Same same but different: A Web-based deep learning application revealed classifying features for the histopathologic distinction of cortical malformations.
Algorithms
Area Under Curve
Cerebral Cortex
/ pathology
Deep Learning
Diagnosis, Computer-Assisted
Epilepsy
/ diagnosis
Humans
Internet
Malformations of Cortical Development, Group I
/ diagnosis
Neural Networks, Computer
Neurons
/ pathology
Neuropathology
Proof of Concept Study
ROC Curve
Reproducibility of Results
Tuberous Sclerosis
/ diagnosis
brain
convolutional neural network
cortical malformations
deep learning
digital pathology
epilepsy
Journal
Epilepsia
ISSN: 1528-1167
Titre abrégé: Epilepsia
Pays: United States
ID NLM: 2983306R
Informations de publication
Date de publication:
03 2020
03 2020
Historique:
received:
12
11
2019
revised:
23
01
2020
accepted:
23
01
2020
pubmed:
23
2
2020
medline:
21
10
2020
entrez:
22
2
2020
Statut:
ppublish
Résumé
The microscopic review of hematoxylin-eosin-stained images of focal cortical dysplasia type IIb and cortical tuber of tuberous sclerosis complex remains challenging. Both entities are distinct subtypes of human malformations of cortical development that share histopathological features consisting of neuronal dyslamination with dysmorphic neurons and balloon cells. We trained a convolutional neural network (CNN) to classify both entities and visualize the results. Additionally, we propose a new Web-based deep learning application as proof of concept of how deep learning could enter the pathologic routine. A digital processing pipeline was developed for a series of 56 cases of focal cortical dysplasia type IIb and cortical tuber of tuberous sclerosis complex to obtain 4000 regions of interest and 200 000 subsamples with different zoom and rotation angles to train a neural network. Guided gradient-weighted class activation maps (Guided Grad-CAMs) were generated to visualize morphological features used by the CNN to distinguish both entities. Our best-performing network achieved 91% accuracy and 0.88 area under the receiver operating characteristic curve at the tile level for an unseen test set. Novel histopathologic patterns were found through the visualized Guided Grad-CAMs. These patterns were assembled into a classification score to augment decision-making in routine histopathology workup. This score was successfully validated by 11 expert neuropathologists and 12 nonexperts, boosting nonexperts to expert level performance. Our newly developed Web application combines the visualization of whole slide images with the possibility of deep learning-aided classification between focal cortical dysplasia IIb and tuberous sclerosis complex. This approach will help to introduce deep learning applications and visualization for the histopathologic diagnosis of rare and difficult-to-classify brain lesions.
Types de publication
Journal Article
Langues
eng
Sous-ensembles de citation
IM
Pagination
421-432Informations de copyright
© 2020 The Authors. Epilepsia published by Wiley Periodicals, Inc. on behalf of International League Against Epilepsy.
Références
Chang HY, Jung CK, Woo JI, et al. Artificial intelligence in pathology. J Pathol Transl Med. 2019;53:1-12.
Ching T, Himmelstein DS, Beaulieu-Jones BK, et al. Opportunities and obstacles for deep learning in biology and medicine. J R Soc Interface. 2018;15:20170387.
Simonyan K, Zisserman A. Very deep convolutional networks for large-scale image recognition. arXiv e-prints. 2014. Available at: https://ui.adsabs.harvard.edu/abs/2014arXiv1409.1556S. Accessed February 13, 2020.
Janowczyk A, Madabhushi A. Deep learning for digital pathology image analysis: a comprehensive tutorial with selected use cases. J Pathol Inform. 2016;7:29.
Coudray N, Ocampo PS, Sakellaropoulos T, et al. Classification and mutation prediction from non-small cell lung cancer histopathology images using deep learning. Nat Med. 2018;24:1559-67.
Ehteshami Bejnordi B, Veta M, Johannes van Diest P, et al. Diagnostic assessment of deep learning algorithms for detection of lymph node metastases in women with breast cancer. JAMA. 2017;318:2199-210.
Arvaniti E, Fricker KS, Moret M, et al. Automated Gleason grading of prostate cancer tissue microarrays via deep learning. Sci Rep. 2018;8:12054.
Yu Y, Wang J, Ng CW, et al. Deep learning enables automated scoring of liver fibrosis stages. Sci Rep. 2018;8:16016.
