Could AI help predict epilepsy in children, leading to earlier treatment?
Epilepsy is thought to affect about 250,000 Australians.
Machine learning has been used to create an algorithm that can diagnose epilepsy more consistently.
The algorithm automatically learns to detect lesions from thousands of MRI scans of patients and has been shown to reliably detect lesions, including many missed by radiologists.
Focal cortical dysplasia
(FCD) are a leading cause of drug-resistant epilepsy. FCDS are typically treated with surgery, but identifying the abnormalities is a challenge. The Multi-centre Epilepsy Lesion Detection (MELD) project used more than 1,000 patient MRI scans from 22 global epilepsy centres to develop an algorithm to find FCDS. Researchers defined cortical features (details in the outer layer of the brain’s cerebrum) from the MRI scans of 300,000 locations across the brain. They trained the AI to look at features like how thick or folded the cortex/brain surface was.
The team then exposed the algorithm to examples labelled by expert radiologists as being either healthy or having FCD. The algorithm detected FCDS in 67% of cases. Out of the cohort, 178 participants were considered by radiologists to not have FCD. Yet the MELD algorithm found the FCD in 63% of these cases.