Vitreous density estimation by Spectral Domain-OCT as a sensitive measure of uveitis activity: defining the technical limits of reliability

Abstract

Purpose : Quantitative measurement of the vitreous by Optical Coherence Tomography (OCT) has been shown to be an objective way of measuring changes in vitreous density such as may occur in inflammation, infection or haemorrhage. This has the potential to transform the monitoring of disease activity in routine clinical practice and to provide a much needed sensitive endpoint for therapeutic trials.
In order to take this technique to the next stage for clinical adoption it is important to assess the limits of reliability of the technique, specifically to assess whether variations in the way the images are taken could introduce error into the vitreous density readings.
We tested the extent to which changes in a novel image acquisition technique might alter the reflectivity signal and cause inaccurate estimations of Vitreous:Retinal Pigment Epithelium (Vit:RPE) signal intensity. Inter- and intra-patient repeatability was also analysed.

Methods : Spectral-domain OCT images were collected prospectively from 40 eyes of 20 visually healthy patients. Isolated adjustments to retina position, Automatic Real Time (ART) level and focus level were made before scans were acquired. Intra-patient repeatability was tested based on 3 repeats of each scan type performed, as was inter-patient repeatability across the 20 participants.

Results : Vitreous:RPE ratio showed good repeatability under standard conditions with a low coefficient of variation (mean (SD) = 0.350 (0.139)). Variation in the vitreous:RPE ratio was seen (mean (SD) = 0.0482 (0.0186)) between individuals which primarily arose due to variation in vitreous reflectivity (mean (SD) = 9.11 (3.83)) rather than RPE reflectivity (mean (SD) = 183.5 (14.7)). Alterations in acquisition settings can significantly alter the absolute measures of reflectivity, with higher vitreous:RPE scores being seen where retina position was lower, where the ART score was higher, and where a positive focus had been applied.

Conclusions : This technique shows good repeatability in estimating Vitreous:RPE ratio within the same patient under standard acquisition conditions. Systematic error can be introduced if there are major deviations in acquisition parameters, but these would be easily avoided in both routine clinical and trial use.

This is an abstract that was submitted for the 2017 ARVO Annual Meeting, held in Baltimore, MD, May 7-11, 2017.