The Three-Dimensional Structural Configuration of the Central Retinal Vessel Trunk and Branches as a Glaucoma Biomarker
Satish K Panda,Haris Cheong,Tin A Tun,Thanadet Chuangsuwanich,Aiste Kadziauskiene,Vijayalakshmi Senthil,Ramaswami Krishnadas,Martin L Buist,Shamira Perera,Ching-Yu Cheng,Tin Aung,Alexandre H Thiery,Michaël J A Girard,Satish K. Panda,Tin A. Tun,Martin L. Buist,Alexandre H. Thiery,Michaël J.A. Girard
DOI: https://doi.org/10.1016/j.ajo.2022.02.020
IF: 5.488
2022-03-01
American Journal of Ophthalmology
Abstract:PurposeTo assess whether the 3-dimensional (3D) structural configuration of the central retinal vessel trunk and its branches (CRVT&B) could be used as a diagnostic marker for glaucoma.DesignRetrospective, deep-learning approach diagnosis study.MethodsWe trained a deep learning network to automatically segment the CRVT&B from the B-scans of the optical coherence tomography (OCT) volume of the optic nerve head. Subsequently, 2 different approaches were used for glaucoma diagnosis using the structural configuration of the CRVT&B as extracted from the OCT volumes. In the first approach, we aimed to provide a diagnosis using only 3D convolutional neural networks and the 3D structure of the CRVT&B. For the second approach, we projected the 3D structure of the CRVT&B orthographically onto sagittal, frontal, and transverse planes to obtain 3 two-dimensional (2D) images, and then a 2D convolutional neural network was used for diagnosis. The segmentation accuracy was evaluated using the Dice coefficient, whereas the diagnostic accuracy was assessed using the area under the receiver operating characteristic curves (AUCs). The diagnostic performance of the CRVT&B was also compared with that of retinal nerve fiber layer (RNFL) thickness (calculated in the same cohorts).ResultsOur segmentation network was able to efficiently segment retinal blood vessels from OCT scans. On a test set, we achieved a Dice coefficient of 0.81 ± 0.07. The 3D and 2D diagnostic networks were able to differentiate glaucoma from nonglaucoma subjects with accuracies of 82.7% and 83.3%, respectively. The corresponding AUCs for the CRVT&B were 0.89 and 0.90, higher than those obtained with RNFL thickness alone (AUCs ranging from 0.74 to 0.80).ConclusionsOur work demonstrated that the diagnostic power of the CRVT&B is superior to that of a gold-standard glaucoma parameter, that is, RNFL thickness. Our work also suggested that the major retinal blood vessels form a "skeleton"—the configuration of which may be representative of major optic nerve head structural changes as typically observed with the development and progression of glaucoma.