Reproducibility of OCT/SLO Measurements in Healthy Eyes

• Published in: European journal of ophthalmology 2010-05, Vol.20 (3), p.552-558
• Main Authors: Labiris, Georgios, Giarmoukakis, Athanassios, Katsanos, Andreas, Gkika, Maria G., Fanariotis, Michael, Pavlidou, Efi, Kozobolis, Vassilios P.
• Format: Article
• Language: English
• Subjects: Cross-Sectional Studies; Female; Fourier Analysis; Humans; Imaging, Three-Dimensional; Male; Middle Aged; Nerve Fibers; Observer Variation; Ophthalmoscopes; Optic Disk - anatomy & histology; Prospective Studies; Reference Values; Reproducibility of Results; Retinal Ganglion Cells - cytology; Tomography, Optical Coherence
• ISSN: 1120-6721; 1724-6016
• DOI: 10.1177/112067211002000303

Purpose. To determine the reproducibility of optical coherence tomography/scanning laser ophthalmoscope (OCT/SLO) measurements within and between observers and within and between sessions in healthy eyes. Methods. Forty eyes of 40 healthy volunteers recruited from the outpatient service of the Department of Ophthalmology at the University Hospital of Alexandroupolis, Greece, were included. Retinal thickness scans were performed 3 times by two experienced operators within each of 3 sessions, which took place within a month, using the spectral OCT/SLO. The OCT/SLO provides retinal nerve fiber layer (RNFL) thickness and volume measurements and a 3D tomography analysis of optic nerve head. Statistical analysis was conducted using Bland-Altman plots, coefficient of variation (CoV), and intra-class correlation coefficient (ICC). Results. AS shown by the CoV values, RNFL parameters were generally found to have high reproducibility (CoV ranging from 4.3% to 10.9% for both operators), whereas optic disc parameters had scores reaching up to 22.5%. On the other hand, as shown by the ICC values, intraobserver and interobserver reproducibility for the OCT/SLO was high for optic disc parameters but low for RNFL thickness and volume measurements. Interobserver reproducibility was found to be highest for the maximum cup depth parameter and lowest for the superior volume. Bland-Altman plots indicate that bias is minimal for practically all parameters and scatterplot homoscedasticity is high. Conclusions. According to our data, SLO/OCT generally has a good reproducibility profile, as differences between the two operators were small. However, additional studies on the reproducibility of measurements with the OCT/SLO are needed.