Background/aims Quantitative analysis of optical coherence tomography angiography (OCT-A) images requires a reproducible approach that accounts for sectoral loss. The objective of this study was to determine whether an index that accounts for both global (perfusion density, PD) and asymmetric loss of perfusion, rather than PD alone, more reliably measures loss of perfusion in patients with glaucoma.
Methods We analysed macular OCT-A scans of 95 glaucoma patients and 59 control subjects. Two-dimensional projection images corresponding to the superficial vascular plexus were exported and analyses performed to calculate global PD and image asymmetry. An unsigned perfusion asymmetry index (PAI) that included PD and asymmetry (with 1:1 wt) was calculated. Perfusion density and PAI were compared with 10-2 visual field mean deviation and ganglion cell layer (GCL) thickness.
Results Median (IQR) visual field mean deviation was −1.73 (−3.76, 0.30) dB for the glaucoma group and 0.67 (0.16, 1.18) dB for the control group. The strength of the correlation with mean deviation was stronger for PAI (r=0.47), compared with PD (r=0.35), whereas with GCL thickness they were comparable (r=0.45 and 0.43, respectively). Compared with controls, mean PD was 12% lower in patients with glaucoma (0.27 vs 0.30), while PAI was 17% lower (0.40 vs 0.48). However, diagnostic accuracy of either PD or PAI was worse than GCL thickness.
Conclusions While PAI yielded better correlation with mean deviation and GCL thickness, and a slightly improved separation between patients with glaucoma and healthy controls, diagnostic accuracy was inferior compared with GCL thickness.
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
Contributors Conception and design: CS, BC. Data acquisition: MW, GPS, DMH. Analysis and interpretation: CS, BC. Drafting and revising the manuscript: all authors.
Funding Canadian Institutes of Health Research (grant no. PJT159564; BC), Alcon Research Institute (BC), Mathers Research Fellowship (CS). Equipment support from Heidelberg Engineering.
Competing interests LMS: Consultant—Alcon, Allergan, Bausch and Lomb, Thea Pharmaceuticals. PER: Consultant—Allergan, Bausch and Lomb; Lecturer—Bausch and Lomb. MTN: Consultant and Lecturer—Allergan, Alcon. BCC: Equipment support—CenterVue, Heidelberg Engineering, Topcon; Lecturer—Santen.
Patient consent for publication Obtained.
Ethics approval Ethics approval was obtained from the Nova Scotia Health Authority Research Ethics Board and the study was conducted in accordance with the Declaration of Helsinki.
Provenance and peer review Not commissioned; externally peer reviewed.
Data availability statement Data are available on reasonable request. Data are not in a repository. Data may be available pending the nature of the request and institutional ethics approval.