ABSTRACT.Purpose: To analyse tomographic changes in eyes classified as 'normal', 'keratoconus-suspect' and 'clinically manifest keratoconus' based on the established KISA% definition of Rabinowitz and Rasheed and to develop the category of 'subclinical keratoconus eyes' to expand the classification into a 'subtopographic' range. Methods: Tomographic and topographic analyses of 670 eyes performed with a rotating Scheimpflug imaging system (Pentacam â , Oculus Inc., Wetzlar, Germany) were retrospectively analysed. Based on the KISA% keratoconus classification system, eyes were assigned to a 'normal', 'keratoconus-suspect' or 'manifest keratoconus' group. In addition, a new group of 'subclinical keratoconus eyes' was analysed, comprising unsuspicious fellow eyes of patients with keratoconus. T-tests, Wilcoxon rank-sum test, receiver operating characteristics (ROC) and robust regression analyses were performed to correlate tomographic parameters with the increasing KISA% index. Results: KISA%-grouped keratoconus eyes showed robust tomographic changes. By adding the subclinical group, although the concurrent topography was unchanged, we were able to demonstrate statistically significant changes for almost all tomographic parameters (parameters with highest sensitivity/specificity: ART_max, [0.69/0.69], BAD_D [0.66/0.66]). The highest coefficient of determination (R 2 ) with the KISA% index was demonstrated for Ele_f_max (R 2 = 0.70), Ele_f_TP (R 2 = 0.69), Ele_b_TP (R 2 = 0.69) and BAD_D (R 2 = 0.68). Conclusion: We recommend the use of the expanded KISA% index (eKISA% index) as the basis for the definition of keratoconus and normal groups in future keratoconus research projects.
