Thanks everyone for being here. My name is Yanan Zhu. I am one of the biotech analysts here at Wells Fargo. It is my great pleasure to be joined by the management team of Belite Bio. With me here is Hao-Yuan Chuang, CFO of the company, and Hendrik Scholl, CMO of the company. Thank you for being with us. I was wondering if you can give a brief overview of the company and its initiatives before we dive into questions. Sure. First of all, thank you for having us. We are Belite Bio. What we have is an oral once-a-day tablet called tinlarebant. We have two indications now. One is Stargardt disease and one is geographic atrophy. Stargardt disease is a gene mutation disease that causes, in almost all cases, legally blind. We have completed our pivotal Phase III study last year, and we have submitted for NDA in the U.S. and Japan. The FDA has confirmed the acceptance of the NDA with priority review, with the PDUFA date being February 12th next year. On GA, we have completed enrollment of the study. It is also a Phase III study globally, and we are just doing the execution right now. Thank you. Great. Congrats on the NDA submission and also importantly, the priority review granted by FDA. Could we start there? Can you talk about what the implication of having this accelerated review timeline designation and also any early feedback from the agency? You want to say it? No, you. Well, I should first, apparently we're very excited and appreciate with the FDA's priority review. I think FDA has been showing a lot of support throughout the process. We do have a lot of patient have been waiting for this drug for a long time, given there is no treatment. We really look forward to continue to work with FDA and provide the drug to the patient as soon as we can. Maybe an extra comment. It just underscores the significant unmet medical need that we have for this rare, but not so rare, severe in pediatric disease. Got it. Can you talk about, for the review, is there any particular focus that you think FDA will have, and also could there be an advisory committee meeting? We have had very constructive discussions with the FDA about the design of the DRAGON trial, how the DRAGON trial was conducted, when we submitted our interim analysis to the FDA that led to breakthrough designation. The agency is very aware about the design of the trial, what was the interim result that led to breakthrough designation, and now with our submitted package, including a very comprehensive document that we call confirmatory evidence, because we seek approval based on one trial, and that, again, was pre-discussed with the agency. It is important to underscore how, and I want to praise the FDA here, how educated the agency is about the primary endpoint that we have been using in both trials. It is just a different nomenclature. In geographic atrophy, it is simply called geographic atrophy. It is not so sophisticated. It is measured on autofluorescence images, and it is again, compared to the optic nerve head in terms of the blackness level. In Stargardt, we call it definitely decreased autofluorescence, meaning very black, again, compared to the optic nerve head. So it is essentially the same outcome measure that is being used for both indications. The FDA already expressed 10 years ago in publicly available papers that if you preserve photoreceptors and you can show that on imaging, that is a good thing. If you cannot preserve them, it is a bad thing. So the FDA is willing and eager to accept efficacy measured precisely on these imaging modalities that are part of standard clinical care and clearly indicate preservation of photoreceptors that eventually, in the long term, correlates with visual function or dysfunction in that case. Therefore, because the agency clearly understands our primary endpoint and perceives this as an approvable endpoint, we do not anticipate necessarily an advisory board, because typically at an advisory board, it would be about the primary endpoint, and we do not believe that there are so many questions about what is the endpoint, what does the endpoint mean, how it is being measured, what is the reproducibility and so forth. Got it. Great. Thank you. That is very helpful. Can I ask you what kind of label are you hoping for or expect to get, for example, in terms of age, disease severity, genetic confirmation along those lines? Yeah. This is We prediscuss with the agency. The agency understands this is a genetically caused disease. It is caused by dysfunction of ABCA4. The question goes to the generalizability of the data, and we believe that given that everybody affected by Stargardt disease is that because there are mutations in the gene, that the patient population 12 to 20 years old can be generalized to all patients out there, regardless of the age of onset. The age of onset shows variability. A latest survey of the literature shows that it can be as early as one year of age, and it can be as late as 84 years of age. We still believe that the data of the DRAGON trial of patients of the age of 12 to 20 is generalizability to all patients out there. The second question is about disease stage. This