Eisai Co., Ltd. (4523) Earnings Call Transcript & Summary

Good afternoon, everyone. I welcome to the symposium on Clarity AD, a Phase III placebo-controlled, double-blind, parallel-group, 18 months study evaluating lecanemab in early Alzheimer's disease. I'm Takeshi Iwatsubo from the University of Tokyo. I'm honored and very thrilled to chair this historic session in Alzheimer's research on therapeutics. Many of you may well remember that on September 27 of this year, the top line results of Clarity AD as a first large completed a positive Phase III trial of disease modifying therapy of AD was announced, which represent a monumental event for patients and their families living with Alzheimer's disease and all scientists in this field, Eisai based in Koishikawa, Tokyo, and then headed by Mr. Haruo Naito, who is also here today, has been at the forefront of Alzheimer's Research over the past 4 decades from developing a symptomatic breakthrough treatment in the 1980s and '90s to the current industry-leading antibody development. The approach rose naturally from the human health care mission. At this present session, 5 distinguished figures from lecanemab and ALOXI's team will present the result of the Clarity AD. First, Dr. Michael Irizarry of Eisai will talk on the background and overview of the study design of lecanemab clinical trials. Next, Dr. Christopher Van Dyck, at the Yale School of Medicine will talk about the top line efficacy results from Clarity AD. Then Dr. Marwan Sabbagh of Barrow Neurological Institute will talk on the safety profiles of lecanemab in early Alzheimer's disease and Dr. Randall Bateman from the Washington University, St. Louis, talks on imaging, plasma and the CSF biomarker assessment from Clarity AD. Lastly, Dr. Sharon Cohen from the Toronto Memory program will wrap up the context of Clarity AD results. We will have some time for panel discussion at the end of the symposium. And I'm very pleased to announce to you all that the primary paper on the trial of lecanemab, in early Alzheimer disease has just been published online in The New England Journal of Medicine at the same time as the opening of this symposium. Now I'm inviting the first speaker Dr. Michael Irizarry of Eisai, please.

