Michael Sughrue Presents Omniscient Neurotechnology at LSI USA '24

Michael Sughrue  0:03  
I am a neurosurgeon by training. My name is Mike Sughrue, I co founded a company called Omniscient Neurotechnology. We are Australian based. We've heard our first employee and had to send them home about two weeks later because COVID broke out. So we've grown under adversity. But despite that, we've we're an early growth stage, post approval neuro technology company that focuses on AI for machine learning, for making things like fMRI and DTI, clinically powerful to answer questions and prove out that we have legit AI chops you can see in the lower right that we won the South by Southwest award for AI machine learning, we also won MCI. And so what are we using AI really to do? And ultimately, as a clinician who's been involved in neuroscience, I've been over 3000 brain surgeries. And I think across neuroscience, the one of the big problems we have is that we don't really know when we see a patient, what is actually wrong in the brain, where's the problem? Well, one of the things that's come out of this is that we have these enormous Tam's with things like depression or, and we just saw even just tinnitus. And these are fundamentally brain problems, but they're highly complex, multi area network problems. And ultimately, these are some of the biggest let diseases that we have left that we don't have effective treatments. One of the main reasons we don't have effective treatments is that we don't entirely understand these. But as we begun to work with a new area called connectomics, which has been my area of research for over a decade now, we can begin to use, hopefully, by starting this video. MRI imaging sequences that have been around for quite a while diffusion tensor imaging on the right is, you'll see it's really great for advertising, but we've never really leveraged it for clinical practice to its potential, in part because the brain has a highly complex structure. On the right, we have things like resting state functional MRI, now function wise, some people may have heard of you lay in the scanner, you move your hand, the brain brings blood to that area that you're using, and you can deduce where function came from. But it turns out, you don't need to actually do a task at all. And what makes that really powerful is it's clinically scalable. What makes it even more powerful, is that mental illnesses across the board show abnormal behavior on fMRI. But in order to mine and use this data in a clinically effective way, you need to feel called connectomics. Think about it like genomics for brain connections, because it's just a highly, highly complex data set. And AI is really good at finding signal in noisy data sets. That's all great. And it's been in the research world for over 15 years. It's a really advanced field. There's large meanings of it, but what no one had ever done. And I began going to these meetings and finding out fact is figure out how do you make a product out of this that actually moves the goalposts, we've done that. So what we've done is we've basically built, cleared, and brought to into the clinical practice trained physicians how to utilize very large complex data sets at the point of care, and get them all the way from the MRI, to a knowledge base question that people can use for a variety of different applications that I'll show a couple in a second, and got them onto image guidance or other medical device systems. More importantly, we've really quickly rolled out this industry, we've made it a household name and least neurosurgery. And we're expanding into Psychiatry and Neurology. And more importantly, we've really, because we are the first people to ever really make a product out of this, we not only have good intellectual property we have, as you can see your 39 patents. But some of these patents are at really critical choke points of how do you show data in the brain from even in ways that are naive to fMRI. So one of the nice things is it's easy to catch infringement. And it's hard to engineer around patents like that. And and that's really where we focused. And again, pretty much everything I'm going to show you has some patent on how the UI works. We've also done a lot of interesting approaches in AI. Again, there's a lot of prior art from the research. But we there's some new and really useful ways to do it that we patented. And more importantly, we've also patented the upstream choke point, how do you get data back and forth from a cloud server in a way that a hospital let you do that? And we published a lot. So as you can see, these are papers with our technology, either by us or other groups. And they're currently over 41. To date. More recently, there was a paper in the journal neurosurgery, which is a landmark study. And what it shows is that our flagship product called Quick tome, when used in neurosurgery can substantially reduce disability. And it's in a way that I think we think will probably move this towards standard of care relatively soon. Now, the quick Tom platform, again, has two components to using diffusion tensor imaging correctly for different applications and surgeries, stroke, traumatic brain injury, and on the right for really getting into more functional neurologic disorders. We really have spent a lot of time thinking about how to make it answer a question. And I'll show why having better knowledge takes other devices, other platforms, and other tools that we've really underutilized in this case, you're gonna see transcranial magnetic stimulation, and actually makes entire new treatment paradigms possible. You see, this patient was treated for a brain tumor was left paralyzed. And we don't make this hardware, what we do is you can actually image guide this hardware. And if you know where to put those stimulator, you can stimulate the brain and often get it to rewire. But where's the problem? This is his motor system. And what you're looking at is the AI telling us how these areas communicate to each other. It's a data set of over 144,000 data points that we're working with, which you can see in the blue is an area that's supposed to be red, it's supposed to be correlated and talking to the rest of the network. And it's not, we can then export that into a treatment system and stimulate it. And for him, this was almost miraculous. So within one day, he's beginning to walk again. And having treated 1000s of brain tumor patients recovery like this, and five days from him, he Priestess is just not what happens. Again, he didn't need to be paralyzed, we would have had