Hello, and thank you for joining us to discuss Belite Bio’s First Quarter 2024 Financial Results. Joining the call today are Dr. Tom Lin, Chairman and CEO of Belite Bio; Dr. Nathan Mata, Chief Scientific Officer; and Hao-Yuan Chuang, Chief Financial Officer. Before we begin, let me point out that we will be making forward-looking statements that are based on current expectations and beliefs. These statements are subject to certain risks and uncertainties, and actual results may differ materially. We encourage you to consult the risk factors discussed in our SEC filings for additional detail. Now I'll turn the call over to Dr. Lin.

Thank you, Judy. Thank you, everyone, for joining our reporting for the fourth quarter. I'm Tom Lin, CEO of Belite Bio. Joining me is our CSO, Nathan Mata and CFO, Hao-Yuan. I'd like to start off with an overview. Tinlarebant is a novel once-a-day oral tablet designed to bind to serum retinol binding protein known as RBP4 as a means to specifically reduce retinol delivery to the eye. This approach is intended to slow or stop the formation of the toxic retinol derived by-products, which are generated in the visual cycle and are implicated in the progression of Stargardt disease and Geographic Atrophy. Belite Bio believes that early intervention directed at emerging retinol pathology, which is not mediated by inflammation, will be the best approach to potentially slow disease progression in Stargardt disease and GA. There's still a significant unmet need for both indications as currently, there is no approved treatment for Stargardt disease, and there are currently no approved oral treatments for GA, and we're already in global Phase 3 trials for both indications. So far, we have been granted fast track designation, rare pediatric disease designation and open drug designation in the U.S., EU, and now Japan. We have several patent families with compositions of matter patents lasting until the 2040s, and with patent term extension and new patents to be filed, we will have patent protection past the 2040s. For Stargardt indication, at ARVO last week we presented further positive findings and treatment results from our end-of-Phase 2 results which our CSO will be presenting. The Phase 3 is fully enrolled and estimated interim results by the end of 2024 or early 2025. We've also initiated a Phase 2 study in Stargardt’s, which will recruit Japanese patients that is required for NDA in Japan. For GA, Dry AMD indication, we currently have about 100 subjects enrolled in our Phase 3 in GA. And with this, I would like to pass it on to our CSO to give the clinical and scientific update. Nathan?