Thank you, Dr. Iwatsubo and thank you all for your attendance today. It's really a pleasure to be presenting on the clinical trial background and design of the Clarity AD Phase III clinical trial of lecanemab in early AD. I'm an employee of Eisai. And -- by way of background, lecanemab is a humanized IgG1 monoclonal antibody that preferentially targets amyloid beta protein, protofibrils, these are soluble aggregated amyloid beta protein species that are believed to mediate the synaptic loss and neuronal toxicity in Alzheimer's disease. Lecanemab has a tenfold higher affinity towards protofibrils versus the amyloid fibrils, which they also bind to, like animals binds to. And the binding of protofibrils is over 1,000 fold higher affinity than amyloid beta protein monomers. Lecanemab has microglial effector function and initiates microglia mediated clearance of protofibrils and amyloid plaques. In terms of where like -- where Clarity AD fits within the Alzheimer's disease continuum, Clarity AD is looking at early AD. So MCI due to AD and mild AD dementia with confirmed elevated amyloid in the brain. And this is the same study population that was looked at in the large Phase II dose-ranging study, Study 201. Lecanemab is also being studied in the AHEAD 3-45 study in individuals with preclinical Alzheimer's disease, cognitively unimpaired individuals with intermediate and elevated amyloid in the brain. And it is also currently being studied as a background anti-amyloid therapy in the DIAN-TU NexGen trial, which is in combination with the microtubule-binding region anti-tau antibody from Eisai. Now Clarity AD really built upon the Phase II study and the Phase II study helped us optimize the study population, the dosage, the outcome measures and the hypothesis by way of reminder, the Phase II study was a large Bayesian Adaptive design that looked at 5 doses of lecanemab versus placebo and had adaptive randomization based on effects on a cognitive outcome measure, a composite outcome measure called ADCOMS. The Clarity AD study took what we learned from the Phase II study to really optimize that study design. So the Clarity AD study rather than being a complicated 5 or 6 dosing arm Bayesian Adaptive design is a simple parallel group study testing the optimal dose that was identified from the Phase II study versus placebo. And that optimal dose was based on the dose that showed the fastest amyloid clearance. So the graphs on this slide are from the Phase II study results. The first graph shows the dose and time-dependent clearance, with the greatest effect on the 10 mg per kg IV biweekly dose. The bottom slide shows the effect on CDR sum of boxes in the phase II study. And that result helped us optimize the sample size for Clarity AD, which use the CDR sum of boxes as the primary outcome measure. Also, from the Phase II study and the experience in the field, we implemented optimized monitoring and management of area within the clinical trial, which I'll also describe. So the Clarity AD study design is a global placebo-controlled double-blind parallel group randomized study. The study population was the same study population that was studied in the Phase II. So we randomized 1,795 patients with early AD, meaning mild AD, MCI due to AD and mild AD dementia with confirmed elevated amyloid in the brain. The MMSE score for the study population was between 22 and 30, the 22 lower limit to exclude more impaired patients in the study. And then we also had Wechsler Memory Scale for -- to confirm a short-term memory deficit for the diagnosis of mild cognitive impairment and mild AD dementia. So patients were randomized to either lecanemab, 10 milligrams per kilogram IV biweekly or matching placebo for 18 months of treatment during the randomization phase. And during the presentation, you'll hear us go back and forth between the randomization phase or the double-blind phase or the core phase. All those terms mean that that 18-month double-blind phase of the study. After 18 months of treatment, subjects had the opportunity to enroll into an open-label extension with lecanemab. The primary outcome measure for Clarity AD was the CDR sum of boxes changed from baseline at 18 months. And the key secondary outcome measures were amyloid PET, the cognitive outcome ADAS-cog14 ADCOMS. This is a composite ADCOM that we had used in Phase II, which we'd included in order to bridge back to the Phase II results, and the functional scale of the ADCS-MCI-ADL scale. The clinical trial included a range of sub studies that looked at biomarkers that we had an amyloid PET substudy, tau PET substudy, a CSF substudy. And Dr. Bateman will go over the results of all the biomarker data. We also implemented efforts to increase the diversity in the clinical trial population with regards to race and ethnicity, but also the range of comorbidities to understand how the results translate across different populations and people with a range of background comorbidities and co-medications. The sample size calculation was based on our Phase II data, the details for the sample size determination for the primary outcome for CDR sum of the boxes are shown here. We estimated a treatment difference of 0.37 in all subjects and from a baseline -- from a standard deviation of 2.031. So we were powered essentially to detect the 25% slowing in the CDR sum of boxes and the overall sample size determination of 1,795 patients was based on that. The statistical testing hierarchy to preserve alpha in the clinical trial is listed here. So first was the CDR sum of boxes changed at 18 months. That had to be statistic significant before moving on to the next step of the hierarchy, which was amyloid PET, that had to be statistically significant before we moved on to the ADAS-cog14, then ADCOMS, then ADCS-ADL-MCI. The safety was monitored throughout the study in a blinded manner by Eisai and also through an unblinded independent data safety monitoring committee. The safety evaluation throughout the study included vital signs, physical exam, adverse events, clinical laboratory parameters and EKGs and also focus on the adverse event of special interest, amyloid-related imaging abnormalities or ARIA. So for ARIA, safety MRIs were performed at 9 weeks, 13 weeks, 6 months, 12 months and 18 months. In order to minimize the potential for functional and blinding due to area, we ensured that investigators responsible for the medical management of patients were different from those that are involved in the clinical outcomes assessments. There was an independent medical monitoring team within Eisai that dealt with all the adverse events related to area or infusion reactions. And this was kept independent from the study team that interacted with the study sites and investigators again, to try to minimize any opportunities for functional unblinding and the independent unblinded DSMB convened at regular intervals with a focus on ARIA, including ARIA-E, ARIA-H as well as the impact of anticoagulation, which I'll discuss as well. In terms of the ARIA management, there were these routine MRIs for mild -- radiographically mild asymptomatic ARIA, individuals could be -- could continue dosing without interruption, and those actually tended to resolve spontaneously. Otherwise, if there was moderate or severe radiographic ARIA or ARIA-E or any symptomatic ARIA-E, dosing was interrupted until radiographic and symptomatic resolution. Overall, almost 6,000 patients were screened and 1,795 patients were randomized. 