Geometric Deep Learning to Identify the Critical 3D Structural Features of the Optic Nerve Head for Glaucoma Diagnosis

center dot PURPOSE: To compare the performance of 2 relatively recent geometric deep learning techniques in diagnosing glaucoma from a single optical coherence tomographic (OCT) scan of the optic nerve head (ONH); and to iden-tify the 3-dimensional (3D) structural features of the ONH that are critical for the diagnosis of glaucoma. center dot DESIGN: Comparison and evaluation of deep learning diagnostic algorithms. center dot METHODS: In this study, we included a total of 2247 nonglaucoma and 2259 glaucoma scans from 1725 par-ticipants. All participants had their ONHs imaged in 3D with Spectralis OCT. All OCT scans were automatically segmented using deep learning to identify major neural and connective tissues. Each ONH was then represented as a 3D point cloud. We used PointNet and dynamic graph convolutional neural network (DGCNN) to diag-nose glaucoma from such 3D ONH point clouds and to identify the critical 3D structural features of the ONH for glaucoma diagnosis. center dot RESULTS: Both the DGCNN (area under the curve [AUC]: 0.97 +/- 0.01) and PointNet (AUC: 0.95 +/- 0.02) were able to accurately detect glaucoma from 3D ONH point clouds. The critical points (ie, critical structural features of the ONH) formed an hourglass pattern, with most of them located within the neuroretinal rim in the inferior and superior quadrant of the ONH. center dot CONCLUSIONS: The diagnostic accuracy of both geo-metric deep learning approaches was excellent. Moreover, we were able to identify the critical 3D structural fea-tures of the ONH for glaucoma diagnosis that tremen-dously improved the transparency and interpretability of our method. Consequently, our approach may have strong potential to be used in clinical applications for the diagno-sis and prognosis of a wide range of ophthalmic disorders. (Am J Ophthalmol 2023;250: 38-48. (c) 2023 Elsevier Inc. All rights reserved.)