brings us back to our primary endpoint, DDF. The agency understands that you have to have something measurable that is an indication of the disease, and that you find an efficacy signal that reflects efficacy for that indication. That means that although patients, number one, needed to have a lesion, and needed to have a lesion that is not too small and not too large, that what we find in the DRAGON trial is still generalizable to all patients out there that have either no or smaller lesions or have lesions that go beyond what was enrolled in the DRAGON trial. To your third question about genetic confirmation, this is a genetic disease, a monogenic disease that is either caused by homozygous or compound heterozygous mutations in the ABCA4 gene. We believe although the diagnosis can be made clinically, it's a pretty obvious diagnosis that if you are at least somewhat familiar with the disease, that you can make on clinical grounds alone. But we believe that the standard of care in the U.S. and most countries around the world is genetic confirmation. You offer patients genetic testing, and at least one mutation in the ABCA4 gene must be found in order to call that Stargardt disease. Got it. Very helpful. On the primary endpoint, I think in geographic atrophy, when a similar endpoint is analyzed, there could be a square root transformation to account for baseline differences. But in your case, you do not have that transformation. Can you talk about the rationale there? Yeah. I am very happy to elaborate on that. This is a debate that is going on for two decades, at least in the GA space, because it was found for the progression of geographic atrophy that the progression rate depended on baseline lesion size. How big the lesion was when a patient is being treated, and what that means, how fast this lesion grows over time. It was found more than a decade ago that if you do a square root transformation, so taking the baseline lesion size into account, that you can get essentially rid of this dependency and you find a linear progression in GA. There has been the same debate in Stargardt. This is about a decade younger in Stargardt disease. In the ProgStar study, when we submitted the ProgStar report number 17 for publication, this is the two year data of the natural history of Stargardt disease in a worldwide natural history study. We did not include square root transformation because we did not think that it is an appropriate measure to get rid of baseline lesion size. But the reviewers wanted to have it. Obviously, maybe they came out of the GA space. We included it in ProgStar report number 17 and showed that even if you include it does not allow to get rid of baseline lesion size. Obviously, since the FDA has approved therapies based in GA on progression with square root transformation, we included in our submission package square root transformation of our data. It looks really good. It just does not allow to at least fully get rid of the dependency of baseline lesion size. The debate is ongoing. At the ASRS meeting in Montreal 2 months ago, there was also a presentation that introduced an even third alternative, so-called perimeter-adjusted progression rate in Stargardt disease. But even that does not allow to get rid, or at least get rid completely, of baseline lesion size. What we did in the DRAGON trial was we used, in my opinion, very appropriate statistical model, MMRM, that allows to account for baseline lesion size and takes all other baseline characteristics into account. Therefore, in my opinion, this is the best approach to adjust for baseline lesion size and take those baseline characteristics into account. Long story short, in my opinion, you have to use an appropriate statistical model and not necessarily use any correction that has not been shown to be efficient to get rid of baseline lesion size. Got it. While we are on the lesion size, I wanted to ask about a competitor-related question. This is the asset that Tarsus recently acquired, ALK-001. I think they have data out there for their lesion size reduction. I think obvious question is when we look at the two companies' data, could we compare these two? If we could, then is the difference, I think your lesion size is greater, is that difference clinically meaningful? If we cannot, what might be the reason that we shouldn't compare this data directly? Before directly answering your question, I would like to say that we would not like to discuss about competitors, but I can comment on what was published, and it was just published last week, and this is publicly available information. I think it is important to understand that we should not compare apples and oranges here. When you read the publication, you will understand that what was done in the so-called TEASE-1 trial is that the placebo arm was augmented by so-called natural history cases. Those natural history cases were not read by the readers in a masked fashion because it is stated in the paper that it was obvious to the readers that these were non-treated subjects. This is one difference. It is also important to understand that we, in our trial, obviously