Thank you, Tom. I thank everyone for attending. What we'd like to share with you today are some new analyses that we have from our open label Phase 2 study in adolescent Stargardt disease. This trial enrolled 13 adolescent Stargardt subjects aged 12 to 18 years of age, and it was a two-year study in which patients took oral Tinlarebant 5 milligrams daily. The first analysis I want to show you is a genetic analysis of all of our patients in the study. This has not previously been reported. You'll see for each subject, there are two entries. That's because there's an allele for each ABCA4 mutation, and so we're looking at two of them, and you can see 11 of 13 subjects have severe biolytic mutations. This analysis, by the way, was conducted by Dr. Rando Allikmets who was the identifier of the Stargardt gene, the ABCA4 gene. So the point here is that our cohort is very prominently affected with the severe mutations. There are only two subjects, subjects three and five, who have a moderate allele of ABCA4 on their mutations. It's important to note the way these genes are classified is through a score called the combined annotated dependent depletion score. In this analysis, scores above 20 are predicted to be among the 1% most deleterious, and you can see all of our subjects, with the exception of those alleles on subjects three and five, have scores well above 20. So this is a very severely affected cohort. The next analysis I'd like to show you is that of these subjects, five of them, subjects one, three, four, 12, and 13, never spawned an atrophic lesion. This is important because in this particular study, subjects came in with only autofluorescent lesions at baseline. So we were monitoring the transition of the autofluorescent lesion to the atrophic lesion type, and then once the atrophic lesion formed, we measured the growth rate of that lesion, and we previously reported a remarkably reduced growth rate in the overall population. Now we're showing you in those subjects that did not grow an atrophic lesion over 24 months, they had very severe mutations. So the absence of lesion growth in these subjects could not be attributed to benign or mild mutations. The next analysis is a very important one because it shows us that four subjects have the exact same allelic mutation. Subjects 9 and 10 are siblings. They are two brothers. You can see there they have the exact same mutations across alleles, and subjects 12 and 13 are a brother and sister. They also have the very same mutations. The reason this is important is because other investigators have suggested that identical genetic mutations predict identical disease course. I can tell you right now that is not what we're seeing in the Phase 2 study, and I'll show you some of that information right now. Here we have the visual acuity analysis of all subjects before they came into the study, and they all have different disease durations. What we did is we looked at subjects that had bilateral vision loss, that is vision loss in each eye, over the period of time between the time they were diagnosed and the time before they came into the study. What we found is there were six subjects who had a very severe bilateral vision loss of 10 letters per eye, a mean annual loss of 10 letters per eye. That's significant because, for two reasons. One, it tells us that autofluorescent lesions in the fovea do in fact cause visual acuity loss. Remember, these subjects have not grown atrophic lesions yet. The other thing that's important is if you look at the sibling subjects again, 9, 10, and 12, and 13, you can see their visual acuity loss is very different. Subjects 9 and 10, again, are their two brothers. They have similar disease duration, but subject 10 lost vision, whereas his brother, subject 9, did not. Same thing for subjects 12 and 13. Subject 12 has an eight year disease duration, whereas the sister has only two years. Yet, the sister with only two years of disease duration lost significant letters, whereas the brother did not. So these data tell us, in fact, that identical genetic mutations do not necessarily predict identical disease progression as far as it pertains to visual acuity loss. This is an analysis of the visual acuity of all subjects on the left-hand side, and then those six subjects shown on the right. I want to first say that in all subjects over 24 months, the mean loss was only five letters. That is, over a whole period of 24 months, all subjects lost only a mean of five letters over 24 months. So that's roughly 2.5 letters per year. Now, if we just focus on the six subjects that came in losing 10 letters per year, we now see over 24 months, they've only lost about 3.8 letters, which is roughly about two letters per year. So we've taken the significant 10 letter per year loss down to two letters per year. That is a substantial preservation of visual acuity in these subjects. The next analysis I'd like to show you is the lesion growth analysis. And it's important to note that these autofluorescent lesions convert to atrophic lesions, and a visual of that is shown on the graphic on the upper right-hand side. This is the actual lesion growth data from subject 10. You can see at baseline, there's this very large autofluorescent lesion, which is referred to as questionably decreased autofluorescence, or QDAF. Over time, atrophic lesions will be spawned within the autofluorescent lesions. The autofluorescent lesions are referred to as definitely decreased autofluorescent lesions. So here we're looking at the transition of the autofluorescent lesion to the atrophic lesion type, and we're seeing if there's a proportionality between the decrease of the autofluorescence and the increase of the atrophic lesion size. And that graphic is shown on the lower right-hand side. As I said before, five of 12 subjects, 42%, never grew an atrophic lesion. We're not including subject 7 in the analysis because that subject was lost to follow-up at month 12. But in those subjects that did grow atrophic lesions, you can see a very clear correlation between the decrease of the autofluorescent area and the increase of the atrophic lesion area, showing that the atrophic lesion does grow, certainly within the autofluorescent zone, but nowhere else. So we have confined the lesion growth to within this zone, and we know that lesion growth from prior analyses is significantly lower than in natural history. There's only one subject, subject 5, that grew an atrophic lesion outside of the initially identified autofluorescent area, and we believe that could be due to reading error of the image software, which I'll get to right now. We typically use, and most people use, a software called the Heidelberg Region Finding Software. This is a software that accompanies the imaging camera that allows ophthalmologists to grade lesions in the back of the eye, retinal lesions. The problem with this software is that it is affected by human error. So humans have to actually look at the lesions and demarcate the zone of atrophy, and sometimes readers don't agree. So you can have a case where you have a difference of opinion, and it has to go to an arbitrator. So this can be a very time-consuming and error-prone process, but this is what is being used today. Our reading center has developed a new algorithm for reading lesions which does not rely on subjective reader bias. In fact, it uses a mathematical classification of lesions based on the proximity of the lesion and the density to that of the optic disc, as well as healthy retinal tissue. And it focuses just on the 6-millimeter macular area, which is the most important to look at when you're talking about visual acuity loss. And as I said before, it's independent of reader bias. So it's looking at just pixel densities across the lesion area, and really all the reader is doing is getting the data back from the software and reporting whatever the computer algorithm provided. So I want to show you a reanalysis of our data using this particular image grading software. What we found, in fact, was that if we looked within the macular area using this new grading algorithm, we found that there are 12 eyes of 8 subjects that did, in fact, have macular lesion involvement at baseline. And we monitored the lesion growth of the macular lesion of these patients over time. And you can see that on the left-hand side. The solid lines and dots show you the actual data from the lesion growth, and the dotted line shows you a third-order polynomial fitted through the data so you can see the trend line. You can see it's very clear. The lesions do grow from baseline out to about month 16, but starting from month 16 to month 24, there was absolutely no further growth of the lesion. And you can see on the right-hand side, the percentage involvement of the atrophy within that 6-millimeter macular zone. The 100% would indicate that all 6 millimeters of that zone are occupied by atrophy. As you can see here, our subjects are showing no more than a 7% encroachment of the atrophic lesion into the macula, and it's static from 16 to 24 months. So we're very pleased to report this data, and it is consistent with the stabilization of visual acuity that I showed you earlier. Finally, I want to talk a little bit about our Phase 3 study in Geographic Atrophy. It's important to note that the Phase 3 trial design in Stargardt disease and the Phase 3 trial design in Stargardt are very, very similar. They have the same dose. They have the same duration, two years, with the same interim analysis. There's the same two-to-one randomization. And, of course, we're looking at all the same safety and efficacy assessments that we're looking at to judge atrophic lesion growth and adverse events. There's only two differences in these studies. First, the indication, GA instead of Stargardt, and then secondly, a higher predicted enrollment population for the GA study to reflect the higher prevalence of the disease in the population. So with that, I'll turn it over to Hao-Yuan to discuss financials. Thank you.

