Abstract
Purpose: :
To evaluate the intra-session repeatability of retinal thickness measurements in patients with diabetic macular edema (DME) using our standard 3D graph search based multilayer OCT segmentation automated algorithm to measure the overall thickness of the central macula.
Methods: :
46 eyes from 29 patients diagnosed with clinically significant DME were included and underwent serial macular-centered spectral domain optical coherence (SD-OCT) scans (Heidelberg Spectralis). The central 5.8x4.7mm2 area were segmented into 4 surfaces and the average thickness between the internal limiting membrane, external limiting membrane, inner/outer segment (IS/OS) junction, and the outer surface of the retinal pigment epithelium (RPE) were determined [1]. The variability between paired scans was analyzed and compared with the central macular thickness obtained from the Heidelberg Spectralis software.
Results: :
The coefficient of repeatability and variation for macular thickness using the Iowa algorithm was 5.79 μm (0.87%) for full thickness retina, 2.56 μm (2.42%) for external limiting membrane to the outer RPE, and 2.52 μm (3.27%) for IS/OS junction to the outer RPE. The coefficient of repeatability and variation of the central macular thickness using the Heidelberg software was 8.2 μm (1.11%). The average thickness of the macula was 340.5 μm using the automated software, while the central macular thickness measured by the Heidelberg software was 375.4 μm.
Conclusions: :
The reproducibility of retinal thickness measurements of the average thickness of the macula in patients with diabetic macular edema is improved by using a graph search multilayer algorithm that incorporates the full 3D information available in SD-OCT. Robust quantification of macular edema allows this to be used as an objective image based secondary endpoint for macular edema treatments, possibly in combination with quantification of External Limiting Membrane disruption.
Keywords: macula/fovea • edema • diabetic retinopathy