Abstract
Purpose :
To determine the feasibility of automated quantification of posterior vitreous inflammation on optical coherence tomography (OCT) and to assess its correlation with changes in clinical grading on examination.
Methods :
This was a prospective, observational study of patients with active posterior uveitis or panuveitis. At each clinic visit, OCT volume scans (512x128) of the posterior vitreous were obtained by centering the scan on the fovea and placing the retina at the lowest part of the scan window (Cirrus 5000 HD-OCT, Carl Zeiss Meditec, Dublin, CA). Volume scans were sectioned into individual frames, and contrast was adjusted in standardized fashion across all included scans to eliminate the fine, granular pattern felt to represent noise in lower signal scans. Larger, hyperreflective foci, felt to represent clumps of inflammatory cells, were quantified using the marching squares algorithm, which encapsulates each identified region in a unique path and approximates the diameter, allowing selective quantification of foci based on pixel size (DotCount, v1.2, LCLN, Cambridge, MA). An OCT grade was assigned based on the average number of hyperreflective foci per OCT frame using the same cut-offs as were applied to the Standardization of Uveitis Nomenclature Classification of anterior chamber cell (0 cells=0, 1-5 cells=0.5, etc). The OCT grade was then compared to the anterior vitreous cell (AVC) grade (0-4), as determined on examination. Pearson’s correlation was used to compare AVC and OCT grade.
Results :
14 eyes from 8 patients were included, with clinical AVC grade and imaging available from 30 unique time points. The change in OCT grade in patients with any increase in AVC grade was 0.25±0.17, while the change in OCT grade with any decrease in AVC grade was 0.12±0.08. The change in OCT grade in patients with no measured change in AVC grade was -0.22±0.17. The Pearson correlation was R=0.06 (p = 0.75).
Conclusions :
Posterior vitreous inflammation can be successfully imaged using OCT and quantified using available image analysis software. However, there is weak correlation between AVC and OCT grade. Further study is needed to determine whether this lack of correlation reflects a true disparity between anterior vitreous and posterior vitreous inflammation and to explore relationships between other measures of clinical measures of inflammation, including vitreous haze.
This is an abstract that was submitted for the 2018 ARVO Annual Meeting, held in Honolulu, Hawaii, April 29 - May 3, 2018.