Abstract
Purpose: :
Fundus autofluorescence imaging is a common diagnostic technique in ophthalmology and could become equally valuable in preclinical studies using animal models. Here we test the reliability and consistency of quantitative measurements of fundus autofluorescence (quantitative AF, qAF) images in mice utilizing a confocal scanning laser ophthalmoscope (Spectralis HRA).
Methods: :
AF images were acquired using a confocal scanning laser ophthalmoscope with 488nm excitation (Spectralis HRA; Heidelberg Engineering, Germany) modified for use in mice by the incorporation of an aperture to restrict the incident laser beam and of an internal fluorescence reference to account for changes in detector sensitivity and laser power (Delori et al. IOVS). AF images were acquired in the high-speed mode (nine frames) with a 55° field (8.9 images/sec) in two groups of mice: Group 1: 5 Abca4-/- mice (10 eyes) at 6 months of age to measure session to session repeatability, and Group 2: 14 Abca4+/+ mice (28 eyes) at 9 months of age to test consistency of qAF measurements in mice of the same age, genotype and coat-color. Repeatability was assessed from images acquired at the same sensitivity in 3 sessions, with repositioning of the mouse and refocus between each session. Consistency was evaluated using different sensitivities. qAF levels in all images were obtained for each eye using Photoshop CS5 in 12 separate, evenly-distributed segments of the image, accounting for reference and zero grey levels. The mean grey scale value of the 12 segments was calculated.
Results: :
The Bland-Altman coefficient of repeatability (95% confidence interval) in Group 1 (Abca4-/-) was ±31% for measurements between sessions. The within session repeatability was ±5%. Consistency of qAF measurements within Group 2 was ± 9.7 % of the mean qAF (95% confidence interval). Mean qAF was 3.2±0.8 for the Abca4-/- mice. This value was significantly higher than 1.3±0.2 for the Abca4+/+ mice (p<0.0001)
Conclusions: :
Conclusions: The qAF method enables reliable measurements of in vivo lipofuscin levels in mice. The use of this tool has the potential to aid in understanding disease processes in retinal degeneration and will facilitate preclinical therapy trials.
Keywords: imaging/image analysis: non-clinical • retina • ipofuscin