A Deep Learning Algorithm to Quantify Neuroretinal Rim Loss From Optic Disc Photographs

Abstract

Purpose: To train a deep learning (DL) algorithm that quantifies glaucomatous neuroretinal damage on fundus photographs using the minimum rim width relative to Bruch membrane opening (BMO-MRW) from spectral-domain optical coherence tomography (SDOCT).

Design: Cross-sectional study.

Methods: A total of 9282 pairs of optic disc photographs and SDOCT optic nerve head scans from 927 eyes of 490 subjects were randomly divided into the validation plus training (80%) and test sets (20%). A DL convolutional neural network was trained to predict the SDOCT BMO-MRW global and sector values when evaluating optic disc photographs. The predictions of the DL network were compared to the actual SDOCT measurements. The area under the receiver operating curve (AUC) was used to evaluate the ability of the network to discriminate glaucomatous visual field loss from normal eyes.

Results: The DL predictions of global BMO-MRW from all optic disc photographs in the test set (mean ± standard deviation [SD]: 228.8 ± 63.1 μm) were highly correlated with the observed values from SDOCT (mean ± SD: 226.0 ± 73.8 μm) (Pearson's r = 0.88; R² = 77%; P < .001), with mean absolute error of the predictions of 27.8 μm. The AUCs for discriminating glaucomatous from healthy eyes with the DL predictions and actual SDOCT global BMO-MRW measurements were 0.945 (95% confidence interval [CI]: 0.874-0.980) and 0.933 (95% CI: 0.856-0.975), respectively (P = .587).

Conclusions: A DL network can be trained to quantify the amount of neuroretinal damage on optic disc photographs using SDOCT BMO-MRW as a reference. This algorithm showed high accuracy for glaucoma detection, and may potentially eliminate the need for human gradings of disc photographs.

Figure 1.

Scatterplot illustrating the relationship between predictions obtained by the deep learning algorithm evaluating optic disc photographs and actual global minimum rim width relative to Bruch’s membrane opening (BMO-MRW) thickness measurements from spectral domain-optical coherence tomography (SDOCT). Data is from the independent test set.

Figure 2.

Activation heatmaps showing the areas of the optic disc photograph that were most important for the deep learning algorithm predictions in an example of a healthy (Top row) and glaucomatous (Bottom row) eye.

Figure 3.

Random examples of optic disc photos from the test sample along with the corresponding global minimum rim width (MRW) relative to Bruch’s membrane opening deep learning (DL) predictions and spectral domain-optical coherence tomography (SDOCT) measurements. (Top) Random examples where predictions closely agreed to observations (within 20μm). (Bottom) Random examples where the predictions disagreed with the observations (greater than 50μm difference).