ABSTRACT
AIM: To evaluate the intraobserver and interobserver reproducibility of macular retinal ganglion cell-inner plexiform layer (GC-IPL) thickness measurement by automated detection on Optical Coherence Tomography (OCT) images in normal, hypertensive (ocular hypertensive (OHT)) and glaucomatous eyes. METHODS: A total of 138 eyes were enrolled in three groups: 69 normal, 35 OHT and 34 primary open-angle glaucoma eyes. All patients underwent a complete ocular examination, 24-2 automated perimetry, biometry and pachymetry. Macular imaging was performed in each eye using the Cirrus HD-OCT 4000 with software V.6.0. (Carl Zeiss Meditec, Dublin, California, USA) three times on the same day by each of two observers, and the GC analysis (GCA) algorithm provided parameters expressed as average, minimum and six sectoral GC-IPL thicknesses. Reproducibility was assessed by intraclass correlation coefficient (ICC), coefficient of variation (CV) and test-retest variability (TRTV) calculated as 1.96 times the SD. RESULTS: Mean GC-IPL thickness was 82.27 ± 7.37 µm, 76.84 ± 7.01 µm and 66.16 ± 11.16 µm in normal, OHT and glaucoma groups, respectively. GC-IPL thickness was significantly lower in glaucomatous eyes than in normal and OHT eyes (p<0.0001 for all parameters). In all groups, ICC ranged from 96.4 to 99.9% and 92.5 to 99.8%, CV ranged from 0.41 to 2.24% and 0.55 to 1.67%, and TRTV ranged from 0.61 to 2.64 µm and 0.83 to 2.22 µm for intraobserver and interobserver reproducibility, respectively. CONCLUSIONS: To the best of our knowledge, this is the first study of GCA algorithm reproducibility in normal, OHT and glaucomatous eyes. The reproducibility of GC-IPL thickness measurements using the Cirrus HD-OCT GCA algorithm was found to be highly satisfactory. GC-IPL thickness may be a promising new OCT parameter for analysis of ganglion cell damage in glaucoma.
