Using OCTA Blood Flow Analysis to Detect Early Diabetic Retinopathy

Researchers have found a biomarker to detect early diabetic retinopathy, which involves evaluating blood flow on optical coherence tomography angiography.

Optical coherence tomography angiography (OCTA) imaging analysis using normalized blood flow index (BFI) of the superficial vascular plexus (SVP) and the deep capillary plexus (DCP) is a valuable biomarker that can detect early diabetic retinopathy (DR). “[Normalized] BFI of the DCP is the most sensitive blood flow parameter to distinguish [patients without DR], mild [nonprofilerative] DR, and control eyes from one another,” said Jennifer Lim, MD, and colleagues.

Dr. Lim and colleagues sought to determine whether quantitative evaluation of blood flow on OCTA imaging enables early detection of DR. She presented their findings at the 2023 Annual Scientific Meeting of the American Society of Retina Specialists held recently in Seattle.

To compare the blood flow parameters of patients with diabetes without DR (NoDR), patients with diabetes and mild nonproliferative DR (mild NPDR), and healthy patients (control group), researchers conducted a retrospective OCTA imaging study. For inclusion in the study, researchers relied on good image quality, foveal centration, and 6 mm x 6 mm scans.

Normalized BFI of the DCP Is Most Sensitive Blood Flow Parameter

To eliminate noise from OCTA images, a thresholding algorithm was used. The study team assessed BFI values for SVP and DCP via enface projections. Normalized BFI was measured by dividing BFI by the standard deviation of the noise removed. Normalized BFI was then able to counteract noise from variable illumination and pigmentation irradiance that could impact quantification of BFI.

Finally, BFI and normalized BFI of the SVP and DCP were compared for all three groups (NoDR, mild NPDR, and control group). Using one-way ANOVA or Kruskal–Wallis one-way ANOVA, researchers were able to perform multiple comparisons among groups. A t test and Mann-Whitney test allowed them to conduct corresponding individual comparisons.

OCTA BFI and normalized BFI analyses were performed on 77 eyes: 21 eyes from 15 patients with diabetes with NoDR; 26 eyes from 22 patients with diabetes and mild NPDR; and 30 eyes from 20 healthy patients (control).

The study team observed that BFI of the DCP could distinguish between controls versus mild NPDR (P=0.0002) and between NoDR versus mild NPDR (P=0.0140). Normalized BFI of the SVP could discern all three groups from one another (ANOVA, P<0.0001) and between controls versus mild NPDR (P<0.0001) and between NoDR versus mild NPDR (P=0.0002). Normalized BFI of the DCP could discern all three groups from one another (ANOVA, P<0.0001) and between control versus NoDR (P=0.0416), control versus mild NPDR (P=0.0093), and NoDR versus mild NPDR (P<0.0001).