Yo Sep Yoon1, Chang Uk Park1, Jun Ho Song1, Jung Kee Min2. 1. Department of Ophthalmology, Ulsan University Hospital, University of Ulsan, College of Medicine, 887 Bangeojinsunhwando-ro, Dong-gu, Ulsan, 44033, Republic of Korea. 2. Department of Ophthalmology, Ulsan University Hospital, University of Ulsan, College of Medicine, 887 Bangeojinsunhwando-ro, Dong-gu, Ulsan, 44033, Republic of Korea. jkmin@uuh.ulsan.kr.
Abstract
PURPOSE: To compare the retinal layer thickness values obtained using two swept-source optical coherence tomography (OCT) wide modes. METHODS: This study included fifty-four healthy eyes. Three-dimensional (3D) wide and 12 radial wide OCT scans were performed in each eye on the same day. Full retinal, retinal nerve fiber layer (RNFL), and ganglion cell-inner plexiform layer (GC-IPL) thicknesses with fully automated segmentation obtained in 3D wide mode, and with semi-automated correction and fully automated segmentation obtained in 12 radial wide scan. RESULTS: The intraclass correlation coefficients for full retinal thickness measurements obtained by the two protocols with fully automated segmentation were from 0.958 to 0.996, 0.754-0.918 for GC-IPL thickness measurements, and 0.013-0.727 for RNFL thickness measurements, in the nine ETDRS subfields. CONCLUSIONS: The full retinal thickness measurements determined using both fully automated segmentation algorithms were reliable and clinically acceptable. However, segmentation errors are shown when using the 12 radial wide scanning protocol with fully automated segmentation for measurement of RNFL and GC-IPL thickness.
PURPOSE: To compare the retinal layer thickness values obtained using two swept-source optical coherence tomography (OCT) wide modes. METHODS: This study included fifty-four healthy eyes. Three-dimensional (3D) wide and 12 radial wide OCT scans were performed in each eye on the same day. Full retinal, retinal nerve fiber layer (RNFL), and ganglion cell-inner plexiform layer (GC-IPL) thicknesses with fully automated segmentation obtained in 3D wide mode, and with semi-automated correction and fully automated segmentation obtained in 12 radial wide scan. RESULTS: The intraclass correlation coefficients for full retinal thickness measurements obtained by the two protocols with fully automated segmentation were from 0.958 to 0.996, 0.754-0.918 for GC-IPL thickness measurements, and 0.013-0.727 for RNFL thickness measurements, in the nine ETDRS subfields. CONCLUSIONS: The full retinal thickness measurements determined using both fully automated segmentation algorithms were reliable and clinically acceptable. However, segmentation errors are shown when using the 12 radial wide scanning protocol with fully automated segmentation for measurement of RNFL and GC-IPL thickness.