Purpose
To compare the differences in the thicknesses of the outer plexiform layer (OPL) and the outer nuclear layer (ONL) in diabetic patients with hard exudates versus those without.
Methods
Underserved diabetic patients from the Alameda County Medical Center, CA, were selected sequentially to match our criteria, using a sample of 1900 who underwent photoscreening in the EyePACS grading system for color fundus photographs (Canon), the Laser Scanning Digital Camera (Aeon Imaging), and Spectral Domain OCT (iVue, Optovue). Two groups of patients were were matched in age, gender, and ethnic status. The 14 cases had hard exudates within one disk diameter of the fovea, seen on color fundus photography and SDOCT by independent graders. The 14 controls had no hard exudates. The subjects, who averaged 58 yr, included 16 males and 12 females, with 12 Asians, 12 Hispanics, 2 African Americans, and 2 non-Hispanic Caucasians. To quantify the thickness of specific retinal layers, manual segmentation of the SDOCT layers was performed with custom software (Matlab, Mathworks). The OPL and ONL thicknesses were sampled at locations 500, 1000, and 1800 microns from the foveal center.
Results
The cases had significantly more variable ONL thickness, p < 0.01 for temporal locations 1800, 1000, and 500 microns from the fovea, and the nasal location 500 microns from the fovea. ONL thickness of up to 5 cases per location was greater than the upper confidence limit for the controls, 1 - 2 cases that were less than the lower confidence limit, i.e. the ONL was thinned. The cases had significantly more variable OPL thickness, p < 0.01 for temporal locations 1800 and 500 microns from the fovea. However, OPL thickness of up to 6 cases per location was greater than the upper confidence limit for the controls, but no cases were thinner.
Conclusions
The presence of hard exudates greatly increases the variability of the thickness of retinal layers. However, hard exudates do not automatically indicate uniform retinal thickening, since several eyes had thinning of the ONL at varied locations. These data indicate that not only are hard exudates associated with increased variability on SDOCT that makes diagnosis of macular edema less accurate, but also that there is photoreceptor damage in eye that may not be classified as having diabetic macular edema.
Keywords: 499 diabetic retinopathy •
549 image processing •
648 photoreceptors