Researchers at the University of Wisconsin–Madison recently assessed the prevalence and severity of technical-related and patient-related artifacts on OCT angiography (OCT-A) and found that they are common and can result in misinterpretation of images. The prevalence of artifacts associated with the reliability of quantitative outputs from commercially available OCT-A systems may be as high as 53.5% in patients with diabetic retinopathy, they noted. Understanding the type of artifacts may improve image quality through algorithm manipulation, hardware upgrade and technician training.

The cross-sectional study assessed 406 OCT-A images from 234 eyes, of which 54.4% were 6mm × 6mm scans and 45.6% were 3mm × 3mm scans. Severe artifacts associated with the reliability were found in 217 images (53.5%). The three most prevalent were shadow (26.9%), defocus (20.9%) and movement (16.0%). Commercially recommended quality index thresholds had an area under the curve value of 0.80 to 0.83, sensitivity of 97% to 99% and specificity of 37% to 41%.

“Knowledge of the artifacts and their association with OCT-A images, along with working toward improving the artifacts, may be necessary before these images can be used reliably for clinical purposes and as a clinical trial end point,” the researchers noted. They believe that prevalence may decrease as technicians become more experienced with OCT-A.

Holmen IC, Konda MS, Pak JW, et al. Prevalence and severity of artifacts in optical coherence tomographic angiograms. JAMA Ophthalmol. December 5, 2019. [Epub ahead of print].