Conclusions: Using a CNN-enhanced platform, 60-4 visual fields identified peripheral glaucomatous defects missed by central testing in mild cases; facial contour correction showed these defects occurred exclusively outside the temporal visual field.

Objective: To develop a methodology to separate facial contour-induced visual field defects from defects related to glaucoma in patients with peripheral field defects.

Methods: Ninety-seven eyes from fifty patients previously diagnosed with glaucoma were enrolled in the study. Thirty-one (62%) participants were male, and 86% were white. The study involved patients with glaucoma who underwent visual field testing using standard automated perimetry, including 10-2, 24-2, 30-2, and 60-4 visual fields. A convolutional neural network (CNN)-enhanced platform, previously developed to identify facial contour-dependent defects in healthy subjects, was used to analyze the visual field data. Methods: The study included ninety-seven eyes from fifty patients with glaucoma, with the majority having primary open angle glaucoma.

Results: In ninety out of ninety-seven eyes, there were 60-4 visual field defects. However, 20 (22%) of these eyes did not have glaucomatous defects on the 10-2 or 30-2 visual fields. All patients with 60-4 defects in the absence of more central defects had mild glaucoma, and these eyes comprised 38% of the mild glaucoma group. 60-4 visual field increased the sensitivity of identifying glaucoma in mild disease. Without facial contour, the 60-4 visual field mean threshold summation was 1099.6±346.6 dB, and decreased to 1057±331 dB when accounting for facial contour, which affected 8.03±4.02% of test points. Facial contour-dependent visual field defects were exclusively found outside the temporal visual field.

Conclusions: The study demonstrated the potential utility of 60-4 visual fields for identifying early functional glaucomatous changes that may not be detected by more central visual field tests (10-2 or 30-2). It may be important to consider peripheral visual field defects, which can be affected by facial contour, in the diagnosis and monitoring of glaucoma, particularly in mild cases.