Clinical features related to optical coherence tomography angiography artifacts in patients with diabetic macular edema

Purpose To investigate the frequency and type of artifacts on optical coherence tomography angiography (OCT-A) images and the relationship with clinical features in eyes with diabetic macular edema (DME). Design Retrospective, cross-sectional comparative study. Subjects 192 eyes of 140 patients with DME were included. Methods Medical records, optical coherence tomography (OCT) and OCT angiography (OCT-A) images (Spectralis, Heidelberg, Germany) and ultra-widefield color fundus photos (Optos California, Optos plc, UK) were evaluated. Main Outcome Measures The frequency of artifact types (segmentation, motion, projection artifact and low signal) was determined. The relationship between artifact types and clinical features such as best corrected visual acuity (BCVA), mean central retinal thickness (CRT), foveal avascular zone (FAZ) area, perimeter, circularity index, perfusion density (PD), vessel density (VD), fractal dimension (FD) in the superficial capillary plexus (SCP), intermediate capillary plexus (ICP) and deep capillary plexus (DCP) and flow voids in the choriocapillaris (FV), presence of hard exudate (HE) and cataract were determined. Results The mean age was 71.6±11.4 years, and 86 out of 140 (61.4%) were males. Artifacts were present in 63 (32.8%) of 192 eyes. 29 (15.1%) eyes had segmentation artifacts, 12 (6.3%) motion artifacts, 11 (5.7%) projection artifacts and 18 (9.4%) had low signal. BCVA, PD, VD, FD in ICP and DCP were significantly lower; and CRT, FAZ area and perimeter in ICP and DCP, presence of cystoid macular edema (CME), HE, and cataract were higher in eyes with artifacts versus eyes without artifacts (p<0.05 for each). Multivariate linear regression analysis showed a significant association between segmentation artifacts and decreased BCVA (Odds ratio (OR)=5.277, p= 0.024), increased CRT (OR=1.015, p<0.001), increased area of FAZ in DCP (OR=6.625 p=0.02), and increased perimeter of FAZ in DCP (OR=1.775 p<0.04); projection artifacts and presence of HE (OR=2.017, p=0.017); motion artifacts and presence of cataract (OR= 4.102, p= 0.01). Conclusions OCT-A artifacts were present in one-third of DME eyes with segmentation artifacts being the most frequent type. Determining OCT-A artifacts is crucial to ensure accurate clinical evaluation. This data could help in developing more standardized clinical protocols for image acquisition and interpretation used in clinical practice and research.