Image Quality as an Important Confounder in Quantitative OCT Angiography: A Review with Quantitative Synthesis.

BACKGROUND

Optical coherence tomography angiography (OCTA) provides quantitative metrics of the retinal microvasculature, most prominently vessel density (VD), that are increasingly used as biomarkers in ocular, systemic, and cerebrovascular disease. Because OCTA relies on the detection of flow-related motion contrast, image quality has emerged as a pervasive determinant of these metrics, yet its effect has been reported in fragmentary and non-comparable ways across the literature.

METHODS

We identified OCTA studies indexed in PubMed that examined the relationship between image quality and quantitative OCTA parameters, and we summarised their methods and findings across the macular, foveal avascular zone, and peripapillary regions and across vascular layers. We integrated directly comparable per-unit scan-quality effects using random-effects meta-analysis and separately synthesized direct cross-sectional Pearson correlations between manufacturer-reported image quality and macular vessel-density outcomes from independent healthy cohorts.

RESULTS

Higher image quality was associated with higher measured vessel density in every contributing dataset. On the Optovue scan-quality (SQ, 0-10) scale, superficial macular VD increased by 3.46% (95% CI 1.85-5.07) per SQ unit in controlled signal-attenuation experiments and by covariate-adjusted observational estimates ranging from 0.90% to 2.16% per SQ unit; a combined order-of-magnitude estimate across both estimator types was 1.64% (95% CI 1.15-2.12). Peripapillary estimates were heterogeneous and derived from only two independent cohorts; they were therefore summarised descriptively rather than pooled. Across three independent externally authored healthy cohorts reporting direct cross-sectional Pearson correlations, higher image quality was strongly associated with higher macular vessel-density outcomes (random-effects pooled r = 0.64, 95% CI 0.50-0.75; I2 = 49%).

CONCLUSIONS

Image quality is an important, directional confounder of VD-based OCTA metrics whose magnitude can rival the biological or physiological signal of interest. Platform-appropriate standardisation, transparent reporting, and consideration of image quality in acquisition and analysis are therefore important for the valid interpretation of vessel-density measurements, particularly in functional and longitudinal studies.