Average 3-Dimensional Models for the Comparison of Orbscan II and Pentacam Pachymetry Maps in Normal Corneas

Abstract

To assess the reliability of Orbscan (Bausch & Lomb, Salt Lake City, UT) and Pentacam (Oculus, Wetzlar, Germany) central corneal thickness (CCT) and peripheral corneal thickness (PCT) measurements based on 2 methodologies. Evaluation of a diagnostic technology. Thirty healthy volunteers were recruited prospectively at the Department of Ophthalmology of the Hôtel-Dieu Hospital, Paris, France. Central corneal thickness and PCT were assessed, using ultrasound pachymetry (USP) as the gold standard. Two methodologies were used: (1) the traditional analysis of pachymetry data from 1 central and 8 peripheral reference positions on the cornea, and (2) a 3-dimensional (3-D) analysis based on average corneal pachymetry maps constructed for each system (Orbscan, Pentacam, and USP), each operator (operators 1 and 2), and each visit (visits A and B). Repeatability, intersystem reproducibility, interoperator reproducibility, reproducibility over time, and accuracy of Orbscan and Pentacam CCT and PCT measurements. Distribution and statistical significance of the differences between 3-D average maps. Repeatability (Orbscan intraclass correlation coefficients [ICCs], 0.967-0.992; Pentacam ICCs, 0.986-0.997), interoperator reproducibility, and reproducibility over time (ICCs, 0.976-0.997) were excellent to almost perfect for both systems. Intersystem agreement was almost perfect for CCT (ICC, 0.980), but less strong for PCT (ICCs, 0.928-0.979). Despite a good to excellent agreement between the optical systems and USP (ICCs, 0.608-0.958), USP CCT readings were thicker (mean difference, up to 15.2 microm; P<0.05), and USP PCT readings were thinner (P<0.05). Orbscan and Pentacam average maps allowed comprehensive interpretation of differences between populations according to the magnitude, distribution, and statistical significance, minimizing the risk of giving excessive weight to few data measured at specific locations on the cornea. Both methodologies showed that Orbscan and Pentacam CCT readings are interchangeable, whereas caution should be used for PCT readings. Interchangeability with USP measurements also was shown to be limited. The high repeatability, interoperator reproducibility, reproducibility over time, the extent of the information generated by a single capture, and the noncontact nature of the Orbscan and Pentacam all suggest that optical systems eventually may replace USP as the gold standard for corneal pachymetry.