Spectral AI, Inc. (MDAI) Earnings Call Transcript & Summary

Good afternoon, and welcome to the Spectral MD Holdings Ltd Final Results Investor Presentation. [Operator Instructions] Before we begin, I would like to submit the following poll. I would now like to hand you over to CEO, Wensheng Fan. Good afternoon to you, sir.

Thank you very much, and good afternoon to all the folks calling in from the U.K., and good morning, America. So we are here to announce our final results for Spectral MD's 2022 financials. And with me today is Mr. Nils Windler, our CFO, and we'll start the presentation now.

Many times we talk about the company Spectral MD, and we are very happy to be reporting that we are a proprietary artificial intelligence company. Right now, our focus is to predict if a wound will heal. We're specialized in predictive analytics and our platform is called DeepView, D-E-E-P-V-I-E-W. And DeepView doesn't come just in one aspect. It includes the following: it's a medical imaging plus AI predictive analytics. So once you fit into that medical imaging and AI, you're the part of the core business of Spectral MD. On the left, you will see the DeepView imaging. That is the imaging kit roaming around the hospitals, collecting clinical data. This is a patented, proprietary, multispectral imaging acquisition. And what it does is acquire data from near UV lights through the visible band all the way into the near IR lights. They capture the full spectrum of that range, and sample it at very specific wavelengths, create a spectral signature of the injured tissue. That's what we collect. But with that, our human brains will now be able to distinguish from the fully injured tissue to partially injured tissue, to the healthy ones. What we used since 2014 will be the AI models, the AI approach, the artificial intelligence approach. So we extract the AI model features from that raw data, raw imaging data, remember, from near-UV light, visible band to the near IR lights. And then we combine that with patient healthcare matrix data and send that into our AI model building. The company's crown jewel is a 263 billion data points that we have collected till date through very rigorous and structured clinical studies in the past few years. I have no doubt the Googles, the Apples in the world, they have all the money, all the resources, and all the talent to build an imaging system like ours. But in order to reach to our AI model performance, they have to go through the similar structured clinical studies, which could take 5, 6, 7 years to perform. And we go through all this hassle. But at the end, our results will be very, very simple. Indeed, it's actually a binary result. We will only show -- highlight the nonhealing wound area. On the upper right corner, you see that it's a picture of a raw wound on a 3-year-old baby. And on the right side, that's a DeepView output. We highlight that areas of the burn wound on his chest that will never heal. This will require surgical removal of the dead tissue, then skin grafts, and other methodologies to help it recover. So in a nutshell, nonhealing wound equal to surgery and for healing wounds goes to routine care. And this is a medical imaging plus AI in our data pipeline, from the imaging, to the data extraction, to the AI model, and then to this meaningful clinical outcome that could guide therapeutical decisions. That's what we do in DeepView. And we focus on a couple of clinical indications at the moment. The first is a burn wound indication. Right now, in America, there are over 1.1 million burns every year according to American Burn Association. For those patients, half of them, about 530,000 of them, will visit the U.S. ERs every year and 40,000 of them will eventually end up in a hospital care -- inpatient care and 8.1 days would be the average length of stay. Length of stay accounts for 72% of a burn patient's journey in the cost aspect, and average cost will be $24,000 per stay. The problem is, right now, for even the best burn specialists, they could have 50% to 70% diagnostic accuracy to determine where the nonhealing wounds are. We humans are not that stupid. When we realize we have a problem, we will always find a workaround. So the workaround today is wait and see. Wait up to 21 days to determine the need for surgery. And this, of course, extended the length of stay, causing a very high infection rate and other complications. DeepView, our technology, after years of development and all the structured clinical studies, we are now sitting at a 92% diagnostic accuracy on day 1, and that is for adults. For pediatrics, we're at 88% diagnostic accuracy on day 1. By a retrospective clinical study performed at Louisiana State, we found that using DeepView could reduce average lengths of stay by over 50%, by 4.9 days, and that's a cost saving of $14,500 per stay. And any companies come to present, they usually have this column of a market size and then the current market problem and their solution being the 1 and only great solution to the wide world. So here, we're no exception. We also follow that format. But I want our audience to walk away from this conversation remembering 1 thing that what in front