Nirschl JJ, Janowczyk A, Peyster EG, et al. A deep-learning classifier identifies patients with clinical heart failure using whole-slide images of H&E tissue. PLoS One. 2018;13:e0192726.
Tang Z, Chuang KV, DeCarli C, et al. Interpretable classification of Alzheimer's disease pathologies with a convolutional neural network pipeline. Nat Commun. 2019;10:2173.
LeCun Y, Boser BE, Denker JS, et al. Handwritten Digit Recognition With a Back-Propagation Network. San Francisco, CA: Morgan Kaufmann; 1990.
LeCun Y, Bengio Y, Hinton G. Deep learning. Nature. 2015;521:436-44.
Blumcke I, Aronica E, Urbach H, Alexopoulos A, Gonzalez-Martinez JA. A neuropathology-based approach to epilepsy surgery in brain tumors and proposal for a new terminology use for long-term epilepsy-associated brain tumors. Acta Neuropathol. 2014;128:39-54.
Harvey AS, Cross JH, Shinnar S, Mathern GW, ILAE Pediatric Epilepsy Surgery Survey Taskforce. Defining the spectrum of international practice in pediatric epilepsy surgery patients. Epilepsia. 2008;49:146-55.
Northrup H, Krueger DA, International Tuberous Sclerosis Complex Consensus Group. Tuberous sclerosis complex diagnostic criteria update: recommendations of the 2012 International Tuberous Sclerosis Complex Consensus Conference. Pediatr Neurol. 2012;2013(49):243-54.
Blumcke I, Thom M, Aronica E, et al. The clinicopathologic spectrum of focal cortical dysplasias: a consensus classification proposed by an ad hoc task force of the ILAE Diagnostic Methods Commission. Epilepsia. 2011;52:158-74.
Muhlebner A, Iyer AM, van Scheppingen J, et al. Specific pattern of maturation and differentiation in the formation of cortical tubers in tuberous sclerosis complex (TSC): evidence from layer-specific marker expression. J Neurodev Disord. 2016;8:9.
Thom M, Martinian L, Sen A, Cross JH, Harding BN, Sisodiya SM. Cortical neuronal densities and lamination in focal cortical dysplasia. Acta Neuropathol. 2005;110:383-92.
Samura K, Morioka T, Yoshida F, et al. Focal cortical dysplasia with calcification: a case report. Childs Nerv Syst. 2008;24:619-22.
Wong M, Crino PB. Tuberous sclerosis and epilepsy: role of astrocytes. Glia. 2012;60:1244-50.
Sosunov AA, Wu X, Weiner HL, et al. Tuberous sclerosis: a primary pathology of astrocytes? Epilepsia. 2008;49(Suppl 2):53-62.
Kulkarni PM, Robinson EJ, Sarin Pradhan J, et al. Deep learning based on standard H&E images of primary melanoma tumors identifies patients at risk for visceral recurrence and death. Clin Cancer Res. 2020. https://doi.org/10.1158/1078-0432.ccr-19-1495
Bankhead P, Loughrey MB, Fernández JA, et al. QuPath: open source software for digital pathology image analysis. Sci Rep. 2017;7:16878.
Chollet F, Rahman F, Lee T, de Marmiesse A, Zabluda O, Pumperla M, et al. Keras GitHub repository. Available at: https://github.com/fchollet/keras. Accessed February 13, 2020.
Zoph B, Vasudevan V, Shlens J, Le QV. Learning transferable architectures for scalable image recognition. arXiv e-prints. 2017. Available at: https://ui.adsabs.harvard.edu/abs/2017arXiv170707012Z. Accessed February 13, 2020.
Chollet F. Xception: deep learning with depthwise separable convolutions. arXiv e-prints. 2016. Available at: https://ui.adsabs.harvard.edu/abs/2016arXiv161002357C. Accessed February 13, 2020.
Huang G, Liu Z, van der Maaten L, Weinberger KQ. Densely connected convolutional networks. arXiv e-prints. 2016. Available at: https://ui.adsabs.harvard.edu/abs/2016arXiv160806993H. Accessed February 13, 2020.
He K, Zhang X, Ren S, Sun J. Deep residual learning for image recognition. arXiv e-prints. 2015. Available at: https://ui.adsabs.harvard.edu/abs/2015arXiv151203385H. Accessed February 13, 2020.
