Increasing scans per examination improves the reproducibility on retinal nerve fiber layer measurements by optical coherence tomography

Purpose. To investigate whether increasing the number of scans per examination would improve the reproducibility of retinal nerve fiber layer (RNFL) measurements by optical coherence tomography (OCT) in healthy and glaucomatous eyes. Methods. A total of 28 eyes of 28 healthy subjects and 20 eyes of 20 patients with primary chronic open-angle glaucoma were recruited. OCT RNFL assessments were obtained in one randomly selected eye. Each subject underwent seven scans on five separate occasions within a 1-month period. Results. In the healthy group, the intervisit mean coefficient of variations (COV's) for the average RNFL thickness (100 sampling points) and the quadrant RNFL measurements (25 sampling points) were similar in a series of three, five, and seven scans per examination (p > 0.05), but COV's for the two clock-hour segment measurements (seven sampling points, superotemporal and inferotemporal) were significantly lower in a series of five and seven scans per examination than that of three scans per examinations (p < 0.01). in the glaucoma groups, COV's for the two quadrant (superior and inferior) and five clock-hour segment RNFL measurements (superior, superotemporal, temporal, inferotemporal, and superonasal) were significantly lower in a series of five and seven scans per examination than that of three scans per examination (p < 0.01). There was no statistical difference in the COV's between a series of five scans and that of seven scans per examination in all OCT RNFL thickness parameters in healthy and glaucomatous eyes (p > 0.05). Similar results were observed with intravisit COV comparisons. Conclusion. Increasing the number of scans per examination may improve the reproducibility of quadrant and clock-hour segment OCT RNFL measurements in healthy and glaucomatous eyes. A series of five scans per examination is recommended for OCT RNFL measurements.