AI and thermal imaging assist rapid disease diagnosis
Peripheral arterial disease (PAD) involves the narrowing or blockage of vessels that carry blood from the heart to the legs. Although less well-known than other vascular conditions such as heart disease or stroke, the condition is widespread – and poorly diagnosed. “More than 230 million people suffer from PAD globally,” says AI-CARE project coordinator Georgi Kadrev, co-founder and CEO of Kelvin Health(opens in new window) in Bulgaria. “Every year, more than 22 million people develop the most severe form of the disease, called critical limb ischemia, which markedly reduces blood flow. This can lead to amputations, and has a fatality rate of 70 % within five years of diagnosis.”
Clinical need for innovations
Current diagnosis methods include taking blood pressure measurements at various points in the ankle, which can be imprecise. In particular, diagnosing patients with diabetes can be challenging due to arterial calcification and a lack of sensitivity in the legs. More invasive methods, including angiography (X-rays used to check blood vessels), are only used in cases suspected to be serious. For Kadrev and his colleagues, there was a clear clinical need for a simple, accurate and affordable diagnosis of this silent disease. “Our concept was to develop a system that applies a portable thermal imaging camera that captures body thermodynamics,” explains Kadrev. “This generates a series of images, which are then segmented and analysed using artificial intelligence (AI) image recognition algorithms.” Kadrev demonstrates the technology by attaching the thermal camera to his mobile phone and taking images of himself. Any anomalies related to the peripheral vascular system would be detected using the system’s machine learning algorithms. “The whole premise of the concept is that our local body temperature is related to blood flow,” explains Kadrev. “Any abnormalities will be picked up by the algorithms, telling us of the presence of any arterial blood supply problems.”
Bringing medical innovation to market
Bringing the technology to market requires traversing what is known as the ‘valley of death’, with significant investment needed to translate an idea with promise into a viable marketable product. “We need to externally validate our prototype in a clinical setting,” says Kadrev. “Just getting to this stage requires a great deal of administrative work, intellectual property analysis, and awareness of issues such as medical data protection. We also need to develop a market access planning strategy.” The AI-CARE project provided Kelvin Health with the support it needed to reach this critical milestone. “We are now preparing for actual clinical validation and starting to collect clinical data,” adds Kadrev. “All this is a prerequisite to commercialisation.”
Regular screening of at-risk patients
The project has also helped to raise the profile of innovation in Bulgaria. “We have been able to work with key opinion leaders,” notes Kadrev. “We were recently invited to the American Heart Association Scientific Session(opens in new window) in Chicago last November, and the LINC conference(opens in new window) in Leipzig, the leading European conference on arterial interventions.” Horizon’s support for European innovation ecosystems is also key to fostering a start-up culture. This is especially true where existing venture capital is not oriented towards research solutions, especially in the healthcare space, because of the time and risks involved. “This has helped to put us on the map,” remarks Kadrev. The ultimate goal is to bring this innovation into clinical settings, and to establish the technique as an effective and cost-effective means of screening for PAD. “We could see a situation where patients at risk of PAD(opens in new window) – heavy smokers, people with diabetes – are screened twice a year, helping to save resources and save lives,” adds Kadrev.
Keywords
AI-CARE, AI, thermal imaging, disease, diagnosis, PAD, Bulgaria
