Relative Flow Index as a Novel Optical Coherence Tomography Angiography Biomarker in Primary Open Angle Glaucoma

Relative flow indices as novel optical coherence tomography angiography (OCTA) biomarkers demonstrated reduced optic nerve head and peripapillary large vessel and capillary perfusion in primary open angle glaucoma (POAG). Large vessel perfusion correlated with disease severity and progression.

To introduce relative flow indices as novel OCTA biomarkers and their pathologic insights in POAG.

This is a retrospective case-control study where 57 POAG and 57 control eyes were included. OCTA 4.5×4.5 mm optic nerve head (ONH) were analyzed using ImageJ 1.53t to calculate global flow indices (GFIs) and relative flow indices (RFIs) for whole image optic nerve head, isolated radial peripapillary capillary plexus, and isolated large vessel angiograms.

Retinal nerve fiber layer thickness (RNFLT), ONH vascular density (VD), except inside disc and large vessel VD, GFIs, and RFIs were lower in POAG than control. There was a positive correlation between RNFLT and both VD and GFIs. Among RFIs, only large vessel RFI (ONHLVRFI) demonstrated a positive correlation with average RNFLT. Linear regression demonstrated a significant positive coefficient for ONHLVRFI with RNFLT as the dependent variable. The area under receiver operating characteristic curve showed diagnostic accuracy ranging fair, good, and excellent for all biomarkers. Inferior RNFLT had the highest area under the curve (0.922) while optic nerve head large vessel density had the lowest (0.523).

POAG showed structural loss of RNFL neurovascular unit manifesting as positively correlated reduction of VD and RNFLT. Also, POAG had lower global perfusion of the optic nerve head and peripapillary area, resulting in the positively correlated reduction of GFIs and RNFLT. Although RFIs were lower in POAG, only ONHLVRFI demonstrated a positive correlation and regression with RNFLT, implying that large vessel hypoperfusion was associated with POAG severity and progression.