A Deep Learning Approach in Rebubbling After Descemet's Membrane Endothelial Keratoplasty.
Aged
Area Under Curve
Corneal Endothelial Cell Loss
/ prevention & control
Deep Learning
Descemet Membrane
/ surgery
Descemet Stripping Endothelial Keratoplasty
Female
Fuchs' Endothelial Dystrophy
/ surgery
Humans
Male
Models, Theoretical
ROC Curve
Reoperation
Retrospective Studies
Tomography, Optical Coherence
Visual Acuity
/ physiology
Journal
Eye & contact lens
ISSN: 1542-233X
Titre abrégé: Eye Contact Lens
Pays: United States
ID NLM: 101160941
Informations de publication
Date de publication:
Mar 2020
Mar 2020
Historique:
pubmed:
20
8
2019
medline:
2
1
2021
entrez:
20
8
2019
Statut:
ppublish
Résumé
To evaluate the efficacy of deep learning in judging the need for rebubbling after Descemet's endothelial membrane keratoplasty (DMEK). This retrospective study included eyes that underwent rebubbling after DMEK (rebubbling group: RB group) and the same number of eyes that did not require rebubbling (non-RB group), based on medical records. To classify the RB group, randomly selected images from anterior segment optical coherence tomography at postoperative day 5 were evaluated by corneal specialists. The criterion for rebubbling was the condition where graft detachment reached the central 4.0-mm pupil area. We trained nine types of deep neural network structures (VGG16, VGG19, ResNet50, InceptionV3, InceptionResNetV2, Xception, DenseNet121, DenseNet169, and DenseNet201) and built nine models. Using each model, we tested the validation data and evaluated the model. This study included 496 images (31 eyes from 24 patients) in the RB group and 496 images (31 eyes from 29 patients) in the non-RB group. Because 16 picture images were obtained from the same point of each eye, a total of 992 images were obtained. The VGG19 model was found to have the highest area under the receiver operating characteristic curve (AUC) of all models. The AUC, sensitivity, and specificity of the VGG19 model were 0.964, 0.967, and 0.915, respectively, whereas those of the best ensemble model were 0.956, 0.913, and 0.921, respectively. This automated system that enables the physician to be aware of the requirement of RB might be clinically useful.
Identifiants
pubmed: 31425350
doi: 10.1097/ICL.0000000000000634
pii: 00140068-202003000-00010
doi:
Types de publication
Journal Article
Langues
eng
Sous-ensembles de citation
IM
Pagination
121-126Références
Price MO, Price FW. Descemet's stripping endothelial keratoplasty. Curr Opin Ophthalmol 2007;18:290–294.
Melles GR, Ong TS, Ververs B, et al. Descemet membrane endothelial keratoplasty (DMEK). Cornea 2006;25:987–990.
Anshu A, Price MO, Price FW Jr. Risk of corneal transplant rejection significantly reduced with Descemet's membrane endothelial keratoplasty. Ophthalmology 2012;119:536–540.
Rudolph M, Laaser K, Bachmann BO, et al. Corneal higher-order aberrations after Descemet's membrane endothelial keratoplasty. Ophthalmology 2012;119:528–535.
Hayashi T, Yamaguchi T, Yuda K, et al. Topographic characteristics after Descemet's membrane endothelial keratoplasty and Descemet's stripping automated endothelial keratoplasty. PLoS One 2017;12:e0188832.
van Dijk K, Droutsas K, Hou J, et al. Optical quality of the cornea after Descemet membrane endothelial keratoplasty. Am J Ophthalmol 2014;158:71–79.e1.
Melles GRJ, Ong TS, Ververs B, et al. Preliminary clinical results of Descemet membrane endothelial keratoplasty. Am J Ophthalmol 2008;145:222–227.
Takahashi H, Tampo H, Arai Y, et al. Applying artificial intelligence to disease staging: Deep learning for improved staging of diabetic retinopathy. PLoS One 2017;12:e0179790.
Kaggle Inc. Diabetic Retinopathy Detection Competition [Kaggle Website]. 2015. Available at: https://www.kaggle.com/c/diabetic-retinopathy-detection. Accessed March 4, 2019.
