To determine the repeatability of retinal nerve fiber layer (RNFL) measurement from 2 spectral domain optical coherence tomography (OCT) instruments and assess the level of agreement. A secondary objective was to assess if pupillary dilatation affected the repeatability of RNFL measurement.
Prospective, observational study.
Fifty normal participants were recruited after being screened for exclusion criteria, which included preexisting glaucoma and significant cataract.
Peripapillary RNFL thickness was measured from either eye of the 50 participants using both the Cirrus and Spectralis OCTs. Three scans were performed consecutively for each instrument and the process was repeated after pupillary dilatation.
Test-retest variability, coefficient of variation, and intraclass correlation coefficient (ICC) for repeatability of RNFL measurement. ICC and Bland-Altman plots for agreement between both devices.
Test-retest variability for the Cirrus and Spectralis OCTs were 4.89 μm and 4.95 μm, respectively. RNFL measurements were less repeatable in the 4 quadrants, ranging from 5.02 to 15.00 μm. The temporal quadrant was the most repeatable quadrant for the Cirrus OCT with a variability of 7.81 μm, whereas the nasal quadrant was the most repeatable for Spectralis OCT with 5.02 μm. Test-retest variability was poorer for the Cirrus OCT in the nasal, inferior, and superior quadrants. Pupillary dilatation increased repeatability of RNFL measurement for Cirrus OCT, particularly in the superior quadrant, whereas not affecting measurements from Spectralis OCT. ICC values for agreement between both instruments were good. For average RNFL as well as RNFL in the superior, inferior, and temporal quadrants, ICC ranged from 0.859 to 0.908. The nasal quadrant had a lower ICC of 0.663. Bland-Altman plots showed that the Spectralis OCT provided the thicker RNFL measurement in general, except in the nasal quadrant when RNFL was thin.
The repeatability of RNFL thickness measurement in normal participants was excellent for both the Cirrus and Spectralis OCTs. Agreement of RNFL measurement between both the devices was generally good, with the exception of the nasal quadrant in which a linear relationship exists. Pupillary dilatation improved the repeatability of RNFL measurement for Cirrus while having minimal influence on Spectralis OCT. More studies will be required to ascertain the relationships of RNFL measurement between the different spectral domain OCT instruments in normal and glaucomatous patients.
Department of Ophthalmology, Tan Tock Seng Hospital, Singapore
Disclosure: The authors declare no conflict of interest.
Reprints: Benjamin B. Tan, MBBS, Department of Ophthalmology, Tan Tock Seng Hospital, Singapore (e-mail: email@example.com).
Received July 30, 2010
Accepted November 14, 2010