did not use natural history. We did real placebo observations, real placebo patients in our trial. When you then try to compare with what was published, then you find it is claimed in the paper that with this Alkeus compound, you find an efficacy signal of 21% when you include natural history. But when you do not include natural history, the efficacy rate is 14%, right? It is much lower without natural history. I think if you compare anything, maybe you should compare this efficacy signal because this is what we did in the DRAGON trial. We did not augment our data with natural history. Having said that, since we are seeking approval from the FDA based on one trial, and I mentioned that earlier, we put together a very comprehensive document called Confirmatory Evidence, along the guidance of the FDA, what is being considered confirmatory evidence. We have included natural history analysis out of the publicly available ProgStar data. By doing that, we can significantly bump the efficacy signal up and the P value down if we do exactly that, right? This is not publicly available information. I am not going to disclose what exactly we found when we include natural history data, but it had obviously a very beneficial effect on our findings. The last comment I want to make is the following. When we compare the placebo group in the DRAGON trial, in our DRAGON trial, with the natural history data out of ProgStar, what we find is that our placebo group progressed somewhat a bit slower than the average that is published in ProgStar. It was even harder for us to find this 36% efficacy signal, while when you look at what was published last week, for whatever reason, the placebo group in this phase II trial conducted by Alkeus Pharmaceuticals at the time, progressed faster than the natural history cohort of ProgStar for whatever reason. It is difficult to compare the two. There are, I think we also learned that from the paper, they used actually one subject with both eyes to do analysis, which at least based on our understanding with the FDA, that is not the way they look at the data. When we do our analysis, one subject is only one eye, on the study eye. That is why we said it is not very easy to really compare the data. Got it. That is very helpful. Thanks for all the insights. I wanted to touch on safety because of the mechanism of action of tinlarebant. There are a few well-known on-target effects. I was just wondering, can you share with us how tolerable those ocular AEs are? I think you also mentioned they were transient and can go away while patients are still on therapy. I just hope you can further elaborate on that. I am happy to. I would like to motivate the audience to listen to what we discussed at our commercial day last week. We had invited key opinion leaders, Dr. Paul Bernstein, professor at the University of Utah, and Professor Michel Michaelides, leading top recruiter in DRAGON I and the DRAGON II trial, professor at Moorfields Eye Hospital in London. They discussed exactly that. You rightly said on-target effects, right? My first comment is the systemic tolerability and safety, in my opinion, was excellent. We only had six SAEs, four of them in the placebo group. None of the SAEs was related to the study drug. I think this is already an important statement. The second are the so-called on-target effects that we see as ocular adverse events that we observe in the DRAGON trial. Almost all were rated mild or moderate. It was discussed at our commercial day. These are very tolerable but need some patient education. Because what we see when a subject, and about a third or so, would have complained in the DRAGON trial about delayed dark adaptation. I think this is really on target, right? Meaning there is less compound in the visual cycle, the regeneration of the photopigment, so it takes more time for the photopigment to form in the photoreceptors. That can be noticed by a subject. The final threshold, how sensitive you are when you wait long enough, is not affected at all with tinlarebant. Therefore, as an example, the term night blindness is not correct, in my opinion, right? Because you do not affect that sensitivity at all, but the time it takes to get to that sensitivity. As Dr. Bernstein and Dr. Michaelides expressed last week, these on-target effects are definitely tolerable, can easily be managed, but will need some patient education in the future. The same is true for an even milder side effect, if I may say so. This, I call it discoloration of the visual scene. There are all sorts of Greek names for it. Xanthopsia if it's yellowish, erythropsia if it's reddish, cyanopsia if it's bluish. They are very fancy names. But what is meant is the following. If you take the drug, again, the photopigment generation is somewhat slowed down. Now we don't talk about adaptation going from light to dark. Now we talk about going from dark to light. This is very fast. We all notice that. We open the door in the morning, all of a sudden there's a lot of light, but it only lasts seconds, maybe a minute or so until we are fully adapted. During that time, there is no color vision deficiency