Abstract
Purpose :
The purpose of the study was to demonstrate the three-dimensional nature of retinal vasculature with separable slabs based on Multi-Layer Segmentation results obtained from optical coherence tomography angiography (OCTA) scans. This model will provide a hands-on visual of retinal layers as identified by MLS. The goal is to create both a virtual view of retinal layers using a virtual reality (VR) headset and a 3D printed model.
Methods :
Reconstructed data of the retinal layers was 3D printed into slabs with a thickness defined by MLS. The layers are then assembled via a mechanical component allowing it to be stacked horizontally to view the different layers. The final model(s) will be generated using angiography scans taken on human eyes with a commercial OCT, the CIRRUS™ 6000 (ZEISS, Dublin, CA). The data will be processed in MATLAB and exported as a 3D dot map of data points. Using the NIH 3D Print Exchange (https://3dprint.nih.gov), the exported data can be converted to 3D model files that can be printed using a commercial 3D printer (e.g. Stratasys Fortus 360MC) or visualized using VR glasses (e.g. ZEISS VR ONE). Initial code in MATLAB was written to input reconstructed data of an OCT scan, locate and isolate retinal layers based on pre-determined offsets, and export each segment into a separate output file.
Results :
We demonstrated in a previous IITE project (Straub et al. ARVO Imagining in the Eye 2017) that the above goal can be achieved using commercially available tools. In the proposed study, we are further developing this model to not only visualize surface retinal vasculature but also visualize the individual retinal layers. Figure 1 shows assembled 2D en face images of segmented layers from an OCT scan. The en face slabs are shown as staggered to show their overlay. 3D printing these slabs and stacking them to create a data cube is the goal of the final model.
Conclusions :
Based on the current workflow of this study, we were able to visualize the three 3D nature of retinal layers using both a VR headset and a physical 3D model. We can use these models to not only study healthy retinal vasculature but also retinas with tumors or staphyloma. This model can also be useful to clinicians for research or teaching purposes to better understand the anatomy of the retina and how to visualize 3D structures based on 2D images.
This is a 2020 Imaging in the Eye Conference abstract.