Efficacy of a Deep Learning System for Detecting Glaucomatous Optic Neuropathy Based on Color Fundus Photographs

To assess the performance of a deep learning algorithm for detecting referable glaucomatous optic neuropathy (GON) based on color fundus photographs. A deep learning system for the classification of GON was developed for automated classification of GON on color fundus photographs. We retrospectively included 48 116 fundus photographs for the development and validation of a deep learning algorithm. This study recruited 21 trained ophthalmologists to classify the photographs. Referable GON was defined as vertical cup-to-disc ratio of 0.7 or more and other typical changes of GON. The reference standard was made until 3 graders achieved agreement. A separate validation dataset of 8000 fully gradable fundus photographs was used to assess the performance of this algorithm. The area under receiver operator characteristic curve (AUC) with sensitivity and specificity was applied to evaluate the efficacy of the deep learning algorithm detecting referable GON. In the validation dataset, this deep learning system achieved an AUC of 0.986 with sensitivity of 95.6% and specificity of 92.0%. The most common reasons for false-negative grading (n = 87) were GON with coexisting eye conditions (n = 44 [50.6%]), including pathologic or high myopia (n = 37 [42.6%]), diabetic retinopathy (n = 4 [4.6%]), and age-related macular degeneration (n = 3 [3.4%]). The leading reason for false-positive results (n = 480) was having other eye conditions (n = 458 [95.4%]), mainly including physiologic cupping (n = 267 [55.6%]). Misclassification as false-positive results amidst a normal-appearing fundus occurred in only 22 eyes (4.6%). A deep learning system can detect referable GON with high sensitivity and specificity. Coexistence of high or pathologic myopia is the most common cause resulting in false-negative results. Physiologic cupping and pathologic myopia were the most common reasons for false-positive results.