Xin Wei1, Sheriel Shannon Choo1, Chun Hau Chua1, William Rojas-Carabali2, Rajagopalan Rajesh1, Rupesh Agrawal1

Purpose: To investigate the use of optical coherence tomography (OCT) based vitreous dots analysis for the quantification of vitreous inflammation.
Methods: This cross-sectional study included 46 patients with anterior, intermediate, posterior or panuveitis and 48 healthy controls. OCT B-scans cutting across the fovea were used for analysis. One eye per patient was included. The COIN website (www.ocularimaging.net), a collaborative portal for ocular image analysis, was used. Segmentation of vitreous was achieved semi-automatically using the inbuilt region of interest tool. Identification of vitreous dots was performed by a trained grader based on their size, shape and reflectivity. The number and total area of vitreous dots were measured. Vitreous dots density index (VDDI) was calculated as the ratio between total area of vitreous dots to total vitreous area. Comparison between uveitis and control group was performed using mutivariate regression, accounting for age and gender.
Results: Mean age in the uveitis and control group was 49.8 ± 12.6 and 50.3 ± 12.9 years, respectively. Compared to control eyes, the mean number of vitreous dots was higher in uveitis eyes (2.15 vs. 0.44, p = 0.001) with a larger total vitreous dots area (8.54E-6 vs. 1.47E-6 mm3, p = 0.007). Mean VDDI was higher in uveitis eyes (0.01% vs. 0%, p = 0.004). For subgroup analysis, non-anterior uveitis group (intermediate, posterior and panuveitis) had more vitreous dots, larger vitreous dots area and higher VDDI when compared to anterior uveitis group.
Conclusion: Vitreous inflammation could be quantitatively assessed using OCT based vitreous dots analysis. Further studies are required for validation and correlation with clinical grading.