Thank you, Nathan. So in Q1 2024, we had R&D expenses of $6.8 million compared to $5.7 million in Q1 2023. The increase was primarily due to increasing expenses due to first, conducting the DRAGON study, second, initiating the DRAGON II study, and third, wage and salary due to our R&D team expansion and share-based compensation granted in the third quarter of 2023. On G&A expenses in Q1 2024, G&A expenses was $1.6 million compared to $1.2 million in Q1 2023. The increase was primarily due to increasing share-based compensation granted in the third quarter of 2023. On net loss, we had a net loss of $7.9 million in Q1 2024 compared to $6.9 million in Q1 2023. In terms of cash, we had a total of $95.5 million in cash and investments in short-term U.S. Treasury bills in Q1 2024 as compared with $77.8 million in cash in Q1 2023. The increase was primarily due to last year's follow-on offering and the exercise of the warrants issued in the follow-on offering and our ATM offering. In Q1 2024, we raised $12.5 million from the exercise of the warrants issued in the follow-on offering and ATM offering. In addition, we raised an additional $25 million from our registered direct offering in April. We expect a cash runway beyond 2026. Thank you. Back to you, Tom.

Thanks, Hao. I would like to conclude with the key milestones to expect for this year. We are still making good progress in the Phase 3 in GA. Many sites have been initiated and are enrolling subjects and more sites will be initiated. Interim data for Phase 3 study in Stargardt disease is expected to be in December this year or early next year depending on data collection and when the DSMB can convene. Thank you for participating and we'll leave some time for Q&A. Thank you.

Our first question comes from Yi Chen with H. C. Wainwright. Yi, your line is now open.

Hi, thank you for taking my question. With respect to the novel grading algorithm, is it being deployed in both DRAGON 1 and DRAGON 2 trials?

Hi Yi, this is Nathan Mata. I'll answer the question. So, we have not yet formally deployed the grading algorithm in either clinical study. We are going to speak with the FDA first about a validation process to determine how they would like us to proceed with this new algorithm in terms of demonstrating its robustness and reliability. That meeting will occur within the next month or so, but there is a chance, there's every chance that obviously we can implement it before the end of the study and retrospectively grade the images once we have FDA's approval to use it as a formal grading analysis. It will always be combined with whatever or be compared to, I should say, with what we obtained with the Heidelberg region finding software.

Okay, and to follow up with the DRAGON I and DRAGON II trials, do you intend to enroll patients with at least some degree of atrophic lesion in both trials?

Oh yes, oh yes, certainly for our Phase 3 studies where the endpoint of course is slowing the growth of the atrophic lesions, all subjects will come in with some measure of atrophy at baseline, but again, as part of our differentiation, we are picking patients with smaller lesions at baseline because we believe that early intervention will be a more appropriate sort of approach for this particular disease, both diseases, GA and Stargardt's.