84% of people randomized to placebo completed the study, slightly fewer 81% completed and the lecanemab study arms. And in terms of the primary analysis, which is the modified intend-to-treat population, that include 875 patients who were randomized to placebo and 859, they are randomized to lecanemab. The safety population included everybody that was randomized. In terms of the baseline characteristics, this slide highlights the baseline characteristics of the overall global population irrespective of what they were randomized to and also the U.S. study population. These were balanced between the 2 treatment arms. I just wanted to highlight that there was a wide age range that was within the clinical trial, about 16% -- 20% overall were under the age of 65 and about 13% were over the age of 80. I also wanted to highlight that within the U.S., we had 4.5% of the population being black or African-American and 22.5% of the population being Hispanic. So underrepresented in terms of the black population, but overrepresented in terms of the Hispanic population. And this had to do with efforts that we implemented within the clinical trial to increase the diversity and representation, this included selecting investigators and clinical trial sites that had access to diverse populations, outreach programs, including community outreach or outreach to faith-based groups, and spurred on by COVID implementing more decentralized clinical trial activities to reduce the burden of participation in the clinical trial. And this included activity such as home infusions or remote safety or outcomes assessments that were really tailored to the situation of the patients or the situation of COVID at the time. We also had relatively liberal inclusion exclusion criteria. We didn't want the results to be able to generalize to the elderly population with Alzheimer's disease and this slide highlights the baseline comorbidities and co-medications within the Clarity AD population. And you can see that, for instance, within the U.S., over 60% of the people had hypertension or hyperlipidemia, about 20% had diabetes or ischemic heart disease and many patients had a combination of these comorbidities. We also allowed patients on anticoagulants to enter the clinical trial. We felt that this was important because up to almost 5% of the people who entered the clinical trial were on anticoagulants and about 3% or 4% more were exposed to anticoagulants during the course of the study. So it's important to understand both the safety and the efficacy in this population and so that they may have the opportunity to benefit from anti-amyloid therapies and appropriately understand the benefit risk. And Dr. Sabbagh will provide some of the safety analysis from anticoagulation. And Dr. Cohen will actually provide the efficacy results within that subpopulation as well. Here's more details on the baseline characteristics split up by the placebo and lecanemab group. Some things to highlight is that approximately 80% of the study population was CDR 0.5, 20% was CDR 1. In terms of the clinical diagnosis of MCI versus mild AD dementia, almost 40% of the patients had mild AD dementia. The remaining 60% had mild cognitive impairment. With regards to the ApoE4 distribution, this did reflect the general AD population with approximately 31 -- about 30% of the population being ApoE non-carriers, the remainder being ApoE carriers. This meant that about 54% or ApoE4 heterozygotes and 15% were ApoE4 homozygotes. Almost half of the participants were on cholinesterase inhibitors or memantine and so the treatment was on top of those symptomatic treatments for Alzheimer's disease. I also list the baseline characteristics of some of the clinical outcome measures as well as amyloid PET. The CDR sum of boxes at baseline was approximately 3.2. So we're really at the low end of the scale, which goes from 0 to 18 and that's consistent with the early AD population. It will be important when looking at the results for amyloid PET to understand that the baseline amyloid PET for the clinical trial is approximately 76 centiloids, so the changes that we present are from that baseline of 76 centiloids, and then the other outcome measures are consistent with the early AD stage of disease. So in summary, the lecanemab, Phase II proof-of-concept study provided a robust framework to optimally design the confirmatory Phase III study Clarity AD, it was really intended to be a definitive Phase III study to assess the safety and efficacy of lecanemab based on the simple clinical trial design, the gold standard, CDR sum of boxes is the outcome is the primary outcome, a comprehensive assessment of clinical biomarker and safety outcomes, which we'll be presenting this evening and efforts to really enhance the enrollment of a global population by race, ethnicity and comorbidities. I'd like to spend the next 20 minutes listing the names of all the investigators. Actually -- so really a huge thanks to all the investigators as well as the patients and their caregivers that contributed to this clinical trial, especially during the very challenging circumstances of COVID, about 2/3 of the clinical trial was enrolled during the COVID pandemic. And I also wanted to give thanks to worldwide clinical trials, the CRO that worked very closely with us help deliver this clinical trial also helped us manage the clinical trial during the COVID pandemic. Thank you for your attention. And I now pass it back to Dr. Iwatsubo.

Thank you very much. So the next speaker is Dr. Christopher van Dyck, and he's presenting the top line efficacy results, Chris please.

