Abstract
Purpose :
The many highly correlated parameters provided by CIRRUS HD-OCT for the objective assessment of structural glaucomatous damage may complicate the clinical interpretation of results. Mwanza et al [1] previously developed a combined OCT structural glaucoma index based on exploratory factor analysis and multivariate logistic regression model (UNC OCT Index) to aid in the diagnosis of early glaucoma that was more effective than any individual parameter. In this retrospective clinical study, we determined the measurement repeatability and reproducibility of CIRRUS HD-OCT for the combined index, as well as for the single parameter input values, in a cohort of glaucomatous eyes.
Methods :
The study was performed in 2 phases (inter-operator and inter-device) using CIRRUS™ HD-OCT 4000 and HFA™ II-i devices (ZEISS, Dublin, CA). A total of 28 eyes from 28 subjects with early to advanced glaucoma were enrolled. Of the included subjects, the mean (SD) age was 73.5 (10.5) years, ranging from 49 to 88 years. The mean (SD) visual field mean deviation was -6.66 dB (5.54), ranging from -21.5 to 1.2 dB. For the inter-device phase, each subject was imaged by a single operator on three devices. For the inter-operator phase, each subject was imaged on a single device by three different operators. Three pairs of Optic Disc and Macular Cube 200x200 scans were obtained per operator or per device for a total of 18 scans per phase. Sixteen standard parameters (rim area, cup-to-disc area ratio (CDR), vertical cup-to-disc diameter ratio (vCDR), average and quadrant RNFL thicknesses, average, minimum, and sectoral GCIPL thickness) from paired scans were used to calculate the UNC OCT Index [1]. A random effects ANOVA model was used to estimate the components of variability for each parameter.
Results :
Repeatability SD is the square root of the worst-case random variance component. Reproducibility includes variations due to random error, operator, device, subject-device interaction, and subject-operator interaction. Reproducibility SD is the square root of the sum of all contributions to variance except subject variance. All coefficient of variation (CV) values were < 4.5%. See Table 1.
Conclusions :
The combined diagnostic index shows good repeatability and reproducibility, comparable to the single input CIRRUS HD-OCT parameters, in a sample glaucoma population. [1] Mwanza et al., 2013. IOVS, 54(13).
This is an abstract that was submitted for the 2017 ARVO Annual Meeting, held in Baltimore, MD, May 7-11, 2017.