Automated detection of early-stage ROP using a deep convolutional neural network.
Automated detection of early-stage ROP using a deep convolutional neural network.
复制标题
DOI：
10.1136/bjophthalmol-2020-316526
复制
发表时间：
2021-08
期刊：
影响因子：
--
通讯作者：
Wu WC
中科院分区：
文献类型：
--
作者：
Huang YP;Basanta H;Kang EY;Chen KJ;Hwang YS;Lai CC;Campbell JP;Chiang MF;Chan RVP;Kusaka S;Fukushima Y;Wu WC
目的：利用深度卷积神经网络(CNN)对早产儿视网膜病变(ROP)的早期阶段进行自动检测和分类。这项回顾性横断面研究是在台湾的一家转诊医疗中心进行的。仅纳入无ROP、第1期ROP或第2期ROP的早产儿。总体而言，11372张视网膜眼底图像被汇编并拆分成10235张(90%)用于训练、1137张(10%)用于验证和244张用于测试。实现了深度CNN，根据ROP分期对图像进行分类。数据收集时间为2013年12月17日至2019年5月24日，分析时间为2018年12月至2020年1月。采用灵敏度、特异度和受试者工作特征曲线下面积等指标来评价算法相对于参考标准诊断的性能。采用五次交叉验证对模型进行训练，训练时的平均准确率为99.93%±0.03，测试时的平均准确率为92.23%±1.39。该模型预测无ROP与ROP、分期1 ROP与无ROP和分期2 ROP、分期2 ROP与无ROP和分期1 ROP的敏感度和特异度分别为96.14%±0.87和95.95%±0.48、91.82%±2.03和94.50%±0.71和89.81%±1.82和98.99%±0.40。该系统可以准确地区分ROP的早期阶段，并有可能帮助眼科医生对ROP进行早期分类。
To automatically detect and classify the early stages of retinopathy of prematurity (ROP) using a deep convolutional neural network (CNN). This retrospective cross-sectional study was conducted in a referral medical centre in Taiwan. Only premature infants with no ROP, stage 1 ROP or stage 2 ROP were enrolled. Overall, 11 372 retinal fundus images were compiled and split into 10 235 images (90%) for training, 1137 (10%) for validation and 244 for testing. A deep CNN was implemented to classify images according to the ROP stage. Data were collected from December 17, 2013 to May 24, 2019 and analysed from December 2018 to January 2020. The metrics of sensitivity, specificity and area under the receiver operating characteristic curve were adopted to evaluate the performance of the algorithm relative to the reference standard diagnosis. The model was trained using fivefold cross-validation, yielding an average accuracy of 99.93%±0.03 during training and 92.23%±1.39 during testing. The sensitivity and specificity scores of the model were 96.14%±0.87 and 95.95%±0.48, 91.82%±2.03 and 94.50%±0.71, and 89.81%±1.82 and 98.99%±0.40 when predicting no ROP versus ROP, stage 1 ROP versus no ROP and stage 2 ROP, and stage 2 ROP versus no ROP and stage 1 ROP, respectively. The proposed system can accurately differentiate among ROP early stages and has the potential to help ophthalmologists classify ROP at an early stage.
登录
查看更多内容
影响因子：
5.1
作者：
Kim SJ;Port AD;Swan R;Campbell JP;Chan RVP;Chiang MF
通讯作者：
Chiang MF
影响因子：
4.1
作者：
Mora, Justin S.;Waite, Christopher;Sloper, John J.
通讯作者：
Sloper, John J.
DOI：
10.1136/bjophthalmol-2017-311789
发表时间：
2018-12
期刊：
The British journal of ophthalmology
影响因子：
--
作者：
Bas AY;Demirel N;Koc E;Ulubas Isik D;Hirfanoglu İM;Tunc T;TR-ROP Study Group
通讯作者：
TR-ROP Study Group
影响因子：
4.1
作者：
Painter, Sally L.;Wilkinson, Andrew R.;Patel, C. K.
通讯作者：
Patel, C. K.
影响因子：
13.7
作者：
Campbell, J. Peter;Ryan, Michael C.;Chiang, Michael F.
通讯作者：
Chiang, Michael F.