Purchase this article with an account.
Jochen Straub, Areg Noshadi, Conor Leahy, Jonathan Bumstead, Arne Ohlendorf, Siegfried Wahl; Impact of retinal dewarping on thickness measurements in widefield optical coherence tomography volumes. Invest. Ophthalmol. Vis. Sci. 2020;61(9):PB00103.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Recent developments in optical coherence tomography (OCT) imaging have demonstrated widefield imaging up to 90° field of view (FOV). The purpose of this study is to develop a two-dimensional dewarping method for retinal OCT volumes and to evaluate the impact of retinal dewarping on thickness measurements.
The method developed by Steidle (Biophotonics: Photonic Solutions for Better Health Care VI, 2018) was expanded to dewarp widefield OCT volumes. A digital model dataset of two retinal layers with a thickness of 200 µm was created. We compared the traditional thickness measurement along the OCT A-scan with a thickness measurement that calculates the distance between the layers normal to a healthy retinal pigment epithelium (RPE) layer as a reference surface. We modeled three eye shapes, normal eyes with a 12mm retinal radius of curvature (ROC) as well as eyes with 11mm and 14mm ROC which model a representative range of human eyes (Atchison, 2002).
We have demonstrated dewarping of retinal OCT volumes with 6x6, 12x12 mm and 21x21 mm (90°) FOV. The traditional way of measuring thickness in OCT volumes is to measure along an A-scan. For small FOV, this method closely resembles the distance between layers normal to the RPE. Figure 1 shows the thickness error when measuring thickness along an A-scan instead of measuring normal to the RPE layer. While the error for a 6x6 mm scan is less than 1%, our data showed that it can reach up 13% for a 21x21mm (90°) FOV in an eye with strong retinal curvature.
We have demonstrated a method to dewarp large FOV retinal OCT volumes. Dewarping is an important step when analyzing structure and shape of the retina. We have further shown that measuring retinal thickness along A-scans introduces an error that increases as a function of FOV angle, reaching up to 13% for a 21x21mm (90°) FOV scan when compared to measurements normal to the healthy RPE layer.
This is a 2020 Imaging in the Eye Conference abstract.
Figure 1: Thickness error in % when measuring axial thickness along an A-scan as compared to measuring normal to a healthy RPE layer. The horizontal axis is the distance x from the center of the scan. The retinal radius of curvature of r=11mm to r=14mm represents the range of ROC found in human eyes (Atchison, 2002).
This PDF is available to Subscribers Only