Agreement between eyes in wide-field fluorescence lifetime imaging ophthalmoscopy measurements at the human retina in healthy volunteers

Abstract

Fluorescence lifetime imaging ophthalmoscopy (FLIO) is a new imaging technique for measuring the timeresolved in vivo autofluorescence generated by endogenous fluorophores in the ocular fundus. The aim is to assess the metabolism of the retina. Based on the fluorescence lifetime, different fluorescent compounds in the eye can be distinguished. Pathologic changes may be observed by detecting changes in the fluorescence lifetime. Until now, FLIO is performed using 30° fundus images. For certain diseases section of the fundus is of interest. In this work, observational clinical study in young healthy volunteers was performed to examine the usability of 55° wide-field fluorescence lifetime images. The time-resolved retina autofluorescence was measured (scanning laser ophthalmoscope: 55° of fundus, 62x62 μm² per pixel; excitation: diode laser with pico-second pulses, 473nm, 80MHz repetition rate; detection: spectral channels 498-560nm (ch1) and 560-720nm (ch2), time-correlated single photon counting method) in both eyes of 11 healthy volunteers (28.7±3.6 years). Three repetitive measurements on different days within one week at a similar time have been performed. All subjects had a crystalline lens and an undilated pupil. A modified 3- exponential approach was applied to determine the mean fluorescence lifetimes τm. τm were computed on the ETDRS grid and in a 15x15 pixel region 25° superior to the fovea. The Wilcoxon rank-sum test was used to test for statistical significant differences between left and right eye. The coefficient of variation in the superior region is 5.6% (right) and 4.7% (left) in ch1 and 2.5% (right) and 2.1% (left) in ch2. No statistical significant differences have been found between right and left eye. FLIO measurements in young healthy volunteers using a 55° lens are repeatable in both eyes and show no significant differences. Spectral channel 2 should be preferred because of its lower variability. The data analysis was done with FLIMX (available at http://www.flimx.de).