firm, a dedicated purpose-built manufacturing facility in Switzerland for the commercial supply of KSI-301. In their Ibex dedicated model, they have built the facility shell, and the dedicated facility itself will be operated within our existing global supply network with the same teams, same batch records, same quality systems. The Ibex dedicate model provides for an accelerated build time and importantly, the ability on our plan to supply millions of doses of commercial supply in year 1 after launch and the ability to flex up this commercial supply to double-digit millions within the same facility in response to market demand as needed. Moving to Slide 55. We have been partnering with Lonza on manufacturing of KSI-301 since 2014, and this agreement represents a stepping in to a successful long-term relationship between the parties. The timing is aligned with planned activities for scale up, BLA readiness and commercial launch. The flex up capability, as I mentioned earlier, is aligned with our world view. When we turn the high-probability data card and we file the BLA and with the potential for a U.S. commercial launch in 2023 and EU shortly thereafter, we don't want you to be waiting for product. We aim to supply millions of doses in year 1 and to have the ability to flex up to double-digit millions shortly thereafter. This is representative of our conviction and our plan to step into the special opportunity that we see with KSI-301 and our ABC platform. Moving to Slide 56. Where are we today with Kodiak and KSI-301? In summary, it is perhaps a simple story. We have a molecule in the platform that we designed over a 10-year period. We now have sufficient characterization of KSI-301 in patients, and we have the strong outline of its safety, efficacy and durability. We have designed high margins of safety into our pivotal clinical studies. DAZZLE is a strong enrolling pivotal today, and we expect our DME and RVO pivotals to be enrolling this quarter, and we are pleased with the level of engagement and interest within the retina specialist community globally for these studies, and we anticipate strong enrollment, and with our concurrent regulatory strategy and our view to build high-quality manufactured product at scale from year 1. And with competing programs, doing the best they can, consistent with their liabilities, be they poor safety and/or poor durability, we are executing on our 2022 vision with conviction commensurate with the opportunity for our special molecule and platform. And also, we are deepening the pipeline based on our ABC platform and its validation. We look forward to sharing this future with you over the next 12, 18 and even 24 months. Thank you very much. We now turn to the operator to lead a question-and-a session.

[Operator Instructions] We'll take our first question from Ronny Gal with Bernstein.

If I may, I have 2. First, on the trial design. I guess why not superiority design? And are you going to do an interim analysis given the powering of the trial in any of those trials? And second, regarding safety, you mentioned those 2 patients out of 130, which come from the early trial. Obviously, you're not including DAZZLE, but obviously you've treated a good number of patients with DAZZLE already. Can you just confirm to us that the rate of information you're seeing in DAZZLE on a blinded basis does not exceed that 1.5%?

Yes. So just to repeat, so a quick question around why we didn't do a superiority design, Jason?

Yes, sure. So in terms of superiority designs, I think a few points to consider there, right? The first is, I think, fundamentally, we're at the top of the efficacy dose response curve with the different doses of anti-VEGFs that were given based on all of the data that have been seen to date. And we know that you can get very good outcomes with the current medicines in a well-conducted study where you really have a lot of efforts to make sure that the patients come back to the -- for every visit. So I don't think in a well-conducted study, there is a reason to think that one anti-VEGF should have superior efficacy fundamentally to another. So to us, the most important thing is to show noninferior efficacy the same safety and then disruptive durability. So that's the outcomes that we're looking for there.

Yes. And then as to a question on a blinded basis, what we're seeing in the DAZZLE study with regards to inflammation? We did have an IDMC or a safety committee meeting, where the mass data was reviewed, and we can confirm that there's no abnormal rates of inflammation, i.e., that the inflammation rates are at or below that level.

And the flexibility of an early readout given the powering of the trial? Are you doing an interim analysis?

Right. So in terms of discussions around an interim, we had a lot of back and forth about that. I think we're really focused on the 2022 vision where really the DME, wet AMD and RVO data are going to come in fairly quick succession. So we don't see the need in the company to do any kind of an early interim analysis. And we're really focused on the manufacturing and the clinical components of the development plan, essentially coming together at the same time in -- from the beginning of 2022 really through the end in BLA submission.

