Retinal vascular geometry detection as a biomarker in diabetes mellitus

Objective: To compare the vessel geometry characteristics of color fundus photographs in normal control and diabetes mellitus (DM) patients and to find potential biomarkers for early diabetic retinopathy (DR) based on a neural network vessel segmentation system and automated vascular geometry parameter analysis software. Methods: A total of 102 consecutive patients with type 2 DM (T2DM) and 132 healthy controls were recruited. All participants underwent general ophthalmic examinations, and retinal fundus photographs were taken with a digital fundus camera without mydriasis. Color fundus photographs were input into a dense-block generative adversarial network (D-GAN)-assisted retinal vascular segmentation system ( http://www.gdcerc.cn:8081/#/login ) to obtain binary images. These images were then analyzed by customized software (ocular microvascular analysis system V2.9.1) for automatic processing of vessel geometry parameters, including the monofractal dimension ( Dbox), multifractal dimension ( D0), vessel area ratio ( R), max vessel diameter ( dmax), average vessel diameter ( dave), arc–chord ratio (A/C), and tortuosity (τn). Geometric differences between the healthy subjects and DM patients were analyzed. Then, regression analysis and receiver operating characteristic (ROC) curve analysis were performed to evaluate the diagnostic efficiency of the vascular geometry parameters. Results: No significant differences were observed between the baseline characteristics of each group. DM patients had lower Dbox and D0 values (1.330 ± 0.041; 1.347 ± 0.038) than healthy subjects (1.343 ± 0.048, p < 0.05; 1.362 ± 0.042, p < 0.05) and showed increasing values of dmax, dave, A/C, and τn compared with normal controls, although only the differences in dave and τn between the groups were statistically significant. In the regression analysis, dave and τn showed a good correlation with diabetes ( dave, OR 1.765, 95% CI 1.319–2.362, p < 0.001; τn, OR 9.323, 95% CI 1.492–58.262, p < 0.05). Conclusions: We demonstrated the relationship between retinal vascular geometry and the process in DM patients, showing that Dbox, D0, dave, and τn may be indicators of morphological changes in retinal vessels in DM patients and can be early biomarkers of DR.

Clinical study of diabetic peripheral neuropathy screening by retinal vascular geometric parameters

To investigate the relationship between geometrical changes of retinal vessels and diabetic peripheral neuropathy (DPN), and to determine the effectiveness of retinal vascular geometry analysis and vibration perception threshold (VPT) for DPN assessment. Type 2 diabetes patients (n = 242) were categorized by stage of DPN. VPT and fundus photography was performed to obtain retinal vascular geometry parameters. The risk factors for DPN and the correlation between DPN stages were analyzed. The efficiency of the retinal vascular geometric parameters obtained with VPT as a diagnostic tool for DPN was examined. Stages of DPN showed a linear correlation with VPT (r = 0.818), central retinal vein equivalent (CRVE) (r = 0.716), and fractal dimension arterioles (DFa) (r = − 0.769). VPT, CRVE, DFa, and fractal dimension veins (DFv) showed high sensitivity (80%, 55%, 82%, and 67%, respectively) and specificity (92%, 93%, 82%, and 80%, respectively) for DPN diagnosis. Good agreement was observed between combined use of geometric parameters (CRVE, DFa and DFv) and VPT (Kappa value 0.430). The detection rate of DPN with combined use of geometric parameters of retinal vessels (64.88%) was significantly higher than that with use of VPT (47.52%). Retinal vascular geometry changes demonstrated significant correlation with DPN severity. VPT, CRVE, DFa, and DFv may provide insights for understanding DPN.