Background
Diabetic Retinopathy (DR) is a severe and widely spread eye disease. Thus, an objective test for the early diagnosis and evaluation of treatment in DR is certainly needed. In this study, the ability of intraretinal layer segmentation to locally detect early retinal changes in diabetic patients is assessed using optical coherence tomography (OCT).
Material and Method
Fifty diabetic patients with no or minimal DR underwent ophthalmic examination, OCT and fundus photography. Automated segmentation of intraretinal layers of the OCT images was performed using a custom-built algorithm. Mean thickness of the macula and intraretinal layers of patients with no DR (DM) was calculated in the fovea, pericentral and peripheral regions and compared with those in patients with mild DR (MDR).
Results
We found reduced retinal nerve fiber layer (RNFL) thickness in the pericentral and peripheral regions (27 + or - 2 versus 18 + or - 5 microm and 42 + or - 3 versus 33 + or - 9 microm, respectively, p<0.001) and reduced thickness of ganglion cell/inner plexiform layer (GCL+IPL) complex in the pericentral region of the macula (92 + or - 7 microm versus 80 + or - 10 microm, p<0.001) in the MDR group. Accordingly, macular thickness was reduced in the pericentral and peripheral region of the macula in the MDR group.
Conclusions
Our results support the view of neurodegeneration in diabetes in the early stage of retinopathy which seems to involve the ganglion cells and cells of the inner plexiform layers (RNFL+GCL+IPL) mostly. Local retinal thickness measures can be obtained from OCT scans using an intraretinal layer segmentation procedure, and these measures could be helpful in finding a surrogate for following development of retinopathy that could affect vision.

Keywords: Retina - pathology, Macula Lutea - pathology, Diabetes Mellitus - pathology, Software, Tomography, Optical Coherence - methods