So, thank you, Takeshi. And I really want to thank the organizers and Eisai. It's really an honor to be part of this presentation, and I want to thank all of you for being here. So as my title says, I'm going to present the top line efficacy results from Clarity AD. These are my disclosures, including my relationships with Eisai. And this slide summarizes the top line efficacy endpoints for Clarity AD, as Michael alluded to, the primary endpoint change from baseline at 18 months in the CDR-SB and key secondary endpoints also change from baseline at 18 months in a biomarker, amyloid PET and then 3 more clinical measures, the ADAS-cog14, the ADCOMS and the ADCS-MCI-ADL. Now this slide summarizes the results for the primary and key secondary endpoints. Again, this trial, Clarity AD, did meet the primary efficacy endpoint, the CDR-SB change from baseline to 18 months. And then below the line, it also met all key secondary efficacy end points for amyloid PET ADAS-cog ADCOMS and ADCS-MCI-ADL. So I can now stop right here. We're done. Actually, I don't want to go through the numbers in this table because we're going to look at these results graphically and individually in the next few slides. So this is maybe the key figure from the study for the primary endpoint, the CDR-SB the clinical dementia rating scale, sum of boxes. And for all of these results that I'm going to show you, remember that they are for the modified intention to treat population that is anyone who was randomized and dosed had a baseline assessment and had at least 1 post-dose primary efficacy assessment with the CDR-SB. And just a little more elaboration on the CDR for anyone not so familiar, this is a global rating scale. It involves an interview with the participant and an informant and it generates scores on 6 different domains, some cognitive, some functional. Those scores are 0 to 3. So the overall score can range from 0 to 18 where higher scores indicate greater impairment. As Michael alluded to, these individuals are more compressed on the low end, starting at 3.2. What this graph shows you, first of all, on the vertical axis, we're looking at, again, adjusted mean change from baseline in the CDR-SB. We're looking at standard error bars, and we're looking at P values that come from the mixed model of repeated measures or MMRM, that's detailed in the footnote. So what do we see here? Well, I -- we see in the placebo group, that people on a placebo were declined by 1.66 points over the course of 18 months. We see that the actively treated group declines by 1.21, a difference of 0.45, which is highly statistically significant and the significant differences emerge as early as the 6-month time point. Now moving on to the key secondary, starting with a biomarker endpoint amyloid PET. And remember, Randy is going to present all the biomarker results, I'm just combined to one the key secondary amyloid PET, which was an efficacy outcome. And this is a substudy involving 39% of the total. These individuals had to be randomized and dosed. They had to have a baseline PET and they had to have at least one post-dose PET-scan. So fibrillar amyloid burden was evaluated from 1 of 3 different approved tracers -- florbetaben, florbetapir, flutemetamol and since we had to pool data from multiple tracers, multiple sites, the SUVr values were converted to centiloids, which you've heard about a few times. I just want to review a little more what it is. So this is a scale that is anchored with a 0 value for what normal young control show on average and 100 is what people with mild to moderate AD dementia show on average. And as Michael said, these folks started at about 76 centiloid. The threshold of positivity, although entry was determined by visual read. We know that entry corresponds some quantitatively to about 30 centiloids so to keep that in line. And what this figure then shows us is that the placebo group actually gains a wee bit of amyloid 3.6 centiloids, and the treated group shows a sizable decline, 55.5, resulting in a drug placebo difference of 59.1 against -- again, highly significant. And in particular, at the end of the study, at the end of 18 months, those in the treated group had an average centiloid value of 23. So below the threshold of amyloid positivity that would get them into the study in the first place and significant differences were evident as early as the very first post-dose PET scan at 3 months. So additional secondaries, now that the clinical secondary starting with the cognitive measure, the ADAS-Cog. This is Score 0 to 90, higher scores are worse and what we see here is that the placebo group declines on the ADAS-Cog by 5.58 points. The lecanemab-treated declines by 4.14, as I recall, a difference of 1.44, again, highly significant and significant differences emerge as early as 6 months. The ADCOMS for those not familiar, this is actually derived from multiple other assessments, the items on the CDR, the ADAS and the Mini-Mental State exam. And it's really engineered to try to produce maximal sensitivity to change in this early AD population. And it scored 0 to 1.97, higher scores or worse. And this figure shows us that the placebo group declines by 0.214 and the actively treated group 0.164, as I recall, producing a difference of 0.05, again, highly statistically significant and significant as early as the 6-month time point again. And finally, the functional measure, the ADCS-MCI-ADL, this is scored 0 to 53. And for ones, it's actually a lower score that's worse. And this graph shows us that for the placebo treated, they declined by -- they worsened by 5.5 points at 18 months, the active group, 3.5 for a difference of 2.0 highly significant and represents a 37% slowing of decline on daily functioning by lecanemab, in 18 months, and significant differences again emerged at 6 months -- by 6 months. So those are the top line efficacy results for Clarity AD. What I want to do with my remaining time is talk about some sensitivity analyses that we did as well as some subgroup analyses, and these are really to look at the robustness of these effects to multiple factors, and then to look at their consistency across the range of subgroups. So here, we're looking at prespecified sensitivity analysis. Remember 1 more time that up until now, we're talking about modified intention to treat population dosed had a baseline and had a post-dose CDR. And these are sensitivity analyses for the CDR-SB itself. So the previous -- the primary analysis is what's shown in the top row above the line for comparison with these 4 different sensitivity analyses. And the first one, we wanted to look at what the impact would be of missing data. So for that, I rank analysis of covariance was performed with missing data imputed by multiple imputation, the second sensitivity analysis instead of modified intention to treat was standard intention to treat, that is all randomized individuals regardless of whether they got dosed. In fact, all of them did. And regardless of whether they had a post-dose CDR-SB and some did not, of course, because they may early terminate before that assessment was done. The third sensitivity analysis was to look at the impact of ARIA-E and primarily for functional unblinding because it could unblind, it certainly unblinds, it could unblind participants and study partners and possibly clinical raters, although they were rigorously kept unaware of safety data. And in that case, what was done was the sensor any assessments that occurred after occurrence of ARIA-E. The fourth sensitivity analysis was to look at the impact of COVID, and again, Michael already spilled the beans but there was this thing a global pandemic. Some of you might have heard about 2020. And so much of the trial was conducted during that's really remarkable, I think, especially what the participants did for the study, but here, we wanted to look at the impact of the pandemic on missed dosing in particular, because some of the sites at the peak of the pandemic were shut down temporarily, my site included. And -- so in this case, individuals who missed at least 3 consecutive doses during the peak of the pandemic were excluded from the analysis. And with that -- all of that in mind then, if you think about these 4 sensitivity analyses, what you can really see is that -- should have been using my pointer. But in this column here, maybe I shouldn't have been using pointer. The treatment difference in 18 months is really pretty similar across sensitivity analysis and similar to the primary, the percent slowing is really pretty similar between 25% and 30%. And of course, the p-values are all highly significant. So in brief, these analyses revealed the results are very robust. So quickly then to go over some subgroup analysis. And this was done to look at the consistency across a number of subgroups. This slide, first of all, shows sensitivity analysis for the CDR. For those subgroups that were used for the randomization stratification. Those are shown on the left column, and these are first parts which show adjusted mean difference in CDR-SB versus placebo with 95% confidence intervals. And if you just scan down the center, the forest plot here, you can see that all of the values favor lecanemab. So all of these factors used for randomization stratification, Alzheimer's meds at baseline clinical subgroup that is MCI versus dementia, ApoE4 carrier status region of the world, all the subgroups showed similar results. And these then are the very same analysis for 2 of the key secondaries, the ADAS-cog, the ADL, and again, if you just scan down the forest plot, you can see again that all of these favored lecanemab for both of these key secondaries. The other subgroup analyses we did were for variables, other kinds of variables of interest that weren't used for randomization stratification, but still important to Alzheimer's disease, and without going into all these in details, if you scan down the forest plots, you can see that nearly all of them favor lecanemab. One exception ApoE4 homozygotes. Now I don't think we should make too much of this. I certainly don't. This is -- with so many subgroups were bound by chance to have some that go the opposite direction. And I think for ApoE4 homozygotes, first of all, it's a small subgroup. It showed unexpectedly slow decline with the CDR-SB in the placebo group, which may have been a factor. And maybe importantly, is that for pretty much all other endpoints, all other assessments, the ApoE4 homozygote group did favor lecanemab, including if we look at the next slide, which is the same analysis with the ADAS-Cog and the ADCS MCI-ADL. Now you scan down the forest plots and every last one of them favors lecanemab, including for the ApoE4 homozygotes. So then just briefly, these subgroup analyses show that the results are very consistent across a broad range of subgroups. So in summary, lecanemab treatment met the primary and secondary endpoints versus placebo at 18 months with highly significant differences starting at 6 months. For the CDR-SB, lecanemab reduced clinical decline by 27% on amyloid PET, lecanemab reduced amyloid burden starting at 3 months. For the ADAS-cog14, lecanemab slowed cognitive decline by 26%, for the ADCOMS by 24%. And for the ADL slowed functional decline by 37%, and the results were consistent across a broad range of endpoints and subgroups. And in conclusion then, lecanemab reduced markers of the amyloid in early AD and resulted in less decline on placebo on all measures of cognition and function at 18 months. These differences were observed as early as 6 months, and these findings encompass a broad range of endpoints and subgroups. And I thank you for your attention.

