Use of Convolutional Neural Networks to Identify Focal Cortical Dysplasia in Patients with Refractory Epilepsy

• Samuel Silva University of São Paulo
• Fabricio Simozo University of São Paulo
• Luiz Murta Junior FFCLRP-USP
• Renato Tinós Universidade de São Paulo

DOI: https://doi.org/10.5753/eniac.2020.12130

Abstract

Focal Cortical Dysplasia (FCD) is a type of brain injury that is the main cause of Refractory Epilepsy in children. Surgery is an alternative for the treatment of patients with FCD. However, the correct identification of the regions with FCD in the brain is necessary. The identification of FCD in Magnetic Resonance Images by using Convolutional Neural Networks (CNNs) is investigated. The CNN classifies small regions of the images (windows), that can allow the identification of the location of the FCD. The method based on the CNN is compared with traditional approach, in which a Multilayer Perceptron (MLP) is employed using predefined filters for the extraction of attributes of the image. The CNN presents better results than the MLP. The CNN presents high sensitivity. However, the specificity is low, indicating that a larger dataset is necessary for training the CNN.

Keywords: Convolutional Neural Networks, Epilepsy, Magnetic Resonance Imaging

References

ABADI, M. et al. (2016). “TensorFlow: large-scale machine learning on heterogeneous systems”, arXiv preprint:1603.04467.

AZEVEDO-MARQUES, P. M. (2001). “Diagnóstico auxiliado por computador na radiologia”, Radiologia Brasileira, 34(5): 285–293.

CAMARGO, D. P. C. et al. (2010). “Displasia cortical focal, aspectos neurofisiológicos, imaginológicos e histológicos”, Journal of Epilepsy and Clinical Neurophysiology, 16(3): 106–111.

FRENCH, J. A. (2007). “Refractory epilepsy: clinical overview”, Epilepsia, 48(SUPPL.1): 3–7.

GARZON, E. (2002). “Epilepsia refratária: conceito e contribuição das novas drogas antiepilépticas e de outras modalidades terapêuticas”, Revista Neurociências, 10(2): 66–82.

GREENSPAN, H.; VAN GINNEKEN, B. & SUMMERS, R. M. (2016). “Guest editorial deep learning in medical imaging: overview and future promise of an exciting new technique”, IEEE Transactions on Medical Imaging, 35(5):1153–1159;

HARALICK, R. M. (1979). “Statistical and structural approaches to texture”, Proceedings of the IEEE, 67(5): 786-804.

KRSEK, P. et al. (2008). “Different features of histopathological subtypes of pediatric focal cortical dysplasia”, Annals of Neurology, 63(6): 758–769.

KWAN, P. & BRODIE, M. J. (2000). “Early identification of refractory epilepsy”, New England Journal of Medicine, 342(5): 314-319.

LECUN, Y.; BENGIO, Y. & HINTON, G. (2015). “Deep learning”, Nature, 521(7553): 436-444.

MAZZOLA, A. A. (2009). “Ressonância magnética: princípios de formação da imagem e aplicações em imagem funcional”. Revista Brasileira de Física Médica, 3(1), 117129.

PALMINI, A. et al. (2004). “Terminology and classification of the cortical dysplasias”, Neurology, 62(6 suppl. 3): S2-S8.

PIEPER, S.; HALLE, M. & KIKINIS, R. (2004). “3D Slicer”, In the 2nd IEEE International Symposium on Biomedical Imaging, 632-635.

SIMOZO, F. H. (2018). “Caracterização e identificação de displasias corticais focais em pacientes com epilepsia refratária através de análise de imagens estruturais de ressonância magnética nuclear”, Tese de Doutorado em Física Aplicada à Medicina e Biologia, FFCLRP, USP.

ABADI, M. et al. (2016). “TensorFlow: large-scale machine learning on heterogeneous systems”, arXiv preprint:1603.04467.

AZEVEDO-MARQUES, P. M. (2001). “Diagnóstico auxiliado por computador na radiologia”, Radiologia Brasileira, 34(5): 285–293.

CAMARGO, D. P. C. et al. (2010). “Displasia cortical focal, aspectos neurofisiológicos, imaginológicos e histológicos”, Journal of Epilepsy and Clinical Neurophysiology, 16(3): 106–111.

FRENCH, J. A. (2007). “Refractory epilepsy: clinical overview”, Epilepsia, 48(SUPPL.1): 3–7.

GARZON, E. (2002). “Epilepsia refratária: conceito e contribuição das novas drogas antiepilépticas e de outras modalidades terapêuticas”, Revista Neurociências, 10(2): 66–82.

GREENSPAN, H.; VAN GINNEKEN, B. & SUMMERS, R. M. (2016). “Guest editorial deep learning in medical imaging: overview and future promise of an exciting new technique”, IEEE Transactions on Medical Imaging, 35(5):1153–1159;

HARALICK, R. M. (1979). “Statistical and structural approaches to texture”, Proceedings of the IEEE, 67(5): 786-804.

KRSEK, P. et al. (2008). “Different features of histopathological subtypes of pediatric focal cortical dysplasia”, Annals of Neurology, 63(6): 758–769.

KWAN, P. & BRODIE, M. J. (2000). “Early identification of refractory epilepsy”, New England Journal of Medicine, 342(5): 314-319.

LECUN, Y.; BENGIO, Y. & HINTON, G. (2015). “Deep learning”, Nature, 521(7553): 436-444.

MAZZOLA, A. A. (2009). “Ressonância magnética: princípios de formação da imagem e aplicações em imagem funcional”. Revista Brasileira de Física Médica, 3(1), 117129.

PALMINI, A. et al. (2004). “Terminology and classification of the cortical dysplasias”, Neurology, 62(6 suppl. 3): S2-S8.

PIEPER, S.; HALLE, M. & KIKINIS, R. (2004). “3D Slicer”, In the 2nd IEEE International Symposium on Biomedical Imaging, 632-635.

SIMOZO, F. H. (2018). “Caracterização e identificação de displasias corticais focais em pacientes com epilepsia refratária através de análise de imagens estruturais de ressonância magnética nuclear”, Tese de Doutorado em Física Aplicada à Medicina e Biologia, FFCLRP, USP.