Orbscan II Machine-to-Machine Repeatability

Abstract

Abstract: : Purpose: Orbscan is an anterior segment topography instrument that uses scanning slit-beams to measure the shape of both corneal surfaces. Orbscan II supplements back-scattered slit data with specular reflective placido data to enhance the accuracy of front surface measurement. Because the cornea is a non-isotropic, non-uniform scatterer, no definitive test objects have ever been successfully made that could test Orbscan posterior surface accuracy. In addition, to test Orbscan front surface accuracy, a different algorithm is used than employed on eyes. The most honest test currently possible is repeatability of human eye exams. This study is unique in that it reports the machine-to-machine repeatability. Methods: Every Orbscan manufactured is tested for repeatability using human eyes. Four consecutive Orbscan exams are typically taken of the same eye, with all raw images saved. Eight eyes of five individuals have been used to final test the last 621 Orbscan II units (serial numbers 965 through 1596). A total of 2461 exams were reprocessed with the latest Orbscan release (version 3.12). Orbscan generated maps for corneal anterior and posterior height, elevation with respect to various reference surfaces, surface curvatures, derived optical powers, and corneal thickness were then sampled in a 0.2 mm or finer mesh grid, and the results analyzed using Mathematica. Single-instrument repeatability was measured as the standard deviation of the four-exam sets. Machine-to-machine repeatability was measured as the standard deviation over all instruments for the same eye. These measurements spanned a period of nearly two years. Because of significant diurnal variation, temporal trends were first removed before the machine-to-machine repeatability was assessed. All measurements were limited to the 2 mm diameter zone about the ocular fixation-reflex axis. Results: Single-machine repeatabilities, averaged over all instruments and eyes, were 0.8 and 2.8 µ m for anterior and posterior corneal elevation from spherical reference surfaces, and 3.0 µ m for corneal thickness. Machine-to-machine repeatabilities, averaged over all instruments and eyes, were 1.5 and 4.0 µ m for anterior and posterior corneal elevation, and 6.5 µ m for corneal thickness. Only the latter showed significant diurnal variation: 0.8 µ m / hour decrease in thickness over the daytime measurement period, which avoided the first hours after awakening when corneal thickness changes rapidly. Conclusion: This is the first study reporting Orbscan II machine-to-machine repeatability on human eyes. It is consistent with previous measurements of single-machine repeatability. Machine-to-machine repeatability is about half as good as single-machine repeatability.

Keywords: 370 cornea: basic science • 599 topography