Purchase this article with an account.
Nitish Mehta, Sarwar Zahid, Bing Q Chiu, Suruchi Bhardwaj, Edmund Tsui, Emma Young, Joshua A Young, Jesse J Jung; Refinement of fractal analysis methodology of swept-source optical coherence tomography angiography images of normal eyes.. Invest. Ophthalmol. Vis. Sci. 2017;58(8):1675.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
We aimed to study the effect of image quality and other paramters on fractal dimension (FD) analysis to characterize the retinal vascular branching pattern in normal eyes using swept-source optical coherence tomography angiography (OCTA).
Macular scans of 3x3, 6x6, 9x9, and 12x12mm sizes of 29 normal eyes were obtained with the investigational Zeiss Plex Elite 9000 (Carl Zeiss Meditec, Inc., Dublin, CA, USA) (Figure 1). The central 3x3mm portions of all images were cropped out for subgroup analysis to account for proportionally different inclusion of the foveal avascular zone (Figure 2). Subgroup analysis was performed using various image resolutions and with various incorporation of vascular elements. Grayscale OCTA images were standardized and binarized using ImageJ (National Institutes of Health, Bethesda, Maryland, USA). Fractal box-counting analyses were performed using Fractalyse (ThéMA, Besançon Cedex, France) and Fraclac ((National Institutes of Health, Bethesda, Maryland, USA). Statistical analysis was achieved with two-tailed Student’s T-test and one-way analysis of variance (ANOVA).
Fractal dimension (FD) and correlation coefficient (CC) were significantly different (p< 0.05) between superficial and deep plexuses in all image sizes and types and between the different image sizes for both superficial and deep plexuses. There was a significant difference (p<0.05) between both FD and CC between the original 3x3mm images and the same 3x3mm region of the modified larger images in the subgroup analysis.
Different FD values between image sizes suggests an effect of inclusion of different anatomy. Cropped images remain statistically different suggesting image quality and size also impact analyses. Further studies may determine the optimal image size/quality to develop FD analysis as an OCTA-derived quantitative tool for assessing retinal vascular structure.
This is an abstract that was submitted for the 2017 ARVO Annual Meeting, held in Baltimore, MD, May 7-11, 2017.
This PDF is available to Subscribers Only