Abstract
Purpose :
Assessment of microstructural abnormalities in the retina in vivo by OCT/OCTA is difficult in a large field of view (FOV) because of eye motion. Here we visualize such abnormalities by using newly developed motion-free, high-dense, and extended FOV Lissajous scan OCT.
Methods :
Thirty-four eyes of 17 diabetic retinopathy patients have been scanned with a convolutional Lissajous scan, which enables high-dense spatial sampling with a large FOV of >6.8-mm diameter, using a prototype OCT device. Eye motion is corrected, and blanks due to eye blinks are interpolated by redundant sampling data. Motion-free en-face images with 6.8×6.8-mm FOV and 820×820 grid points are reconstructed without eye tracking. All eyes were also scanned by a low-dense (6×6 mm, 320×320 samples) and high-dense (3×3 mm, 320×320 samples) raster scans by a commercial OCT device (DRI-OCT, Topcon, Japan). Both OCT devices are 1-µm swept-source OCT and have similar specifications, i.e., spatial resolutions, sensitivity, and A-scan rate.
To assess the effects of high-dense sampling, metrics of the foveal avascular zone (FAZ) were compared in five eyes. The FAZ was segmented by a semi-automatic method with the same segmentation parameters. The FAZ areas and perimeters were compared between high-dense raster and Lissajous scans and between high-dense and low-dense raster scans. Paired t-test has been applied with a significant level of 0.05.
Results :
Fig. 1 shows OCTA images obtained by the Lissajous and raster scans from an example case. The Lissajous scan shows smooth vasculature. Near the fovea (white ellipse), smooth vasculature similar to the high-dense raster scan can be visualized compared to the low-dense raster scan.
The difference between high-dense and low-dense raster scans in FAZ metrics obtained from 5 eyes was statistically significant (area: p=0.039, perimeter: p=0.033). Meanwhile, the difference between high-dense and Lissajous scans was not statistically significant (area: p=0.165, perimeter: 0.881).
Conclusions :
The Lissajous scan OCTA enables high-dense en-face posterior eye vasculature imaging with a large FOV and achieves FAZ metrics similar to the high-dense raster scan. This technique will enable the assessment of biomarkers accurately with a large FOV.
This abstract was presented at the 2023 ARVO Annual Meeting, held in New Orleans, LA, April 23-27, 2023.