Evaluating data acquisition and smoothing functions of currently available videokeratoscopes

Purpose: To compare the accuracy of computerized videokeratography systems using identical, calibrated rest objects. Setting: Lions Eye Institute, Albany, New York. Methods: We evaluated the accuracy and smoothing of raw data acquisition (axial solution) of seven commercially available videokeratoscopes: Alcon EyeMap, Computed Anatomy TMS, EyeSys GAS, Humphrey MasterVue, Topcon CM-1000, Optikon Keratron, and TechnoMed C-Scan. We used six calibrated test objects to simulate clinical settings: spherical, spherocylindrical, simulated myopic ablation, hyperopic ablation, and a simulated central island. Results: None of the systems accurately imaged all objects. Although all systems imaged spherical objects with reasonable accuracy, errors greater than 4.0 diopters (D) frequently occurred in the central 6.0 mm optical zone (maximum error 10.0 D). Sources of error included excessive raw data smoothing, inability to read large transitions, loss of accuracy in the periphery, and poor central coverage. Conclusion: The clinician should be aware of the potential limitations of corneal topography when making clinical decisions.