Decoding of X-ray examinations using AI will be introduced in Moscow clinics / City News / Moscow Website

A decision was made to introduce autonomous artificial intelligence in city clinics to analyze X-ray studies. From May 2024, patients will begin to receive conclusions from the neural network in electronic medical record after performing fluorography or chest x-ray.

“In Moscow clinics, artificial intelligence will analyze X-rays without the participation of a doctor. An average of two million chest x-rays and chest x-rays are performed annually, most of them for preventive purposes. For example, in 99 percent of fluorographs no pathologies are detected, and doctors are engaged in describing the norm. Given the vast accumulated experience in this field, smart algorithms are able to accurately determine the absence of signs of disease. Therefore, we are ready to transfer the analysis and description of research to AI services configured for maximum sensitivity. For this purpose, a special compulsory medical insurance tariff has been introduced,” Sergei Sobyanin wrote in his Telegram channel.

Source: telegram channel of Sergei Sobyanin @mos_sobyanin

The development of smart algorithms in Moscow medicine has reached such a level that it is now possible to switch to using artificial intelligence services without the participation of a doctor. For the first time, when describing radiation studies, neural networks will be able to independently issue conclusions. They are configured for maximum sensitivity and are able to detect even the smallest deviations that are inaccessible to the human eye.

If the research results do not reveal any pathologies, the AI’s conclusions will automatically be entered into the patients’ electronic medical records.

At the same time, to ensure that the risks of diagnostic errors are reduced to zero, until September the results of the AI ​​analysis will be rechecked by the country’s leading radiologists. Specialists of the clinic named after Professor Yu.N. Kasatkin Russian Medical Academy of Continuing Professional Education will conduct additional checks within 48 hours and form their own conclusions, which will also be saved in electronic medical records. Thus, patients will have access to two protocols – from the AI ​​service and from a medical specialist.

Studies in which artificial intelligence identifies pathologies will be sent to doctors at the capital’s radiology reference center for analysis and preparation of a conclusion. In this case, the patient will receive the result within 24 hours.

The work of autonomous AI services will be paid for within the framework of the compulsory health insurance system; a special payment tariff has been approved for this purpose. Thus, these technologies will continue to remain free for Muscovites.

The introduction of autonomous artificial intelligence will improve the quality of research and at the same time free doctors from the routine work of analyzing similar studies without pathologies. Their results will be available much faster, and specialists will be able to devote more time to those patients who really need it.

Every year, clinic doctors perform an average of two million such studies, the vast majority of which are preventive in nature. For example, when performing fluorography in 99 percent of cases, doctors do not detect any pathologies and are engaged in describing the norm.

The quality of the AI ​​service will be monitored by a team of experts from the capital’s Center for Diagnostics and Telemedicine (doctors, scientists, engineers and analysts).

If the results are successful, the Moscow Government will decide to introduce automated interpretation of X-ray data using AI technologies on an ongoing basis without additional review by a doctor.

Artificial intelligence in Moscow healthcare

Radiation specialists from the Moscow healthcare service have been using AI services since 2020. To date, more than 12 million studies have been processed with their help. Moscow doctors have over 50 services at their disposal computer vision. Neural networks help to find signs of malignant lung tumors, pneumonia, osteoporosis, aortic aneurysm, coronary heart disease, stroke, pulmonary hypertension and hydrothorax (a total of 37 different diseases) in medical images – mammograms, CT, MRI and X-rays. The diagnostic accuracy of AI services exceeds 95 percent.

Thanks to the introduction of artificial intelligence, the time to describe a photo has been reduced by an average of 30 percent. 26 development companies cooperate with the capital’s healthcare system, providing services that allow them to identify signs of one or more diseases.

Medical organizations from nine regions of Russia have submitted applications to connect to the Moscow platform with medical AI services. Sobyanin: Doctors from all over Russia have gained access to Moscow AI services