Purchase this article with an account.
E. J. Sarver, R. A. Applegate, J. D. Marsack; Errors in Reflection Ring Corneal Topography Measurements Due to Skew Ray Surfaces. Invest. Ophthalmol. Vis. Sci. 2008;49(13):1045. doi: https://doi.org/.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
To determine the extent to which the theoretical principles detailed by Rand et al (Optom Vis Sci. 1997; 74(11):926-30) cause measurement errors in reflection ring corneal topographers.
Four surfaces were cut using a DAC 2X-ALM lens lathe. Surface 1 was an axis-symmetric asphere (apical radius = 7.5 mm, conicity = 0.5503). Surfaces 2 - 4 were aspheres plus an overlaid sinusoid. Surface 2 had peak to valley (PV) of 5 microns and surfaces 3 and 4 had PV of 10 microns. Surfaces 2 and 3 had the sinusoid added from the 3 mm zone to the periphery. Surface 4 had the sinusoid added from the center to the periphery. Each surface was measured three times with a Keratron reflection ring corneal topographer as well as a secondary reflection ring corneal topographer currently under development. We developed a program to read in the exams, compare the data with analytical values, and compute error statistics.
The mean, standard deviation (SD) and maximum (absolute) errors for the exams are given in Table 1. The mean and SD of the axial curvature errors were relatively independent of the surface. For the elevation errors, the SD of the asphere (over the entire surface) was about 1 micron. The addition of the sinusoid caused the elevation error SD to increase to around 50% of the PV value. The maximum elevation errors also increased and occurred in the periphery. Values from the second topographer were in agreement with these values from the Keratron exams.
As predicted, the addition of a sinusoid overlaid onto otherwise smooth, symmetric surfaces leads to an increase in elevation errors with SD around 50% of the PV of the sinusoid. If we replace the measured axial curvature values with theoretical values from the analytic models and perform an arc-step reconstruction (Optom Vis Sci. 1991; 68:960-5) to compute the elevation data per meridian, the elevation error SD values only decreases by about 10%. This shows the source of the increase in elevation error is due to the meridian-by-meridian processing typically used in commercial reflection ring corneal topographers.
This PDF is available to Subscribers Only