Name

Plexus capillaires rétiniens dans la zone de transition dans la rétinopathie diabétique : mesure de la perte de densité vasculaire de la fovéa jusqu'en moyenne périphérie en OCTA Swept-Source

Capillary occlusions in diabetic retinopathy (DR) can occur over the whole retinal surface, but it is not well documented whether it is homogeneous between the center and the periphery. Ultra-wide field imaging seems to show that the development of retinal non-perfusion starts in the mid-periphery. One explanation may lie in the distribution of the different capillary plexuses in a transition zone beyond 8 mm eccentricity of the fovea where the intermediate capillary plexus (ICP) disappears. The purpose of this study was to explore the capillary loss in this area of interest on OCT-A montages up to the temporal mid-periphery in eyes with DR compared to healthy control eyes using SS-OCTA. 

An observational study was conducted in Lariboisiere Hospital (Paris, France) including treatment-naïve diabetic patients with different degrees of retinopathy severity and healthy age-matched volunteers. As single widefield acquisitions have a lower resolution, we analyzed the capillary network from the macula to the temporal mid-periphery using 3 x 3 mm volume scans with PlexElite 9000 OCTA to create a horizontal band passing through the fovea. Binarized images were obtained on ImageJ and capillary densities (CD) were calculated in consecutive adjacent areas of 0.1 x 2.0 mm2. We compared CD between diabetic and control eyes separately in each vascular plexus according to the distance from the foveal centre using automatic slabs for the superficial vascular plexus (SVP) and the deep vascular complex (DVC), and custom slabs for the intermediate (ICP) and deep (DCP) capillary plexuses. 

A total of 30 eyes were included: 20 eyes with diabetic retinopathy (11 classified as moderate non-proliferative DR, 9 as severe NPDR and 2 as proliferative DR) and 10 control eyes. OCT-A montages averaged up to 9.6 &plusmn; 0.9 mm (range 8.8-10.2) temporal to the fovea. CD values were highest at 1.5 mm in each plexus and then gradually decreased towards the periphery along with the retinal thinning. CD were significantly lower from the fovea to the mid-periphery in diabetic eyes than control eyes in both SVP and DVC. In the SVP, CD decreased by about 55.0% at 9 mm from the foveal center in diabetic and control groups (p=0.8). In the ICP, CD were lower in diabetic eyes in the center (30.6 vs 36.2, p<0.001), and then strongly decreased in both groups towards the periphery becoming undetectable from 6 to 8mm of eccentricity. In the DCP, the reduction in mid-periphery was higher in diabetic group (56.0 vs 48.0, p<0.001). Capillary loss in diabetic eyes was assessed as the percentage of CD value in the healthy eye in the same area. In the SVP, CD loss did not differ between mid-peripheral and central areas (10.4 and 8.8 respectively, p=0.1). In contrast, CD loss in the DCP was higher at 9 mm (28.4 vs 15.3 in the center, p<0.001).  

CD gradually decreased from the center to the mid-periphery in both plexuses. As observed in the healthy subjects, we found a disappearance of the intermediate plexus between 6 and 8mm of eccentricity. Capillary loss in DR increased from the fovea to the mid-periphery in the DCP, whereas it was stable regardless of distance from the fovea in the SVP.

Whereas capillary rarefaction in diabetic retinopathy seems homogeneous in the SVC, it predominates in mid-periphery in the DCP. This difference could be due the increasing vulnerability to hypoxia of the DVC, which decreases in the mid-periphery with the disappearance of the intermediate plexus.