Racial differences in texture analysis features of the choriocapillaris

Texture analysis extracts high-order mathematical features from an image with the potential to reveal subtle image differences. Although racial variations have been demonstrated in traditional optical coherence tomography angiography (OCTA) parameters, texture analysis has not been well explored.

This is a retrospective, cross-sectional study of 226 eyes from 127 nondiabetic subjects (60 Black, 67 White). We collected 3 × 3 mm en face OCTA images of the superficial choriocapillaris (CC), which were processed using computerized texture analysis to extract the following features: average grey level, grey level 30% threshold, grey level 70% threshold, balance, skewness, energy, entropy, contrast, mean gradient, root-mean-square (RMS) variation, and first moment of power spectrum.

There were no significant baseline differences in the demographics and clinical characteristics of the 2 racial groups, other than a higher rate of diagnosed hypertension in Black subjects (20% in Black vs 6% in White subjects; p = 0.017). Using a mixed-effects linear regression model, we found statistically significant differences in the absolute value of the grey level, skewness, and root-mean-square variation, which remained significant in a subgroup of nonhypertensive patients (p < 0.001 for all).

These findings demonstrate racial variation in CC texture features, underscoring the importance of diversity when building clinical datasets. Future studies may better explore how these differences correlate with underlying chorioretinal pathogenesis.