We'll take our next question from Michael Yee with Jefferies.

Thanks for the very in-depth presentation. Appreciate it. Two-part question. One is, when you look at your data in AMD, the swimmer plots as well as all the BCVA and OCT data and you think about handicapping the Phase III results, just so people don't get too out of hand looking at this remarkable Phase II data, how do you think about how the time to second injection falls off just a little bit? Is there any comments about what's going on in just a couple of those patients? And does that impact anything on the time to second injection? Because there's a drop down, right, in the Phase III but -- year 1, but you can actually go up in year 2. So maybe just talk about how to think about that. And then on OCT, same question, is there any reason to think that it could actually be better than a comparator on OCT or the same? Or how do you think about that? And then lastly, just you made a bunch of comments on port delivery. I think we heard a bunch of comments out of Roche. They obviously don't think endophthalmitis is too big of an issue, but maybe just talk to your overall perspective on that data today and maybe give some more color on that.

Sure. Let's see maybe work it backwards, a couple of comments or 2 around the port delivery system, and then we can talk a little bit about the durability of the durability on vision and maybe a quick comment on the OCT. I think on the port delivery system described as generally well tolerated, I think 20% adverse events that I think the community deems significant, and there's -- with any surgical device that's implanted, that's designed to bridge exterior into the interior of the eye, it's pretty serious. And so they do have early SAEs or early adverse events and then of course, there's further worry about later adverse events, as the device ages, as the eye ages, as the patient continues. Putting aside the early adverse events, of course, this was a highly curated set of surgeons with a highly witnessed set of placements and so probably represents very close to an optimal best set of early safety data. I think, more importantly, in a 9-month end point and a fairly short study, I think we were perhaps surprised to see the number of conjunctival blebs, which represent a leaking from inside the eye in and around the placement and then also a fairly high number of conjunctival erosions, which represent sort of a weakening of that epithelia, which is covering and protecting the device from the outside world and I think fairly worrisome for the long-term safety. So whether or not those led to endophthalmitis in the short study, the point is that when you open an outside-in or an inside-out chamber across the eye, that's pretty serious. So I think the question is, it's an elegantly designed device, and it probably fits certain patients. It may not be clear today exactly which patients. It's for -- in any case, it's a very large market opportunity. And as Jason and I have mentioned, our objective is to try and build KSI to be a product really for everybody. And I think a bit of a downside surprise on the safety data, again, creates more white space for KSI within the competitive environment. Moving a little bit to a discussion of the durability of the durability with KSI in the wet AMD swim lanes, Jason?

Yes. Thanks, Victor, and thanks, Mike. So I think it's very early to say in the study in terms of the number of people who've had multiple sequential retreatments over time, but 41 patients have received their first retreatment and only 30 patients have received a second retreatment, and actually, I'd say -- actually see some improvement in durability, right, 49% of patients achieved a first 6-month interval and 68% have achieved the same interval during follow-up. And I think also, what you see in the swim lanes, this is -- it's a real life, unselected treatment-naïve patient population where you see variability over time, right? And so it's important that we capture that. Most studies, there's no readout of that. You just say, well, there was an average of 3 injections over this month period, right? So that implicit heterogeneity in the patient population is like nicely represented in this data set. So I think in terms of reading that through into DAZZLE, again, we allow a drop-down, but not a sort of slide-forward in the treatment regimens in the first year. Again, we want to think about in the context of a noninferiority study, making sure that we hit the key outcome with really great durability. So I'm quite optimistic about that overall. And I think it's overall like a quite positive set of data for what we've seen on retreatment so far and then of course, the data will continue to mature over the next period of time. I think your last...

Okay. On OCT, yes.

Yes. It was on OCT. In terms of reasons to think, are we -- how do we stack up against the comparator? I think what we see across all of the different indications that we've looked at, 3 different diseases, AMD, DME and RVO, is really, I'd say, potent and appropriate impacts on the OCT, right, when you look at all of the different data across each disease. How that would compare to EYLEA or in a head-to-head study? I think really, you need to have the data in the head-to-head study all read at the same reading center by masked graders to really game that out, right? It's like the -- some of the differences between, say, Avastin and EYLEA or EYLEA and brolucizumab in these studies are never really apparent until you have the head-to-head data. And then the second question is, well, what's really the impact of any of those potential differences relative to what the regulators really think about, which is functional vision for the patients.

