Improved Automated Detection of Diabetic Retinopathy on a Publicly Available Dataset Through Integration of Deep Learning
Improved Automated Detection of Diabetic Retinopathy on a Publicly Available Dataset Through Integration of Deep Learning
复制标题
DOI：
10.1167/iovs.16-19964
复制
发表时间：
2016-10-01
影响因子：
4.4
通讯作者：
Niemeijer, Meindert
中科院分区：
文献类型：
--
作者：
Abramoff, Michael David;Lou, Yiyue;Niemeijer, Meindert
目的.为了比较用于自动检测糖尿病视网膜病变（DR）的深度学习增强算法的性能，与之前发表的该算法的性能，爱荷华州检测计划（IDP）-无深度学习组件-由三位美国委员会认证的视网膜专家在相同的公开可用眼底图像集和之前报告的共识参考标准集上进行。我们使用了先前报道的一致参考标准，即糖尿病视网膜病变中度、重度非增殖性（NPDR）、增殖性DR和/或黄斑水肿（ME）的国际临床分类。Messidor-2图像和设置Messidor-2参考标准的三位视网膜专家均未用于训练IDx-DR版本X2.1。计算敏感性、特异性、阴性预测值、曲线下面积（AUC）及其置信区间（CI）。敏感性为96.8%（95%CI：93.3%-98.8%），特异性为87.0%（95%CI：84.2%-89.4%），假阴性6/874，阴性预测值为99.0%（95%CI：97.8%-99.6%）。无严重NPDR、PDR或ME病例漏诊。AUC为0.980（95% CI：0.968-0.992）。敏感性与发表的IDP敏感性无统计学差异，其CI为94.4%至99.3%，但特异性明显优于发表的IDP特异性CI 55.7%至63.0%。用于DR自动检测的深度学习增强算法实现了比先前报告的、在其他方面基本相同的、不采用深度学习的算法显著更好的性能。深度学习增强算法有可能提高DR筛查的效率，从而防止这种毁灭性疾病造成的视力丧失和失明。
PURPOSE. To compare performance of a deep-learning enhanced algorithm for automated detection of diabetic retinopathy (DR), to the previously published performance of that algorithm, the Iowa Detection Program (IDP)-without deep learning components-on the same publicly available set of fundus images and previously reported consensus reference standard set, by three US Board certified retinal specialists.METHODS. We used the previously reported consensus reference standard of referable DR (rDR), defined as International Clinical Classification of Diabetic Retinopathy moderate, severe nonproliferative (NPDR), proliferative DR, and/or macular edema (ME). Neither Messidor-2 images, nor the three retinal specialists setting the Messidor-2 reference standard were used for training IDx-DR version X2.1. Sensitivity, specificity, negative predictive value, area under the curve (AUC), and their confidence intervals (CIs) were calculated.RESULTS. Sensitivity was 96.8% (95% CI: 93.3%-98.8%), specificity was 87.0% (95% CI: 84.2%-89.4%), with 6/874 false negatives, resulting in a negative predictive value of 99.0% (95% CI: 97.8%-99.6%). No cases of severe NPDR, PDR, or ME were missed. The AUC was 0.980 (95% CI: 0.968-0.992). Sensitivity was not statistically different from published IDP sensitivity, which had a CI of 94.4% to 99.3%, but specificity was significantly better than the published IDP specificity CI of 55.7% to 63.0%.CONCLUSIONS. A deep-learning enhanced algorithm for the automated detection of DR, achieves significantly better performance than a previously reported, otherwise essentially identical, algorithm that does not employ deep learning. Deep learning enhanced algorithms have the potential to improve the efficiency of DR screening, and thereby to prevent visual loss and blindness from this devastating disease.