Comparison of Reproducibility and Repeatability of Corneal Topography Measured by ATLAS 500 and ATLAS 9000 in Average Patient Population

Pedro Tañá-Rivero, Paz Orts-Vila, Pedro Tañá-Sanz, Santiago Tañá-Sanz, Maria Ramos, Cristina Tello-Elordi Department of Ophthalmology, Oftalvist, Alicante, SpainCorrespondence: Pedro Tañá-Rivero, Department of Ophthalmology, Oftalvist, C/Angel Lozano 11, Alicante, 03001, Spain, Tel +34 965 14 15 00, Email [email protected]: To compare repeatability, reproducibility, and agreement of the Placido-based ATLAS 500 topographer (straight and swivel-out positions) with its predecessor, ATLAS 9000.Design: Prospective, single-center, post-market observational study.Methods: In this prospective, single-center post-market study involving 100 eyes with various corneal profiles, two operators obtained randomized, repeated measurements using the ATLAS 500 and ATLAS 9000 topographers. Parameters included corneal power, astigmatism, white-to-white distance, meridian powers, aberrations, and Placido coverage. Primary outcomes were repeatability and reproducibility; secondary analyses assessed coefficient of variation (CV), inter-device agreement, and keratoconus classification (Cohen’s κ).Results: The ATLAS 500 showed equal or superior repeatability and reproducibility compared to ATLAS 9000 across most parameters, with lower standard deviations (SDs) for central power (K1: 0.025 vs. 0.062 D; K2: 0.023 vs. 0.058 D) and astigmatism (0.015 vs. 0.018 D). The only exception was white-to-white distance, where ATLAS 9000 had lower variability (SD 0.032 mm vs. 0.066 mm), albeit with more missing data (41 vs. 10). The ATLAS 500 yielded lower coefficients of variation for corneal power and astigmatism (e.g. CV 0.058% vs. 0.146% for K1). Bland–Altman analysis showed minimal bias for K1/K2, with narrower limits in intra-device comparisons. Inter-device variability was greater for astigmatism and WTW, with ATLAS 500 measuring smaller WTW values. The swivel-out configuration increased Placido coverage by approximately 4%. Keratoconus classification showed moderate agreement (κ = 0.50) between devices.Conclusion: ATLAS 500 showed measurement precision comparable to or better than ATLAS 9000, with strong intra-device repeatability and broader corneal coverage in the swivel-out mode, confirming its suitability for routine corneal topography and refractive assessment.Keywords: corneal topography, placido disk, repeatability, reproducibility, bland–altman agreement, keratoconus, ATLAS 500, ATLAS 9000