Researchers claim they have developed an artificial intelligence (AI) algorithm that can detect and identify different types of brain injuries.

The research team from Cambridge University and Imperial College London, clinically validated and tested the AI on large sets of CT scans and found it was effective in identifying, segmenting, quantifying and differentiating various types of brain lesions.

The findings, published in the journal 'The Lancet Digital Health,' may be useful in large-scale research trials to create more tailored therapies for head injuries and may be useful in certain clinical contexts, such as those where radiological experience is at a premium, with further clarification.

"CT is an extremely valuable diagnostic device, but it is seldom used quantitatively," said co-senior study author David Menon, professor at the University of Cambridge in the UK.

"Many of the rich information available in a CT scan is often overlooked and as researchers, we know that the type, magnitude, and position of a brain lesion is important for patient outcomes," added Menon.

The researchers wanted to design and create a method that could recognize and measure the various types of brain lesions automatically, so that we could use it in research and explore its potential use in a hospital environment.

The team developed a machine-learning method based on a neural artificial network. They trained the device on more than 600 different CT scans, showing various sizes and forms of brain lesions.

Additionally, the researchers said it could potentially be used in emergency rooms, helping patients get home sooner. Only between 10 and 15 per cent of all patients who have a head injury have a lesion that can be identified on a CT scan.