of your eyes in this burn wound indication development is being supported by a $122.5 million U.S. federal government contracts. And to date, this is over $80 million plus spending to produce this wonderful results as of today. And that federal funding is granted to us by the agency called BARDA, B-A-R-D-A. They are part of the Department of Health and Human Services of the U.S. federal government. We have been working with them for the past 9-plus years, and they are the cornerstone in supporting this burn indication development. Back in 9/11, you folks might remember, when the Twin Tower fell, it killed over 3,000 Americans right away. If those Twin Towers didn't fall, what we have on our hand would be over 3,000 burn victims. And our country is not ready to handle that many burn patients right away. COVID taught us how important to have a quick test kit to decide the severity of the patient's suffering and the situation in order to save the healthcare resources. I sometimes think our DeepView in burn indication as a quick COVID test kit like for the burn wound assessment. We provide a very objective digital burn wound assessment diagnostic tool. That's what we're doing with the federal government support. Besides this, we actually also developed into other indications. For example, the other one that we're focusing on right now is diabetic foot ulcer. Why? Because these are very similar problems with a much larger market, 5.2 million DFU patients per year. These are only in the U.S., U.K., and EU. 15.5 visits per year costs up to $63,000 per DFU patient per year. And the problem is worse, no diagnosis available. 30 days, instead of the 21 days wait and see, here 30 days is what usually they do as a wait and see go through a thing called standard wound care for DFUs first, before the payers can approve advanced wound care products for the diabetic foot ulcer patients. The sad fact here is 40% of that 5.2 million will not heal. They do not respond, and they will require advanced therapy all the way to revascularization. Again, DeepView, as we announced in this January, through the data we have collected via very structured and rigorous clinical studies, where now our model's sitting at 86% diagnostic accuracy on day 1. For chronic wounds, there's no true day 1. What it means is the first encounter with the physicians that using DeepView can predict whether the diabetic foot ulcer will close by more than 50% 30 days later or not. And to do so, we provide a quicker time to advanced therapy, better wound healing and reduce overall hospital visits and utilization. This DFU indication is what we call the commercial indication in contrast to the burn as we call it a government indication in some of our references. If we look at this chart, this is actually on the cover of our AIM IPO package. We have the imaging kit, we have the data, we have the AI model, we have the burn indication. To me, we call this the data pipeline. I do not think the U.S. government just developed indications through this process with over $80 million spending, but rather they create a very, very strong, young, vibrant team working on this cutting edge of AI in medicine. That's why in 2021 when we came to the AIM market, we successfully raised $16 million to develop the DFU indication. Back then I told the IPO investors during the roadshows that if I come to you asking $16 million to do the DFU indication as what we plan to do, please, that is not possible. The reason we can do it is because we're sitting on the shoulders of [ giants ]. We have all this infrastructure already fully developed, and we know how to develop all these add-on indications through clinical trials and FDA and other regulatory body submissions and approval process. Besides AI DFU and AI burn, those are the flagship indications we're working on, we also have AI 3D in a POC-ready situation, proof-of-concept ready. What this provide is the wonderful measurement of technology development breakthrough in the wound size measurement. The wound is usually measured 3 to 9 o'clock, 6 to 12 o'clock in the dimensions, also the area and in volume of a wound. So with our DeepView imaging technology right now, we can actually measure the wound without requiring placing the reference markers into the wound. We don't require that. Other technologies, they require multiple imaging of the same wound in order to achieve a size measurement. We don't need that either. In our case, it's a 200 millisecond snapshot, and we're done with our imaging collection along with our DFU and burn assessment. So this is a true technology breakthrough in our development, and we will definitely move this forward to product passage, and we hope to get it registered with FDA before the end of 2023. Other indications we're working on, including venous leg ulcer, critical limb ischemia, digital-guided therapy, and AI cosmetics, and we're working on those as horizon indications right now. Commercialization strategy for the burn; we are actually going through the similar pathways as our other BARDA sister companies, who has gone from advanced research and development type of contracts, which is what we're in, into the federal procurement