# Assessing Agreement between Blue-Light and Green-Light Autofluorescence of Macular Hyperautofluorescent Rings in Inherited Retinal Diseases

## TL;DR

This study compares blue and green autofluorescence imaging for measuring hyperautofluorescent rings in inherited retinal diseases, finding high correlation but slight differences in measurements.

## Contribution

The study provides a detailed comparison of blue and green autofluorescence imaging for inherited retinal diseases, highlighting their agreement and discrepancies.

## Key findings

- Blue autofluorescence measured slightly larger ring areas than green autofluorescence.
- High correlations were observed between the two imaging modalities for all metrics.
- Discrepancies exceeding 10% were observed in a significant percentage of cases for area and diameters.

## Abstract

To assess the agreement between blue autofluorescence (BAF) and green autofluorescence (GAF) fundus imaging in measuring the macular hyperautofluorescent (hyperAF) ring in patients with inherited retinal diseases (IRDs).

A prospective, within-subject agreement study.

A total of 124 patients with IRD (67% with retinitis pigmentosa, 33% with cone–rod/macular dystrophies) from Moorfields Eye Hospital were included. Mean age was 36.5 years; 60% were male. Most (99%) had a genetically confirmed diagnosis.

Participants underwent BAF (Heidelberg) and GAF (Optos) imaging during the same visit. Hyperautofluorescent ring area and horizontal and vertical diameters were independently measured by 2 ophthalmologists. Agreement between imaging modalities was assessed using intraclass correlation coefficients, Pearson correlation, paired t tests, linear and mixed-effects regression, Bland–Altman plots, and 1-sided z-tests for equivalence within a ±10% margin.

Comparison of hyperAF ring area and diameters between BAF and GAF.

Intergrader agreement was excellent (intraclass correlation coefficient: 0.93–0.98). Blue autofluorescence measured slightly larger ring areas than GAF (mean difference: 0.7 mm2, P = 0.006), while horizontal and vertical diameters were nearly equivalent (mean differences: 0.03 μm and 0.08 μm, P = 0.96 and P = 0.04, respectively). Discrepancies >10% were observed in 17% of cases in horizontal diameter, 27% of cases in vertical, and 42% of cases in area. Correlations were high (r = 0.98) for all metrics. Mixed-effects models including both eyes (n = 217) estimated that the area measured in GAF is 4.1% smaller than in BAF, the horizontal is approximately 0.5% smaller, and the vertical is around 2.2% smaller, with evidence of agreement within a ±6% margin at a 5% significance level.

High correlation and consistent regression slopes were seen between BAF and GAF macular hyperAF ring. These results suggest that while the 2 modalities yield broadly comparable measurements, with BAF yielding larger values, these modalities should ideally not be used interchangeably.

Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.