Inter and intra-observer variabilities in the measurement of aqueous flare with a custom-made ocular spot fluorometer

Abstract

Purpose : In our previous study, the intensity of light scatter (ILS) from the anterior chamber was measured in cataract subjects as a measure of aqueous flare using a custom-made ocular fluorometer (TVST; 2018). While the data was sensitive, we observed a poor correlation between ILS and SUN grades for patients with mild to moderate inflammation (i.e., SUN grades 0 and 1+). To further explore the variability, we are continuing to establish the repeatability and reproducibility of ILS in uveitis patients.

Methods : Uveitis (n = 111; with SUN scores of 0 to 2+) and healthy subjects (n = 93) were included in the study. One clinician performed SUN grading. For repeatability, ILS was measured twice by the same observer. For reproducibility, ILS was recorded by two observers. We calculated the intra-observer variability (expressed as the repeatability coefficient (Rc)) and intra-class correlation coefficient (ICC) to assess intra- and inter-observer variability. The latter was further analyzed by Bland- Altman (BA) plots.

Results : The intra-observer measurements in healthy subjects revealed no statistical difference in the mean ILS (0.1480 ± 0.0700 mV; n = 93; p > 0.98). The corresponding within-subject standard deviation (Sw) is 0.0272 mV, which corresponds to a repeatability coefficient (Rc) of 0.0754 mV (= 2.77 * Sw). Uveitis patients with SUN scores of 2+ showed much higher ILS (0.3740 ± 0.1090 mV; n = 5) with Sw = 0.0418 mV, and Rc of 0.1159 mV. The inter-observer analysis with uveitis patients (n = 72 for SUN 0 and n = 8 for SUN 1) revealed an ICC of 0.63 and 0.95 for measurements with SUN grades 0 and 1, respectively. In healthy subjects (SUN grade 0), ICC was 0.73 (n = 76). Corresponding Bland-Altman plot analysis in uveitis subjects indicated 0.0044 mV and 0.0046 mV as the difference between the mean ILS between observers for SUN grades 0 and 1, respectively.

Conclusions : ILS in control and uveitis are ~ 6 and ~ 9 times larger than the maximum of Sw, respectively. The lowest inter-observer variability in ILS measurement is ICC > 0.6 in both healthy and uveitis patients. Thus, ILS measurements using the fluorometer are highly reliable and can be employed to provide a higher granularity in recording aqueous flare. Thus, we attribute the apparent variability in ILS of uveitis patients with 0 and 1+ to SUN scores' subjective grading.

This is a 2021 ARVO Annual Meeting abstract.