A non-Euclidean phasor approach for distinction of fluorescent compounds using two-photon fluorescence lifetime imaging microscopy in ex vivo human skin (Conference Presentation)

2018 
Two-photon fluorescence lifetime imaging microscopy (FLIM) is a technique that not only probes the intensity of fluorophores, but also provides the temporal decay trace of said fluorophores on a pixel-by-pixel basis. These traces can then be transformed into the frequency domain for subsequent analysis, resulting in a scatterplot of phasor coordinates where each phasor corresponds to a single image pixel. With this in mind, it follows that individual fluorophores result in distinct clusters in the phasor plot, and a mixture of two fluorophores results in phasors that fall somewhere along a line linking the two clusters depending on the relative fluorophore concentrations. Until now, distinction of fluorescent species has relied mainly on computing the Euclidean distance between a given phasor and the mean coordinates of reference phasor clusters. However, this approach becomes inadequate in cases where one fluorophore has a much wider lifetime distribution than the other. As such, we propose the use of the Mahalanobis distance as an alternative to the Euclidean distance, as this metric additionally factors in the relative spread of each reference phasor cluster. This method has been applied to studying the oxidative response of ex vivo human skin via endogenous NADH fluorescence as it is exposed to chemical sun filters, the active ingredients in sunscreens. Given that both NADH and sun filters are fluorescent under the same excitation and emission conditions, the proposed Mahalanobis distance approach was used to distinguish the source of fluorescence in images of human skin. This allowed for the assessment of oxidative response as well as the tracking and monitoring of the sun filter formulation as it permeated throughout the skin.
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    0
    References
    0
    Citations
    NaN
    KQI
    []