Bruch's Membrane Opening-Based Peripapillary Optical Coherence Tomography Angiography Analysis

Peripapillary optical coherence tomography angiography (OCT-A) scans are a challenge for vessel density (VD) analysis, being dependent on demarcation of the optic disc. Longitudinal VD analysis requires that each pixel of the OCT-A scan must be at the exact same location during follow-up scans in order to see inter-visit differences. The aim of the present study was to investigate reliability of Bruch's membrane opening (BMO)-based peripapillary OCT-A analysis with and without the implementation of the anatomical positioning system (APS) compared to manual analysis.

Thirty-seven eyes were measured twice by OCT-A (Heidelberg OCT II Spectralis) and analyzed by the Erlangen-Angio-Tool (EA-Tool) with an APS- and BMO-based analysis of the macula and peripapillary region. APS allows alignment of OCT-A scans according to each individual FoBMOC (fovea-to-Bruch's membrane opening center). Peripapillary OCT-A scans were analyzed: (1) manually (by the shortest distance of the vertical or horizontal diameter of the optic disc), (2) BMO-based, and (3) BMO-based with APS information. Coefficients of variation (CVs) were calculated.

Peripapillary mean VD was 42.52 ± 4 and 41.69 ± 4 (manually), 50.29 ± 4 and 48.8 ± 4 (BMO-based), and 44.73 ± 3 and 44.39 ± 4 (BMO-based and APSified) for 1st and 2nd scan, respectively. Peripapillary mean VD yielded a significant difference between the 1st and 2nd scan for manual (p = 0.04), but not for BMO-based (p > 0.05) and BMO-based and APSified analysis (p > 0.05). CVs were 10.0% (manually), 8.0% (BMO-based), and 8.0% (BMO-based and APSified).

The integration of BMO and APS information into the EA-Tool allows a BMO-based peripapillary VD analysis for long-term OCT-A analysis.