The relationship between corneal topography and visual performance

Computer-aided videokeratoscopes have become the instrument of choice to investigate corneal topography because of their ability to provide a comprehensive description of the corneal shape. The work reported in this thesis is primarily related to the analysis of corneal topography in groups with wide ranging corneal shapes and the relationship between corneal shape and visual performance. The performance of different types of videokeratoscopes in measuring test surfaces and real eyes was initially evaluated. Three placido-disk videokeratoscopes (Keratron, Medmont and TMS) and the PAR-CTS videokeratoscope which uses the raster-stereogrammetry technique were evaluated. The accuracy and precision performance of these instruments in measuring the central 7 .0 mm horizontal chord of a spherical, an aspheric, a multicurve and three bicurve (5.0, 6.5, 8.5) test surfaces was compared with Talysurf readings. The results of the study indicated that the Keratron and Medmont videokeratoscopes were more accurate and precise than the TMS and PAR-CTS videokeratoscopes, especially in measuring the regular test surfaces (sphere and asphere). However, videokeratoscopic measurements for the irregular test surfaces were less accurate and less precise for all instruments. Videokeratoscopic measurements of real eyes are expected to be less repeatable than test surfaces due to factors such as the quality of pre-corneal tear film and small eye movements. The precision of videokeratoscopic measurements for real eyes was investigated for 54 subjects over two visits. The recruited subjects with normal and abnormal corneal shapes were classified into the control, corneal graft, keratoconus, orthokeratology and refractive surgery groups for analysis. The refractive surgery group of subjects were further divided into those who had undergone the radial keratotomy (RK), photorefractive keratectomy (PRK) and laser in-situ keratomileusis (LASIK) procedures. Precision of the corneal topography maps were evaluated for the 4 mm and 7mm diameters, which are approximations of photopic and mesopic pupil sizes. In general, measurements taken by the Keratron and Medmont videokeratoscopes of real eyes were found to have good precision. However, the precision of the elevation data (for the horizontal meridians of the standard deviation corneal maps) was found to become increasingly poorer away from the centre of the map. The worst repeatability (maximum error= 3.9 μm) occurred at a point which was 3.4 mm from the centre of the standard deviation map. In comparison, elevation data of spherical and aspheric test surfaces were also less repeatable at points which were further away from the centre of the maps and the precision of videokeratoscopic measurements of these test surfaces were better than real eyes, as anticipated. Corneal topography maps of subjects were fitted using Zernike polynomials to investigate the defining topographical characteristics of normal and abnormal corneal shapes. In addition, a root mean square (RMS) index of corneal topography was also derived to provide an overall measure of higher-order corneal aberrations. Topographical characteristics of 65 subjects with a range of corneal shapes were evaluated by fitting the elevation data of corneal topography maps using 30 terms of the Zernike polynomials for 4 mm and 7mm diameters. The values of the Zernike coefficients for control subjects were significantly different to subjects with abnormal corneas, especially for the corneal graft, keratoconus and RK subjects. Corneal shapes of control subjects were found to be adequately described using 11 Zernike terms or less. In comparison, topography maps of subjects with highly distorted corneas were found to require many higher-order Zernike terms such as primary coma, trefoil, tetrafoil, pentafoil and sextafoil-like shapes for their descriptions. The results of the analysis also showed that corneal shapes of orthokeratology and PRK/LASIK subjects were similar to normal subjects, especially in the central corneal regions. Six vision tests based on the high and low contrast Bailey-Lovie and PelliRobson charts were used to evaluate the visual performance of the 65 subjects under glare, photopic and mesopic conditions. Test scores of the subjects were found to be repeatable for the different vision tests, even though visual performance of subjects with abnormal corneas was observed to be more variable than control subjects. Visual performance of subjects with abnormal corneal shapes was found to be significantly worse than the control subjects, especially for the low contrast Bailey-Lovie and Pelli-Robson letter charts. This finding was consistent with previous studies where visual performance of corneal graft, keratoconus and refractive surgery subjects for the low contrast vision tests has been found to be reduced more than their high contrast visual acuity. In addition, differences in test scores of subjects for the high and low contrast Bailey-Lovie vision tests were also investigated. It was found that evaluating test score differences for these vision tests may be a useful method of assessing visual performance of subjects with abnormal corneas. Using a ray tracing technique, RMS values (representing the total residual refractive powers of the corneas) were derived from corneal topography maps to determine the relationship between corneal topography and visual performance of subjects with a range of corneal shapes. It was found that visual performance of the subjects for the six vision tests was significantly correlated to RMS, especially for the low contrast Bailey-Lovie vision test performed without glare (r = 0.60). Predictions of visual performance were found to be even better when logarithm of the RMS values was used for the correlation, and test scores of subjects with corneal opacities were excluded from the data analysis. The correlation value for the low contrast Bailey-Lovie vision test ascertained under these conditions was found to have improved (r = 0.83). Standard errors of estimate analysis further suggested that the RMS of subjects with abnormal corneas could be used to predict their visual performance for high and low contrast letter charts. RMS values are directly related to the higher-order aberrations of the corneas and can potentially be used as an index of overall higher-order aberrations.