Repeatability of choriocapillaris flow deficit measurements in PLEX Elite 9000 for normal eyes

Abstract

Purpose : Quantitative measurement of flow deficits in the choriocapillaris have been reported using swept-source optical coherence tomography angiography (SS-OCTA). Changes in the perfusion of the choriocapillaris may be important in understanding the pathophysiology of retinal diseases such as age-related macular degeneration. This study evaluates the repeatability of fully automated choriocapillaris flow-deficit measurements on PLEX^® Elite 9000 (ZEISS, Dublin, CA) in normal eyes.

Methods : In this retrospective study, repeated PLEX Elite 9000 SS-OCTA Angio 6 x 6 mm scans were evaluated in a group of normal eyes (15 eyes from 15 subjects, the eye with the best aggregate image quality was selected when both available). Data was uploaded to the Advanced Retinal Imaging (A R I) Network Hub, a cloud-based co-development platform supporting open innovation with ZEISS, and evaluated with the two versions of a choriocapillaris flow deficit algorithm from the University of Washington [Quant Imaging Med Surg. 2018 Aug;8(7):658-666]. The algorithm evaluates the choriocapillaris angiography en face images and provides several quantitative outputs, including flow deficit area (FDA), flow deficit density (FDD), flow deficit number (FDN), and average flow density size (FDS).
The two versions include local normalization to bring up signal within shadows, with Method A being more aggressive than Method B.
Repeatability standard deviation (RSD) of the measurement was calculated as the square root of the average of the variance for all scans on an eye, and coefficient of repeatability (CoR) by dividing RSD by the mean of the measurements.

Results : Table 1 shows the population mean with standard deviation, and the RSD and CoR for all parameters for both versions of the algorithm. Repeatability was best for flow deficit size for both methods. Repeatability was better for Method A.

Conclusions : Repeatability of flow deficit measurements is good in normal eyes for both methods. It may be useful in detecting early damage to the choriocapillaris in the presence of disease. The A R I Network Hub allows efficient comparisons among different algorithms. Performance of the two methods in a set of diseased eyes will be required to determine if one method is better than the other for disease management.

This abstract was presented at the 2019 ARVO Annual Meeting, held in Vancouver, Canada, April 28 - May 2, 2019.