Motion Artifacts

Since OCTA detects erythrocyte movement to generate contrast, it is extremely motion sensitive and prone to motion artifacts.⁶ Movements that cause artifacts can occur in the axial direction (from heartbeat and respiration) or in the transverse direction (as in saccadic movements of the eye). As a result, the acquired volume might not accurately correspond to the real retinal architecture.⁸ Algorithms, such as the split-spectrum amplitude-decorrelation angiography (SSADA), lower the resolution of axial OCT in order to minimize sensitivity to axial movement.⁴ One advantage of this approach is that it does not affect the analysis of OCTA images, since most of the flow signal in the ocular fundus is in the transverse rather than axial dimension.⁹ However, transverse movements are still a major problem. These movements are recognized as white lines in the OCTA en face scans. Some systems use software-based motion correction by creating two orthogonal volumes from a Fast-X (fast B-scans acquired in horizontal direction) and a Fast-Y (fast B-scans acquired in vertical direction) raster scans. The two orthogonal volumes are then merged to obtain a single volume with better signal quality.⁸ Summarily, the software tries to estimate and correct the eye motion for each A-scan in the two volumes ( ▶ Fig. 2.1). Unfortunately, while correcting small saccades and fixation losses, the system can introduce artifacts of its own, such as double vessels, stretching, and crisscrossing defects ( ▶ Fig. 2.2).^4,​8 In addition, gross eye movements cannot be compensated and might result in poor-quality images ( ▶ Fig. 2.3). In a recent analysis of image quality when using the RTVue XR Avanti with AngioVue software (Optovue, Fremont, CA), we found that motion artifacts counted for the majority of poor-quality exams (i.e., those scans for which neither qualitative nor quantitative assessment could be made) and was significantly higher in patients with a vision acuity of worse than 20/70, probably due to their difficulty in maintaining fixation during the acquisition process (PRC, Oliveira, MD; DR, Chow, MD, FRCS; 2016). Faster scanning speeds and eye-tracking technology might help improve these issues.

Blinking is another source of artifacts while obtaining images and appears as black lines on OCT angiograms ( ▶ Fig. 2.4). The OCT signal is blocked from getting to the retina and no flow is detected.

2.4 Projection Artifacts

This is probably the most common artifact and is virtually present in every OCTA exam. It corresponds to the fluctuation of light that passes through moving blood in the inner retinal vessels and is projected back to the deeper reflective layers such as the retinal pigment epithelium (RPE). Since that light also changes over time, those deeper layers will seem to have blood vessels with the pattern of the overlying retinal vessels ( ▶ Fig. 2.5).^{4,​5,​10,​11} Projection artifacts can especially cause falsely positive results during identification and quantification of choroidal neovascular (CNV) membranes. Some OCTA systems include the option of eliminating the projection artifacts. But in so doing, they can produce artificially low signal intensity in areas of pathologic vessels, leading to underestimation of the real magnitude of the lesions. When this secondary artifact happens, the projection artifact is actually substituted by a shadow artifact, leaving gaps in the neovascular network. Zhang et al published a “projection-resolved” algorithm, which improves the quality of the resulting OCTA after the suppression of the projection artifact. But, as stated by the authors, this system remains imperfect, leaving minor gaps in the deep retinal plexus and residual projection artifact in the RPE layer.¹⁰ As the OCTA technology continues to improve, it will be a matter of time before issues related to projection artifacts are resolved.

Stay updated, free articles. Join our Telegram channel

Join