Among the participants in the Moscow experiment on the use of computer vision in radiation diagnostics, leaders were chosen. The rating of services based on artificial intelligence (AI), the use of which can increase the speed of diagnosis and reduce the number of medical errors, is posted on the experiment's website and will be updated monthly.
“We have prepared a list of the best solutions for an experiment in the study of computer vision in radiation diagnostics. Publication of the rating in the open access will help stimulate competition between artificial intelligence developers for the development of promising technologies and objectively evaluate existing solutions, for example, before being introduced into the healthcare system of other regions. Patients themselves will also be able to use the rating data outside of the experiment: many developers allow users to upload their studies to get a second opinion from the AI service, and an independent comparison will help them to adequately relate to the results, ”said Moscow Deputy Mayor Anastasia Rakova.
Assessing the work of AI, in addition to the main parameters – sensitivity, specificity and diagnostic accuracy – the experts took into account the adaptability of the service to a single city information system and the ability to work with large flows of research.
Services for automatic analysis of beam studies are being tested since spring 2020. They are connected to a unified radiological information system that unites all departments of radiological diagnostics of Moscow polyclinics and hospitals. Radiologists use 19 services to diagnose lung and breast cancer, COVID-19, and other chest pathologies, including osteoporosis and coronary heart disease.