Evaluation of Retinal Microcirculation Heterogeneity with Optical Coherence Tomography Angiography in Glaucoma

Authors: Hoyoung Jung¹, Cameron Oliver¹, Mahsa Siadati¹, Henry Hodaly², Yusi Miao¹, Myeong Jin Ju¹, Kirk Stephenson¹, Kulbir Gill¹, Steven Schendel¹, Zaid Mammo¹. ¹University of British Columbia, ²Simon Fraser University.

Author Disclosures: H. Jung: None. C. Oliver: None. M. Siadati: None. H. Hodaly: None. Y. Miao: None. M. Ju: None. K. Stephenson: None. K. Gill: None. S. Schendel: None. Z. Mammo: None.


Abstract Body: 

Purpose: Structural and functional dysregulation of the papillary and peripapillary microvasculature have been implicated in the pathogenesis of glaucomatous optic neuropathy. To detect relative blood perfusion heterogeneity in the macular circulation in subjects with glaucoma and healthy controls, we used a commercial optical coherence tomography angiography (OCTA) system with post-acquisition motion-correction software. The goal was to define and validate the use of OCTA in the assessment of depth-resolved perfusion heterogeneity, as a marker of vasomotor dysregulation, in patients with glaucoma and healthy subjects. 

Study Design: Cross-sectional observational cohort study. 

Methods: Research ethics board approval and written informed consent from all patients were obtained. Sixteen control patients (male:female 11:5, mean age 32.9 years) and thirty-six patients with glaucomatous optic neuropathy (male:female 16:20, mean age 65.6 years) were included. Exclusion criteria included a history of vascular and/or ophthalmic
disease or treatment known to affect retinal microcirculation. For each eye, ten 3x3 mm field-of-view volumes centered at the fovea were sequentially acquired using the Zeiss Plex Elite swept-source OCTA platform. These volumes were aligned using a MATLAB-based (Mathworks; 2019b) registration algorithm and the coefficient of variation (CoV) was computed for each pixel in the registered image stack. A deep learning-based vessel segmentation algorithm was applied to increase the signal-to-noise ratio. This process was repeated for both the superficial and deep vascular complexes (SVC and DVC). The mean CoV value was calculated for each eye and was used to compare the differences in perfusion heterogeneity between patient groups. 

Results: Of 52 eyes imaged, 11 control and 11 glaucomatous (7 open-angle glaucoma, 4 normal-tension glaucoma) eyes without significant motion artifacts were included in our analysis. Mean CoV in the DVC was significantly higher than in the SVC in both control ( p =0.0013) and glaucoma groups ( p <0.001). Mean CoV of both the SVC and DVC were significantly higher ( p <0.001) in glaucomatous eyes than in control eyes. Capillary shunting was frequently detected in the inner foveal avascular zone (FAZ) capillary ring. 

Conclusions: Serial acquisition, motion correction, and pixel variation analysis of OCTA images can be performed to identify relative perfusion heterogeneity in the macular microcirculation of healthy and glaucomatous eyes. Greater relative perfusion heterogeneity was noted in our glaucoma subjects in both the macular SVC and DVC compared to control subjects. Future directions include increasing sample size with a wider range of glaucoma types and severity, and further refining our understanding of the underlying pathophysiology of glaucoma and associated perfusion changes.

X

X