VIDEOKERATOGRAPHY DATABASE OF NORMAL HUMAN CORNEAS

Aim-To form a database of videokeratography patterns and quantitative indices describing normal human corneas using the absolute scale. Meth ods-Both eyes of 195 normal subjects were examined with a TMS-1 videok eratoscope. Videokeratographs were divided into 10 categories based on a classification scheme devised from the absolute scale and analysed with 10 quantitative indices devised to describe phenotypic features o f keratoconus videokeratographs. Correlations were sought between vide okeratograph patterns and quantitative indices. Additionally, data wer e analysed for differences in age, sex, and ethnicity. Results-For sym metric videokeratography patterns, analysis in the absolute scale was similar to a previous study done in the normalised scale. In the asymm etric categories, analyses differed markedly. Using the absolute scale and our classification scheme more variation in normal videokeratogra phy patterns could be appreciated. There was good correlation between quantitative indices and videokeratography patterns. Neither videokera tography patterns nor indices differed significantly between sex, ethn ic groups, or age using two way analysis of variance. Conclusions-Patt ern analysis of videokeratographs in the absolute scale using, a stand ard classification scheme, may be more useful in trying to determine w hether a cornea is normal or represents subtle early disease than anal ysis in the normalised scale. Quantitative indices could remove the su bjectivity from the decision making process thus facilitating universa l reproducibility of videokeratography data interpretation.