Endoscopic autofluorescence micro-spectroimaging of alveoli: comparative spectral analysis of amiodarone-induced pneumonitis patients and healthy smokers

Fibered confocal fluorescence microscopy (FCFM) with spectroscopic analysis capability was used during bronchoscopy, at 488nm excitation, to record autofluorescence images and associated emission spectra of the alveoli of 5 healthy smoking volunteers and 7 non-smoking amiodarone-induced pneumonitis (AIP) patients. Alveolar fluorescent cellular infiltration was observed in both groups. Our objective was to assess the potential of spectroscopy in differentiating these two groups. Methods: We previously demonstrated that in healthy smokers alveolar elastin backbone and tobacco tar contained in macrophages contribute to the observed signal. Each normalized spectrum was modeled as a linear combination of 3 components: S exp (λ) = C e .S e (λ)+C t .S t (λ)+C G .S G (λ), C e , C t and C G are amplitude coefficients. S e (λ) and S t (λ) are respectively the normalized elastin and tobacco tar emission spectra measured experimentally and S G (λ) a gaussian spectrum with tunable width and central wavelength. Levenbergt-Marquardt algorithm determined the optimal set of coefficients. Results: AIP patient autofluorescence spectra can be uniquely modelized by the linear combination of the elastin spectrum (C e = 0.61) and of a gaussian spectrum (center wavelength 550nm, width 40nm); the tobacco tar spectrum coefficient C t is found to be zero. For healthy smoking volunteers, only two spectral components were considered: the tobacco tar component (C t = 1,03) and the elastin component (C e = 0). Conclusion: Spectral analysis is able to distinguish cellular infiltrated images from AIP patients and healthy smoking volunteers. It appears as a powerful complementary tool for FCFM.