Key points from article :
A new artificial intelligence (AI) tool, MELD Graph, has been developed by researchers at King’s College London and University College London to detect brain abnormalities responsible for epilepsy. The tool has been shown to identify two-thirds of the lesions that doctors typically miss on MRI scans, offering hope for more effective diagnosis and treatment.
Focal cortical dysplasia—tiny brain abnormalities often undetectable by the human eye—can cause uncontrolled seizures in around 30,000 people in the UK. These seizures, which may involve jerking movements, stiffness, or loss of awareness, can significantly impact daily life. Surgery to remove the affected brain tissue can be an effective treatment, but delays in identifying lesions can postpone intervention.
Published in JAMA Neurology, the study involved MRI scans from 1,185 patients across 23 hospitals. MELD Graph processed the images faster and in greater detail than doctors, identifying previously undetected abnormalities. While the AI system still missed one-third of cases, its ability to spot lesions could lead to faster diagnoses and fewer unnecessary medical tests. In one case, the tool identified a lesion in a 12-year-old boy who had suffered from daily seizures despite trying nine medications.
Experts in childhood epilepsy, including Prof. Helen Cross from Great Ormond Street Hospital, see the tool’s potential to transform treatment by quickly pinpointing abnormalities that could be surgically removed. However, further studies are required before it can be officially approved for clinical use. While the tool represents a major step forward, epilepsy charities caution that it does not address broader issues such as the shortage of specialist epilepsy nurses in the UK.
For now, MELD Graph has been made available as open-source software, allowing hospitals worldwide to use it for clinical research while further trials continue. If approved for wider use, it could be life-changing for many people with epilepsy.