and deployment type of contracts. And in that way, the federal folks will really support the late-stage clinical studies, the federal procurement, and federal deployment for this mass casualty countermeasures. We're here to make Americans better prepared and ready for those unfortunate events. If such event happens, the U.S. folks will have access to this type of technologies and not just to look for any other things in that situation where the healthcare resources will be very, very key to protect. And the burn will be using ERs, burn centers, inpatient, new technology, add-on payments type of approach, along with a bundled pricing in the DRG situation. That is a capital sale and will go through the U.S. government federal procurement and deployment type of contract to help the market penetration. That's the thought process. And for that diabetic foot ulcer, because all the wound care clinics are way more sensitive to the pricing points and they use this way more often, so we are actually going to offer it at a very low entry price, low entry cost for the imaging kit. And instead of focusing on the system, we actually have an ongoing software as a service, SaaS model, a per-click-based model. An average wound care clinics based upon CMS data would see 220 new DFU patients per year. And let's do a very ultra-conservative estimation. Let's say 4 imaging sessions for each of the DFU patients, then per year you can calculate that up to 880 billable imaging sessions. Well, let's round it up to 1,000 imaging sessions that's billable. And if I charge $12, $15, I don't think that's overcharging at all for what we can provide. In other words, our plan is to offer a licensing annual fee around say $15,000 per year. So this will be used at ERs, wound care clinics, podiatrist office. These are outpatient CPT code based, and these are very SaaS model driven. In other words, low entry capital, but focus on the software-as-a-service cost later. We talked about the DFU commercialization strategy. And then 1 thing to note is for our burn indication, we have the U.S. federal government procurement and deployment side of [ venue ] to look for. And in very early February of this year, 2023, the U.S. federal government sent out this Source Sought Notice, I have a problem pronouncing that second word S-O-U-G-H-T, so it's SSN, that U.S. government announced to the world that they would love to acquire burn wound imaging technologies in order to provide the burn wound imaging technology could enable physicians to efficiently triage burn patients and make more informed treatment decisions. Sounds familiar? That's what DeepView does. And I think the DeepView would be an excellent candidate to answer for this SSN. What I think is irrelevant. Let's look into the details of the SSN so that we can have a good assessment. The system that the U.S. government need would require to measure burn size and depth. In other words, severity. And DeepView definitely can do that, in combination with AI 3D and AI burn algorithm module that we have already at 92% adults, 88% pediatrics with the data collected over the years that you cannot buy or download online. You have to acquire that data through our imaging modality. And the U.S. government asks for ease of use, high accuracy, and interpretation. And that's also what DeepView does. It's a noninvasive, easy to interpret, actionable information for clinical decisions. And our federal sponsor even highlights in this SSN that 50% accuracy read of nonburn specialists as a benchmark. Again, 92% for adults, 88% for pediatrics. And they ask for regulatory readiness. The product needs to be FDA approved or near FDA approval. We're going to embark our validation study for AI burn indication. We are close. We're near FDA approval. Demographic data diversity; our data are currently acquired from 13 leading U.S. research hospitals across the country, and it covers multiple conforming parameters such as race, age, gender. All those considered, Spectral MD is FDA breakthrough designated technology indication for burn by U.S. FDA. We have maintained open, transparent communication with U.S. FDA and those conforming parameters were actually decided by both FDA, BARDA, and Spectral MD early on in our efforts. The U.S. government also asked for commercialization and infrastructure readiness. The products need to be developed and ready for commercialization with U.S. manufacturing. We use our contract manufacturers located not only in U.S., but in the wonderful state of Texas, in the most brilliant city called Dallas. We are fulfilling all these requirements 1 by 1 by 1. So this is why we consider Spectral MD would be an excellent candidate for this SSN. The deadline is tomorrow. We're already submitting our application. When I IPOed on AIM for our presentation to the investors, some of the investors asked me what keep you up at night. I honestly told them that I would worry about how to grow our team to transition the company from an R&D-centric team into a commercialization-ready firm. And you guys already heard me talking about how close we're getting the DFU moving forward, how the burn is performance improved constantly. Now we're