Retinal Photoreceptors and Arterioles Alterations Detected using Adaptive Optics Scanning Laser Ophthalmoscopy in Diabetes and Diabetic Retinopathy

Thirty subjects were recruited in the Guangdong Diabetic Retinopathy Multiple-omics Study (GD-RMOS). Adaptive optics scanning laser ophthalmoscopy (AOSLO) was adopted for image acquisition. An AI-based algorithm conducted quantitative analyses of photoreceptor morphology, including cone density, spacing, and regularity. Vessel morphology (inner diameter, outer diameter, wall thickness, wall-to-lumen ratio [WLR]) was manually assessed. The arteriolar index ratio (AIR) was derived to assess arteriolar remodeling. Generalized estimating equation was utilized to compare the differences of inner diameter, outer diameter, wall thickness, WLR and the AIR among healthy control, DM and DR groups. Thirty patients, consisting of 9 healthy controls, 11 with DM, and 10 with DR, were evaluated (mean [SD] age, 66.33(7.02). A significant reduction in photoreceptor cell density in macular areas correlated with the presence of DM (P=0.033) and DR (P=0.006). Notably, a significant increase of the wall thickness was observed. These variations were pronounced in the arteriole segments of diabetic patients (P<0.001) and escalated with the presence of DR (P=0.018). Accounting for WLR, the DR group showed a significantly larger ratio comparing to control group (P=0.011). The AIR in small retinal arterioles differed significantly among DR (P<0.001), DM (P=0.024) and control groups. AOSLO imaging showed delineating retinal photoreceptor and arteriolar alterations throughout DR progression, potentially serving as a predictive measure for anatomical or functional retinal changes in DM and DR patients.