Does that suggest that earlier stage patients without atrophic lesion may not be a suitable population for the treatment?

That's not the case. The case is actually a more regulatory concern because the agency, FDA and even EMA are not convinced that autofluorescence in and of itself is actually causing photoreceptor cell death or dysfunction, whereas atrophic lesions represent dead retina and of course that will lead to loss of vision. So until there's more clinical data to demonstrate that autofluorescence by itself is actually affecting photoreceptor function, and we have some of that data as I mentioned in the presentation, we have these kids who have autofluorescent lesions at baseline, they don't have atrophy, many of them, but they're losing vision. So we'll need more data like that to convince the agency that in fact the autofluorescence is impacting photoreceptor function, then it would be a valid endpoint. But we've shown you here a slowing of the transition from the autofluorescent lesion to atrophic lesions, and then once the atrophic lesions are formed, a slowing of growth of those lesions. So we believe in both contexts that slowing the transition as well as slowing the growth of the atrophic lesions, this treatment approach is appearing to be effective at both of those different scenarios.

Got it, thank you.

Our next question comes from Marc Goodman with Leerink Partners. Mark, your line is now open.

Hi, good afternoon. This is Basma on for Marc. We have a question regarding the interim readout. Can you remind us again of your regulatory plans if the interim is positive? And again, a positive interim doesn't mean that the separation between the treatment arm and the placebo arm has to be statistically significant. That's our first question, and we do have a follow-up question.

Yes, hi Basma, this is Nathan. I'll answer the question. So the way the interim analysis will be handled, it will be independently assessed by our DSMB. Ourselves on the sponsor side will remain masked, so we will not know the treatment effect, nor can we announce a particular treatment effect. The DSMB will do analysis based upon a predefined statistical window that tells us whether or not we need to add patients. So if we don't need to add patients, there's a very good chance we've fallen within the statistical window in which the treatment effect would be consistent with something the FDA would approve that is 20% or greater. If we are asked to add patients, that means we're not at that particular point yet, so we would be adding more patients to sort of increase the conditional power. I should mention that that statistical analysis was developed when we planned the enrollment at 90 subjects. We had since over-enrolled the study at 104 subjects, so we actually already have a buffer of subjects that probably we won't need to add, so I'm pretty confident that number, that is the treatment effect number, will be consistent with our statistical window. There is an outcome where we are not in the statistical window and it would be futile to add patients, but this is a non-binding futility, so if we do get that outcome, we will still continue the trial to the end of two years because it is very likely that you could get a better treatment effect during the second year of the study. So those are the three different outcomes. I hope that was clear for you, Basma.

Right, so like you're not going to, if, okay, so my understanding now is that if there's a positive outcome, there's no change in plans, there's no change in enrollment, and, okay, thank you.

Correct.

Our follow-up question regarding the five out of 12 patients with pathogenic variants that you showed that did not develop any atrophic lesions throughout the 24 months of the Phase 2 trial.

Yes.

Can you remind us again, a patient with these kinds of variants based on natural history studies, how long does it take them really to develop atrophic lesions from QDAF to DDAF, just to know the treatment effects, to have a better understanding of the treatment?

It really depends on their baseline status. So if you look at the baseline status of our subjects, the size of their autofluorescent lesions predicts growth within two years. That means predicts atrophic lesion growth within two years. So the majority of subjects should have had atrophic lesion growth. There's probably two subjects, I can't identify them by number because I don't have the data in front of me, but I believe there were two that had relatively small areas of autofluorescent lesions, but it was encroaching the fovea. So they're still losing vision. But the short answer is in our cohort, it was expected that the majority of subjects, at least 75% of subjects would have converted by two years, would have converted to atrophic lesions. So the fact that we're seeing 42% of subjects not convert is actually a very, very promising finding in addition to actually slowing the growth of atrophic incident lesions once they are formed.

Thank you. Thank you, Nathan.

This now concludes our Q&A portion of the call. I would like to turn the call back over to Tom for final remarks.

Well, thank you everyone for joining this call and asking questions on the trial. Please look forward to more updates from us. As we're making good progress on both studies. Thank you very much and we'll see you soon.