Abstract
Purpose :
Recent research has demonstrated that retinal texture is a valuable source of biomarkers for neurodegeneration, as it can detect minute differences in the central nervous system. Previous studies have examined the layer thickness of the neuroretina in women and men as measured from optical coherence tomography (OCT) data. However, the effect of sex-based retinal interocular differences has been relatively unexplored and unaddressed. This study looked into the texture metrics from OCT data of healthy adult Portuguese individuals to determine if there are interocular differences and if they are related to sex.
Methods :
Gray-level co-occurrence matrix features were computed from mean-value fundus images representing each of the six layers of the neuroretina. These features were tested separately for group mean differences between the female and male participants’ right and left eyes. Since many features (456) were being tested, a pairwise correlation analysis was first performed to exclude similar features. The non-correlated features were tested for normality, using the Shapiro-Wilk test at a 10% significance level for a more conservative normality decision. If the feature distribution was normal for both the right- and the left-eye groups, a paired-sample t-test was used to evaluate the group mean differences; otherwise, a signed rank test was applied. Due to the large number of pairwise tests, three different pairwise correction methods were applied: Bonferroni, Benjamini–Hochberg, and the False Discovery Rate.
Results :
Initial uncorrected results showed that interocular retinal texture asymmetries were mainly present in women and with a larger effect size than in men. After correction, except for the highly conservative Bonferroni method, the results revealed that texture differences were only present in the female group, while the eyes of men were similar.
Conclusions :
Our work suggests that structural patterns in right and left eyes manifest differently in females and males, raising the question of why such differences occur and how they might vary over the lifespan, both in healthy individuals and in patients diagnosed with neurodegenerative disorders. In addition, it further emphasizes the need for accurate control in studies that use both eyes.
This abstract was presented at the 2023 ARVO Annual Meeting, held in New Orleans, LA, April 23-27, 2023.