even looking at procurement and deployment type of contracts. Commercialization is coming very, very soon. So I'm happy to report on this leadership chart, we have 60% of the folks are hired after the London IPO on AIM market. This money has helped us to move this forward. That's why we got all these great veterans joining our team, and I'm forever grateful to our AIM IPO investors. In summary, in a nutshell, we have these 2 substantial high-growth wound healing diagnostic market opportunities both in burn and in DFU. And we have the AI 3D following right up very strong, already turned from a horizon into a product development now. And we have the FDA breakthrough designated AI technologies for wound healing diagnostics, and we have a robust, ready to execute commercialization strategy. Again, the burn SSN is an outside independent validation for what we're doing as a great meaning in the U.S. federal mindset. With that, I would like to turn the presentation to our CFO, Mr. Nils Windler, to report our 2022 financial results. Nils, please. Thank you, Wensheng. Yes, also for me, thank you very much for joining today. I'm excited to be here to talk about the 2022 final results for Spectral MD. I will start off with talking about the operational highlights when we look at our burn application, our DeepView AI burn, that is the application that is funded through BARDA, BARDA in August last year awarded us an additional $8.2 million in addition to our current contract. That brings total funding awarded by BARDA to roughly $123 million. And this has been for us a very strong sign that BARDA keeps highly committed to the progress of our burn application. We also released the Burn Image Assessment Study results. In that study, we asked physicians to draw on a picture where they believe where the nonhealing areas of the wounds are and give a treatment recommendation. The results of that clinical study was very impressive and reinforced our value proposition and the clinical need. In 31% of healing wounds, those physicians incorrectly [ selected ] immediate referral to a burn center or suggested surgery, so in other hands, the wound which would heal with standard care and doesn't require that extra step, which is more expensive, and in 74% of nonhealing wounds, the physicians fail to select immediate referral or surgery. We also made significant progress in the enrollment of our burn AI Training Study. At the end of 2022, we finished the adult enrollment in that study. From a DFU perspective, at the end of 2022, we are on track to complete the enrollment in our current DFU validation study by the end of June this year. The goal for 2023 is to submit for regulatory approval in the United States. We also made significant progress on the development of our fully handheld and wireless device called DeepView Snapshot M, and it's important to report that this handheld version has very similar performance to our current cart-based version. Looking at the financials, we ended 2022 with a very strong cash position of $14.2 million. This strong cash position, along with future monthly cash inflows that we are receiving from BARDA, we believe we are very well positioned to execute our commercial strategy for the DFU and burn technology and are well funded to support our current commercial goals all the way through 2023 into 2024. Average revenue was $2.1 million per month. That is mainly coming from BARDA. And I want to, again, just point out what is really, really important, the R&D revenue that we are receiving is a monthly sustainable cash inflow to the company. It is driven by the R&D expenses that we are incurring for our burn indication that is invoiced on a monthly basis to BARDA and marked up with a fee for administrative purposes. R&D expenses have significantly increased from $9.5 million to $16.5 million. The main driver was our burn indication, but we also significantly invested more money in our DFU application. Other operating expenses, so all the expenses excluding R&D expenses, increased by 16% as we're further scaling up the organization to prepare for commercial readiness. Wensheng spoke earlier about building up the leadership team. We also made significant investments in parts of the organization such as marketing, quality, regulatory, manufacturing. Adjusted EBITDA for the period was better than planned, and we ended with minus $1.5 million of adjusted EBITDA. This slide illustrates how the BARDA revenue works. So again, it's very important. What we do on a monthly basis is we invoice BARDA for the expenses incurred related to our burn application, mark that up, submit it, and receive payments within up to 5 days. So what that does is provides a sustainable monthly cash inflow to the company. Out of the $25.4 million of revenue reported in 2022, $24.9 million was coming directly from BARDA. The remaining $500,000 is the funding received from U.S. military for the development of our handheld device. Now switching over to the 2022 post period highlights. I already spoke about finishing the adult enrollment in our burn AI training study, but we've also been able to finish the pediatric enrollment in that study. So in other words, the