Department of Endocrine and Metabolic Diseases, Shanghai Institute of Endocrine and Metabolic Diseases, Ruijin Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, China; Shanghai National Clinical Research Center for metabolic Diseases, Key Laboratory for Endocrine and Metabolic Diseases of the National Health Commission of the PR China, Shanghai National Center for Translational Medicine, Ruijin Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, China.
Department of Internal Medicine, The Second People's Hospital of Yuhuan, Yuhuan, China.
BMJ Open Diabetes Res Care. 2020 Oct;8(1). doi: 10.1136/bmjdrc-2020-001596.
Early screening for diabetic retinopathy (DR) with an efficient and scalable method is highly needed to reduce blindness, due to the growing epidemic of diabetes. The aim of the study was to validate an artificial intelligence-enabled DR screening and to investigate the prevalence of DR in adult patients with diabetes in China.
The study was prospectively conducted at 155 diabetes centers in China. A non-mydriatic, macula-centered fundus photograph per eye was collected and graded through a deep learning (DL)-based, five-stage DR classification. Images from a randomly selected one-third of participants were used for the DL algorithm validation.
In total, 47 269 patients (mean (SD) age, 54.29 (11.60) years) were enrolled. 15 805 randomly selected participants were reviewed by a panel of specialists for DL algorithm validation. The DR grading algorithms had a 83.3% (95% CI: 81.9% to 84.6%) sensitivity and a 92.5% (95% CI: 92.1% to 92.9%) specificity to detect referable DR. The five-stage DR classification performance (concordance: 83.0%) is comparable to the interobserver variability of specialists (concordance: 84.3%). The estimated prevalence in patients with diabetes detected by DL algorithm for any DR, referable DR and vision-threatening DR were 28.8% (95% CI: 28.4% to 29.3%), 24.4% (95% CI: 24.0% to 24.8%) and 10.8% (95% CI: 10.5% to 11.1%), respectively. The prevalence was higher in female, elderly, longer diabetes duration and higher glycated hemoglobin groups.
This study performed, a nationwide, multicenter, DL-based DR screening and the results indicated the importance and feasibility of DR screening in clinical practice with this system deployed at diabetes centers.
NCT04240652.
由于糖尿病的流行日益加剧，需要一种高效且可扩展的方法来早期筛查糖尿病视网膜病变（DR），以降低失明率。本研究旨在验证一种人工智能驱动的 DR 筛查方法，并调查中国成年糖尿病患者 DR 的患病率。
该研究在中国 155 家糖尿病中心前瞻性进行。每只眼采集一张非散瞳、黄斑中心的眼底照相，并通过基于深度学习（DL）的五阶段 DR 分类进行分级。从三分之一的随机参与者中抽取图像用于 DL 算法验证。
共纳入 47269 例患者（平均（SD）年龄 54.29（11.60）岁）。随机抽取的 15805 例患者由专家组进行 DR 分级算法验证。DR 分级算法对检出可治疗性 DR 的敏感性为 83.3%（95%CI：81.9%至 84.6%），特异性为 92.5%（95%CI：92.1%至 92.9%）。五阶段 DR 分类性能（一致性：83.0%）与专家的观察者间变异性相当（一致性：84.3%）。通过 DL 算法检测到的糖尿病患者的任何 DR、可治疗性 DR 和威胁视力的 DR 的估计患病率分别为 28.8%（95%CI：28.4%至 29.3%）、24.4%（95%CI：24.0%至 24.8%）和 10.8%（95%CI：10.5%至 11.1%）。女性、老年人、糖尿病病程较长和糖化血红蛋白较高的患者患病率更高。
本研究开展了一项全国性、多中心的基于 DL 的 DR 筛查，结果表明该系统在糖尿病中心临床应用中进行 DR 筛查的重要性和可行性。
NCT04240652。
