Advanced Characterization of Vitreous Hyperreflective Dots via OCT-Derived Metrics: A Cross-Sectional Study

The use of optical coherence tomography (OCT) as a potential tool for the measurement of vitreous inflammation has been previously described as a more objective and reproducible method when compared to historically known subjective scales. In this study, our objective is to evaluate OCT's ability to characterize vitreous hyperreflective dots (VHDs) across eyes with varying conditions, including healthy controls, vitreous degenerations, intraocular inflammation, and others.

We utilized a purpose built semiautomated software comprising an image binarization tool to segment OCT scans of 61 eyes, comprising 15 eyes with vitreous degenerations, 20 uveitic eyes, 17 healthy controls, and 9 with other eye conditions. The vitreous dot index (VDI) was computed by determining the number of dots (VDI-N) and the dot area (VDI-A). VHDs were identified as the hyperreflective shadows observed in OCT images within segmented areas of the vitreous, stratified as zones I, II, and III. We compared the difference between groups using analysis of variance (ANOVA). Intergrader reliability was evaluated by comparing results obtained by two trained independent graders, employing intraclass correlation coefficient (ICC) analysis.

When the medians of VDI-N and VDI-A were compared in healthy controls, patients with uveitis, patients with vitreous degeneration, and others, patients with vitreous degeneration had the highest VDI-N median (2.61 ± 2.76 mm³ p < 0.001) followed by healthy controls (0.48 ± 0.87 mm³ p < 0.001) in zone l. As for VDI-A in the same zone, healthy controls had the greatest median (0.71 ± 0.96, p < 0.001) among the different groups. In zone II, uveitis and the healthy control group had similar medians for VDI-N (0.03 ± 0.36 and 0.03 ± 0.29, p < 0.001 respectably) and VDI-A was greater in the vitreous degeneration group (0.40 ± 0.50 p < 0.001). Zone III had lower VDI-N and VDI-A; patients with uveitis and patients with vitreous degeneration had equal VDI-N (0.00 ± 0.03 p < 0.001) and patients with uveitis had the higher VDI-A among the rest of the groups (0.00 ± 0.65 p < 0.001). For the total vitreous (TV), the highest VDI-N was found in patients with vitreous degeneration (2.92 ± 2.85 p < 0.001) while the highest VDI-A was in the uveitis group patients (0.66 ± 1.31) p < 0.001. The average vitreous dot density index and the average vitreous dot reflectivity index (VDRI) in the TV were greater in patients with vitreous degeneration (2.15 × 10^-5 ± 1.52 × 10^-5) and patients with uveitis (0.13 ± 0.08), respectively. When comparing VDI markers using a Kruskal-Wallis nonparametric one-way ANOVA test, we found that only the average vitreous dot reflectivity index in zone I and VDI-A in TV were statistically significant. However, only the reflectivity index was significant when comparing patients with vitreous degeneration and healthy controls in a pairwise analysis.

While vitreous inflammation scales must evolve toward more objective metrics, our findings suggest that VHDs on OCT can act as confounders, as they may represent normal vitreous cells or even the presence of vitreous degeneration. The reflectivity index appears to have better reproducibility than simple count; however, when searching for a more objective parameter for measuring vitreous inflammation, vitreous degeneration must be considered.