Abstract
Purpose :
To investigate the repeatability and agreement between local and global thresholding strategies for the quantification of choriocapillaris (CC) flow deficits (FD) using spectral domain (SD) optical coherence tomography angiography (OCTA).
Methods :
Eighteen eyes from nine healthy volunteers aged>18 years were imaged four consecutive times using a SD-OCTA system (Heidelberg Engineering, Heidelberg, Germany) using a 10x10 high-resolution protocol centered on the fovea. Projection artifacts were removed, and CC was bracketed between 10 and 30µm below the Bruch’s membrane. Three local (Phansalkar, mean and Niblack) and three global (mean, default, Otsu) thresholding strategies were applied to each of the four consecutive CC scans. We used the intraclass correlation coefficient (ICC) to compare the repeatability and agreement between global and local thresholding strategies for quantifying FD number, mean FD size, total FD area and FD density.
Results :
A total of 72 OCTA examinations were considered. We verified a high amount of variation in the quantification of FD number, mean FD size, total FD area and FD density) according to the thresholding strategy used (Figure 1). The agreement among the thresholding strategies for measuring each quantitative parameter in the same scan was low (ICC ranging from 0.000 to 0.339). Regarding the repeatability of global versus local thresholding strategies for evaluating FD number, mean FD size, total FD area and FD density in four consecutive scans, we verified that local strategies achieved a significantly superior ICC for measuring all the four quantitative parameters, as seen in Table 1.
Conclusions :
FD quantification varies according to the thresholding strategy that is used and direct comparisons should not be performed. Local thresholding strategies are significantly superior to global ones for quantifying FD and should be preferred.
This is a 2020 ARVO Annual Meeting abstract.