Automated retinal microvascular velocimetry based on erythrocyte mediated angiography

Retinal blood flow is an emerging biomarker in ocular and systemic disease. Erythrocyte mediated angiography (EMA) is a novel technique that provides an easily interpretable blood flow velocity quantification by directly tracing individual moving erythrocyte ghosts over time in vivo, imaged using a scanning laser ophthalmoscope (Heidelberg Retina Angiograph platform). This tracking procedure, however, requires time-consuming manual analysis to determine blood flow. To overcome this current bottleneck, we developed an objective and automated velocimetry approach, EMA - Automated Velocimetry (EMA-AV), which is based on a modified sequential Monte Carlo method. The intra-class correlation coefficient (ICC) between trained human graders and EMA-AV is 0.98 for mean vessel velocity estimation and 0.92 for frame by frame erythrocyte velocity estimation. This study proves EMA-AV is a reliable tool for quantification of retinal microvascular velocity and flow and establishes EMA-AV as a reliable and interpretable tool for quantifying retinal microvascular velocity.