AI-Based Stroke Diagnosis from CT Scans

AI solution that analyzes CT scans to quickly and accurately identify signs of stroke, its type and area

How does it work?

To classify and mark up images, image recognition uses ultra-precise neural networks that are trained on tens of thousands of images and conclusions. On input, they take in and analyze CT images of the head without contrast At the output, according to the results of the analysis, the AI ​​algorithm marks the images, highlighting areas with acute cerebrovascular accident – signs of early or late stroke. 

The following algorithms may be used together with the CT Stroke service:

– ASPECTS scale to evaluate early CT changes during stroke

– CT Angiography to identify blood clots on contrast CT



Accuracy of the model is ~96%


CT Stroke -
diagnosis of stroke using artificial intelligence

The service provides

Reduced burden on the doctor due to the automatic marking of CT scans
Minimization of medical errors: AI notices abnormalities that the doctor can miss due to overload
Fast and accurate diagnosis: enables immediate assessment of stroke damage by quantifying the volume of ischemia and determining the type
Streamlined clinical decision making - through automation and standardization of diagnostics

Reviews from doctors

“Each diagnosis used to take from 10 minutes to an hour. Thanks to the AI, the examination takes 2 minutes, which reduces lots of stress on the radiologist."

Anton Polyakov, Head of the department, doctor of the highest category of the Regional Clinical Hospital No. 2, Khabarovsk