Estimation of Diabetic Retinal Microaneurysm Perfusion Parameters Based on Computational Fluid Dynamics Modeling of Adaptive Optics Scanning Laser Ophthalmoscopy

Miguel O. Bernabeu, Yang Lu, Omar Abu-Qamar, Lloyd P. Aiello, Jennifer K. Sun

Diabetic retinopathy (DR) is a leading cause of vision loss worldwide. Microaneurysms (MAs), which are abnormal outpouchings of the retinal vessels, are early and hallmark lesions of DR. The presence and severity of MAs are utilized to determine overall DR severity. In addition, MAs can directly contribute to retinal neural pathology by leaking fluid into the surrounding retina, causing abnormal central retinal thickening and thereby frequently leading to vision loss. Vascular perfusion parameters such as shear rate (SR) or wall shear stress (WSS) have been linked to blood clotting and endothelial cell dysfunction, respectively in non-retinal vasculature. However, despite the importance of MAs as a key aspect of diabetic retinal pathology, much remains unknown as to how structural characteristics of individual MAs are associated with these perfusion attributes. MA structural information obtained on high resolution adaptive optics scanning laser ophthalmoscopy (AOSLO) was utilized to estimate perfusion parameters through Computational Fluid Dynamics (CFD) analysis of the AOSLO images. The HemeLB flow solver was used to simulate steady-state and time-dependent fluid flow using both commodity hospital-based and high performance computing resources, depending on the degree of detail required in the simulations. Our results indicate that WSS is lowest in MA regions furthest away from the feeding vessels. Furthermore, areas of low SR are associated with clot location in saccular MAs. These findings suggest that morphology and CFD estimation of perfusion parameters may be useful tools for determining the likelihood of clot presence in individual diabetic MAs.