Thank you very much. The next speaker is Dr. Marwan Sabbagh. He will talk about the safety profile. Marwan, please.

So today, I'm going to spend time walking you through the data as it comes to safety. I know there's been a lot of discussion, a lot of buzz, a lot of energy around safety-related issues. So I'm going to walk you through it in a deep dive of the data so we can understand what we -- what the science tells us. I do not have proprietary stake in the company or the investigational product. I'm going to walk you through the adverse event rates, the laboratory abnormalities, vital signs, ARIA-E and ARIA-H which are events of interest and infusion-based reactions. As Mike Irizarry told you, there was a very carefully monitored study with the DSMB that had continuous input at all times. And so the rigor of the science and the safety we're always there, and we'll kind of work through that. My first thing I want to talk to you about is the -- is this slide. And I'll actually ask you to first go to the footnote. And what I want to show you here is that the causes of death are in the footnote. And if you read the lecanemab group, it's death, cerebrovascular accident, myocardial infarction, respiratory failures, metastasis, meninges, COVID-19. And if you look at the placebo group, there was respiratory failure, myocardial infarction, metastasis of bone, macro hemorrhage, endocrinal hemorrhage, COVID and pancreatic cancer. I'm starting to hear because I want to make the point that in the randomized placebo-controlled trial, the 301 study, there were no ARIA related deaths in the treated group. Okay? And no cause of death related to ARIA. If you actually then look back to the rest of it, you see that the SAE rate is pretty similar between the treated and the placebo group except for the ARIA related SAEs. And if you take out the ARIA related SAEs and infusion-related SAEs, you see that the safety -- the SAE rate is effectively the same in both groups. Let me now talk to you about the most common event rates. And let me first show you look at the right-hand side of the slide. In baseline and in all the treated arm -- treated visits, there were no differences in laboratory events, EKG, vital sign parameters or any other notable events. So the perception, of course, is that lecanemab at least from a vital science standpoint and laboratory standpoint is generally safe. If you look at the bottom of this slide, you see that the incidence of common things like headache, falls, UTIs, COVID were the same amongst both groups. Again, no heavy position to either 1 group. And if you look at the top, you see, as expected, there are more infusion-based reactions ARIA-E, ARIA-H, and I want to point out that fourth line, which is that if you take out ARIA-E combined with ARIA-H, just looking at ARIA-H alone, there is no difference between treatment and placebo in terms of lecanemab having more ARIA-H. If we look at the immunogenicity and the infusion-related reactions, I'll first direct you to the bottom of the slide. First thing you notice is that there are very few neutralizing antibodies. Second, that there are only 10% treatment-emergent positive antidrug antibodies with very low titers and that there were no impact of immunogenicity on efficacy or safety endpoints, meaning that generally, there was very little in terms of immunogenicity. When we look at the top right-hand side, you see that the lecanemab group had more infusion-related reactions, 26% of the cohort. Almost all of them occurred in the first treatment dosing and were mild in mild to moderate, 96%. They only occurred mostly in the first treatment and did not occur in subsequent treatments. And that whether you pretreated or didn't pretreat with things like acetaminophen or antihistamine, it did not change the recurrence rate for subsequent treatment. So if you had infusion-based reactions, they almost always occurred with the first dose. This is the slide that everybody wants to talk about, which is the ARIA-E incidence rate. So I'm going to kind of walk you through this because I like numbers. The first thing I'm going to show you is that the denominator for ApoE4 carriers is 620 out of 898. That means that the overwhelm majority of the sample were ApoE4 carriers. I say this to you because the 201 study had fewer ApoE4 carriers. And therefore, when you see this, that was always suggesting that we would have differential effects or concerns about safety, and the answer is no. We do not have differential concerns about safety. If you take the ApoE4 carriers that 620 and break it down, you see that non-carriers only had an ARIA-E incidence rate of 5.4% and heterozygotes only had a ARIA-E incidence rate of 10.9% and that, that 15% was driven a lot by the homozygotes. When you look at it in totality, the ARIA-E rates were mild to moderate radiographically, which I'll show you in another slide, and were clinically asymptomatic 78% of the time. When they were symptomatic, they were symptomatic less than 3% of the time, and it was a mild, which we'll talk about, and we include headache, visual disturbance and some confusion. Another thing we want to think about is when the time to event analysis. When does the ARIA rate occur? And the answer is in lecanemab treatment, 90% of them occurred within the first 180 days, and a majority of them were in the first 90 days. So we can predict when patients will have -- if they're going to have it, it will be in the first 90 to 180 days. And important other thing is that regardless of the ApoE4 carrier status, the majority of them resolved, 54% in 90 days and 81% in 120 days. So we know, again, when are the ARIA events likely to occur? ARIA-E events likely to occur? And when are they likely to resolve? Now let's spend a minute talking through this slide, again, a lot of numbers in front of you. So you see here that the denominator is 898. The ARIA-E rate for the numerator is 113. And if you look at the radiographic severity that on the right -- on the middle of right, that's 37, 66, and 9. That means that the overwhelming majority of the ARIA events were mild or moderate in radiographic severity. And clinically, when they were symptomatic, only 25% were clinically symptomatic, mild to moderate, 10, 12 and 3. So again, the majority of the sample were mild to moderate in clinical severity and mild to moderate in radiographic severity. And when you actually look at in the ApoE4 carriers, you see the same thing that 90% of the -- 98 out of 620 had an ARIA-E--related event, and the majority of them were either low 9.2% for homozygote or 1.7% for heterozygote, again, slow numbers. And so I want to point out to you that it is mild to moderate radiographically, mild to moderately clinically and is related to ApoE4 carrier status. Microhemorrhages or ARIA-H is a little different. As I said before, if you did not have an ARIA-E event, the microhemorrhage rate or ARIA-H was essentially unchanged from placebo. And when it did occur, 17% versus 9% in the placebo group occurred. And that most of them were clinically asymptomatic, as you see, symptomatic ARIA-H was only 1.4% of the samples in lecanemab group versus 0.2% in the placebo group. So again, very minimally symptomatic and was driven by the ApoE4 carrier status. Now when we look at the deaths. This is important because this is where all the controversy and excitement and reporting has been. Let me point out that there was 1 cerebral macro hemorrhage related death that occurred in the randomized study, and it occurred in the placebo group, not in the treated group. So when you actually now add -- and all the comments I've made so far only talk about the randomized study. When we look at the open-label study, and we added -- we see that there were 2 deaths that occurred with cerebral macro hemorrhage. The first of them was a 65-year-old female ApoE4 homozygote with left middle cerebral artery territory infarction or occlusion, I should say. This occurred in the open-label extension. That patient received tPA -- tissue plasminogen activator, my neurology colleagues know that it's a thrombolytic, which means it has a known complication of a hemorrhage and the hemorrhage did occur and the patient expired. But it can -- it was in the setting of TPA in lecanemab. The second one was an 87-year-old female, 87-year old male, ApoE4 noncarrier on Eliquis for atrial fibrillation who had many falls and pneumonia and developed a macro hemorrhage 4 months after initiation of lecanemab open-label treatment. Eliquis was stopped. The patient subsequently because he had atrial fibrillation had a myocardial infarction and died from cardiopulmonary arrest. These cases show that the causality with lecanemab is little difficult. And I understand that people want to say that there's some causality, these things are continue to be explored. So the point here is that the rate of macro hemorrhage is low. It's only 0.6% to 0.7% compared to placebo. It does go up with anticoagulation and that might be a relative risk that needs to be managed, might need to be discussed. We know that there was no clear relationship between macro hemorrhage and ApoE4 carrier status baseline MRI or timing of treatment. And it will be a caution that patients on anti-coagulation might need some further consideration. In summary, lecanemab was generally well tolerated. Most adverse events were infusion-related reactions, ARIA-H and ARIA-E and headache. The incidence of ARIA incidence was within expectations and was low with the ARIA-E rate of about 12.6% in the lecanemab group versus 1.7% in the placebo. Symptomatic ARIA-E was less than 3%. Most of them occurred within the first 90 to 120 days or 180 days and most of them occurred -- resolved within the first 120 days. ARIA-H was in combination with ARIA-E with a 17% incidence rate and symptomatic was only 0.7% in the treated group. There was no imbalance when it was isolated ARIA-H alone. And when you looked at them, it was less common ARIA-E and ARIA-H in ApoE4 noncarriers with higher frequency occurring expectedly in ApoE4 carrier homozygotes versus heterozygotes. And in summary, infusion-related reactions, ARIA-E and rare macro hemorrhage are important adverse events that can be seen with lecanemab treatment and if approved, clinicians, patients and caregivers will need to understand the incidence, monitoring and management of these events. Thank you.

Thank you very much. So the next speaker is Dr. Randall Bateman and he will talk on imaging and biomarkers. Randy, please.