We'll take our next question from Anupam Rama with JPMorgan.

This is Matt on for Anupam. So one from us just as it relates to the pivotal DAZZLE trial. It looks like the readout is now guided for 2022 compared to 2021 previously. So just wondering what the source of this "delay" was. And are you referring to it previously as a potential readout in 2021? You did mention that 80% of patients would be from the U.S. So just wondering if there's any geographic-specific slowdown. Any additional color there would be great.

Sure. Yes. Thanks for the question. I think as we indicated in our last quarterly discussion, because of COVID, we were concerned about a slowdown on potential for missed patients in DAZZLE. And so we delayed the activation of Europe enrollment for DAZZLE. Since that time, enrollment has spiked back up in the United States, and we've also activated a number of countries in Europe. As Jason mentioned, we're nearing the close of enrollment in the United States. We want to preserve a minimum of 80 -- or 20% ex U.S. patients. And so that's really going to drive the closeout of enrollment for DAZZLE. We do think it's possible that we will complete DAZZLE enrollment in U.S. and ex U.S. by the end of this year. Although because our real focus is on the 2022 vision, we're not rushing the completion of the DAZZLE enrollment. That's moving, and we're also very focused on getting the DME and RVO studies started and to try and have a more rapid ramp in terms of U.S. and ex U.S. in terms of study sites for the 2 DME and the RVO studies. And I think, as we mentioned, we're also optimistic that we'll begin enrolling in perhaps treating patients this quarter. And so we're looking to complete enrollment potentially for DAZZLE, in summary, by the end of this year, but that may push into the first quarter of next year, and with the 12-month end point, that gives you a view into the time for top line data for DAZZLE. And that will come, probably DAZZLE will read out, then presumably a BEACON RVO study with a 6-month end point depending a bit on time line for RVO enrollment and then the 2 DME studies, we believe, can enroll actually quickly on a global basis with their 12-month end point. That coming together with our manufacturing will feed BLA readiness on the 2022 vision.

We'll take our next question from Graig Suvannavejh with Goldman Sachs.

Just wanted to follow up on the comments you made earlier about the port delivery system, about the challenged economics that you guys are currently envisioning. Can you just maybe provide a little bit more specific color on what exactly you mean by that and whether that then presents a certain obstacle for adoption of that product relative to the current intravitreal anti-VEGF injections? And then a second question just has to do with retinal specialists who are currently giving anti-VEGF injections. Like what percent of those specialists can also do the PDS surgical implant themselves? Or is there a view that they truly need to refer that to someone else? And then just the last question that just has to do with, would you perceive the current level of awareness of 301 amongst the specialist community? In other words, is there already a very broad awareness of 301 in the promise of the extended duration? Or do you believe that there's still additional kind of educational efforts that are needed to really optimize kind of what the launch could look like?

Right. Thanks. I'd say we're not going to propose that we're experts in the economics of surgical solutions, right? The beauty of what we're building is kind of, as we mentioned, the same where it matters; in other words, an anti-VEGF biologic dosed in the same way and reimbursed and distributed in the same way. I think when we do speak with retina specialists about the port, I think one of the questions is, as a retina specialist, do you own your own surgical center or you don't? And if you don't, it's fairly complex complexities, even as simple, for example, if there's a LUCENTIS injection that's required at the beginning, is that going to be given by the retina specialists in their unit or is that going to be given right as part of the surgical theater, who's going to capture that or who's going to eat that? And is that going to be a part of an integrated cost; also, complexities around what the co-pay would be; complexities around how you keep inventory; complexities around who's going to pay for rescue injections if they're needed; or complexities around who's going to pay for the repairs and some of the complexities? So we think, potentially, there could be patients for this new system. And maybe it will be bispecifics or other things -- other agents down the road for niche populations. And we look forward to seeing more for the time line for the commercialization. Obviously, Roche/Genentech also have faricimab, which is an anti-VEGF biologic, which sort of represents a higher molar dose LUCENTIS and perhaps it may be easier for them to try and build market share with faricimab rather than the port system. So that's one answer. I think another one is the question of, well, is there broad awareness of KSI within the retina specialist community? It's a really great question. Certainly, there's broad awareness within the clinical investigator community, of course, especially, in the United States where we have, I think, close to 50 sites for DAZZLE, and we're going to grow that, I think, to across the 2 DME studies as many as, say, 80 sites. I think -- and then in Europe, I think we're just building awareness now as we bring DAZZLE to Europe and then we bring the 2 DME and RVO studies to Europe. I think, actually, though, what we've heard is the overall awareness level is fairly low, and that really represents an opportunity, I think, for Kodiak and KSI. And we do plan, as we shift from this sort of earlier phase now into the registrational study phase and into what we consider ourselves to be a little bit of a pre-commercial stage to really build a set of coherent messages around what is Kodiak or what is KSI. And I think that's probably also true from an investor community standpoint as well, where I think we're well known within the small group of very loyal investors and long-term investors that we have, but we're probably -- and we know we're not well known outside of that. And I think we have our work cut out both for patients, physicians, but also from an investor standpoint to really build that awareness now that we're sort of moving into this next chapter of the company. Thanks.

