To determine whether monoscopic vs stereoscopic viewing impacts evaluation of optic disc photographs for glaucoma diagnosis in an expert population.

Prospective observational study.

Twenty pairs of high-quality monoscopic and stereoscopic photographs of similar size and magnification (ie, 40 images), were selected to demonstrate a range of optic disc features from a total of 197 eyes of 197 patients with glaucoma and normal subjects recruited from a tertiary clinic. These were presented in randomized order via an interactive platform (stereo.gone-project.com/). Participants assessed 9 topographic features and estimated glaucoma likelihood for each photograph. Main outcome measures were intra- and inter-observer agreement.

There was good intra-observer agreement between monoscopic and stereoscopic assessments of glaucoma likelihood (κ_w = 0.56). There was also good to substantial agreement for peripapillary atrophy (κ_w = 0.65), cup shape (κ_w = 0.65), retinal nerve fiber layer loss (κ_w = 0.69), vertical cup-to-disc ratio (κ_w = 0.58), and disc shape (κ_w = 0.57). However, intra-observer agreement was only fair to moderate for disc tilt, cup depth, and disc size (κ_w = 0.46–0.49). Inter-observer agreement for glaucoma likelihood in monoscopic photographs (κ_w = 0.61, 95% confidence interval [CI] = 0.55–0.67) was substantial and not lower than in stereoscopic photographs (κ_w = 0.59, CI = 0.54–0.65). Monoscopic photographs did not lead to lower levels of inter-observer agreement compared to stereoscopic photographs in the assessment of any optic disc characteristics, for example disc size (mono κ_w = 0.65, stereo κ_w = 0.52) and cup-to-disc ratio (mono κ_w = 0.72, stereo κ_w = 0.62).

For expert observers in the evaluation of optic disc photographs for glaucoma likelihood, monoscopic optic disc photographs did not appear to represent a significant disadvantage compared to stereoscopic photographs.