Section of Ophthalmology, King's College London, London, UK.
King's Ophthalmology Research Unit, King's College Hospital, London, UK.
Sci Rep. 2022 Jul 1;12(1):11196. doi: 10.1038/s41598-022-15491-1.
Diabetic retinopathy (DR) screening images are heterogeneous and contain undesirable non-retinal, incorrect field and ungradable samples which require curation, a laborious task to perform manually. We developed and validated single and multi-output laterality, retinal presence, retinal field and gradability classification deep learning (DL) models for automated curation. The internal dataset comprised of 7743 images from DR screening (UK) with 1479 external test images (Portugal and Paraguay). Internal vs external multi-output laterality AUROC were right (0.994 vs 0.905), left (0.994 vs 0.911) and unidentifiable (0.996 vs 0.680). Retinal presence AUROC were (1.000 vs 1.000). Retinal field AUROC were macula (0.994 vs 0.955), nasal (0.995 vs 0.962) and other retinal field (0.997 vs 0.944). Gradability AUROC were (0.985 vs 0.918). DL effectively detects laterality, retinal presence, retinal field and gradability of DR screening images with generalisation between centres and populations. DL models could be used for automated image curation within DR screening.
糖尿病性视网膜病变（DR）筛查图像具有异质性，包含不希望出现的非视网膜、不正确的视野和不可评估的样本，这些都需要进行审查，这是一项繁琐的手动任务。我们开发并验证了用于自动审查的单输出和多输出侧别、视网膜存在、视网膜视野和可分级性分类深度学习（DL）模型。内部数据集由来自 DR 筛查（英国）的 7743 张图像和来自葡萄牙和巴拉圭的 1479 张外部测试图像组成。内部与外部多输出侧别 AUROC 分别为右眼（0.994 对 0.905）、左眼（0.994 对 0.911）和无法识别（0.996 对 0.680）。视网膜存在 AUROC 为 1.000。视网膜视野 AUROC 分别为黄斑（0.994 对 0.955）、鼻侧（0.995 对 0.962）和其他视网膜视野（0.997 对 0.944）。可分级性 AUROC 分别为 0.985 和 0.918。DL 能够有效地检测 DR 筛查图像的侧别、视网膜存在、视网膜视野和可分级性，并且在中心和人群之间具有很好的泛化能力。DL 模型可用于 DR 筛查中的自动图像审查。
