Meeting Banner
Abstract #0587

Automated lesion prediction and characterisation of focal cortical abnormalities: a MELD study

Konrad Wagstyl1, Mathilde Ripart2, MELD Consortium3, Hannah Spitzer4, Torsten Baldeweg2, and Sophie Adler2
1Wellcome Centre for Human Neuroimaging, UCL, London, United Kingdom, 2Great Ormond Street Institute for Child Health, UCL, London, United Kingdom, 3UCL, London, United Kingdom, 4Institute of Computational Biology, Helmholtz Zentrum M√ľnchen, Munich, Germany

Synopsis

Identification of subtle epilepsy-causing focal cortical dysplasias (FCD) on MRI remains an outstanding challenge during presurgical assessment of patients. The Multi-centre Epilepsy Lesion Detection (MELD) project created an MRI lesion detection algorithm using a large, heterogenous cohort. The algorithm had a sensitivity of 67%, performing well on unseen test sites. Analysis of MRI lesions revealed distinct subgroups, with differing histopathological subtypes, imaging features and detection rates. Individual patient reports highlight a predicted lesion’s location, imaging features and their relative saliency to the classifier. This tool has the potential to improve presurgical lesion identification on MRI from patients with epileptogenic FCD.

This abstract and the presentation materials are available to members only; a login is required.

Join Here