From the Department of Ophthalmology (A.B., K.E., A.H., A.F., L.B., K.Y., K.V., N.C., L.R.P.), Icahn School of Medicine at Mount Sinai, New York, New York, USA; Department of Ophthalmology (A.B., K.E., A.H., A.F., L.B., K.Y., K.V., N.C., L.R.P.), New York Eye and Ear Infirmary of Mount Sinai, New York, New York, USA.
Biomedical Data Science (H.C., R.B.A.), Stanford University, Stanford, California, USA.
Am J Ophthalmol. 2023 Nov;255:161-169. doi: 10.1016/j.ajo.2023.07.007. Epub 2023 Jul 23.
To develop an automated deep learning system for detecting the presence and location of disc hemorrhages in optic disc photographs.
Development and testing of a deep learning algorithm.
Optic disc photos (597 images with at least 1 disc hemorrhage and 1075 images without any disc hemorrhage from 1562 eyes) from 5 institutions were classified by expert graders based on the presence or absence of disc hemorrhage. The images were split into training (n = 1340), validation (n = 167), and test (n = 165) datasets. Two state-of-the-art deep learning algorithms based on either object-level detection or image-level classification were trained on the dataset. These models were compared to one another and against 2 independent glaucoma specialists. We evaluated model performance by the area under the receiver operating characteristic curve (AUC). AUCs were compared with the Hanley-McNeil method.
The object detection model achieved an AUC of 0.936 (95% CI = 0.857-0.964) across all held-out images (n = 165 photographs), which was significantly superior to the image classification model (AUC = 0.845, 95% CI = 0.740-0.912; P = .006). At an operating point selected for high specificity, the model achieved a specificity of 94.3% and a sensitivity of 70.0%, which was statistically indistinguishable from an expert clinician (P = .7). At an operating point selected for high sensitivity, the model achieves a sensitivity of 96.7% and a specificity of 73.3%.
An autonomous object detection model is superior to an image classification model for detecting disc hemorrhages, and performed comparably to 2 clinicians.
开发一种自动深度学习系统，用于检测视盘照片中盘状出血的存在和位置。
深度学习算法的开发和测试。
根据是否存在盘状出血，由专家分级员对来自 5 个机构的视盘照片（至少有 1 个盘状出血的 597 张图像和没有任何盘状出血的 1075 张图像，共 1562 只眼）进行分类。将图像分为训练集（n=1340）、验证集（n=167）和测试集（n=165）。基于对象级检测或图像级分类的两种最先进的深度学习算法在数据集上进行了训练。将这些模型相互比较，并与 2 位独立的青光眼专家进行比较。我们通过接收者操作特征曲线下的面积（AUC）来评估模型性能。AUC 与 Hanley-McNeil 方法进行了比较。
在所有保留图像（n=165 张照片）中，对象检测模型的 AUC 为 0.936（95%CI=0.857-0.964），明显优于图像分类模型（AUC=0.845，95%CI=0.740-0.912；P=0.006）。在选择高特异性的工作点时，模型的特异性为 94.3%，灵敏度为 70.0%，与专家临床医生的结果无统计学差异（P=0.7）。在选择高灵敏度的工作点时，模型的灵敏度为 96.7%，特异性为 73.3%。
自主对象检测模型在检测盘状出血方面优于图像分类模型，并且与 2 位临床医生的表现相当。
