Artificial intelligence is actively penetrating healthcare

Unlock business potential through effective first dataset management solutions.
Post Reply
tanjimajuha20
Posts: 563
Joined: Thu Jan 02, 2025 7:15 am

Artificial intelligence is actively penetrating healthcare

Post by tanjimajuha20 »

AI is increasingly penetrating healthcare. For example, it is used in all medical institutions of the Moscow Health Department as part of an experiment to introduce computer vision. "Neural networks are becoming an increasingly popular assistant for radiologists," said a representative of the press service of the Center for Diagnostics and Telemedicine.

The practice is also spreading indonesia telegram to the regions: in February, Moscow provided access to a special platform with artificial intelligence services for Russian medical institutions, and medical institutions from 20 regions are already using them.

Some organizations are already conducting successful tests on the use of AI in radiation diagnostics. At the IT Dialogue forum, Irina Kamyshanskaya, a radiologist at the Mariinsky Hospital, spoke about her experience working with AI in radiation diagnostics. Together with the company Care Mentor AI, the hospital developed a program for analyzing images to identify the degree of flatfoot. To do this, they collected a database of about 5,000 images over the course of a year, on which they trained the AI. "The program works very well, is easy to learn, and makes few mistakes when calculating the angle of the arch of the foot - no more than 1.3 degrees," said Irina Kamyshanskaya.

Together with the company "Esper", the Mariinsky Hospital developed the program "Esper.Scoliosis" for detecting scoliosis. The program can correctly recognize vertebrae, designate them, and see the degree of scoliotic changes in the spine. At the same time, when compared with a doctor, the program often more accurately estimates the angles of scoliotic deformation.

According to research by scientists at the Center for Diagnostics and Telemedicine, the diagnostic accuracy of some artificial intelligence services reaches 95% - the researchers studied the diagnostic accuracy of AI and compared it with the indicators of radiologists. They came to the conclusion that AI is not inferior to a doctor. However, they also noted that the degree of agreement between a specialist and a neural network varies widely and depends on their quality and the type of changes in the images obtained during radiation studies.

Vladimir Frolov claims that OxyTech solutions provide results with an accuracy of 94-99%. "We recently conducted a clinical trial of one of our models. Out of 1,000 images of the spine of people of both sexes from 18 to 90 years old, there were only five false positive and seven false negative research results," said Vladimir Frolov.

He believes that the use of AI in healthcare has many advantages: it does not get tired, is not affected by emotions, and can work at any time. However, the final word should still be with the doctor, and AI should always remain only a specialist's backup.

Read also

AI-powered medical image analysis now available under compulsory medical insurance policy
Artificial intelligence (AI) technologies in healthcare have learned to identify pathologies when analyzing medical images with an accuracy of up to 99.98%. AI can currently only identify 37 groups of pathologies out of 120 most significant ones. For further training of AI models, Moscow-based EMIAS is ready to provide scientists and developers with labeled
Post Reply