Evaluation of Macular Ganglion Cell Layer Measurements Obtained Using a New Retinal Layer Segmentation Algorithm in Glaucoma Patients

Abstract

Purpose: Using a novel segmentation algorithm, SD-OCT can now measure the thickness of the ganglion cell layer (GCL) independently from other retinal layers. The purpose of this study was to evaluate the ability of macular GCL thickness for glaucoma diagnosis and compare it to peripapillary retinal nerve fiber layer (RNFL) thickness.

Methods: A cross-sectional study was conducted including 278 eyes of 155 participants from the Diagnostic Innovations in Glaucoma Study. A total of 90 subjects had primary open angle glaucoma and 67 were healthy. Glaucoma was defined by the presence of repeatable abnormal standard automated perimetry (SAP) results and/or progressive glaucomatous optic disc change on masked grading of stereophotographs. All patients underwent spectral-domain optical coherence tomography imaging (Spectralis SD-OCT; Heidelberg Engineering, Heidelberg, Germany). Based on contrast differences between layers, a novel software algorithm allowed the SD-OCT to segment and measure macular GCL thickness independently from other macular layers. Areas under the receiver operating characteristic (ROC) curves were used to investigate the ability of each parameter in discriminating glaucomatous from healthy eyes. ROC curves were adjusted for age using a ROC regression model.<br />

Results: The mean±standard deviation (SD) age of participants was 60.1±16.8 years. The average SAP mean deviation in glaucoma eyes was -4.7 ± 5.7 dB. The global average GCL thickness in glaucomatous eyes was significantly lower than in healthy eyes (37.7±7.7 μm vs. 46.9±3.8 μm, respectively; P<0.001). Glaucomatous eyes also had thinner GCL thickness areas than healthy eyes in all quadrants. The average peripapillary RNFL thickness in glaucomatous eyes was 72.7±15.0 μm, which was significantly lower than the 100.4±9.1 μm in healthy eyes (P<0.001). There was strong correlation between global GCL thickness and RNFL thickness (R² = 0.83; P < 0.001). Peripapillary RNFL thickness performed better than global GCL thickness in discriminating healthy and glaucomatous eyes with ROC curve areas of 0.91 and 0.72, respectively (P = 0.002).

Conclusions: Measurements of macular GCL thickness showed strong relationship with peripapillary RNFL thickness. However, peripapillary RNFL thickness still performed significantly better in discriminating glaucomatous from healthy eyes.