Purchase this article with an account.
Ik Soo Byon, Marco Nassisi, Srinivas R Sadda; Impact of slab selection on quantification of the choriocapillaris on optical coherence tomography angiography. Invest. Ophthalmol. Vis. Sci. 2019;60(9):3056.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
To assess the impact of slab selection on quantitative measurements of choriocapillaris (CC) flow voids using optical coherence tomography angiography (OCTA).
Using a swept-source OCTA system (Zess Plex Elite 9000), two repeat 3 x 3 mm volume scans were acquired in both eyes of healthy subjects at the Doheny - UCLA Eye Centers in this IRB-approved study. For each OCTA acquisition, en face slabs to isolate the CC were obtained at three different settings: Setting A was a 20 μm thick slab starting 31 μm posterior to the automatically segmented RPE band; Setting B was a 10 μm thick slab starting 31 μm posterior to the RPE; Setting C was a 10 μm thick slab starting 41 μm posterior to the RPE. Corresponding en face slabs from the two repeat acquisitions were averaged and compensated for signal loss using the corresponding structural en face images. Resultant CC OCTA images were binarized and the extent of flow voids was computed as a percentage of the total scan area. Flow void extent was compared amongst the three CC slab selections.
Twenty seven eyes of 27 normal subjects were prospectively enrolled and included in this analysis. Mean age of subjects was 42.0 years old. Mean subfoveal choroidal thickness (SCT) was 264.48 μm. The CC flow void extent was significantly higher for setting A (20.67 %) compared to both setting B (11.62 %, p<0.001) and C (10.17 %, p<0.001). The CC flow void extent for setting B was also significantly higher compared to setting C (P=0.016).
Quantitative CC parameters may be significantly influenced by small differences in the level and thickness of the en face slab selection. In particular, inclusion of deeper slab components appears to be associated with a greater flow void extent. These findings are of relevance as automated RPE segmentation can be susceptible to errors in the setting of disease, and this may impact the reliability of quantitative CC metrics.
This abstract was presented at the 2019 ARVO Annual Meeting, held in Vancouver, Canada, April 28 - May 2, 2019.
This PDF is available to Subscribers Only