Abnormalities causing epileptic seizures can be detected by AI

Over a thousand patient MRI scans from twenty-two international epilepsy centres have been utilised by the Multicentre Epilepsy Lesion Detection project (MELD) to develop the algorithm. The algorithm provides reports of where abnormalities are in cases of drug-resistant focal cortical dysplasia (FCD), a leading cause of epilepsy.
FCDs, often causing drug-resistant epilepsy, are areas of the brain that have developed abnormally. It is normally treated with surgery; however, FCDs can look normal, so recognising the lesions from an MRI scan is a constant challenge for clinicians.
Dr Sophie Adler, the co-senior author, stated: “We hope that this technology will help to identify epilepsy-causing abnormalities that are currently being missed. Ultimately it could enable more people with epilepsy to have potentially curative brain surgery.”
The team quantified cortical characteristics from the MRI scans to develop the algorithm, such as how folded or thick the brain surface was, using approximately three hundred thousand areas across the brain.
Researchers then trained the algorithm on samples labelled by radiologists, depending on their features and patterns, as either having FCD or being a healthy brain.
The findings found that from a total of five hundred and thirty-eight participants, the algorithm could detect the FCD in sixty-seven per cent of cases.
Previously, radiologists had been incapable of finding the abnormality, as 178 participants had been considered MRI negative. Still, the MELD algorithm identified the FCD in sixty-three per cent of these cases. This is especially vital; if doctors can find the irregularity in the brain scan, then surgery to remove it can offer a cure.
In the UK over half a million people are affected by epilepsy, and although drug treatments are available for most people living with epilepsy, 20-30 per cent do not respond to medications.