A stratified analysis of a deep learning algorithm in the diagnosis of diabetic retinopathy in a real‐world study. Issue 2 (9th December 2021)

• Record Type: Journal Article
• Authors: Li, Na; Ma, Mingming; Lai, Mengyu; Gu, Liping; Kang, Mei; Wang, Zilong; Jiao, Shengyin; Dang, Kang; Deng, Junxiao; Ding, Xiaowei; Zhen, Qin; Zhang, Aifang; Shen, Tingting; Zheng, Zhi; Wang, Yufan; Peng, Yongde
• Is Part Of: Journal of diabetes. Volume 14:Issue 2(2022)
• Page Start: 111
• Page End: 120
• Publication Date: 2021-12-09
• Subjects: deep learning algorithm -- diabetic retinopathy -- referable DR -- retinal fundus images; 深度学习算法 -- 糖尿病视网膜病变 -- 可参考DR -- 视网膜眼底图像; Diabetes -- Periodicals; 618.3646005
• DOI: 10.1111/1753-0407.13241
• Languages: English
• ISSNs: 1753-0393

Abstract

Abstract: Background: The aim of our research was to prospectively explore the clinical value of a deep learning algorithm (DLA) to detect referable diabetic retinopathy (DR) in different subgroups stratified by types of diabetes, blood pressure, sex, BMI, age, glycosylated hemoglobin (HbA1c), diabetes duration, urine albumin‐to‐creatinine ratio (UACR), and estimated glomerular filtration rate (eGFR) at a real‐world diabetes center in China. Methods: A total of 1147 diabetic patients from Shanghai General Hospital were recruited from October 2018 to August 2019. Retinal fundus images were graded by the DLA, and the detection of referable DR (moderate nonproliferative DR or worse) was compared with a reference standard generated by one certified retinal specialist with more than 12 years of experience. The performance of DLA across different subgroups stratified by types of diabetes, blood pressure, sex, BMI, age, HbA1c, diabetes duration, UACR, and eGFR was evaluated. Results: For all 1674 gradable images, the area under the receiver operating curve, sensitivity, and specificity of the DLA for referable DR were 0.942 (95% CI, 0.920‐0.964), 85.1% (95% CI, 83.4%‐86.8%), and 95.6% (95% CI, 94.6%‐96.6%), respectively. The DLA showed consistent performance across most subgroups, while it showed superior performance in the subgroups of patients with type 1 diabetes, UACR ≥ 30 mg/g, and eGFR < 90 mL/min/1.73m 2 . Conclusions: This study showed that the DLA was a reliable alternative method for the detection of referable DR and performed superior in patients with type 1 diabetes and diabetic nephropathy who were prone to DR. Abstract : Highlights Our deep learning algorithm (DLA)‐based diabetic retinopathy (DR) grading software showed excellent performance in detecting referable DR, especially in patients with type 1 diabetes and diabetic nephropathy who were prone to DR. The DLA showed consistent performance in the detection of referable DR across different subgroups stratified by blood pressure, sex, BMI, age, glycosylated hemoglobin, and diabetes duration in a real‐world diabetic center. 摘要: 背景: 前瞻性探索在中国的一个真实世界糖尿病中心中, 深度学习算法 (DLA) 在按糖尿病类型、血压、性别、BMI、年龄、HbA1c、糖尿病病程、肾小球滤过率(UACR)和估计肾小球滤过率(eGFR)水平分层的不同亚组中检测可参考糖尿病视网膜病变 (DR) 的临床价值。 方法: 从 2018 年 10 月至 2019 年 8 月, 招募1147 名糖尿病患者。视网膜眼底图像由 DLA进行分级, 并将可参考 DR(中度非增殖性 DR 或更差)的检测与一位经认证拥有十二年以上经验的眼科专家的参考标准进行比较。评估按糖尿病类型、血压、性别、BMI、年龄、HbA1c、糖尿病病程、尿白蛋白与肌酐比和肾小球滤过率分层的不同亚组 DLA 的表现。 结果: 对于所有 1674 个可分级图像, DLA 对可参考 DR 的 AUC、敏感度和特异度分别为 0.942(95% CI, 0.920‐0.964)、85.1%(95% CI, 83.4%‐86.8%)和 95.6% (95% CI, 94.6%96.6%)。 DLA 在大多数亚组中有一致的表现, 而在1型糖尿病、UACR≥30 mg/g和eGFR <90mL/min/1.73m 2 的患者亚组中有优异的表现。 结论: 本研究表明, DLA是检测可参考DR的一种可靠的替代方法, 并且在易发生DR的1型糖尿病和糖尿病肾病患者中表现更好。