Bastian David1, Judith Kröll-Seger2, Fabiane Schuch1,3, Jan Wagner4, Jörg Wellmer5, Friedrich Woermann6, Bernhard Oehl7, Wim Van Paesschen8, Tobias Breyer9, Albert Becker10, Hartmut Vatter11, Elke Hattingen12, Horst Urbach13, Bernd Weber14, Rainer Surges1, Christian Erich Elger1, Hans-Jürgen Huppertz2, Theodor Rüber1,15,16. 1. Department of Epileptology, University Hospital Bonn, Bonn, Germany. 2. Swiss Epilepsy Clinic, Klinik Lengg AG, Zurich, Switzerland. 3. Department of Neurology, St. Johannes Hospital Troisdorf, Germany. 4. Department of Neurology, University Clinic Ulm, Ulm, Germany. 5. Department of Neurology, Ruhr-Epileptology, University Hospital Knappschaftskrankenhaus, Ruhr-University, Bochum, Germany. 6. Epilepsy Center Bethel, Mara Hospital & Society for Epilepsy Research, Bielefeld, Germany. 7. Ortenau Klinikum, Offenburg, Germany. 8. Laboratory for Epilepsy Research, Department of Neurology, University Hospitals and KU Leuven, Leuven, Belgium. 9. Department of Radiology and Neuroradiology, Klinikum Dortmund, Dortmund, Germany. 10. Department of Neuropathology, University Hospital Bonn, Bonn, Germany. 11. Department of Neurosurgery, University Hospital Bonn, Bonn, Germany. 12. Department of Neuroradiology, Goethe-University Frankfurt, Frankfurt am Main, Germany. 13. Department of Neuroradiology, University of Freiburg, Freiburg, Germany. 14. Institute of Experimental Epileptology and Cognition Research, University Hospital Bonn, Bonn, Germany. 15. Department of Neurology, Epilepsy Center Frankfurt Rhine-Main, Goethe-University Frankfurt, Frankfurt am Main, Germany. 16. Center for Personalized Translational Epilepsy Research (CePTER), Goethe-University Frankfurt, Frankfurt am Main, Germany.
Abstract
OBJECTIVE: Focal cortical dysplasias (FCDs) are a common cause of drug-resistant focal epilepsy but frequently remain undetected by conventional magnetic resonance imaging (MRI) assessment. The visual detection can be facilitated by morphometric analysis of T1-weighted images, for example, using the Morphometric Analysis Program (v2018; MAP18), which was introduced in 2005, independently validated for its clinical benefits, and successfully integrated in standard presurgical workflows of numerous epilepsy centers worldwide. Here we aimed to develop an artificial neural network (ANN) classifier for robust automated detection of FCDs based on these morphometric maps and probe its generalization performance in a large, independent data set. METHODS: In this retrospective study, we created a feed-forward ANN for FCD detection based on the morphometric output maps of MAP18. The ANN was trained and cross-validated on 113 patients (62 female, mean age ± SD =29.5 ± 13.6 years) with manually segmented FCDs and 362 healthy controls (161 female, mean age ± SD =30.2 ± 9.6 years) acquired on 13 different scanners. In addition, we validated the performance of the trained ANN on an independent, unseen data set of 60 FCD patients (28 female, mean age ± SD =30 ± 15.26 years) and 70 healthy controls (42 females, mean age ± SD = 40.0 ± 12.54 years). RESULTS: In the cross-validation, the ANN achieved a sensitivity of 87.4% at a specificity of 85.4% on the training data set. On the independent validation data set, our method still reached a sensitivity of 81.0% at a comparably high specificity of 84.3%. SIGNIFICANCE: Our method shows a robust automated detection of FCDs and performance generalizability, largely independent of scanning site or MR-sequence parameters. Taken together with the minimal input requirements of a standard T1 image, our approach constitutes a clinically viable and useful tool in the presurgical diagnostic routine for drug-resistant focal epilepsy.
OBJECTIVE: Focal cortical dysplasias (FCDs) are a common cause of drug-resistant focal epilepsy but frequently remain undetected by conventional magnetic resonance imaging (MRI) assessment. The visual detection can be facilitated by morphometric analysis of T1-weighted images, for example, using the Morphometric Analysis Program (v2018; MAP18), which was introduced in 2005, independently validated for its clinical benefits, and successfully integrated in standard presurgical workflows of numerous epilepsy centers worldwide. Here we aimed to develop an artificial neural network (ANN) classifier for robust automated detection of FCDs based on these morphometric maps and probe its generalization performance in a large, independent data set. METHODS: In this retrospective study, we created a feed-forward ANN for FCD detection based on the morphometric output maps of MAP18. The ANN was trained and cross-validated on 113 patients (62 female, mean age ± SD =29.5 ± 13.6 years) with manually segmented FCDs and 362 healthy controls (161 female, mean age ± SD =30.2 ± 9.6 years) acquired on 13 different scanners. In addition, we validated the performance of the trained ANN on an independent, unseen data set of 60 FCDpatients (28 female, mean age ± SD =30 ± 15.26 years) and 70 healthy controls (42 females, mean age ± SD = 40.0 ± 12.54 years). RESULTS: In the cross-validation, the ANN achieved a sensitivity of 87.4% at a specificity of 85.4% on the training data set. On the independent validation data set, our method still reached a sensitivity of 81.0% at a comparably high specificity of 84.3%. SIGNIFICANCE: Our method shows a robust automated detection of FCDs and performance generalizability, largely independent of scanning site or MR-sequence parameters. Taken together with the minimal input requirements of a standard T1 image, our approach constitutes a clinically viable and useful tool in the presurgical diagnostic routine for drug-resistant focal epilepsy.
Authors: Horst Urbach; Marcel Heers; Dirk-Matthias Altenmueller; Andreas Schulze-Bonhage; Anke Maren Staack; Thomas Bast; Marco Reisert; Ralf Schwarzwald; Christoph P Kaller; Hans-Juergen Huppertz; Theo Demerath Journal: Neuroradiology Date: 2021-10-09 Impact factor: 2.804