AI automatically determines the types of brain injuries. Researchers note that the tool may open up new forms of therapy.
The new AI tool automatically detects various types of head injuries by analyzing the results of computed tomography. The system uses a machine-learning algorithm to identify types of brain damage and tissue areas damaged by trauma or illness.
The differences she finds can help develop personalized treatments for brain injuries that affect up to 60 million people each year. This will save time: usually patients with traumatic brain injuries are sent for computed tomography. Radiologists then scan the images to decide which drugs or operations are best used for treatment.
The system is based on a neural network developed by researchers at Imperial College London. Scientists have trained the tool in more than 600 computer images with various types of brain lesions. They found that AI successfully classified the individual parts of each image.
Researchers note that this tool can show how head injuries develop, revealing invisible changes that even doctors can skip. The developers believe that AI will even allow a new interpretation of brain injuries and will help in the development of forms of their treatment.
“So far we use this AI only for research, but with further development it can be implemented in medical procedures. You can start from countries with limited resources, where there are few radiologists, ”said co-author of the study Miguel Monteiro.