Researchers from the Instituto de Salud Carlos III, the Mundo Sano Foundation, the Universidad Mayor de San Simón and the startup SpotLab have developed a new system for analysing blood samples using mobile phones, a telemedicine platform and artificial intelligence. This is demonstrated by a paper to be presented on 27 October at the International Congress of Tropical Medicine and Malaria (24-28 October, Bangkok).
Microscopy is the gold standard for the diagnosis of several diseases caused by blood parasites, such as malaria, which causes more than 241 million cases per year, Chagas disease or filariasis. This multi-centre study proposes to use artificial intelligence in smartphones and real-time assistance from a telemedicine platform to revolutionize microscopy diagnosis, helping the microscopist to provide an accurate and faster diagnosis. For this purpose, 145 blood samples from two different centres were digitized and more than 5200 labels were created. An algorithm was developed with a sensitivity of 92.88%, 80.85% and 84.24% for filariasis, malaria and Chagas diseases respectively. The trained algorithms were deployed on both the telemedicine platform and the smartphone to provide real-time assistance.
This work, funded by the European Union, transforms a conventional optical microscope into an intelligent digital microscope, using artificial intelligence algorithms embedded in a mobile application on a smartphone, which helps in the diagnosis of these parasitic blood diseases. This system can be extended to other infectious diseases, making universal diagnosis more accessible.
SpotLab is a social enterprise that seeks to democratise access to quality diagnostics by bringing artificial intelligence to diseases that are not yet in the digital world. To make this possible, we use exponential and affordable technologies – smartphones, AI and 3D printing – to reduce the time, cost, distance and CO2 footprint of medical diagnostics. Our technology is present in Europe, Latin America and Africa. For more information, please visit us at https://spotlab.ai/