Intrasession repeatability and intersession reproducibility of peripapillary OCTA vessel parameters in non-glaucomatous and glaucomatous eyes.

Objective To compare intrasession repeatability versus intersession reproducibility of the peripapillary vessel parameters using optical microangiography-based optical coherence tomography angiography (OCTA) in non-glaucomatous and glaucomatous eyes. Methods In an observational, longitudinal study, peripapillary OCTA scans were collected to evaluate intrasession repeatability and intersession reproducibility using within-eye coefficient of variation (CV_W) and intraclass correlation coefficient (ICC). Images were quantified using a custom research-oriented quantification software calculating vessel area density (VAD) and flux and a commercially developed, clinicoriented quantification software (Cirrus 11.0, Carl Zeiss Meditec) calculating perfusion density (PD) and flux index (FI). Effect of signal strength on the reliability of OCTA parameters was also evaluated. Results Among 120 non-glaucomatous eyes, intrasession CV_W were 4.2% for VAD, 5.3% for flux, 1.5% for PD and 2.0% for FI. The intersession CV_W were 6.5% for VAD, 8.0% for flux, 2.0% for PD and 3.2% for FI. The intrasession ICC ranged from 0.928 to 0.945, and intersession ICC ranged from 0.811 to 0.866. From 118 glaucomatous eyes, intrasession CV_W was 9.0% for VAD, 10.3% for flux, 1.7% for PD and 2.3% for FI. The intersession CV_W was 12.1% for VAD, 14.2% for flux, 2.3% for PD and 3.5% for FI. The intrasession ICC ranged from 0.904 to 0.972, and intersession ICC ranged from 0.855 to 0.955. Signal strength was significantly positively associated with OCTA vessel parameters (p<0.0001) for both groups. Conclusion Peripapillary OCTA vessel parameters had greater intrasession repeatability compared to intersession reproducibility in both non-glaucomatous and glaucomatous eyes. The built-in commercially developed quantification software demonstrated greater agreement than the custom research-oriented quantification software. [ABSTRACT FROM AUTHOR]