Effects of Changing Operators and Instruments on Time-Domain and Spectral-Domain OCT Measurements of Retinal Nerve Fiber Layer Thickness

Abstract

To determine the amount of interoperator and interinstrument variability introduced into retinal nerve fiber layer (RNFL) thickness measured with Stratus OCT and Cirrus HD-OCT (Carl Zeiss Meditec, Dublin, CA) in normal and glaucomatous eyes. Thirty-seven eyes from 37 subjects (20 normal and 17 with glaucoma) were included in the study. Each eye was scanned with one Stratus OCT and one Cirrus OCT by two operators. The eyes were then scanned with a different Stratus OCT and Cirrus OCT by one of the operators. Average, quadrant, and clock-hour RNFL measurements were compared between operators and between instruments after determining the proportion of variance due to subjects, operators, and instruments by means of variance component analysis. Significant differences were observed between Stratus OCT instruments, but not between Cirrus OCT instruments, in average (P = .01), inferior (P = .006), and nasal (P < .001) quadrant RNFL thickness in glaucomatous eyes. Interoperator differences in RNFL thickness measured with both types of instruments were non-significant. The interoperator variability ranged from 0% to 1% on Stratus OCT and was 0% on Cirrus OCT, whereas the interinstrument variability ranged from 0% to 3% for Stratus OCT and from 0% to 1% for Cirrus OCT. The interoperator variability of RNFL thickness measured with Stratus OCT and Cirrus OCT, as well as the interinstrument variability of measurements obtained with Cirrus OCT, are non-significant. However, there is significant variability between RNFL measurements performed on two different Stratus OCT instruments.