to hope that he made it a recovery on his own, which often they don't do, or they do very slowly. But we didn't, we wouldn't have known that he didn't, that didn't have to happen. So we really think this is a revolutionary technology, we went from a standing start to really moving very quickly. Obviously, being a well known neurosurgeon myself, I can get the product in people's hands and at least get them to talk to us. That's helpful. And but more importantly, we're growing as quick as the hospital systems will allow us to grow. So this is what we're calling for the next two quarters. Again, we think we're moving very fast given that we wrote our first line of code about four years ago. More importantly, we have a strategy that we think is very, very lucrative from a unit economics standpoint, and as fast as the hospital system allows us to go. So we get in the door and land with brain surgery, which has the highest gross margin procedure in the hospital. So they listen to us about that. In proposed COVID era, almost every other capital purchase gets put into a much different sales cycle. But we can show them with current economic evidence. So we make the money and we improve care and reduce costs. We have all that data. What we then do though, is because we sell as a SAS, we can basically begin to spit band out into other parts of the neuroscience which are lucrative, but not as lucrative as brain surgery. And this, again, begins to scale that account and increase revenue. You can see this is one of our very first accounts, we haven't been in market long enough to get all of this proven out. But you can see that a hospital starts out initially with neurosurgery. But eventually, you have eight or 10 Different people who use this on an almost daily basis. And these accounts go from $100,000, your recurring revenue account to a $500,000, your recurring revenue account. And this is without having to really go through a capital sales process over and over again. More importantly, we have gross margins that are about 85%, which is pretty typical. And these are actually improving over time. Because we have all the value of software based services, we don't make hardware. In the last year, again, these are Australian fiscal years because we're an Australian company. So the 2023 revenue grew 300% year on year, we're on pace to pass that our current quarter is somewhere going to land about 1.4 million, this is actually getting pretty typical for us, we hope to start passing 2 million a quarter. So we're growing quickly because we have strong product market fit because we move the goalposts pretty significantly. But how are we doing on expanding? Okay, so we start out at all of our accounts, these are the accounts that we have renewals on and been around long enough to have data with one person. But over time, many times we have eight to 10 people or even more. And we've started to sign up entire ad ends to the service because again, three to four hospitals. And then some of these, this is get really quickly getting more. And what you're seeing is now we've done about 7000 cases to date. And that's growing pretty quickly because we're doing about 500 cases a month. So more importantly, it's battle tested in the real world, we know that our software works and it tells the truth to the extent that is verifiable.

We've also the begin to jump onto the generative AI angle. And how do we use generative AI in a way that actually is helpful. So you know, there's not lots of bots and things and open AI and chatty video and these are all great tools and people will find the right place for it. Well, we thought generative AI was most useful for is to actually help people who aren't experts like general practitioners, psychiatrist, etc. Who aren't not brain surgeons and neurologists interface with our tools and find the circuit of interest. So this is a tool we just we've just launched it luckily comes with its own codes, we're actually dovetailing off of existing code. And again, we're just beginning to get into our first hospitals with that, essentially, the patient just tells me what's wrong with me what is their psychiatric function problems. And we're not here to make a depression diagnosis, we don't believe that disease really exists as a single entity, it's too heterogeneous. But we can then begin to look at the cognitive and emotional circuits, we know where they are from our large dataset. We've been in training on large hyper cluster models, and we have over 110,000 data sets, it's probably the largest in the world, we can then begin to use really sophisticated tools to benchmark their conductivity. And then quickly go from where do you look in the brain? Is it actually a normal? Is there an actual cause for the patient's symptom, and more importantly, linking it to a treatment plan that generative AI can say that provides evidence based treatments on how to rewire circuits rewire circuits in the reward system, either through medicines, therapies, or even things like doing psychotherapy that better way, and again, explaining this here, I can't do it justice in this time period. We're, again, cleared, we have good solid evidence. But what we're now doing is going more aggressively and expanding our code list and reimbursement. We have several studies that are currently ongoing, we've learned how to do this cost effectively. Where, again, a lot of times trial, randomized trials aren't the best way to answer our questions. But more importantly, we've really begun to build registries and really get scaling to the level of entire hospitals that are doing large amounts of studies with us. We anticipate that in the relatively near future, we should have a more broad reimbursement across the entire neuroscience line and hopefully into psychiatry in the next few years. Finally, we're, as I pointed out, early growth stage, we're growing quickly. We're we just finished an equity round, but we will be raising again very soon. And our goal is to get to profitability, which we think are about a year and a half off, at which point we have a lot of different options. Most importantly, we don't necessarily have to exit with with an acquisition, we can exit a lot of different ways, including IPO. And that's currently our default strategy, which we think we'll be ready for in the next few years. As our revenue begins to go up, and we begin to further build up our base. So we're also interested in partnerships because we think there are a lot of different ways that people can use our tools to accelerate their own business models. Thanks, everyone.