Evaluating the reproducibility of topography systems on spherical surfaces.

Newly devised software was used to compare the ability of the Topographic Modeling System-1 (Computed Anatomy, New York, NY) and the Corneal Analysis System (EyeSys Laboratories, Houston, Tex) to reproduce power measurements on spherical surfaces. Reproducibility results were compared for spheres of 40.00, 42.50, and 44.00 diopters. The program determines the absolute difference in corneal power at defined keratoscope positions for paired examinations of the same eye. Four examinations of each sphere were obtained with each instrument. Individual points were sampled at specific keratoscope locations at 30 degrees-semimeridional intervals. The program compared variability of measurements at four defined ranges of distance from the vertex normal: within 0.60 mm, 0.61 to 1.5 mm, 1.51 to 2.5 mm, and 2.5 mm or greater. The Corneal Analysis System showed significantly greater variability of readings obtained within 0.60 mm of the vertex normal for all three spheres (P = .001 by Duncan's multiple comparison procedure), whereas the Topographic Modeling System-1 showed equally consistent readings within 0.60 mm as it did between 0.61 and 1.5 mm from the vertex normal.