In American journal of ophthalmology ; h5-index 67.0
PURPOSE : To develop an artificial neural network model incorporating both spatial and ordinal approaches to predict glaucomatous visual field (VF) progression.
DESIGN : Cohort study.
PARTICIPANTS : From a cohort of primary open-angle glaucoma patients, 9,212 eyes of 6,047 patients who underwent regular reliable VF examinations for >4 years were included.
METHODS : We constructed all possible spatial-ordinal tensors by stacking three consecutive VF tests (VF-blocks) with at least 3 years of follow-up. Trend-based, event-based and combined criteria were defined to determine the progression. VF-blocks were considered "progressed" if progression occurred within 3 years; the progression was further confirmed after 3 years. We constructed six convolutional neural network (NN) models and two linear models: regression on global indices and pointwise linear regression (PLR). We compared area under the receiver operating characteristic curve (AUROC) of each models for the prediction of glaucomatous VF progression.
RESULTS : Among 43,260 VF-blocks, 4,406 (10.2%), 4,376 (10.1%), and 2,394 (5.5%) VF blocks were classified as progression based on trend-based, event-based and combined criteria. For all three criteria, the progression group was significantly older and had worse initial MD and VFI than the non-progression group (p < 0.001 for all). The best-performing NN model had an AUROC of 0.864 with sensitivity of 0.42 at specificity of 0.95. In contrast, an AUROC of 0.611 was estimated from sensitivity of 0.28 at specificity of 0.84 for the PLR.
CONCLUSIONS : The NN models incorporating spatial-ordinal characteristics demonstrated significantly better performance than the linear models in the prediction of glaucomatous VF progression.
Shon Kilhwan, Sung Kyung Rim, Shin Joong Won
artificial intelligence, glaucoma, machine learning, visual field