Ioffe S, SzegedyC. Batch normalization: accelerating deep network training by reducing internal covariate shift. arXiv e-prints. 2015. Available at: https://ui.adsabs.harvard.edu/abs/2015arXiv150203167I. Accessed February 13, 2020.
Wu R, Yan S, Shan Y, Dang Q, Sun G. Deep image: scaling up image recognition. arXiv e-prints. 2015. Available at: https://ui.adsabs.harvard.edu/abs/2015arXiv150102876W. Accessed February 13, 2020.
Wong SC, Gatt A, Stamatescu V, McDonnell MD. Understanding data augmentation for classification: when to warp? arXiv e-prints. 2016. Available at: https://ui.adsabs.harvard.edu/abs/2016arXiv160908764W. Accessed February 13, 2020.
Jung A. imgaug. Github. Available at: https://github.com/aleju/imgaug. Accessed February 2019.
Smith LN. Cyclical learning rates for training neural networks. arXiv e-prints. 2015. Available from at: https://ui.adsabs.harvard.edu/abs/2015arXiv150601186S. Accessed February 13, 2020.
Zhou B, Khosla A, Lapedriza A, Oliva A, Torralba A. Learning deep features for discriminative localization. arXiv e-prints. 2015. Available at: https://ui.adsabs.harvard.edu/abs/2015arXiv151204150Z. Accessed February 13, 2020.
Selvaraju RR, Das A, Vedantam R, Cogswell M, Parikh D, Batra D. Grad-CAM: why did you say that? arXiv e-prints. 2016. Available at: https://ui.adsabs.harvard.edu/abs/2016arXiv161107450S. Accessed February 13, 2020.
Gildenblat J. Implementation of Grad CAM with Keras. GitHub. Available at: https://github.com/jacobgil/keras-grad-cam. Accessed April 2019.
Yosinski J, Clune J, Bengio Y, Lipson H. How transferable are features in deep neural networks? arXiv e-prints. 2014. Available at: https://ui.adsabs.harvard.edu/abs/2014arXiv1411.1792Y. Accessed February 13, 2020.
Iizuka T, Fukasawa M, Kameyama M. Deep-learning-based imaging-classification identified cingulate island sign in dementia with Lewy bodies. Sci Rep. 2019;9:8944.
Komura D, Ishikawa S. Machine learning methods for histopathological image analysis. Comput Struct Biotechnol J. 2018;16:34-42.
Madabhushi A, Lee G. Image analysis and machine learning in digital pathology: challenges and opportunities. Med Image Anal. 2016;33:170-5.
Kothari S, Phan JH, Stokes TH, Osunkoya AO, Young AN, Wang MD. Removing batch effects from histopathological images for enhanced cancer diagnosis. IEEE J Biomed Health Inform. 2014;18:765-72.
Louis DN, Perry A, Reifenberger G, et al. The 2016 World Health Organization classification of tumors of the central nervous system: a summary. Acta Neuropathol. 2016;131(6):803-20.
Mühlebner A, van Scheppingen J, Hulshof HM, et al. Novel histopathological patterns in cortical tubers of epilepsy surgery patients with tuberous sclerosis complex. PLoS One. 2016;11:e0157396.
Li S, Yu S, Zhang C, et al. Increased expression of matrix metalloproteinase 9 in cortical lesions from patients with focal cortical dysplasia type IIb and tuberous sclerosis complex. Brain Res. 2012;1453:46-55.
Talos DM, Sun H, Kosaras B, et al. Altered inhibition in tuberous sclerosis and type IIb cortical dysplasia. Ann Neurol. 2012;71:539-51.
Adamson AS, Welch HG. Machine learning and the cancer-diagnosis problem-no gold standard. N Engl J Med. 2019;381:2285-7.
Deng J, Dong W, Socher R, Li L, Kai L, Li F-F. ImageNet: A Large-Scale Hierarchical Image Database. Miami, FL: 2009 IEEE Conference on Computer Vision and Pattern Recognition, 2009; pp. 248-255.
Blumcke I, Spreafico R, Haaker G, et al. Histopathological findings in brain tissue obtained during epilepsy surgery. N Engl J Med. 2017;377:1648-56.
Beghi E, Giussani G, Nichols E, et al. Global, regional, and national burden of epilepsy, 1990-2016: a systematic analysis for the Global Burden of Disease Study 2016. Lancet Neurol. 2019;18:357-75.
Newton CR, Garcia HH. Epilepsy in poor regions of the world. Lancet. 2012;380:1193-201.