Kovács I, Miháltz K, Kránitz K, et al. Accuracy of machine learning classifiers using bilateral data from a Scheimpflug camera for identifying eyes with preclinical signs of keratoconus. J Cataract Refract Surg 2016;42:275–283.
Souza MB, Medeiros FW, Souza DB, et al. Evaluation of machine learning classifiers in keratoconus detection from orbscan II examinations. Clinics (Sao Paulo). 2010;65:1223–1228.
Smadja D, Touboul D, Cohen A, et al. Detection of subclinical keratoconus using an automated decision tree classification. Am J Ophthalmol 2013;156:237–246.e1.
Lopes BT, Ramos IC, Salomão MQ, et al. Enhanced tomographic assessment to detect corneal ectasia based on artificial intelligence. Am J Ophthalmol 2018;195:223–232.
Hayashi T, Oyakawa I, Kato N. Techniques for learning Descemet membrane endothelial keratoplasty for eyes of Asian patients with shallow anterior chamber. Cornea 2017;36:390–393.
Dapena I, Moutsouris K, Droutsas K, et al. Standardized “no-touch” technique for descemet membrane endothelial keratoplasty. Arch Ophthalmol 2011;129:88–94.
Fukuda S, Kawana K, Yasuno Y, et al. Repeatability and reproducibility of anterior chamber volume measurements using 3-dimensional corneal and anterior segment optical coherence tomography. J Cataract Refract Surg 2011;37:461–468.
Dirisamer M, van Dijk K, Dapena I, et al. Prevention and management of graft detachment in descemet membrane endothelial keratoplasty. Arch Ophthalmol 2012;130:280–291.
Simonyan K, Zisserman A. Very deep convolutional networks for large-scale image recognition. 2014. Preprint at: https://arxiv.org/pdf/1409.1556.pdf.
He K, Zhang X, Ren S, et al. Deep residual learning for image recognition. Proc IEEE Comput Soc Conf Comput Vis Pattern Recognit 2016:770–778.
Huang G, Liu Z, Van Der Maaten L, et al. Densely connected convolutional networks. CVPR 2017;2.
Szegedy C, Vanhoucke V, Ioffe S, et al. Rethinking the inception architecture for computer vision. IEEE Comput Soc Conf Comput Vis Pattern Recognit 2016:2818–2826.
Szegedy C, Ioffe S, Vanhoucke V, et al. Inception-v4, inception-resnet and the impact of residual connections on learning. AAAI 2017;4:12.
Chollet F. Deep learning with depthwise separable convolutions. Preprint at: https://arxiv.org/pdf/1610.02357.pdf.
Deng J, Dong W, Socher R, et al. Imagenet: A large-scale hierarchical image database. CVPR 2009:248–255.
Russakovsky O, Deng J, Su H, et al. Imagenet large scale visual recognition challenge. Int J Comput Vis 2015;115:211–252.
Lee CY, Xie S, Gallagher P, et al. Deeply-supervised nets. InAISTATS 2015;2:5.
Scherer D, Müller A, Behnke S. Evaluation of Pooling Operations in Convolutional Architectures for Object Recognition. Berlin, Heidelberg, Springer, 2010, pp 92–101.
Krizhevsky A, Sutskever I, Hinton GE. Imagenet classification with deep convolutional neural networks. Adv Neural Inform Process Syst 2012;1:1097–1105.
Srivastava N, Hinton G, Krizhevsky A, et al. Dropout: A simple way to prevent neural networks from overfitting. J Machine Learn Res 2014;15:1929–1958.
Qian N. On the momentum term in gradient descent learning algorithms. Neural networks 1999;12:145–151.
Nesterov Y. A method for unconstrained convex minimization problem with the rate of convergence O (1/k^ 2). Doklady AN USSR 1983;269:543–547.
Agrawal P, Girshick R, Malik J. Analyzing the performance of multilayer neural networks for object recognition. Proc ECCV 2014;8695:329–344.
Schisterman EF, Faraggi D, Reiser B, et al. Youden Index and the optimal threshold for markers with mass at zero. Stat Med 2008;27:297–315.
Treder M, Lauermann JL, Alnawaiseh M, et al. Using deep learning in automated detection of graft detachment in descemet membrane endothelial keratoplasty: A pilot study. Cornea 2019;38:157–161.