Artificial Intelligence Helps Russian Doctors Diagnose Stroke
In Russia’s Novosibirsk region, neural networks are being used to process medical images, speeding up diagnosis and helping physicians detect strokes at an earlier stage.
Since January 2025, AI systems in the region have analyzed about 250,000 patient images: roughly 63,000 mammograms, more than 17,000 CT scans, and nearly 170,000 fluorography studies, according to Alexander Kolupaev, Deputy Minister of Health of the Novosibirsk Region.
The technology has been integrated into the Central Archive of Medical Images (CAMI), ensuring that every image uploaded to the archive is automatically returned with neural network markings. In 2024, the region purchased three imaging modalities—chest CT, mammography, and fluorography—all running on Russian software developed by the company 'Third Opinion.'
In 2025, a new 'brain' modality was introduced by MosMedII. Doctors are now testing this function to analyze brain scans for early stroke detection. So far, the system has assisted in diagnosing two patients, while ruling out stroke in another case.
In these situations, time is critical. With AI, results were in doctors’ hands within 20 minutes, whereas additional examinations would have taken much longer.
