Performance of a Deep-Learning Algorithm vs Manual Grading for Detecting Diabetic Retinopathy in India
Performance of a Deep-Learning Algorithm vs Manual Grading for Detecting Diabetic Retinopathy in India
复制标题
DOI：
10.1001/jamaophthalmol.2019.2004
复制
发表时间：
2019-09-01
影响因子：
8.1
通讯作者：
Webster, Dale R.
中科院分区：
文献类型：
--
作者：
Gulshan, Varun;Rajan, Renu P.;Webster, Dale R.
深度学习模型在印度糖尿病患者队列中的表现如何？在这项中度或重度糖尿病视网膜病变和可诊断性糖尿病黄斑水肿的观察性研究中，自动糖尿病视网膜病变系统的性能等于或超过手动分级。意义深度学习模型在印度糖尿病患者群体中表现良好。重要性印度有超过6000万人患有糖尿病，并有患糖尿病视网膜病变（DR）的风险，这是一种威胁视力的疾病。视网膜眼底照片的自动化解释可以帮助支持和扩展一个强大的筛选程序，以检测DR。目的前瞻性地验证在印度的2个站点的自动DR系统的性能。设计、设置和患者这项前瞻性观察性研究在印度的2家眼科护理中心（Aravind眼科医院和Sankara Nethralaya）进行，包括3049例糖尿病患者。数据收集和患者入组于2016年4月至2016年7月在Aravind以及2016年5月和2017年4月在Sankara Nethralaya进行。该模型于2016年3月进行了训练和修复。干预自动DR分级系统与由来自每个研究中心的1名经过培训的分级员和1名视网膜专家进行的手动分级进行比较。在意见不一致的情况下，由3名视网膜专家组成的专家组的裁定作为参考标准。主要结局和指标中度或重度DR或可缓解的糖尿病黄斑水肿的敏感性和特异性。结果在3049例患者中，1091例（35.8%）为女性，Aravind和Sankara Nethralaya患者的平均（SD）年龄分别为56.6（9.0）岁和56.0（10.0）岁。对于中度或重度DR，由个体非裁定分级者手动分级的敏感性和特异性范围分别为73.4%至89.8%和83.5%至98.7%。自动DR系统的性能等于或超过手动分级，灵敏度为88.9%（95% CI，85.8-91.5），92.2%特异性（95%CI，90.3-93.8），Aravind眼科医院数据集的曲线下面积为0.963，灵敏度为92.1%（95% CI，90.1-93.8），95.2%特异性（95% CI，94.2-96.1），以及Sankara Nethralaya数据集上的曲线下面积为0.980。结论和相关性本研究表明，自动DR系统推广到这一人口的印度患者在一个前瞻性的设置，并证明了使用自动DR分级系统，以扩大筛选programmes.This研究的可行性评估的有效性，自动糖尿病视网膜病变系统相比，在印度的2个网站手动分级。
Key PointsQuestionWhat is the performance of a deep-learning model in a cohort of patients with diabetes in India? FindingsIn this observational study of moderate or worse diabetic retinopathy and referable diabetic macular edema, the automated diabetic retinopathy system's performance was equal to or exceeded manual grading. MeaningDeep-learning models performed well within a population of patients with diabetes from India.ImportanceMore than 60 million people in India have diabetes and are at risk for diabetic retinopathy (DR), a vision-threatening disease. Automated interpretation of retinal fundus photographs can help support and scale a robust screening program to detect DR. ObjectiveTo prospectively validate the performance of an automated DR system across 2 sites in India. Design, Setting, and ParticipantsThis prospective observational study was conducted at 2 eye care centers in India (Aravind Eye Hospital and Sankara Nethralaya) and included 3049 patients with diabetes. Data collection and patient enrollment took place between April 2016 and July 2016 at Aravind and May 2016 and April 2017 at Sankara Nethralaya. The model was trained and fixed in March 2016. InterventionsAutomated DR grading system compared with manual grading by 1 trained grader and 1 retina specialist from each site. Adjudication by a panel of 3 retinal specialists served as the reference standard in the cases of disagreement. Main Outcomes and MeasuresSensitivity and specificity for moderate or worse DR or referable diabetic macula edema. ResultsOf 3049 patients, 1091 (35.8%) were women and the mean (SD) age for patients at Aravind and Sankara Nethralaya was 56.6 (9.0) years and 56.0 (10.0) years, respectively. For moderate or worse DR, the sensitivity and specificity for manual grading by individual nonadjudicator graders ranged from 73.4% to 89.8% and from 83.5% to 98.7%, respectively. The automated DR system's performance was equal to or exceeded manual grading, with an 88.9% sensitivity (95% CI, 85.8-91.5), 92.2% specificity (95% CI, 90.3-93.8), and an area under the curve of 0.963 on the data set from Aravind Eye Hospital and 92.1% sensitivity (95% CI, 90.1-93.8), 95.2% specificity (95% CI, 94.2-96.1), and an area under the curve of 0.980 on the data set from Sankara Nethralaya. Conclusions and RelevanceThis study shows that the automated DR system generalizes to this population of Indian patients in a prospective setting and demonstrates the feasibility of using an automated DR grading system to expand screening programs.This study assesses the validity of an automated diabetic retinopathy system compared with manual grading across 2 sites in India.