Abstract
Abstract: :
Purpose: The study was designed to compare the subjective quality of stereoscopic viewing and quantitative mapping of retinal thickening areas in diabetic macular edema by evaluation of stereo fundus photographs using polarizing shutter goggles as compared to an autostereoscopic display system. Methods: 32 stereo fundus photographs of patients with clinically significant diabetic macular edema were evaluated by 2 methods:polarized liquid crystal shutter goggles (Stereographics Crystal Eye CE–3 goggles with Oxygen GVX1 videocard and a Dell ultrascan P1110 21" screen) and an autostereoscopic screen (Dimension Technology Inc.2015 XCS virtual window 15" screen); both at resolution 1024 x 768. 2 observers viewed the set of images independently. Each observer assessed the quality of stereoscopic perception across the 2 modalities giving an image score from 1–10.Both observers used the 2 methods to plot the areas of retinal thickening detected stereoscopically. The measurments were then repeated to compare the reproducibility of both methods. The plotted maps from the 2 observers were aligned and calculated using MATLAB software. The areas were evaluated for overlapping and agreement between the observers and the results were compared to the slit lamp biomicroscopy plotted areas. Results There was a definite, subjective improvement in stereo–viewing agreed by the 2 observers, both scoring consistently higher with the autostereoscopic display system.(p–values 0.0021 & 0.0074 Wilcoxon signed rank test). The inter–observer agreement of retinal thickening plotted areas was assessed across the 2 modalities; the observers' results were more reproducible with the polarised goggles as compared to the autostereoscope screen (0.64:0.44). When the intra–observer reproducibilty comparing both modalities was tested, observer 1 showed consistent results across both modalities (0.82) yet with observer 2 the level of consistency measured only (0.45). The repeatability results were then assessed,it showed high reproducibility in both modalities( scoring slightly higher with the goggles as compared to the autostereoscope screen.)(0.83:0.79). Conclusions: This is the first published study to use the autostereoscopic display screen to evaluate macular edema. Despite subjective inprovement in stereo viewing, and within the systems tested, there was no objective advantage with the autostereoscopic screen compared to the polarised goggles.
Keywords: diabetes • imaging/image analysis: non–clinical