Abstract
Purpose :
To evaluate hyperreflective retinal spots (HRS) in diabetics with and without macular edema (DME vs no DME) on linear B scans and corresponding en-face imaging obtained using spectral domain OCT (SD OCT).
Methods :
A retrospective detailed evaluation of OCT images of 38 eyes/subjects (19 diabetics without DME and 19 with DME) was performed. A population of normal 19 subjects/eyes served as control. On B scan SD OCT the following characteristics of HRS were evaluated: location (inner retina: IR or outer retina: OR); size (<30μ or >30μ); reflectivity (similar to retinal nerve fiber layer: RNFL or to retinal pigment epithelium: RPE); presence/absence of back-shadowing. On en-face SD OCT: absence of vessel or any other lesion; presence of vessel or microaneurysm; presence of hard exudates (confirmed on fundus color photo) were recorded. All gradings, were performed twice by two graders in a masked fashion.
Results :
HRS size <30μ, reflectivity similar to RNFL, absence of back-shadowing and location in both IR and OR (mainly in DME patients), were associated with absence of vessels or any other lesion on en-face imaging (p=0.0001, for each evaluation). Size >30μ, reflectivity similar to RPE, presence of back-shadowing and location in the OR were all associated with the presence of hard exudates on en-face image, (p<0.0001, for each evaluation). Multiple logistic regression analysis showed that HRS presence in the IR (p<0.0001), size >30μ (p=0.0075), reflectivity similar to RNFL (p=0.02) and presence of back-shadowing (p<0.0001) are directly associated with the presence of microaneurysms on en-face imaging. Intra-grader and inter-grader repeatability were excellent for all evaluations, for both B scans and en-face images.
Conclusions :
HRS can be easily evaluated on OCT images in normals and diabetics, even with DME. HRS with specific characteristics may have different origin. This data may help to better understand the pathophysiology of HRS in diabetic eyes, even vs controls.
This is an abstract that was submitted for the 2016 ARVO Annual Meeting, held in Seattle, Wash., May 1-5, 2016.