Abstract
Abstract: :
Purpose: To compare test–retest variability of neuroretinal rim area measurements in normal subjects and glaucoma patients by Retinal Thickness Analyzer (RTA) and Optical Coherence Tomography (OCT). Methods: One eye was randomly selected from each of 25 normal subjects and 31 glaucoma patients and scanned using OCT3 and RTA. Glaucoma diagnosis was based on SITA 24–2 visual field test repeatable defects. Eyes were imaged by two experienced observers over two visits. Time between visits ranged from 1 to 3 weeks. Repeatability for Rim Area measurements was calculated for observer one and two, and visits one and two by means of Bland & Altman plots and Intraclass Correlation Coefficient (ICC) for the sequences interobserver–intravisit (1 – 2, 1’– 2’), interobserver–intervisit (1’– 2, 1 – 2’) and intraobserver–intervisit (2 – 2’, 1–1’). OCT scans were acquired using the Fast Optic Disc scan protocol and analysed with the Optic Nerve Head protocol. No corrections were made to the automatic recognition of the disc borders. RTA scans were reviewed and image quality was graded by a masked observer (recognizability of disc margin). Four images were excluded from the analysis based on this selection. Each observer drew contour lines independently for each scan. Results: The ICC values for Rim Area measurements for the sequence (1 – 2, 1’– 2’), (1’– 2, 1 – 2’) and (2 – 2’, 1–1’) with the OCT were 0.907, 0.894 and 0.912 respectively. The mean differences between measures (95% limits of agreement) for the same sequence were 0.06 (–0.01 to 0.13), –0.03 (–0.10 to 0.05) and 0.06 (–0.01 to 0.13). OCT Disc Area ICC ranged from 0.70 to 0.79. For the RTA the ICC values for Rim Area measurements for the same sequence were 0.894, 0.899 and 0.879. Mean difference between measures were 0.03 (–0.04 to 0.10), –0.03 (–0.10 to 0.05) and –0.01 (–0.10 to 0.07). RTA Disc Area ICC ranged from 0.80 to 0.81. Conclusions: The RTA measurements show comparable reproducibility to OCT measurements. The RTA required a new contour line to be drawn on each scan, potentially adding variability. OCT contour line drawing is performed automatically by the software, but the variability in Disc Area calculation may affect the reproducibility of the Rim Area measurements. Image quality during acquisition may be a further source of variability on each scan as the automatic disc area detection will be impaired. As Rim Area is related to Disc Area, any variability of this measurement will induce variability on rim area calculations. The use of a fixed, transferable contour line may reduce variability on rim area measurement.