Okay. And if I just could, just do you have an idea of what percent of current anti-VEGF specialists who give those injections can also themselves do the PDS surgical implantation?

Jason, do you have any ideas? I mean how many of them own their own surgical center, I think, is the question.

Right. So about 20% of the retina specialists in the United States are "medical retina" physicians, i.e., they do not do surgery. I think, Victor, you really identify what I think is the critical question in terms of market access for a surgical device like the port, which is not just who is doing surgery, but what are the economics behind that surgery, what are the opportunity costs of doing the surgery versus other surgeries or the surgery for the port versus treating a bunch of patients with anti-VEGF in the office and all of the other economic complexities that you raised.

We'll go next to Gena Wang with Barclays.

I have 2 sets of questions. First is regarding the DAZZLE study. Can you remind us your trial assumption for patient allocation for each of the q 12-, 16- and 20-week treatment interval? My second question is regarding the 2 inflammation. Just wondering, were both at 5 milligram and also think both patients were on RVO trial? Do you think the RVO treatment interval play a role here? I think RVO patient treatment interval is a little bit shorter.

Right. So the first question is around whether we have an assumed outcome in terms of the proportion of patients in DAZZLE that will read out at 12 weeks, 16 weeks or 20 weeks? And the second question is whether or not on the 2 cases of mild inflammation that we had in the 1b, both of them were RVO patients, and does that somehow provide some caution for RVO? I think when we think about where we may read out for DAZZLE in terms of the durability proportions, I think we walked through, Gena, on the call a little bit what we had thought the benchmarking might be for a next-generation agent, right, with 80% to 100% of patients on q 12 and what we -- somebody might want for a disruptive agent, which would be maybe north of 50% on q 16-week dosing. I think our data is very strong in the durability, and we don't know. I think it seems very likely that we should be able to deliver 50% of patients on 16-week dosing and probably substantially more than that. I think we'll continue to do what we showed before at Angiogenesis, Gena, which is we can apply the DAZZLE criteria to the Phase Ib data as it emerges, right? When we did that before, I think we showed that 75% of the patients would be on q 20-week dosing. Having said that, I don't want to throw that out as our internal objective for the study. So I think we should be able to do very well because our overall durability view is that we're delivering durability data well beyond our expectation. So what we show in the pivotal is, the key is to show that the molecule is safe, effective and that our durability sort of meets what we call that Generation 2.0, right, which is we hope to be very differentiated from EYLEA and in the eyes of the clinical community. I think as to whether the 2 cases of mild inflammation that we've had, which is a remarkably low number, actually, in the Phase Ib and also given how intensively we're monitoring the patients, which is monthly across all 3 indications, Jason, perhaps you'd like to weigh in as to whether or not that may provide any kind of caveats for the molecule across any of the indications. And I think my quick answer would be that we're extremely pleased with the safety profile of the molecule, and we're clearly at or below the stage a LUCENTIS or an EYLEA profile, and we hope to continue to track at that level. Jason, do you have any quick comments?

