×
AI software Meld Graph outperforms radiologists in early epilepsy detection
Written by
Published on
Join our daily newsletter for breaking news, product launches and deals, research breakdowns, and other industry-leading AI coverage
Join Now

The increasing sophistication of medical imaging analysis has opened new frontiers in neurological diagnosis, particularly in the field of epilepsy treatment. A groundbreaking AI tool called Meld Graph, developed by scientists at King’s College London and University College London, aims to detect subtle brain abnormalities that human radiologists might miss.

The breakthrough technology: Meld Graph was specifically designed to identify focal cortical dysplasia (FCD), a type of brain malformation that occurs when neurons develop incorrectly and can cause drug-resistant epilepsy.

  • The AI system analyzes MRI scans to detect subtle abnormalities that might be overlooked during conventional radiological review
  • Testing involved MRI data from 1,185 patients, including 703 with FCD, across 23 global epilepsy centers
  • The system successfully detected 64% of brain malformations in the study

Clinical impact and potential: This innovation could significantly improve treatment options for the approximately 126,000 UK patients whose epilepsy is caused by brain lesions that may be surgically treatable.

  • Early detection of FCDs could help patients avoid years of unsuccessful drug treatments
  • Surgical intervention, when appropriate, could potentially cure certain types of epilepsy
  • The tool could reduce the need for multiple costly tests and procedures

Real-world applications: Medical institutions are already beginning to explore the practical implementation of Meld Graph in clinical settings.

  • A 12-year-old patient who had tried nine different anti-seizure medications without success had their condition correctly identified by Meld Graph after radiologists missed the subtle lesion
  • Great Ormond Street Hospital is among the institutions participating in the implementation
  • The software has been released as open source, allowing medical professionals worldwide to access and utilize the technology

Global adoption and accessibility: The development team is actively working to expand the tool’s reach and usability across different healthcare systems.

  • Doctors from various countries including the UK, Chile, India, and France are already implementing the tools
  • The research team is conducting workshops to train clinicians on proper usage
  • The system is designed to complement rather than replace radiologists’ expertise

Expert perspectives: Leading medical professionals have expressed optimism about Meld Graph’s potential to transform epilepsy diagnosis and treatment.

  • Dr. Konrad Wagstyl of King’s College London emphasizes the tool’s ability to support overwhelmed radiologists and improve NHS efficiency
  • Professor Helen Cross of Great Ormond Street Hospital highlights the potential to accelerate diagnosis and treatment for pediatric patients
  • Dr. Luca Palma notes the tool’s value in surgical planning and risk reduction

Future implications: While Meld Graph represents a significant advance in epilepsy diagnosis, its success could pave the way for similar AI applications across other neurological conditions, potentially transforming how we approach brain imaging analysis and surgical planning in multiple medical contexts.

How AI could help detect epilepsy before radiologists

Recent News

Microsoft reiterates plans to make $80B AI investment, albeit with potential adjustments

Microsoft adjusts $80 billion AI investment strategy by canceling third-party data center leases while expanding internal capacity.

AI musicians take center stage at Mip London Korea Showcase

Korean studios unveil AI music artists and virtual production methods to combat rising drama production costs that have tripled since 2022.

Protecting human intellectual property in the AI era

AI's disruption of creative industries drives urgent calls for revenue-sharing models to protect human artists and content makers.