1. Multicenter Validation of a Deep Learning Detection Algorithm for Focal Cortical Dysplasia
- Author
-
Hyo Min Lee, Fernando Cendes, Ravnoor S. Gill, Mira Semmelroch, R. Edward Hogan, Benoit Caldairou, Neda Bernasconi, Andrea Bernasconi, Maxime Guye, Andreas Schulze-Bonhage, Graeme D. Jackson, Matteo Lenge, Vanessa Cristina Mendes Coelho, Francesco Deleo, Kyoo Ho Cho, Renzo Guerrini, Dewi V. Schrader, Carmen Barba, Ludovico D'Incerti, Fabrice Bartolomei, Seok-Jun Hong, Horst Urbach, Neuroimaging of Epilepsy Laboratory, Montreal Neurological Institute, McGill University, Montreal, Pediatric Neurology Unit and Laboratories Children's Hospital A. Meyer-University of Florence, Epileptology and Experimental Neurophysiology Unit, Fondazione IRCCS Istituto Neurologico 'Carlo Besta', Fondazione IRCCS Istituto Neurologico 'Carlo Besta', Department of Neurology University of Campinas, The Florey Institute of Neuroscience and Mental Health, University of Melbourne, Department of Pediatrics British Columbia Children's Hospital, Vancouver, Institut de Neurosciences des Systèmes (INS), Institut National de la Santé et de la Recherche Médicale (INSERM)-Aix Marseille Université (AMU), Service de neurophysiologie clinique [Hôpital de la Timone - APHM], Hôpital de la Timone [CHU - APHM] (TIMONE), Centre de résonance magnétique biologique et médicale (CRMBM), Assistance Publique - Hôpitaux de Marseille (APHM)-Aix Marseille Université (AMU)-Centre National de la Recherche Scientifique (CNRS), Centre d'Exploration Métabolique par Résonance Magnétique [Hôpital de la Timone - AP-HM] (CEMEREM), Assistance Publique - Hôpitaux de Marseille (APHM)-Aix Marseille Université (AMU)-Centre National de la Recherche Scientifique (CNRS)-Assistance Publique - Hôpitaux de Marseille (APHM)-Aix Marseille Université (AMU)-Centre National de la Recherche Scientifique (CNRS)- Hôpital de la Timone [CHU - APHM] (TIMONE), Freiburg Epilepsy Center Universitätsklinikum Freiburg, Freiburg Epilpesy Center Universitätsklinikum Freiburg, Yonsei University College of Medicine [Seoul, Republic of Korea], Department of Neurology Washington University School of Medicine, St. Louis, Università degli Studi di Firenze = University of Florence (UniFI), Aix Marseille Université (AMU)-Institut National de la Santé et de la Recherche Médicale (INSERM), Aix Marseille Université (AMU)-Assistance Publique - Hôpitaux de Marseille (APHM)-Centre National de la Recherche Scientifique (CNRS), Centre d'Exploration Métabolique par Résonance Magnétique [Hôpital de la Timone - APHM] (CEMEREM), Hôpital de la Timone [CHU - APHM] (TIMONE)-Centre de résonance magnétique biologique et médicale (CRMBM), and Aix Marseille Université (AMU)-Assistance Publique - Hôpitaux de Marseille (APHM)-Centre National de la Recherche Scientifique (CNRS)-Aix Marseille Université (AMU)-Assistance Publique - Hôpitaux de Marseille (APHM)-Centre National de la Recherche Scientifique (CNRS)
- Subjects
Adult ,Male ,Adolescent ,[SDV.IB.IMA]Life Sciences [q-bio]/Bioengineering/Imaging ,Neuroimaging ,Temporal lobe ,03 medical and health sciences ,Epilepsy ,Young Adult ,0302 clinical medicine ,Text mining ,Deep Learning ,Imaging, Three-Dimensional ,Image Interpretation, Computer-Assisted ,False positive paradox ,Medicine ,Humans ,Generalizability theory ,Child ,030304 developmental biology ,0303 health sciences ,business.industry ,Deep learning ,Cortical dysplasia ,Middle Aged ,medicine.disease ,Magnetic Resonance Imaging ,Malformations of Cortical Development ,Child, Preschool ,Cohort ,Female ,Neurology (clinical) ,Artificial intelligence ,business ,Algorithm ,030217 neurology & neurosurgery ,Research Article - Abstract
Background and ObjectiveTo test the hypothesis that a multicenter-validated computer deep learning algorithm detects MRI-negative focal cortical dysplasia (FCD).MethodsWe used clinically acquired 3-dimensional (3D) T1-weighted and 3D fluid-attenuated inversion recovery MRI of 148 patients (median age 23 years [range 2–55 years]; 47% female) with histologically verified FCD at 9 centers to train a deep convolutional neural network (CNN) classifier. Images were initially deemed MRI-negative in 51% of patients, in whom intracranial EEG determined the focus. For risk stratification, the CNN incorporated bayesian uncertainty estimation as a measure of confidence. To evaluate performance, detection maps were compared to expert FCD manual labels. Sensitivity was tested in an independent cohort of 23 cases with FCD (13 ± 10 years). Applying the algorithm to 42 healthy controls and 89 controls with temporal lobe epilepsy disease tested specificity.ResultsOverall sensitivity was 93% (137 of 148 FCD detected) using a leave-one-site-out cross-validation, with an average of 6 false positives per patient. Sensitivity in MRI-negative FCD was 85%. In 73% of patients, the FCD was among the clusters with the highest confidence; in half, it ranked the highest. Sensitivity in the independent cohort was 83% (19 of 23; average of 5 false positives per patient). Specificity was 89% in healthy and disease controls.DiscussionThis first multicenter-validated deep learning detection algorithm yields the highest sensitivity to date in MRI-negative FCD. By pairing predictions with risk stratification, this classifier may assist clinicians in adjusting hypotheses relative to other tests, increasing diagnostic confidence. Moreover, generalizability across age and MRI hardware makes this approach ideal for presurgical evaluation of MRI-negative epilepsy.Classification of EvidenceThis study provides Class III evidence that deep learning on multimodal MRI accurately identifies FCD in patients with epilepsy initially diagnosed as MRI negative.
- Published
- 2021