3D image registration and retinal perfusion heterogeneity analysis of clinical optical coherence tomography angiography

Blindness is a debilitating detriment to human quality of life. In the pursuit of investigating the pathogenesis of blindness-causing retinal diseases, innovative tools for imaging the human retina are needed. Optical coherence tomography (OCT) and OCT angiography (OCTA) are recently-established imaging modalities that are safe, non-invasive, and provide 3D images superior than the traditional dye injection-based 2D imaging modalities. However, artifacts induced by patient motion degrade OCTA data quality, and patients with retinal diseases like diabetic retinopathy (DR) exhibit these artifacts more often than their healthy counterparts. A novel methodology is proposed for software-based OCTA motion correction achieved through serial acquisition, volumetric registration, and averaging. This method permits clear 3D visualization of retinal structures and pathologies superior than that of traditional ophthalmic imaging. The method is then deployed to investigate retinal perfusion heterogeneity which, if properly understood, may improve our understanding of the pathogenesis of retinal diseases like DR.