Abstract
Purpose :
Optical coherence tomography angiography (OCTA) is an exciting, maturing imaging technology which may provide valuable information for assessing retinal disease. However, OCT angiograms have shadowing artifacts, making the visualization of the data challenging. Our goal was to explore a software workflow to visualize the three-dimensional (3D) architecture of inner retinal vessels from OCT angiography in both two-dimensional (2D) en face and 3D display formats.
Methods :
We used the RTVue XR Avanti Optical Coherence Tomography Angiography (OCTA) instrument (Optovue Inc, Fremont, California, USA) to obtain 3D angiograms from healthy volunteers. The angiograms were exported from the OCTA device, enabling us to explore artifact removal and 2D/3D visualization techniques. For 2D en face display, we developed a simple method to produce composite images of the superficial and deep plexuses of the inner retinal network. Maximum amplitude projections (MAP) of the two plexuses were exported from the Optovue software. We then produced composite images in two steps: subtraction and color merging, which were performed in MATLAB (R2015b, MathWorks) and Fiji (National Institutes of Health, Bethesda, MD, USA), respectively. For 3D display, we exported the angiography data and performed 3D layer segmentation to separate the superficial and deep plexus. We then converted the data into a standard 3D medical image format, which could then be loaded into various software packages for 3D visualization, including the MRIcroGL software package.
Results :
Figure A shows a 2D MAP of the superficial plexus in a healthy control, and Figure B shows an example of a 2D MAP deeper plexus in the same patient. Figure C shows the final result of the deeper plexus after the subtraction process. We compared color composite images before and after subtraction as shown in Figure D and Figure E, respectively. The contrast between the superficial and deep plexi in Figure E is higher. Figure F shows an example of 3D display of the two plexuses (from a different patient) after our software pipeline.
Conclusions :
We have successfully explored several methods for the visualization of OCTA data from the Optovue system, which will be useful in the future for better communication of the data among clinicians and scientists.
This is an abstract that was submitted for the 2016 ARVO Annual Meeting, held in Seattle, Wash., May 1-5, 2016.