Okay. Shown here are my disclosures. I'm going to be presenting the imaging and the biomarker results from the Clarity AD study of lecanemab. As part of the lecanemab's development, there are a number of ways in which biomarkers are used, and that includes in patient selection, for example, in the Clarity AD study with amyloid PET or CSF 42/40 in the prevention studies, for example, AHEAD 3-45 using blood plasma 42/40 measuring the target engagement and pharmacodynamic measures, we saw the results -- the top line results of amyloid PET and also the CSF and plasma measures for amyloid beta 42 to 40. And in development as a protofibril assay, and in order to measure the pathophysiology of the disease and the pathology that exists in the brain, there's a number of different measures, including PET scans, amyloid and tau-PET as well as CSF and plasma amyloid measures, tau measures, measures of synaptic function and astrocyte function and then more traditional measures of brain volume by MRI. And finally, the clinical measures were already reviewed, and I'll put that in context to these biomarkers. As shown here, the various measures that were performed in the trial. We're going to be reporting on all of these measures as well as the way they were measured shown in the footnotes here. The amyloid biomarkers are shown on this slide for the soluble measures. You see on the top 2 graphs, the cerebrospinal fluid amyloid beta 42 and the cerebrospinal fluid amyloid beta 42 to 40 ratio -- and the bottom left is the cerebrospinal fluid amyloid beta 40 measures. What you'll note is that lecanemab had no effects on amyloid beta soluble levels for A beta40 however, significantly increased the values of CSF amyloid beta 42 at 12 and 18 months. Correspondingly, in the top right graph, you see the 42 to 40 ratio also improved at 12 and 18 months and the corresponding plasma amyloid beta 42 to 40 ratios in the bottom right. Overall, this indicates early and sustained amyloid reversal effects and likely slowing the process of aggregation of amyloid beta in the central nervous system. Shown here is the amyloid PET graph that was reviewed before. I'll make a few other points here. One is that the population that came into the study had centiloid levels of around 75 to 77. And already at 3 months, there was a substantial drop of 20 centiloids towards normal. By 6 months, the drop was closer to 35 centiloids. And by 1 year, the drop was 50 centiloids, putting them close to the threshold to cut off. By the end of the study at 18 months, a further drop was seen, and you can see the percent of amyloid negative, the people who fell below the amyloid cutoff of 30 centiloids in the bottom table, -- you can see that percent increase so that by 1 year, most people, 54% were amyloid negative by quantitative cutoff. And by 18 months, 2/3 of the people were amyloid negative. Shown on the right, placebo and lecanemab values for representative individuals at baseline in 18 months are the images where you can see the relative reduction of amyloid removal in this trial. For the tau biomarkers, there are both soluble measures of tau and CSF and plasma as well as insoluble measures by tau-PET. These measure 2 different kinds of tau processes. One is the phosphorylation of soluble in terminal to mid-domain tau. It seems to be a response to amyloid plaques in a market of early tau pathology. You can see the changes here over time that as people progress in their disease, they tend to increase their soluble levels of phospho tau 181, both in the CSF and the plasma, but with treatment at 1 year and at 18 months, you see substantial declines of soluble phospho tau 181 towards normal. And so there's a large delta here throughout the time course both in CSF and plasma, and that this indicates that the downstream pathophysiology of removing amyloid changes the response that the brain has in phosphorylating tau. Tau PET is another measure of Tau. This measures a different form of tau that's aggregated in the brain, and it's largely comprised of the microtubule binding region in the brain in the form of neurofibrillary tangles and dystrophic neurites. The different regions were measured in this study and a comparison across all regions was done in a prespecified way to look at the changes. Although there were trends in most regions of less talent accumulation over time, it was the temporal lobe in the study that reached significant levels of less tau accumulation. And that's shown on the right in the medial temporal, the meta temporal and the full temporal lobe on the right. So this data indicates that in certain regions of the brain, notably the temporal lobe and suggestive changes in other areas that the accumulation of tau pathology over time with amyloid removal is blunted and slowed. This is quite important because in Alzheimer's disease, we define the disease as plaques, tangles and dementia and 2 of these pathologies have been slowed or reversed. The neurodegeneration markers are shown here. This includes characterization of cerebrospinal fluid and plasma neurofilament light chain as well as cerebrospinal fluid total tau levels, similar to the phospho tau levels, you see total tau increases over time as the disease progresses. But with amyloid removal with lecanemab, you see decreases in the amount of total tau in the CSF at 12 and 18 months that are highly significant. Neurofilament was not significant in the differences, although there were some trends noticed in the plasma NFL towards 18 months, subsequent follow-up will be needed in order to determine if this is a sustained or a robust effect. And finally, for the soluble measures, we have astrocytic and synaptic biomarkers of the disease process. More recently activated astrocytosis and the secretion of glial fibrillar associated protein GFAP has been shown to be an accurate measure of disease progression. And you see that here during the trial, plasma GFAP has increased and stay sustained throughout the trial. But with treatment with lecanemab, reversal is seen and plasma GFAP lowers at highly significant levels at 6, 12 and 18 months. Similarly for synaptic measure, neurogranin, you see the CSF neurogranin levels are largely stable during the trial, although these are altered at high levels. And over time, that also lowers towards normal at 12 and 18 months in a significant way. Finally, the volumetric MRI, the measures of attributed free are shown here, representative areas, including whole brain volume, cortical thickness and lateral ventricular volumes are shown on the top panels while the hippocampal areas are shown on the bottom areas. Across the brain in general, increased atrophy was seen during treatment with lecanemab at 6, 12 and 18 months for whole brain, cortical thickness and ventricular volumes. In contrast, hippocampal volumes did not show that same increased atrophy and actually were significantly less atrophied by the end of the trial at 18 months. So now I'm going to discuss my interpretation of these results. Lecanemab has impacted various measures in CSF plasma and imaging biomarkers of Alzheimer's disease pathology and pathophysiology in the domains of amyloid, tau neurodegeneration, astrocyte and synaptic pathophysiological measures. Lecanemab had beneficial effects on a multitude of these biomarkers including the amyloid, tau and other pathophysiology measures. The neurodegeneration markers of NFL and the volumetric MRI gave a mixed picture, potentially because these measures may take more time to respond, they may be compounded by having amyloid removal and will be followed in ongoing studies. The drug directly removed amyloid plaques as measured by soluble and amyloid PET measures, it also had downstream effects on tau pathologies measured by PET, soluble measures of tau pathophysiology in the CSF and the blood, synaptic dysfunction and astrocyte activation. I believe that these findings do indicate a biological disease modification by improving the amount of amyloid and tau pathology during -- in the course of the disease, while also impacting the pathophysiology of Alzheimer's disease. Disease modification is further supported by the clinical benefits shown in cognitive and functional scales and also the Phase II data where it was shown that during a gap period, the benefit is sustained during that time. This will be further explored in the ongoing open-label extension with biomarkers, clinical and cognitive measures to better understand the full effects of lecanemab. Thank you.

Thank you very much. The last speaker of this session is Dr. Sharon Cohen, and she will talk about the context of Clarity AD results. Sharon, please.