there, but there can be a discoloration of the visual scene. You see everything. You can recognize faces, you can read, but the visual scene can appear a bit yellowish, then it's xanthopsia, or a bit bluish, then it's cyanopsia or what have you. And this will fade away after a few seconds or a minute or so. Again, it's a patient education issue. You have to educate your patient this can happen. It's a very benign side effect. It doesn't affect your vision really, and you will be fully adapted after a few seconds or minutes or so. And interestingly, and I will conclude with that statement. In the trial, what we found was that if there were these adverse events, and about a third of patients complained about it or mentioned it, maybe that's a better term here, then it was typically in the beginning of the trial. Obviously, patients adapted to these adaptation adverse events, and at the end of the trial, there was virtually no complaint left. Great. That's super helpful. I was also wondering about the rate of patients rolling over to open label as an indicator for whether patients are bothered by anything- Yeah. From a drug treatment. Can you comment on that? Yeah. It is important to note before I respond what's the fraction of patients that was rolled over to our long-term extension study called CT08, is that none in the U.S. and the U.K. That's intended because in the U.S. and the U.K., we opened the DRAGON II trial, and at the time when we launched DRAGON II, we would not like to adversely affect our enrollment by offering patients this long-term extension study. Therefore, patients in the U.S. and in the U.K. were offered to be enrolled into DRAGON II. Many patients opted to do that. But in the other countries where we do not have DRAGON II, as an example in Europe, we offer that to all patients in the DRAGON trial. I think that I have all the numbers, is for example, in our phase II trial, all patients rolled over into CT08, and that also means that we meanwhile have more than four-year observation of patients on the drug in our long-term extension study. Great. With the February PDUFA approaching, let's talk about commercial launch and preparedness. I was wondering, could you comment on pricing strategy, or what do you expect the pricing might be? Yeah. Well, it is still too early to comment about the pricing, but we do expect it is going to be a rare disease priced drug. Yeah. Okay. Got it. So along that line, have you done any payer work and see if where you are thinking about could be acceptable to payers? Yeah. We have done, I think more than three rounds of payer research. We have seen the payers being very supportive. Especially this is a treatment that almost all cases patient become legally blind, so there is no treatment. So we are very encouraged by the payers' response. Okay, great to hear. I think there are 20,000 clinically diagnosed patients. I was wondering, how are they distributed in the healthcare system? How concentrated is care? Is there a group of physicians that takes care of vast majority of the patients, for example, or not really? How do you plan to reach these 20,000 patients? Stargardt disease is a monogenic disease that would predispose you, so to speak, to be seen by an IRD specialist. But then again, it's not so rare, meaning that many Stargardt patients are being seen by retinal specialists, even general ophthalmologists. When we talk about the distribution, it's likely a skewed distribution, meaning that although IRD specialists are a minority of all retinal specialists, and all retinal specialists are a minority of all ophthalmologists that are out there, IRD specialists obviously see many Stargardt patients. Retinal specialists see also many when you calculate the sum of them. General ophthalmologists also see Stargardt patients. The typical referral pattern would be that if a general ophthalmologist sees a Stargardt patient, he or she would typically, not necessarily, but typically refer to a retinal specialist. If a retinal specialist sees a Stargardt patient in his or her own practice, he or she would likely manage that patient on her or his own. If it's a large practice and there's an IRD specialist in this practice, and large practices have IRD specialists, would rather refer to that IRD specialist. That this leads to a skewed distribution, that there is a fair number of IRD specialists and a pretty large number of Stargardt patients that are regularly being seen, genetically confirmed with these IRD specialists that can be treated right away when tinlarebant hits the market. Could you give a sense of how many of those IRD specialists are out there, and how many salespeople do you need to target that population? Well, I think what we know is probably there's a list on the FFB website. I think they have about 100 IRD experts there. So they probably are the most active IRD experts out there that could be the focus group for us to target. Thank you. Of the 20,000 patients diagnosed, half of them have a genetic confirmation, half of them don't. I was wondering, first of all, for the label, do you think it will specify genetic confirmation in the