AI Tool Detects 64% Of Missed Brain Abnormalities In Children With Epilepsy

Scientists have developed an AI tool called MELD Graph to detect brain abnormalities linked to epilepsy missed by human radiologists in 64% of cases. Published in JAMA Neurology by researchers from King’s College London and UCL, the study highlights its ability to improve detection of focal cortical dysplasia (FCDs), a leading cause of epilepsy.

By analyzing MRI data from over 1,000 participants across 23 global centers, MELD Graph aims to accelerate diagnosis, reduce NHS costs by up to £55,000 per patient, and enhance treatment outcomes for children with epilepsy. The tool is available as open-source software, facilitating international collaboration.

AI Tool Detects Subtle Brain Abnormalities in Children With Epilepsy

Scientists have developed an AI-powered tool called MELD Graph that detects 64% of brain abnormalities linked to epilepsy that human radiologists often miss. This breakthrough could significantly improve care for children with epilepsy, a condition affecting 1 in 100 people in the UK and millions worldwide. The study, published in JAMA Neurology by researchers from King’s College London and University College London (UCL), highlights how MELD Graph enhances the detection of focal cortical dysplasia (FCDs), a leading cause of epilepsy.

The tool was trained on MRI data from 1185 participants, including children, across 23 epilepsy centers globally. FCDs are challenging to detect due to their subtle nature, and up to half of these lesions are missed by radiologists. Delays in diagnosis can lead to prolonged seizures, disrupted lives, and increased healthcare costs. MELD Graph addresses this challenge by identifying abnormalities that might otherwise go unnoticed.

Dr. Konrad Wagstyl from King’s College London emphasized the tool’s potential to streamline workflows for radiologists, reduce unnecessary tests, and lower NHS costs by up to £55,000 per patient. Dr. Luca Palma from Bambino Gesù Children’s Hospital shared a case where MELD Graph detected a subtle lesion in a 12-year-old boy with daily seizures, highlighting the tool’s life-changing impact.

While not yet clinically available, MELD Graph is open-source and being trained worldwide, showcasing its potential for global adoption. The research underscores the importance of international collaboration, as highlighted by Dr. Sophie Adler from UCL, who noted that such advancements require collective efforts to achieve the goal of “no missed epilepsy lesions worldwide.”

This innovation marks a significant step forward in improving outcomes for children with epilepsy, offering hope for faster and more accurate diagnoses.