Thank you, Takeshi. Thank you all for being here with us tonight. So as you've heard as the last speaker in this panel, my job is to put things in context and make some sense of this. So let's begin. Here are my disclosures and relevant to lecanemab. I was a site investigator for the 201 study and [indiscernible] for the Clarity AD study as well as for the AHEAD Prevention Study, and I serve on the aducanumab -- sorry, I serve on the lecanemab, Global Strategic Counsel excuse me. Sorry about that. I'm glad that was comic relief. Okay. So no discussion of context of results of a clinical trial can take place without us reflecting on this huge unmet need. And I know that many of you can recite this in your sleep, but it's important that we keep this front of mind. This is an urgent health care priority. It is 10 years since the World Health Organization declared it such a chronic, progressive, disabling and fatal disease, 6 leading cause of death in some countries, accounts for 60% to 80% of all cases of dementia and the numbers are rising 55 million according to Alzheimer's Disease International rising to 75 million, just around the corner, 2030. And Alzheimer's disease is a significant cause of economic burden globally. And beyond the costs, this disease has severe impact on patients, families and health care systems, and established treatments, largely symptomatic treatments are insufficient. They target distal neuropathology, they are indicated primarily for the dementia stage of the disease. So 1 must get worse before 1 is eligible, if 1 is diagnosed early at the MCI stage. And the symptomatic treatment effect is modest if 1 achieves 1 and is transitory as the disease continues to progress. So if we reflect on what some near-term treatment goals would be let alone the long-term treatment goals of preventing the disease in those who don't have it and curing it in those who do, let's be realistic for the moment and say, we would like greater impact on symptoms, yes, greater magnitude of improvement and longer duration of improvement. And with respect to disease slowing, we would like to maintain or stabilize abilities for longer periods than the natural history and delay or prevent onset of more disabling and costly stages of disease. And perhaps above all, we want to maintain or improve quality of life for patients and families. Now it might be interesting to reflect on the FDA and the EMA terminology with respect to disease modification because there are many discussions going on about this. So what the FDA has to say is disease modification in Alzheimer's disease means persistent effect on disease course and direct effect on the underlying disease pathophysiology. EMA, fairly similar, persistent delay in the underlying neuropathological process delay of clinical decline. So now turning to the Clarity AD results, of which you've heard quite a bit. On the primary outcome measure, the CDR sum of boxes, a global measure of cognition and function, we see at 18 months, this 27% slowing of disease relative to placebo. But keep in mind that, number one, we're seeing this treatment effect as early as 6 months. It's growing in magnitude until the duration of the trial, 18 months and the CDR is not made up of 1 domain. There are 6 domains, all of them relevant for individuals with Alzheimer's disease, memory, orientation, judgment problem solving, community affairs, home and hobbies, personal care. And we see on an item level analysis that there is a positive treatment effect, a slowing of decline across all of these domains in the order of 21.2 to 29.9% slowing depending on which domain. And we're not seeing inconsistent results with some domains showing a treatment effect and others not. We love that as clinicians consistency. And we're not seeing any 1 domain driving the overall effect. If we look across comorbidities, as you've seen this data, and over 50% of the Clarity AD population had 2 or more comorbidities, common health problems. We see that there is a treatment benefit across all of these groups and in the groups who are anticoagulated, although there may be more risk associated, you heard that from Marwan Sabbagh in terms of the anticoagulation increased risk, but we also see a treatment benefit when we're looking at efficacy. So now if we look at the specific measures of cognition and function, 2 separate measures from the primary outcome measure CDR sum of boxes here, we're looking at the ADAS-cog14 administered to the patient, the ADCS-MCI-ADL, informant based questionnaire. We see 26% slowing on 1, 37% slowing on the functional scale. I like that again. I'm wearing my clinician hat, 37% slowing of decline on things that are meaningful to patients in day-to-day life. And we see the same trend of 6 months, we start to see the separation of the curve statistically significant and this grows over time. Now here, I'm talking about the global CDR as a staging scale, not the sum of boxes. CDR time to worsening. So this is a time-to-event analysis here. And we see that those patients who entered and most of them, you heard from Michael Irizarry, were at CDR Global of 0.5. The others were 1.0 or so equivalent, if you will, to mild cognitive impairment and mild stage AD dementia, the time to progression to next stage or to worsening on the global CDR was slower and continue to slow over the 18-month course of the disease. So 31% lower risk of converting to the next global stage on CDR, hazard ratio of 0.69. If you look at the extrapolated slope projection, this is an analysis looking at observed data and extrapolating out from 18 months to 2 years. We see 32% slowing of slope annually. And therefore, the projected treatment effect at 25.5 months, based on slope is 0.68 whereas at 18 months based on slope that is 0.48. And so again, this trend towards increasing benefit or what some refer to as cumulative benefit, the longer you're treated, the greater the treatment effect. And here, I'm referring now to published data on the 201 study, and this will be updated with a publication on Clarity AD, but right now, you're looking at the 201 study. This is lifetime health outcomes, trial data modeling simulation. The first author is [indiscernible] And what this shows is slower rates of disease progression with extended duration in the MCI and mild AD stages of disease in those treated with lecanemab plus standard of care versus those on standard of care alone. And -- sorry, forgive me if I didn't say the shortened duration of moderate and severe stages. In the meantime, to mild, moderate or severe dementia in the lecanemab and standard of care treated group versus standard of care only was 2.51, 3.13 and 2.34 years, respectively. So now we're not talking about weeks or months, we're talking about gaining years in milder stages of disease, on this projected model. And this projection also predicted lower rates of institutionalization. Again, this will be updated with the 301 study. And finally, something dear to my heart and to those of you who are clinicians, Quality of life is so important, and it isn't necessarily captured on cognitive or even in ADL scale. How do you feel about your relationship in the family, where you're living about your friends? And here, we have patient responding to quality of life surveys and also caregiver on the [Sarah Burton] interview, sorry. And we see consistent benefit statistically significant and meaningful in my eyes across all of these scales. And for the quality of life AD, I just want to point out the 56% less decline. This is the scale where patients get to weigh in on how happy they are in their life. Is there a fun factor, how are they with their friends, really important things that we don't measure generally. And [Sarah Burton] interview 38% less decline. So in summary, the convergence of evidence across multiple measures of cognition, function, disease progression health-related quality of life, caregiver burden and biomarkers demonstrates that lecanemab treatment may offer meaningful benefits to patients, care partners and society. And I won't repeat the numbers here, although they all go in the right direction. But the results of Clarity AD, support the role for treatment for therapy in early Alzheimer's disease stages of an amyloid protofibril targeted therapy. The results of Clarity AD are consistent with the FDA and EMA terminology for disease modification, and they provide momentum to our field at large, it behooves us to continue to push these treatments forward. And these results really also ask us to focus on a paradigm shift of diagnosing earlier in order to achieve disease slowing and to extend milder stages of disease. Thank you so much for your attention.

Thank you very much, Sharon for a wonderful presentation. So unfortunately, the closing time is approaching that on behalf of the audience, I would like to make a few questions to the panel that I will take. So I have a question to Dr. Irizarry. Given the information presented, it is important to understand the incidence and monitoring of the ARIA, how is the ARIA monitored and managed?

Thank you, Takeshi Iwatsubo. Ultimately, so ARIA is a very important adverse event that's associated with anti-amyloid treatments and with lecanemab. And the key features will the benefit risk discussion of the physician with the patient, the monitoring and the management. And that would depend both on the label if it were approved and of the appropriate use recommendations from the field. Dr. Sabbagh presented a lot of data, the part of the benefit risk discussion is understanding the incidence of area, the 12.6% incidence, the increased incidence in ApoE4 by ApoE4 allele and also the incidence of symptomatic area, which is 2.8% overall, but again, higher in ApoE4 homozygotes. We also had discussed the incidence of ARIA-H and the most important form of ARIA-H is the incidence of macro hemorrhage, so bleeds into the brain greater than 1 centimeter. We had been following that closely during the clinical trial. Dr. Sabbagh presented a comprehensive analysis across all the lecanemab studies and the open-label extension up through last month. Overall, the incidence rate of macro hemorrhage was 0.6 or 0.7 in the lecanemab-treated group, which was higher than in the placebo group, which was about 0.2 and importantly, in people that were on both anticoagulation and lecanemab, the incidence rate of macro hemorrhage was about 2.4% to 3.6%. So that has to be balanced with the benefit that has been shown in those particular subgroups. And this is a discussion that will need to be had between the clinician, the caregiver and the patient as they take the drug. And then in terms of monitoring, we expect that MRI monitoring will be weighted toward early on in treatment, where that's the highest risk of ARIA. Clinicians and patients and caregivers will need to be vigilant for symptoms of ARIA because that would trigger a clinical assessment and MRI. And then understanding if the MRI does show ARIA how is that managed in terms of interrupting treatment or whether to continue treatment through a symptomatic mild ARIA and the type of monitoring that's required before initiating treatment again, if treatment is paused. So I think all these factors will be in consideration in terms of the appropriate safety monitoring and risk discussions with patients and caregivers.

