Diagnose and identify a disease: how neural networks help doctors in Moscow

Today, the Moscow healthcare employs 12 registered domestically developed AI-based medical tools approved by the Federal Service for Surveillance in Healthcare (Roszdravnadzor); 11 of them are neural networks that help radiologists find signs of diseases on CT images, mammograms, fluorograms, and X-rays. Artificial intelligence has already processed more than 9 million radiation tests and made about 10 million preliminary diagnoses based on patient complaints, as announced by Anastasia Rakova, Deputy Moscow Mayor for Social Development.

“Currently, in Moscow healthcare we are already using 12 AI-based medical tools, which are approved by Roszdravnadzor. They include 11 artificial intelligence applications for analyzing medical images and the Medical Decision Support System; this is 70 per cent of all tools registered in Russia. The use of neural networks in medicine reduces the test time and also provides medical staff with information for making more precise diagnoses and prescribing more efficacious treatment, while algorithms significantly increase the availability of healthcare for patients. Artificial intelligence helps radiologists find signs of diseases on medical images, while physicians and general practitioners are able to make a preliminary diagnosis when seeing the patient,” Anastasia Rakova said.

In accordance with the order of the Russian Ministry of Health, in 2023 constituent entities of the Russian Federation are required to introduce at least one of the AI-based medical products mentioned in the approved list.

The first one is the Medical Decision Support System (MDSS) used to make a preliminary diagnosis; the development has already been integrated into the Moscow Unified Medical Information and Analysis System. When seeing a patient, the doctor enters the patient’s complaints to examination records, and artificial intelligence analyzes them and suggests three highly likely diagnoses for the doctor to choose one of them or make an alternative judgment, while it is always the doctor who makes the final decision regarding further testing and treatment. The Medical Decision Support System was developed by the Moscow Healthcare Department and major Russian experts in accordance with official clinical protocols.

Other 11 are AI-based tools that identify signs of diseases on radiation images and help radiologists find signs of lung cancer, coronavirus, stroke, dropsy of chest, aortic aneurysm, etc. 150 healthcare institutions in Moscow make use of neural networks, including software for visualizing and processing images, diagnosing novel coronavirus infection and analyzing digital CT brain scans. They also widely employ modules for studying fluorograms, chest X-rays and CT scans.

The use of AI technologies in medicine makes the diagnostics process faster and more precise and helps prescribe efficacious treatment. Every day, neural networks make preliminary diagnoses based on the patient’s history or find abnormalities on medical images.

Moscow has been digitizing its healthcare system for over 10 years, while smart technologies are widely used in daily medical practices to accelerate and simplify diagnosis and treatment, providing excellent medical care to residents.