Validation of a Deep Learning Model to Screen for Glaucoma Using Images from Different Fundus Cameras and Data Augmentation
Validation of a Deep Learning Model to Screen for Glaucoma Using Images from Different Fundus Cameras and Data Augmentation
复制标题
DOI：
10.1016/j.ogla.2019.03.008
复制
发表时间：
2019-07-01
影响因子：
2.9
通讯作者：
Kiuchi, Yoshiaki
中科院分区：
文献类型：
--
作者：
Asaoka, Ryo;Tanito, Masaki;Kiuchi, Yoshiaki
目的：验证深度残差学习算法用于青光眼的诊断。设计：横断面研究。参与者：训练数据集，包括1364张具有青光眼适应症的彩色眼底照片和1768张不具有青光眼特征的彩色眼底照片。两个测试数据集包括(1)95张青光眼图像和110张正常眼图像；(2)93张青光眼图像和78张正常眼图像。方法：使用一种称为残差网络(ResNet)的深度学习算法对训练数据集进行青光眼诊断。这两个测试数据集是使用多个研究所的不同眼底相机(不同制造商)获得的。训练数据的大小被人为地增加，方法是对原始数据进行微小的修改，即所谓的“图像增强”。主要观察指标：受试者工作特征曲线下面积。结果：在不使用图像增强的情况下，第一个测试数据集中的AROC为94.8%(90.3~96.8)，第二个数据集中的AROC为99.7%(99.4~100.0)。这些AROC值在没有增强的情况下显著减小(P<；0.05)(在第一个测试数据集中为87.7%[82.8-92.6]，在第二个测试数据集中为94.5%[91.3-97.6])。结论：以前开发的深度残差学习算法在多个研究所的不同眼底摄像机下获得了较高的诊断性能，尤其是在使用图像增强时。(C)2019年由美国眼科学会主办
Purpose: To validate a deep residual learning algorithm to diagnose glaucoma from fundus photography using different fundus cameras at different institutes.Design: Cross-sectional study.Participants: A training dataset consisted of 1364 color fundus photographs with glaucomatous indications and 1768 color fundus photographs without glaucomatous features. Two testing datasets consisted of (1) 95 images of 95 glaucomatous eyes and 110 images of 110 normative eyes, and (2) 93 images of 93 glaucomatous eyes and 78 images of 78 normative eyes.Methods: A deep learning algorithm known as Residual Network (ResNet) was used to diagnose glaucoma using a training dataset. The 2 testing datasets were obtained using different fundus cameras (different manufacturers) across multiple institutes. The size of the training data was artificially increased by adding minor alterations to the original data, known as "image augmentation." Diagnostic accuracy was assessed using the area under the receiver operating characteristic curve (AROC).Main Outcome Measures: Area under the receiver operating characteristic curve.Results: When image augmentation was not used, the AROC was 94.8% (90.3-96.8) in the first testing dataset and 99.7% (99.4-100.0) in the second dataset. These AROC values were significantly (P < 0.05) smaller without augmentation (87.7% [82.8-92.6] in the first testing dataset and 94.5% [91.3-97.6] in the second testing dataset).Conclusions: The previously developed deep residual learning algorithm achieved high diagnostic performance with different fundus cameras across multiple institutes, in particular when image augmentation was used. (C) 2019 by the American Academy of Ophthalmology