enrollment of subjects in that burn AI training study has been completed and upon completion of the study, that will take us to the next step in the regulatory pathway, which is clinical validation, which will eventually be used for FDA submission. We spoke about the federal contract opportunity already, which was initiated by the Department of Health and Human Services, to which BARDA belongs, for burn wound imaging technology. As Wensheng Fan mentioned before, the company responded, and will now be evaluated for contract fulfillment. We also developed our AI 3D wound measurement that allows wound size measurement with a single picture and without any external reference markers needed. We believe that this will seamlessly fit into the clinical workflow and will provide benefits for the healthcare system. From a DFU perspective, we were able to increase the accuracy of our algorithm from 81% to 86%. Our U.S. clinical study remains to be on track with additional sites being incorporated. We're planning to finish enrollment by the end of Q2 2023 and that study will provide us with the necessary data needed for our FDA and UKCA regulatory submission. We also initiated a clinical study in the European Union with the Royal College of Surgeons in Ireland, which is conducted at Connolly Hospital in Dublin, Ireland. The data collected here will further support our regulatory submissions for the United Kingdom, the United States, as well as the European Union. We continue to develop our handheld device, and we're also working on continued funding from the U.S. government. We will keep the market updated on the developments here in due course. To summarize and to give everyone an outlook of where we are today and what we're planning for the future, this slide summarizes future regulatory milestones and what we mean with regulatory milestones. Those bubbles on this chart here indicate where we believe or when we will believe we will receive regulatory approval for our different indications. I'm starting with AI burn. We believe that we will file for UKCA mark in 2023 and anticipating to get UKCA mark approval for our burn application. Following a potential federal procurement contract award in 2024, FDA and CE mark approval for our burn applications are planned in 2025. Our DeepView indication, same as for burn, we're planning to submit for UKCA in 2023 and plan to receive the approval in 2023. We will submit for regulatory approval in the U.S. in 2023 and anticipating the clearance from FDA in early 2024. DeepView Snapshot, that is our handheld device, that is the device that's been funded through the U.S. military. I mentioned earlier we are working on a follow-up contract, but regardless, our plan is to have an FDA regulatory approval ready product in 2025 and will then file for UKCA and CE mark approval in 2026. Our AI 3D application to wound measurement application, our plan is to register this AI 3D wound measurement with the FDA in 2023, and then followed in 2024 and 2025 in the United Kingdom and the European Union. This will be a phased approach as the 3D measurement goes hand in hand with the indications of wounds that we will get approved in the future. Looking further out for horizon applications, so those are wound indications that we can look at, we're already working on proofs of concept with the plans of getting regulatory approval in the United States, U.K., and European Union in 2026. What's next? What's the big picture? This slide illustrates the wound therapy progression, all the way from prevention, to the [ rights ], to follow up. So we're Spectral MD right now. We're a diagnosis company. We're diagnosing if a burn wound will heal or not, or if a DFU wound will close or not. We're currently targeting burn centers, A&Es or Eds, wound care clinics and hospitals. Looking into the future, we will add indications to our current platform, such VLU or CLI, and we're thinking about expanding the sites of care to medical clinics, plastic surgery, or even at-home use. Very critical in this journey is the development of our handheld device, which is illustrated on the left, because the handheld device will eventually allow us to enter the homecare or enable the home use, with the potential to then play for chronic wounds and preemptive wound screening. Treatment decision; our device, our output is binary. It's very simple, although very, very meaningful. We determine if a wound will heal or not. Going forward, we might be able to also have our AI predict what kind of wound therapy has the best outlook for an outcome. We are not competing with current medical device manufacturers, but going forward we might partner with those companies as we're all working on the same goals. From a follow up perspective, follow up means evaluating the therapy efficiency as well as providing needed reimbursement support for insurance, now that we have our AI 3D wound measurement application coming, that is one area that we're also focusing on. That concludes the end of our presentation. I would like to thank everyone again for joining, for interest in Spectral MD, and I hope to see you all very, very soon. And with this, I'm handing back to the host.

