Purchase this article with an account.
Jennifer K. Sun, Sonja Prager, Salma Radwan, David J. Ramsey, Paolo S. Silva, Hanna Kwak, Stephen A. Burns, Lloyd P. Aiello; Photoreceptor Mosaic Changes in Diabetic Eye Disease Assessed by Adaptive Optics Scanning Laser Ophthalmoscopy (AOSLO). Invest. Ophthalmol. Vis. Sci. 2012;53(14):4647.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
To evaluate differences in photoreceptor (PR) density and arrangement between subjects with a range of diabetic retinopathy (DR) and macular edema (DME) severity as assessed by AOSLO.
Our AOSLO imaging system (Boston Micromachines, MA) permits imaging of 1°x1.2° areas of retina at a resolution limit of 2µm, capable of resolving individual PRs. A 10x10 block grid (50 frames/block) was centered on the fovea. Foveal, superotemporal (ST) and inferotemporal (IT) blocks (foveal eccentricity=1.7mm) were corrected for distortion and averaged (Matlab program, Burns lab, IN). PR density (PRD) and Voronoi tessellation particle areas (VTPA, a measure of area occupied by each PR) were evaluated using ImageJ and assessed by 2 masked graders. PRD discrepancies >20% were adjudicated by consensus. Nearest neighbor (NN) calculations were performed using WinDRP. ETDRS fundus photographs and OCT were assessed by a masked grader.
Twenty-one eyes of 21 subjects were assessed. For subjects, mean + SD age was 47 + 14 yrs, 57% (N=12) were male and 71% (15) had diabetes (DM) of 22 + 17 yrs duration. Of diabetic eyes, 19% (4) had no DR, 14% (3) mild nonproliferative DR (NPDR), 29% (6) moderate NPDR, 10% (20) severe NPDR or PDR, and 33% (5) had DME, of which only 1 was center-involved. As expected, PRD was higher at the fovea as compared with ST & IT regions (p< 0.001). ST lower PRD (p=0.004), higher VTPA (p=0.004), and decreased NN regularity index (p=0.0001) were associated with DME presence even though 3/5 eyes with DME did not have ST DME lesions on fundus photos. IT lower PRD (p=0.02) and higher VTPA (p=0.02) were also associated with DME, despite 4/5 eyes with DME lesions outside the IT area. There was no association between DME and foveal PR paramenters. Increasing DR severity was associated with increased VTPA (p=0.03) and decreased NN regularity index (p=0.03) in the ST region only, but these trends were not significant after adjusting for DME. No significant relationship was detected between DM presence or duration and any PR parameter.
The AOSLO system allows in vivo characterization of the photoreceptor mosaic of the diabetic retina. Interestingly, alterations in PR density and arrangement were most evident in the ST quadrant in eyes with DME, even when DME lesions were not present in that area. This finding may relate to early changes in blood flow and later propensity to develop neovascularization preferentially in the ST area. If confirmed in larger patient numbers, such AOSLO findings may prove useful as early biomarkers of DME or DR, and may help elucidate an association between neural and vascular pathology in the diabetic eye.
This PDF is available to Subscribers Only