Yes. I don't think I have anything else really to add, Victor. I think you really hit it on the head. I think in terms of the dosing frequencies, Gena, I don't think that that RVO would be like a particular concern for more or less inflammation than the other ones.

Okay. And were these patients dosed at 5 milligram?

Were both on the 5-milligram dose? Well, the study is randomized 1 to 3 to the 5-milligram dose. So most of everything that happens to everybody is in the 5-milligram dose. But I don't think that that's -- like we haven't seen any particular safety concerns with either dose.

We'll go next to Robyn Karnauskas with SunTrust.

And great job in the presentation. Two quick ones. So number one, as you've collected more data, is there any understanding as to what might define a patient who could do less frequent dosing? Or have there been any learnings from what type of patients are getting less frequent dosing on EYLEA? Anything that you could prospectively define maybe that would help uptake? The second question is on the bispecifics. So I think you're filing your IND next year. I'm just curious if it's possible for us to get data on the bispecifics next year or if it's too early to tell.

Yes. Thanks, Robyn. Let's see. I think on the bispecific on KSI-501, the time line there is really driven by the GMP manufacturing, and that will go into manufacturing early fourth quarter of this year. And so with toxicology and the IND, I think it will be early, I mean, to get clinical data that we would be able to talk about in 2021. I think 2022 is shaping up to be a pretty big year for the company, and we're very excited about the bispecific, and we're also very, very excited about the new triplet therapies built on the ABC platform as well, I think, as we begin to really drive leadership in 3 mechanisms of action, including small molecule and biologic for things like dry AMD and glaucoma as they relate to the retina. So I think our pipeline expansion is very exciting and obviously, a lot of the focus is on KSI-301. To your other question about, well, how could we predict which patients can go long, I think, as Jason mentioned, it's really nice that we're actually in the broad set of treatment-naïve patients because I think it provides like a breadth of data and the ability to have like this high margin of kind of safety as we think about the pivotals. There is no way that we're aware of. Obviously, we have people in the company that are mining the data and trying to better understand as we look for patients that go short versus patients who go very long, right, what could be different. I think in the broad community across Roche and Novartis as well, people are trying to do the same thing. I do think the one thing the PDS people folks did well was to try and identify a fairly probably narrow and homogeneous group of patients that went into their Phase III study and so maybe that's a question to ask them, what are the particular criteria that they used to winnow out all of the other patients that are not really a good fit for their device and whether they really understand what set of patients they actually got because maybe the answer to your question is somewhere in there. Of course, our product works for everybody. So that's the goal.

We'll take our last question from Matthew Luchini with BMO Capital.

Congrats on the data, and thank you for the detailed presentation today. So 2 from me. First, on DAZZLE. At the last update, you made a mention that you'd seen fewer than 5% skipped visits in the spring, and I was just wondering if the study is continuing to see high rate of visit compliance and if there's anything of note that you can share at this point as the pandemic has expanded geographically to a broader swath of the U.S.? And then the second question is on the DME and RVO pivotal. Can you just talk a little bit more about other than perhaps site activation, if there's -- what's left to do in terms of taking steps to getting those studies actually up and running, sounds like, potentially by the end of this quarter?

Thanks. Jason?

Yes. Sure. So first question was around skip visits and COVID. Yes. I mean I think things have stayed very much the same for us, less than 5% missed visits continuing. I think the retina specialists have really worked out, I think, good systems for getting patients to the clinic and working people through the clinic in as safe a manner as possible, obviously, with the spread of the COVID pandemic in the U.S. outside of the Northeast. That's something that we've been continuing to put a lot of attention on doing everything that we can do to support the sites. And thankfully, so far, that number of missed visits has basically been in the very low single digit level. With respect to DME and RVO, obviously, there's always many things to do at the study start-up. I would just say that bringing together all of the different vendors and the systems and all of these pieces of the jigsaw puzzle are all nicely on track, the team is really doing a great job towards our goal of first patient in.

At this time, we have no further questions. I would like to turn the conference back to your presenters for any additional or closing remarks.

No. Thank you.

Thanks so much. Thanks for joining us.

Thank you. Ladies and gentlemen, this does conclude today's conference. We appreciate your participation. You may now disconnect.