I think there are questions raised throughout the presentation and there are many questions about the timeline of the process, of the progress kind of questions. They were already addressed by Nils showing the chart of each indication because there are quite a few questions about the burn timeline, DFU timeline and those nature. So those are already answered by the very latest slides of Nils. And there was a question asking about other competing technologies on the market. And I do want to address that to a little bit of detail that for the imaging-related technologies being used in the wound care industry, I actually admire all our peers who are using imaging solutions to do, for example, the size measurement, even though they might require a reference marker, even though they might require multiple images taken. I admire those companies who can measure wound temperature, the blood perfusion, the oxygen saturation, the biomarkers left behind, the bacterial residue left behind by the biofilm, and all those companies, I really admire their efforts and I consider them as our peers and great trailblazers for bringing imaging as a tool in the wound care industry. Our thought process at Spectral MD is wound healing is a very complicated matter. It is going beyond just 1 physiology measurement. That's why we go from near UV light through the visible band all the way into near IR lights. We measure many, many physiology signals, and we actually connect all those individual dots to make the injured tissue spectral signature. So to us, we provide a really coherent solution with many physiologies measured at the same time, and our human brain will [ not ] be able to process all that information anymore since the AI introduction. So in a nutshell, to say what our competing edge over the other folks would be our AI processing engine, give us a very, very strong support in really figuring out which parts of the wound are truly nonhealing with all those physiologies combined and included together. And also, asks, with this, what will be the entry barriers or adaptation barriers when we bring the product to the market? This is a brilliant, brilliant question. We are an FDA breakthrough designated technology in burn. And for diabetic foot ulcer, another vicious problem, this is also very, very new. So on 1 side we are truly a disruptive technology. On the other hand, it requires a lot of good market education and promotion and great data to support our claims, and also a very, very important workflow to reach market users. Actually there is also a question says, your DFU accuracy improved from 81% to 86%. What percentage do you see being achievable here? That is actually -- at 86% compared to nonexistence. The 92% from burn compared to the current standard of care of 50% to 70%, these are already well meeting the requirements for the product to be FDA cleared. We're continuing our structured clinical studies for both burn and DFU because we see there's no better way to reach the market than having the market to be involved as part of our development. Our burn study is running at 13 leading U.S. research hospitals. All these wonderful doctors wonderful doctors, they use our products, they actually publish the data, eventually they will do the publications [indiscernible]. To me that is the best way to actually educate the market, bring data to our targeted users. We will evaluate from a business perspective and gradually build a solid case for our DFU clearance FDA submission package. And once we clear the products, we will be able to bring the product to even more broad user base in order to get more data so that to improve the AI performance. And there are questions asking us about whether we have the aspiration to go to the U.S. market or the timing of that. We actually consider AIM treating us very well, and we got the $16 million during the IPO. We're forever grateful to our investors. With that kind of money, it will support us over the DFU clinical studies and regulation submissions. We believe sky is the limit for Spectral MD. When the time and opportunity is right, we might consider U.S. market as well. And I want to emphasize point that U.K. and EU, this is not only a launching pad for our financial side, but also good technology launching pad and a market launching pad. For example, in Middle east, the diabetic population for some countries are above 20%. That is a very, very bad situation. I think those market, regions would be much better served for our team to be in the U.K. and EU and using the CE mark and the UKCA to really approach those regions. So to us, we are growing. Our indication is a worldwide problem. We provide a solution to the worldwide problems in both burn and DFU and U.K. and EU is a very, very critical component in Spectral MD's overall global development strategy. One last thing I want to say is last Friday marks the 1-year anniversary for the Ukraine war, and I gave an interview to Medical Technology a few weeks before, full article published by Catherine and she provided a good description about Spectral MD in the burn indication in the mass casualty situation. When you have a war started, you ask for some kind of technology that's much needed, that's already too late. It takes 10, 15, 20 years of development for those technology for the mass casualty countermeasure. For that, I applaud the great efforts and great vision from the U.S. federal government in developing the burn-related mass casualty countermeasures. My other sister companies in the U.S. burn portfolios are all therapeutical companies. Spectral MD is the only diagnostic solution for the U.S. burn mass casualty countermeasures. So with that, I thank you very much for all your great interest in Spectral MD. And this concludes the company's presentation to this webinar broadcast. Thank you very much. Have a good day.

Wensheng, Nils, thank you once again for updating investors today. Could I please ask investors not to close this session as you will now be automatically redirected to provide your feedback in order the management team can better understand your views and expectations. This will only take a few moments to complete, and I'm sure will be greatly valued by the company. On behalf of the management team of Spectral MD Holdings Limited, we would like to thank you attending today's presentation. That concludes today's meeting. Good afternoon to you all.