Development and prospective clinical validation of a convolutional neural network for automated detection and segmentation of focal cortical dysplasias.

Journal: Epilepsy research
PMID:

Abstract

PURPOSE: Focal cortical dysplasias (FCDs) are a leading cause of drug-resistant epilepsy. Early detection and resection of FCDs have favorable prognostic implications for postoperative seizure freedom. Despite advancements in imaging methods, FCD detection remains challenging. House et al. (2021) introduced a convolutional neural network (CNN) for automated FCD detection and segmentation, achieving a sensitivity of 77.8%. However, its clinical applicability was limited due to a low specificity of 5.5%. The objective of this study was to improve the CNN's performance through data-driven training and algorithm optimization, followed by a prospective validation on daily-routine MRIs.

Authors

  • Vicky Chanra
    Hamburg Epilepsy Center, Protestant Hospital Alsterdorf, Department of Neurology and Epileptology, Hamburg, Germany.
  • Agata Chudzinska
    theBlue.ai GmbH, Hamburg, Germany.
  • Natalia Braniewska
    theBlue.ai GmbH, Hamburg, Germany.
  • Bartosz Silski
    theBlue.ai GmbH, Hamburg, Germany.
  • Brigitte Holst
    University Hospital Hamburg-Eppendorf, Department of Neuroradiology, Hamburg, Germany.
  • Thomas Sauvigny
    University Hospital Hamburg-Eppendorf, Department of Neurosurgery, Hamburg, Germany.
  • Stefan Stodieck
    Hamburg Epilepsy Center, Protestant Hospital Alsterdorf, Department of Neurology and Epileptology, Hamburg, Germany.
  • Sirko Pelzl
    theBlue.ai GmbH, Hamburg, Germany.
  • Patrick M House
    Hamburg Epilepsy Center, Protestant Hospital Alsterdorf, Department of Neurology and Epileptology, Hamburg, Germany; theBlue.ai GmbH, Hamburg, Germany; Epileptologicum Hamburg, Specialist's Practice for Epileptology, Hamburg, Germany. Electronic address: house@epileptologicum.de.