Repeatability of autofluorescence lifetime imaging at the human fundus in healthy volunteers.

PURPOSE: We aim to evaluate the repeatability of a new fluorescence lifetime imaging (FLIM) technique which measures time-resolved autofluorescence to assess metabolism of the retina.
MATERIALS AND METHODS: We performed FLIM with two spectral channels (channel 1: 490-560 nm and channel 2: 560-700 nm) on 10 healthy volunteers, with 10 replicates per volunteer. From the 30° fundus FLIM images, we selected three regions: the fovea, the optic disc and the papillo-macular bundle. For each channel in these regions, we determined an average multi-exponential approximation with three components, and the six resulting parameters, α1-α3 (amplitudes) and τ1-τ3 (fluorescence lifetimes), were analyzed in terms of the coefficient of variation (CV).
RESULTS: Repeatability was highest in the papillo-macular bundle, followed by the fovea and the optic disc. Repeatability was higher in channel 1 (mean CV of 7.9%) than in channel 2 (mean CV of 17.7%). The average CV for the diagnostically most relevant channel 1 and the most relevant parameters was as follows: τ1 (5.5%) and τ2 (4.7%) in the papillo-macular bundle, and τ1 (6.8%) and τ2 (6.9%) in the fovea.
CONCLUSIONS: We demonstrated repeatability of FLIM measurement results within acceptable ranges of variation. Based on the detailed coefficients of variation, we derived recommendations for parameter ranges suitable for diagnostic applications.