UC Davis

Purpose

Although commercial optical coherence tomography angiography (OCTA) machines quantitate retinal vascular density (VD) by dividing the vasculature into superficial and deep capillary plexus (SCP, DCP), histology reveals three distinct plexus layers. This study tested the hypothesis that the VD measurement of three distinct retinal plexus layers obtained using custom segmentation has high repeatability comparable to that of automatically segmented SCP and DCP layers.

Materials and methods

Forty-four participants (86 eyes) were enrolled - 54 eyes with retinal vasculopathy and 25 eyes with macular edema. Macular OCTA images (3x3 mm and 6x6 mm) were obtained twice within 30 minutes by the same personnel using the same instrument (AngioVue, Optovue, version 2018.0.0.18). The intraclass correlation coefficient (ICC) was calculated to access repeatability.

Results

The repeatability of VD for SCP and DCP was good-to-moderate (ICC=0.65-0.85) and minimally affected by image quality, retinal vasculopathy, or macular edema. The repeatability of the VD of the custom-segmented intermediate and deep plexus layers (cICP and cDCP) was poor/moderate (ICC=0.40-0.74) but better in the subset without macular edema using 3x3 mm scans with good images quality (ICC=0.58-0.93). Repeatability of cICP and cDCP VD measurement for 6x6 mm scans was poor (ICC≤0.5) in eyes with retinal vasculopathy and/or macular edema.

Conclusion

Although repeatability of the VD measurement is high for the automatically segmented SCP and DCP, repeatability of VD is poor for the cICP and cDCP using larger scans in eyes with retinal vasculopathy and/or macular edema.