Comparison of retinal nerve fiber layer thickness measured by Cirrus HD and Stratus optical coherence tomography.

Abstract

To study the relationship between retinal nerve fiber layer (RNFL) thickness measured by time domain (Stratus) and spectral domain (Cirrus HD) optical coherence tomography (OCT). Cross-sectional study. Sixty healthy participants, 48 glaucoma suspects, and 55 subjects with glaucoma. Participants were imaged by a single trained operator using the Stratus OCT (fast RNFL scan mode) and Cirrus HD-OCT (optic disc cube mode) at the same visit. The RNFL thickness as measured by the Stratus OCT (fast RNFL scan mode) and Cirrus HD-OCT (optic disc cube mode) was compared (paired t test). The relationship between RNFL thickness measurements of the 2 OCTs were evaluated using a Pearson correlation analysis. The presence of abnormal classification, as determined by using an internal normative data base in each machine was compared (chi-square test). The sensitivity and specificity of normative classification of 2 OCT measurements were calculated. Average RNFL thickness as determined by the 2 OCT machines was correlated (r = 0.94; P<0.001), but was significantly different with the 2 machines (Stratus, 98.0 mum, standard deviation [SD] 18.0; Cirrus, 85.6 mum, SD 14.6; P<0.001). The Cirrus HD-OCT classified a significantly higher percentage of eyes as abnormal (Stratus, 12.9%; Cirrus, 23.3%; P<0.001) in average RNFL thickness. Cirrus OCT demonstrated higher sensitivity and specificity (63.6% and 100%) than Stratus OCT (40.0% and 96.7%) in normative classification of average RNFL thickness. There were significant differences in RNFL thickness and normative classification as determined by Stratus OCT and Cirrus HD-OCT despite an excellent correlation of RNFL thickness measurement. Overall sensitivity and specificity were higher with Cirrus OCT. These findings are particularly relevant when an individual undergoes longitudinal follow-up with different OCTs.