Purchase this article with an account.
Jamie Lea Schaefer, Gina Martorana, Zachary L Lukowski, L. Jay Katz, Joel S Schuman, Andy Leoncavallo, Anthony Greer, Jonathan Shuster, Mae O Gordon, Mark B Sherwood, ; Comparison Among Experts and Trainees for Detection of Glaucomatous Disc Change Using the Computerized MatchedFlicker® Software Verses Stereo Disc Viewing. Invest. Ophthalmol. Vis. Sci. 2015;56(7 ):1024.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Prospective, masked, randomized study to compare the accuracy of the computerized MathedFlicker® software verses manual viewing of the same stereoscopic disc photos to detect glaucoma progression among glaucoma experts and trainees.
Three glaucoma experts and three ophthalmology trainees each separately evaluated stereoscopic pairs of disc photographs of 100 eyes taken at two time points. Fifty pairs were determined by the OHTS Optic Disc Reading Center and the OHTS end-point committee to show glaucoma progression while the other 50 photos were taken within a few minutes of each other, and thus had no change. Twenty of the 50 eyes in each group were randomly selected to have repeat pairs in order to determine intra- and inter-observer variability in detecting progression, making a total of 140 image pairs (280 photos) examined. Each observer analyzed the photographs using two different methods to judge for glaucomatous disc progression; a handheld stereo-viewer and computerized MatchedFlicker® which rapidly alternates the two different timed images on a computer screen to simulate a sensation of movement if there is structural alteration. The method of examining the disc pairs was divided into alternating blocks of 70 paired images and the order of starting randomized between the MatchedFlicker® and stereo-viewing. The ONH photos were assessed for disc progression and the total evaluation time for each method for each participant was measured.
The expert observers correctly interpreted 79.0% of the slides when analyzing with the stereo-viewer and 86.7% with the MatchedFlicker® software. The trainees identified correctly 76.0% with stereo-viewer and 87.6% with MatchedFlicker®. Evaluator viewing time of the 140 images averaged 74.7±25.6 and 79.4±11.4 minutes with conventional stereophoto viewing verses 56.0±19.8 and 58.1±13.2 minutes with MatchedFlicker® program for the experts and trainees respectively. Intra-observer variability showed 84.2% consistency for both experts and trainees when using the stereo-viewer, and respectively 89.2% and 85.8% for MatchedFlicker®.
The detection of glaucoma progression in ONH photographs was quicker and generally more accurate with the MatchedFlicker® software when compared to stereo-viewers for both experts and trainees.
This PDF is available to Subscribers Only