Manténgase sano!
- Posted February 27, 2025
AI Can Guide Surgery For Childhood Epilepsy
Artificial intelligence (AI) might help treat childhood epilepsy by detecting brain abnormalities that are causing kids’ seizures, a new study suggests.
The AI tool, called MELD Graph, found 64% of brain lesions linked to epilepsy that human radiologists had previously missed, researchers report in JAMA Neurology.
Surgery to remove these lesions can be an effective way to stop epileptic seizures, but they are difficult to see with the human eye, researchers said.
“Many of the children I see have experienced years of seizures and investigations before we find a lesion,” researcher Helen Cross, director of the University College London Great Ormand Street Institute of Child Health in the U.K., said in a news release.
About 1 in 5 people with epilepsy have seizures caused by lesions in the brain, which are called focal cortical dysplasias (FCDs), researchers said in background notes.
To develop the new AI, researchers pooled MRI data from nearly 1,200 people treated at 23 epilepsy centers around the world. The initiative was called the Multicenter Epilepsy Lesion Detection Project, or MELD.
“The epilepsy community is searching for ways to speed up diagnosis and treatment,” Cross said. “Initiatives such as MELD have the potential to rapidly identify abnormalities that can be removed and potentially cure the epilepsy.”
Researchers trained the AI using the half of the MRI scans, teaching it to detect subtle lesions that might otherwise be overlooked.
The team then tested the AI’s effectiveness on the other half of the scans.
In the independent testing, the AI accurately detected 82% of known lesions that had already been successfully removed in surgeries that rendered patients seizure-free, results show.
The AI also revealed another 64% of lesions that radiologists previously missed when reviewing MRI scans, researchers said.
“MELD Graph identified a subtle lesion missed by many radiologists in a 12-year-old boy who had daily seizures and had tried nine anti-seizure medications with no improvement to his condition,” researcher Dr. Luca de Palma, a neurologist with Bambino Gesu Children’s Hospital in Rome, said in a news release.
Overall, the AI accurately identified about 70% of the brain lesions in the MRIs, researchers said. There also was a fourfold reduction in false positives -- brain areas erroneously identified as seizure-causing lesions.
“This tool could identify patients with surgically operable epilepsy and help with surgical planning -- reducing risks, saving money, improving outcomes,” de Palma added.
The tool is not yet available to doctors, but the research team has released it as open-source software. They currently are running workshops to train doctors and researchers in how to use it.
More information
The Epilepsy Foundation has more on epilepsy caused by abnormal brain structures.
SOURCES: King’s College London, news release, Feb. 24, 2025; JAMA Neurology, Feb. 24, 2025