label? I do think you mentioned for payer, you probably will need it anyway, right? Yeah. In that case. Yeah. Please talk about how to get those half who don't have it yet to convert them to have a confirmation. Yeah. It's important to understand that genetic testing is something that is being regularly done in clinical practice. There's also the profession in the U.S. I say that because in Switzerland, I still see patients one day a week, that this profession does not exist, the so-called genetic counselor, but we have that profession in the U.S. This just shows that genetic testing and counseling is something that is regularly being done. Now that there's a treatment coming, this is a game changer for genetic confirmation, obviously, because in the past, there was no immediate consequence other than confirming the diagnosis. Which is important in my opinion, but it had no therapeutic consequence. It will have a therapeutic consequence in the future, and this will motivate the remaining roughly half of that clinically diagnosed patients, both from the HCP side and the patient side, to actually seek confirmation to become a candidate for the treatment. Got it. Would you be able to subsidize the testing or not really by regulation? Well, there are free testing in a way that I think My Retina Tracker is one of them, that the patient can get the testing free. There are other labs that are available, but the cost is also not very high. I think we talk about this on the commercial day, but the cost and the time it takes for genetic testing have come down so much for the last years. The understanding about genetic testing has also improved a lot, and I think what is missing is really a treatment that does make that testing to become a game changer for the patient. We will not worry about the cause or the time it takes, be it the burden on the genetic testing. Got it. Maybe a small comment here. I do not want to scare anybody in the audience, but given the prevalence of the carrier rate of mutations in ABCA4, I just calculated the probability, and the probability is 50% that we have at least one carrier of mutations ABCA4 in the room. Right. Just those diagnosed patients could support very large opportunities already. But I was also wondering, there are also 33,000, sounds like, patients who are undiagnosed, but by epidemiology, there should be that many in the U.S., right? At what point of the launch does it become a priority to identify those patients, or is that the thing that you could start on day one? Yeah. That is a very important question because this is, in the bottom of my heart, I am a healthcare provider, always have been. And we want to bring that therapy to all patients affected by Stargardt disease. And so how do we get to all patients? You said that there is 33,000 patients undiagnosed. We have to be more precise. It is a severe disease. It would be a very rare instance if somebody would not carry any diagnosis. If you cannot see and it cannot be corrected with glasses, you likely have sought clinical care once in your life, and there must have been a diagnosis, maybe not as precise as Stargardt disease, but a diagnosis. Meaning that those remaining 33,000, 35,000 patients likely carry some diagnosis. But how do we activate them? How do we bring them back to clinical care? There are all sorts of activities that can be done, and we are actively pursuing such as patient advocacy programs because these patients stay connected. Many patients stay connected through social media channels, or they follow Foundation Fighting Blindness or Prevent Blindness or Stargardt's Connected, which is a worldwide group of patients affected, and families by Stargardt disease. And we are trying to activate those channels in order to get to all patients out there. Great. Great to hear. Lastly, I wanted to ask about geographic atrophy for that effort. I think you have a study ongoing. There is supposed to be an interim analysis. The question is, what is the updated timing for that analysis? Also, what will happen after that analysis? What are the scenarios once that analysis is done? Well, I think we talked about this on the recent earnings call, that we expect the interim analysis to be closer to end of the year or maybe even early next year. We are working very hard on the Stargardt submission right now and try to get the drug to the patient as soon as we can. At the moment, we put the guidance of the GA probably what depends on the workload that the team has. We cannot really speculate what we're going to learn from the interim. You will have it designed to be agreed with the FDA, and it depends on what the FDA would allow us to know. But within the chance of that, the same thing, I think the DRAGON study was a pleasant surprise with the ethical reason behind it. That's why we learned about the data. We're not going to speculate anything about the GA interim analysis, but we'll let the data speak for itself. Yeah. Great. With that, I think we're out of time. I wanted to thank you for all the very helpful insights, and thank you for your time. Thank you. Thank you very much. Thank you. Great.