Background and aim of the work: retinal thickness values obtained by automated analysis with new spectral-domain optical coherence tomography (OCT) devices exceed those measured by old time-domain OCTs. Aim of the present work is to assess reproducibility and comparability of manual measurements performed on both time-domain and spectral-domain OCT scans.
Methods: 6 eyes were scanned using Stratus OCT3 and Cirrus hd-OCT. Raw exported images were analyzed by ordinary computer software; multiple manual measurements of retinal thickness were taken at different eccentricities. Stratus Retinal Thickness (SRT) was measured from Internal Limiting Membrane (ILM) to the photoreceptors Internal-Outer Segment interface (IS/OS), while two series of measurements were performed in Cirrus images: CRT = Cirrus-RT (from ILM to the Retinal Pigment Epitelium, RPE) and cCRT = corrected-CRT (from ILM to IS/OS). Measurements were repeated twice in two eyes and reproducibility was assessed by Intra-Class correlation (ICC) and Coefficient of Variation (CV). Bland-Altman plots, paired t-test and ICC were used for comparative analysis between Stratus and Cirrus measurements.
Results: Mean SRT, CRT and cCRT values +/- SD were respectively 244.75 +/- 34.78 microm, 275.78 +/- 34.36 microm and 244.95 +/- 33.78 microm. Paired t-test resulted in p<0.0001 comparing SRT and CRT series, versus p=0.6544 between SRT and cCRT series. ICC was 0.65 between SRT and CRT and 0,94 between SRT-cCRT. Reproducibility of the measurements was excellent (CV=2,13% for Stratus and 1,62% for Cirrus; ICC=0,994 for both devices).
Discussion: while a systematic error affects comparison of Stratus and Cirrus macular thickness maps, manual linear measurements result interchangeable, hence allowing comparison of images acquired with either OCT systems. (www.actabiomedica.it).