A primer on imaging and quantifying choriocapillaris with swept source OCT angiography

Abstract

Purpose : To demonstrate appropriate protocols for choriocapillaris (CC) visualization and CC flow deficits (FDs) quantification using OCT angiography (OCTA).

Methods : Normal subjects and subjects with drusen secondary to AMD were imaged using both clinically available swept source OCTA (SS-OCTA) (PLEX® Elite 9000 (ZEISS, Dublin, CA)) and lab built high resolution SS-OCTA. Both OCT and OCTA data were analyzed to determine the appropriate position and thickness of the CC slabs. After generating OCTA CC en face images, radially averaged power spectrum analyses were used to determine inter-capillary distance (ICD) and two previously published algorithms, fuzzy C-means approach (FCM method) and Phansalkar’s local thresholding (Phansalkar method), were used to segment and quantify FDs. With the Phansalkar method, different parameters were tested and local window radii of 1-15 pixels were used. FD density (FDD), mean FD size (MFDS) and FD number (FDN) were calculated for comparison.

Results : Accurate visualization of CC slab depends on correct selection of CC slab. Using the centerline of retinal pigment epithelium (RPE)/ Bruch’s membrane (BM) complex as a reference, the averaged distance from segmentation line to CC centerline is 19.85(±1.59) µm, the averaged thickness of CC slab is 14.93(±3.95) µm. The averaged ICD is 24.5µm, using a 15µm thickness slab, starting 16µm underneath the RPE/BM complex. FD metrics were highly dependent on the selection of the local window radii using the Phansalkar method. Larger window radii resulted in FDD values not physiologically meaningful. FDN increased with the increase of the window radii but then decreased, with an inflection point at about 1-2 ICDs. MFDS decreased then increased with increasing window radii.

Conclusions : To ensure accurate CC visualization and quantification with SS-OCTA, CC slab selection must be physiologically sound; selected slab must show similarity in appearance compared to histology; repeated scans of same and different size should be qualitatively similar. To ensure accurate CC quantification, selected parameters must be physiologically and physically meaningful; measured ICD should be consistent with known measurements using histology; repeated scans of same and different size should be repeatable and correlated. If the Phansalkar method was used, it is recommended that the window radius should be related to the expected ICD in normal eyes.

This is a 2020 Imaging in the Eye Conference abstract.