PURPOSE: Previously, we have shown that age-related macular degeneration (AMD) influences glaucoma classification with the ganglion cell complex (GCC) parameters of the RTVue-100 optical coherence tomograph (RTVue-OCT) in nonglaucomatous eyes. Now, we reevaluated the influence of AMD on GCC image segmentation and classification for glaucoma on the same eyes, using a new version of the software. METHODS: GCC images of nonglaucomatous eyes [30 healthy, 19 with early/intermediate AMD, 16 with subfoveal choroidal neovascularization (CNV), and 19 after intravitreal antiangiogenic treatment of CNV, CNV-anti-VEGF] were reanalyzed with software versions 6.3 (the currently available version) and 6.12 (a version not yet commercially released). RESULTS: Global loss volume (GLV) was significantly reduced with version 6.12 in all groups (P≤0.0416). Segmentation errors were seen in 2 versus 0 of the normal eyes (P=0.500), 8 versus 0 of the early/intermediate AMD eyes (P=0.0312), 16 versus 6 of the CNV eyes (P=0.0080), and 18 versus 3 of the CNV-anti-VEGF eyes (P=0.0004) with software versions 6.3 and 6.12, respectively. For focal loss volume the distribution of the classification results differed significantly between the software versions in the CNV and CNV-anti-VEGF groups (P=0.0312 and 0.0160, respectively). For both groups more eyes were classified as "within normal limits," and less as "outside normal limits" with software version 6.12 than with version 6.3. CONCLUSIONS: For nonglaucomatous AMD eyes the frequency of GCC segmentation errors was significantly reduced, GLV was significantly lower (more normal), and the classification for glaucoma was more correct with software version 6.12 than with version 6.3.
PURPOSE: Previously, we have shown that age-related macular degeneration (AMD) influences glaucoma classification with the ganglion cell complex (GCC) parameters of the RTVue-100 optical coherence tomograph (RTVue-OCT) in nonglaucomatous eyes. Now, we reevaluated the influence of AMD on GCC image segmentation and classification for glaucoma on the same eyes, using a new version of the software. METHODS: GCC images of nonglaucomatous eyes [30 healthy, 19 with early/intermediate AMD, 16 with subfoveal choroidal neovascularization (CNV), and 19 after intravitreal antiangiogenic treatment of CNV, CNV-anti-VEGF] were reanalyzed with software versions 6.3 (the currently available version) and 6.12 (a version not yet commercially released). RESULTS: Global loss volume (GLV) was significantly reduced with version 6.12 in all groups (P≤0.0416). Segmentation errors were seen in 2 versus 0 of the normal eyes (P=0.500), 8 versus 0 of the early/intermediate AMD eyes (P=0.0312), 16 versus 6 of the CNV eyes (P=0.0080), and 18 versus 3 of the CNV-anti-VEGF eyes (P=0.0004) with software versions 6.3 and 6.12, respectively. For focal loss volume the distribution of the classification results differed significantly between the software versions in the CNV and CNV-anti-VEGF groups (P=0.0312 and 0.0160, respectively). For both groups more eyes were classified as "within normal limits," and less as "outside normal limits" with software version 6.12 than with version 6.3. CONCLUSIONS: For nonglaucomatous AMD eyes the frequency of GCC segmentation errors was significantly reduced, GLV was significantly lower (more normal), and the classification for glaucoma was more correct with software version 6.12 than with version 6.3.