Artificial intelligence reduces the time to diagnose COVID-19 on CT scans by a third

May 10

For a year now, Moscow doctors have been using AI-based computer vision technologies in radiology. In April 2020, a new tool was introduced to recognize signs of coronavirus pneumonia on CT scans of the lungs. It allows doctors to reduce the time to process images by an average of 30 percent and helps to detect COVID-19 faster.

In urgent cases, the attending physician receives the radiologist's report within five minutes after the examination and quickly makes the correct diagnosis. The technology became standard for the COVID-19 diagnosis in Moscow at the beginning of the pandemic, and later it was used to detect other diseases.

To date, artificial intelligence has already processed more than 1.5 million studies, more than 730 thousand of them for the diagnosis of COVID-19.

Until the end of 2020, the experiment was conducted in four types of studies. In addition to detecting coronavirus, neural nets analyze images to detect signs of breast cancer, lung lesions in various diseases, such as lung cancer. In 2021, the number of areas where AI is used to process research results has been expanded to 11. They now include magnetic resonance imaging of the brain, knee joint, prostate, X-ray of the musculoskeletal system and abdominal organs, and computed tomography of the abdominal organs and brain.

Computer vision capabilities are available for doctors of all Moscow medical institutions connected to the unified radiological information service of the Unified Medical Information and Analytical System (URIS UMIAS). The cloud storage of medical images unites more than 1,300 X-ray diagnostics units into a single network. When the tests are uploaded to the system, the algorithms apply color markings that draw the radiologist's attention to possible deviations. The images are divided into groups: the first ones to be processed are those where pathological changes are more likely to be detected. When processing an image, the doctor can see the original test, as well as the version processed by AI.

Neural nets improve the quality and speed of diagnosis, help doctors not to miss pathologies in a large flow of research and identify groups of patients with a high risk of disease progress.

Such services became available due to a single digital healthcare platform developed by the Moscow Social Development Department and the Information Technology Department of the city. They allowed the capital to cope with the challenges of the COVID-19 pandemic, ensuring continuity of cases management at all stages of treatment.


If you continue to use our website, you are agreeing to accept the use of cookies on your device. Cookie files ensure the website’s efficiency and help us provide you with the most interesting and relevant information. Read more about cookie files.
Accept ccokies