ImportanceMachine learning (ML) algorithms have the potential to identify eyes with early diabetic retinopathy (DR) at increased risk for disease progression.ObjectiveTo create and validate automated ML models (autoML) for DR progression from ultra-widefield (UWF) retinal images.Design, Setting and ParticipantsDeidentified UWF images with mild or moderate nonproliferative DR (NPDR) with 3 years of longitudinal follow-up retinal imaging or evidence of progression within 3 years were used to develop automated ML models for predicting DR progression in UWF images. All images were collected from a tertiary diabetes-specific medical center retinal image dataset. Data were collected from July to September 2022.ExposureAutomated ML models were generated from baseline on-axis 200° UWF retinal images. Baseline retinal images were labeled for progression based on centralized reading center evaluation of baseline and follow-up images according to the clinical Early Treatment Diabetic Retinopathy Study severity scale. Images for model development were split 8-1-1 for training, optimization, and testing to detect 1 or more steps of DR progression. Validation was performed using a 328-image set from the same patient population not used in model development.Main Outcomes and MeasuresArea under the precision-recall curve (AUPRC), sensitivity, specificity, and accuracy.ResultsA total of 1179 deidentified UWF images with mild (380 [32.2%]) or moderate (799 [67.8%]) NPDR were included. DR progression was present in half of the training set (590 of 1179 [50.0%]). The model’s AUPRC was 0.717 for baseline mild NPDR and 0.863 for moderate NPDR. On the validation set for eyes with mild NPDR, sensitivity was 0.72 (95% CI, 0.57-0.83), specificity was 0.63 (95% CI, 0.57-0.69), prevalence was 0.15 (95% CI, 0.12-0.20), and accuracy was 64.3%; for eyes with moderate NPDR, sensitivity was 0.80 (95% CI, 0.70-0.87), specificity was 0.72 (95% CI, 0.66-0.76), prevalence was 0.22 (95% CI, 0.19-0.27), and accuracy was 73.8%. In the validation set, 6 of 9 eyes (75%) with mild NPDR and 35 of 41 eyes (85%) with moderate NPDR progressed 2 steps or more were identified. All 4 eyes with mild NPDR that progressed within 6 months and 1 year were identified, and 8 of 9 (89%) and 17 of 20 (85%) with moderate NPDR that progressed within 6 months and 1 year, respectively, were identified.Conclusions and RelevanceThis study demonstrates the accuracy and feasibility of automated ML models for identifying DR progression developed using UWF images, especially for prediction of 2-step or greater DR progression within 1 year. Potentially, the use of ML algorithms may refine the risk of disease progression and identify those at highest short-term risk, thus reducing costs and improving vision-related outcomes.

中文翻译：

---

自动机器学习从超宽视野视网膜图像预测糖尿病视网膜病变进展

重要性机器学习 (ML) 算法有潜力识别患有早期糖尿病视网膜病变 (DR) 且疾病进展风险增加的眼睛。目标根据超宽视野 (UWF) 视网膜图像创建并验证用于 DR 进展的自动化 ML 模型 (autoML)。设计、设置和参与者使用具有轻度或中度非增殖性 DR (NPDR) 的去识别 UWF 图像以及 3 年纵向随访视网膜成像或 3 年内进展证据来开发自动化 ML 模型，用于预测 UWF 图像中的 DR 进展。所有图像均来自三级糖尿病特定医疗中心视网膜图像数据集。数据收集时间为 2022 年 7 月至 9 月。ExposureAutomated ML 模型是根据基线轴上 200° UWF 视网膜图像生成的。基线视网膜图像根据临床早期治疗糖尿病视网膜病变研究严重程度量表对基线和后续图像的集中阅读中心评估进行标记。用于模型开发的图像被分割为 8-1-1，用于训练、优化和测试，以检测 DR 进展的 1 个或多个步骤。使用来自模型开发中未使用的相同患者群体的 328 个图像集进行验证。主要结果和测量精确回忆曲线 (AUPRC) 下的面积、灵敏度、特异性和准确性。结果总共 1179 个未识别的 UWF 图像，具有轻度(380 [32.2%]) 或中等 (799 [67.8%]) NPDR 被纳入。一半的训练集中出现 DR 进展（1179 组中的 590 组[50.0%]）。该模型的基线轻度 NPDR 的 AUPRC 为 0.717，中度 NPDR 的 AUPRC 为 0.863。在轻度 NPDR 眼睛的验证集上，敏感性为 0.72（95% CI，0.57-0.83），特异性为 0.63（95% CI，0.57-0.69），患病率为 0.15（95% CI，0.12-0.20），并且准确率为64.3%；对于中度 NPDR 眼睛，敏感性为 0.80（95% CI，0.70-0.87），特异性为 0.72（95% CI，0.66-0.76），患病率为 0.22（95% CI，0.19-0.27），准确度为 73.8% 。在验证集中，9 只眼睛中的 6 只 (75%) 患有轻度 NPDR，41 只眼睛中的 35 只 (85%) 中度 NPDR 进展了 2 步或更多。确定了所有 4 只在 6 个月和 1 年内进展的轻度 NPDR 眼睛，并分别确定了在 6 个月和 1 年内进展的中度 NPDR 的 9 只眼睛中的 8 只 (89%) 和 20 只眼睛中的 17 只 (85%)。结论和相关性本研究证明了使用 UWF 图像开发的用于识别 DR 进展的自动化 ML 模型的准确性和可行性，特别是对于预测 1 年内 2 步或更大 DR 进展的预测。机器学习算法的使用可能会改善疾病进展的风险，并识别短期风险最高的人群，从而降低成本并改善与视力相关的结果。