Georgi Kadrev Presents Kelvin Health at LSI USA '24

Georgi Kadrev  0:04  
Hello, everybody, it's pleasure to be here. I'm Georgi from Kelvin Health and our mission is to end preventable deaths and suffering from conditions like limp ischemia by combining the power of mobile thermography and Di. We alert a team of professionals with the long lasting experience in machine learning medical science and venture building. And this venture is very personal to us as it actually happened as a consequence of multiple serendipities in our founding team imagine hearing that both your legs needs to be amputated in 2018, shortly before my son was born, my wife and I were devastated when her father who nearly lost both of his legs because of a late diagnosis peripheral arterial disease, PDS result of diabetes, and unfortunately, more than 200 million people worldwide alike who suffer from PTSD. And because of the hard access to precise and timely diagnostics, more than 20 million, which is above the population of Florida, or the average population of a whole European country, develop critical limb ischemia condition that first takes your limb then kills you in less than five years. In almost half of the cases. Patients with diabetes like you are very exposed to very high risk of developing cardiovascular disease and complications, and ironically, are among the hardest to diagnose because of their typical higher tolerance to pain and medial arterial calcification. So this is Professor Petrov head of on geology and cardiology of city clinic. And when patients finally end up at his at his office, it's typical at quite advanced stage. Also the needs to be examined fastly rapidly. Under the current diagnostic methods are not good enough, especially for patients with diabetes. And in general as well, they're quite timely. Even the AVI the TBI take quite quite some time and training. The same goes for the cinematography and the Doppler ultrasound. On the other end of the spectrum, we have the X ray angiography, it's quite timely, it's quite invasive required injection of a die radiation of multiple hours in typically hospitalization. So provoked actually also, by the COVID 19 pandemic four years ago, we started thinking, can we use our long lasting more than 10 years experience in image recognition for generic images, and actually start building a diagnostic solution that is as simple as taking an image with additional thermal camera. Basically, using a pocket sized thermal camera attached to your phone like this one, we can capture the heat of the body detect local anomalies, and notify the doctor if needed. And this is something that can be done by by the specialists, but it can also be done by the nurse practitioner, by general practitioner, or even his wife. So this whole concept is possible because our local body temperature is directly related to blood flow. And if there is an anomaly in our temperature, this means that there are some anomaly in our blood flow. And there's a relatively limited amount of reasons that might affect it, such as typically inflammation, or directly a vascular problem, typically some type of blockage, like stenosis or thrombosis. And if we can collect enough samples, we can train good enough machine learning models that can assist in precise diagnostic process. There are naturally multiple potential therapeutic areas of them, some have been more explored or less explored in the literature. But we decided to focus a lot on vascular disease. As you can see here, it's pretty visible even with the naked eye a difference between let's go with a vascular healthy pair of legs, and a pair of legs that have critical limb ischemia below the knee line of the patient. We started collecting and annotating thermal images, also doing the other typical clinical pathway for the patient to examine them. And we were able to train machine learning models that can both segment the different zones of the legs, and also detect the level of profusion problem in each one of them with certain level of confidence. So the process again, is as simple as that capturing the thermal image, and then suggesting certain scores for the different zones. Together with the team of Professor Petrov. We decided to dig as further as possible and we witnessed that there's a strong correlation between what's visible within geography, which as I said, is very invasive. And what's positionally visible or with Kelvin. This is application, something that we see a lot of concordance along specialists for pairing and post procedure application, even in real time or close to real time. You can see how the profusion changes based on the distribution of heat. And we've witnessed more than a few cases unfortunately, where the doctor specialist might miss a severe vascular problem and say that this is normal, maybe typically after other like coronary procedure, but then just doing this thermal image and sending the patient to thermography to angiography to confirm that actually they need rapid intervention. Again, the global flow here is visible. In conditions like the ray syndrome or our to blockage, you can see even with the naked eye, what can be visible, our specialists got so excited that they decided to start taking photos of other territories vascular in the carotid area. And they saw that hypoperfusion in the internal and external might be visible based on thermography as well. So this is a new additional research area of ours currently. So encouraged by this application to vascular, we decided to focus a lot at first on building Kelvin Health as a vascular diagnostic assistive device that comes with the benefit of being non invasive, and non operator dependent, which is very important given the lack of specialists and the throughput that might handle but it's also rapidly applicable, and eventually even in a remote patient context in the future. So if we need to drill down in the value proposition, again, optimizing part of the workflow, saving time for the clinician, saving costs eventually for the hospital, and ultimately, of course, bringing better quality of life and healing for the patient. So, we already collected more than 12,000 patient data or sorry, 1200 patient data more than 9000 images actually of today. And we saw pretty high both sensitivity and specificity when we need to detect PhD presence in the different or to zones of the legs. We saw very excited specialists from various parts of the world that aim to you know, apply Kelvin in their their daily practice once it's regulatory, clear. And we also witnessed that if we do the calculation, and if we have a certain amount of charge on a diagnostic is the service business model, the total the serviceable addressable market for PhD diagnostics, based on the 14 point 1 million PD patient visits in the US NTU alone is more than 800 million. We saw we started two months ago, we started conversations with your specialist as well. And we witnessed that they are very excited also about the application in primary care eventually by podiatrist as well and in emergency response sitting. So this again encouraged us to think about the multistage approach where we start by after clearance to apply this in the hands of the specialist, but in a enabling the general practitioner and maybe the distant or not so distant future, also the patient to have such type of device for multiple conditions so they can be practically screened in their home. We are currently raising 4.8 million to speed up the regulatory clearance process and to prepare for commercialization. About 1/3 of it is already soft, committed by European investors. And they're multiple other companies aiming to build some type of wellbeing devices based on thermography. But none of them to the best of our knowledge has been building something for a socially significant device based on a well trained machine learning dataset. So if you also believe that a true unicorn company should positively affect the lives of many people, we are happy to discuss how we can join forces and help save the lives and lives of people like you together.