Thank you very much. So I have another question to Dr. Cohen. This may be the last question. So who would be the appropriate patient to consider treatment with lecanemab?

Thank you, Takeshi. So we are heading towards precision medicine here. We need a biologic diagnosis of Alzheimer's. So someone must be amyloid positive to qualify. I think that's a given regardless of the fact that we don't have a label yet. And we need to be in the mild stage of disease. So we need to have some way to stage people. And by mild stage I'm talking here about MCI or mild dementia with amyloid positivity. And I would add to that from a safety perspective, we must have people who can undergo MRIs, that's to exclude some people because MRI is going to be imperative not to gauge benefit but to mitigate safety. I think that we're very fortunate that the Clarity AD trial approximated as best as these trials can a real world population with an age range that's broader than most of these Alzheimer's trials, 50 to 90 a larger inclusivity of diversity than what we often see and multiple comorbidities, which are often exclusionary in similar trials. So patients had an opportunity to join this trial, but we had the opportunity to learn that they can actually benefit from this therapy rather than this being a question mark. So I think, again, where my clinician had this gives me confidence that there's a broad swath of people out there who have common variety of Alzheimer's, not the ideal as trial participant will benefit. But of course, there will be subgroups at more risk, and it will be important to have risk-benefit discussions with the skilled clinician to inform patients and families about individual risk.

Thank you very much. So the organizer kindly allowed us to take very few questions from the floor. Anyone who has -- yes, please.

I'm [Sohail Rasool] from [indiscernible] so I'm going to repeat the same question, which I posted at Eisai's forum. So what particular this lecanemab binds to protofibrils. So what may be particular -- protofibrils, mixtures, [indiscernible] So do you have any idea about what particular aggregates this lecanemab binds, molecular weight and does lecanemab has any effect on normal tau and hyper-phospho tau as well?

Yes. So I'll try to address that, although there are better people that can answer that. So lecanemab binds the small and large protofibrils that are aggregates that range from 50 to 200 kilodaltons and it may vary in terms of the absolute size of that. The other question was the impact on tau or phospho tau and I believe Dr. Bateman had presented the data in terms of those markers that were assessed within the clinical trial.

Thank you, Raymond Schwartz, a neurologist and geriatrician from Sydney, Australia. Congratulations to Eisai and to team Clarity. My question perhaps to Sharon Cohen is what should we as clinicians tell the individual patients and their carers when considering going on to lecanemab down the track. What the benefit is and the risk is to them as individuals mindful of relative and absolute risk reductions?

Yes. Thank you very much for your question. So what I would say, to patient in front of me is that we're trying to keep you the best you. The things that you enjoy that you can still do now. My hope is that in months from now, maybe this time next year, you will still be doing these or most of these things. And if that is the case, then we are winning because you have a chronic progressive disease, and we want to buy time, while you're still well to enjoy this.

[indiscernible] from UCSF. Congratulations to all of you on a very transparent presentation and extremely exciting results. My question was about the correlation between amyloid lowering and clinical benefit. Is that something that you could comment on?

So we have looked at that in the Phase II and at a population group level, which is the most efficient way to look at it because you can adjust for placebo decline and just look at the treatment effect. The most efficient way for us to look at that will be through PK/PD outcomes modeling, which is in progress, and we hope to be able to report soon.

[Anton] from the University of Rochester. So question to the panel. Maybe the biggest surprise to me was the ApoE4 homozygote group and the clinical response on the CDR Sum of Boxes. So if you can speak a little bit to that and particularly then kind of the comparison of clinical benefit and risk for that group in particular.

So [Anton] nice to hear from you. I thought I said about all I know on that. But -- so let me just reiterate, I think that, first of all, in subgroup analysis, we really have to take with caution, given just the huge number of them, and some are going to go surprising directions by chance. I think that, again, as I alluded to, it is a small subgroup, so maybe more susceptible to those issues. When we did look at the placebo decline, the decline in the placebo arm of the ApoE4 homozygotes, there was surprisingly slow decline on the CDR, lower-than-expected decline in the placebo Group for the CDR-SB, but looking at all the other endpoints like the ADAS, like the ADL and others, including even biomarker endpoints, every last 1 of them that I saw favored lecanemab. So I just don't think at this point, we jump up and down and say, this isn't the drug for homozygotes.

Yes. And it was different than the Phase II study as well -- in the Phase II study, 4 carriers, what actually better?

Well, in the Phase II study, we really can't comment on the critical dose, right, because they just weren't enough carriers at all in the relevant dose.

So I'll add and reinforce that. We looked in detail at all the outcome measures in ApoE4 carriers and the CDR Sum of Boxes is the only 1 that favored placebo, and that included all the quality of life and caregiver burden and time to event these other outcome measures that we looked at, all the plasma biomarkers. If we look at the placebo progression of the homozygotes versus the heterozygotes and the noncarriers, the heterozygotes and the noncarriers are right on top of each other, and the placebo progression for the CDR sum of boxes is really shallow for the noncarriers for the homozygotes. So that's been the only factor that we could identify so far that might be able explain that discrepancy, but we'll continue to look at it.

Last question who will be?

[indiscernible] at University of Wisconsin. I'm using lecanemab currently with my patients, and I'm looking forward to using the lecanemab after presentation. The question is, how are you going to get Medicare to pay for it.

So we have been engaged with conversations with CMS since before the results. And we've -- there have been very positive engagements in terms of presenting to them the design of the study and the type of outcomes that we were assessing and also getting their feedback in terms of the type of information that they would want to see from a treatment for Alzheimer's disease, including the clinical significance of the findings, the safety and efficacy over time, the impact in subgroups. So we'll be reviewing this data with CMS. We believe that this data provides a high quality of evidence regarding the safety of efficacy in lecanemab, and we hope that CMS takes this into consideration as they evaluate lecanemab if it is approved.

Okay. So I'm very sorry. We are in short of time and I am closing this session. So thank you very much for your active participation.