В Петербурге разработали нейросеть для выявления гематом мозга по КТ-снимкам

The software has already passed clinical trials

A software for detecting hematomas in the brain using computed tomography (CT) scans has been developed at the St. Petersburg Federal Research Center of the Russian Academy of Sciences (SPC RAS) in collaboration with scientists from the National Medical Research Center (NMIC) named after Almazov, SPC RAS reported.

The research center explained that the software is an application that helps doctors quickly identify chronic subdural hematomas.

The program supports the doctor's decision-making and allows monitoring changes in the volume of a chronic hematoma.
Alexey Kashevnik, employee of SPC RAS

Diagnosing a hematoma takes a lot of time for doctors, because due to its characteristics, it requires careful analysis and accuracy. A neural network trained on 4,000 CT scans of patients does this much faster. Based on the information received, the doctor can prescribe treatment.

Clinical trials have shown that the developed application allows analyzing brain CT scans on average 7 times faster.
Arkady Ivanov, neurosurgeon at the NMIC named after V. A. Almazov

The project was implemented as part of a cooperation agreement between SPC RAS and NMIC named after Almazov. In the future, the use of a neural network for evaluating CT scans will be introduced into medical practice.

Earlier www1.ru reported that Russia is planning to detect strokes using AI.

Read materials on the topic:

Faster and more accurate: in Moscow, doctors make a diagnosis based on the assessment of artificial intelligence

AI helped doctors in Moscow make 14 million preliminary diagnoses

Russian scientists have created a neural network to improve the diagnosis and removal of tumors

Sources
TASS

Now on home