Vessel density and En-face segmentation of optical coherence tomography angiography to analyse corneal vascularisation in an animal model

Kavya Devarajan; Wen Di Lee; Hon Shing Ong; Nyein C Lwin; Jacqueline Chua; Leopold Schmetterer; Jodhbir S Mehta; Marcus Ang

doi:10.1186/s40662-018-0128-8

Vessel density and En-face segmentation of optical coherence tomography angiography to analyse corneal vascularisation in an animal model

Eye Vis (Lond). 2019 Jan 8:6:2. doi: 10.1186/s40662-018-0128-8. eCollection 2019.

Authors

Kavya Devarajan¹, Wen Di Lee¹, Hon Shing Ong^{1

2}, Nyein C Lwin¹, Jacqueline Chua^{1

3}, Leopold Schmetterer^{1

4

5

6}, Jodhbir S Mehta^{1

2

3

6}, Marcus Ang^{1

2

3}

Affiliations

¹ 1Singapore Eye Research Institute, Singapore, Singapore.
² 2Singapore National Eye Center, Singapore, Singapore.
³ 3Eye-ACP, Duke-NUS Graduate Medical School, Singapore, Singapore.
⁴ 4Department of Clinical Pharmacology, Medical University of Vienna, Vienna, Austria.
⁵ 5Center for Medical Physics and Biomedical Engineering, Medical University of Vienna, Vienna, Austria.
⁶ 6Nanyang Technological University, Singapore, Singapore.

Abstract

Background: Optical coherence tomography angiography (OCTA) is a novel non-invasive angiography technology that has recently been extensively studied for its utility in anterior segment imaging. In this study, we compared a split-spectrum amplitude decorrelation angiography (SSADA) OCTA and an Complex OCT signal difference angiography [corrected] (CODAA SD) [corrected] OCTA system to current angiographic technique, indocyanine green angiography (ICGA), to assess corneal vascularisation in an animal model.

Methods: We imaged 16 rabbits, (one eye per animal) with corneal vascularisation using SSADA OCTA (AngioVue; Optovue Inc., USA), CODAA OCTA [corrected] (Angioscan; RS-3000 Nidek Co. Ltd., Japan) and ICGA in the same region of interest of the cornea at successive time-points. We then analysed all scanned images for vessel density measurements and used paired t-tests and Bland-Altman plots to examine for significant differences. The en-face segmentation images from each of the OCTA scans were also extracted and were matched at every 50 μm segmentation to be compared for vessel density at the respective depths.

Results: Bland-Altman plots revealed a good agreement between all three imaging techniques (P > 0.05) for all vessel density measurements computed, and the ranges of 95% limit of agreement were acceptable from a clinical perspective. No significant difference was reported, with ICGA (μ = 16.52 ± 8.94%) being more comparable to the CODAA [corrected] OCTA (μ = 16.23 ± 9.51%; p = 0.50) than the SSADA OCTA (μ = 17.09 ± 7.34%; p = 0.33) system. Also, a good correlation value (r > 0.9) was obtained when comparing the vessel density measurements of the en-face segmentations between the OCTA systems.

Conclusions: Comparable vessel density quantification between the two OCTA systems, and with ICGA was obtained. Segmentation analysis of the vasculature at different depths showed varied performance in the two OCTA systems relative to each other. The implications of the study may help to aid in the development of better OCTA algorithms for the anterior segment and its use in clinical translational research.

Keywords: Anterior segment; Corneal vascularisation; En-face OCTA; OCTA; Optical micro angiography; Split-spectrum amplitude decorrelation angiography; Vessel density.