cornea advanced imaging

Autofluorescence intensity (top row) and corresponding FLIM images (bottom row) of corneal epithelium wing cells, stroma,

Descement’s membrane, and endothelial cells. In the autofluorescence intensity images the signal intensity represents

the number of photons detected in each pixel. Epithelial and endothelial FLIM images are pseudo-color coded for NAD(P)H mean 

autofluorescence lifetime. The stroma and Descement’s membrane FLIM images are pseudo-color coded for the mean autofluorescence lifetime.

Lifetime values ranges are indicated in the color bars. Keratocytes in the corneal stroma are indicated by the white arrows. Scale bar = 20 µm.

Background

Corneal diseases are the second major cause of blindness worldwide, according to the World Health Organization. Moreover, being a highly innervated tissue, corneal diseases cause severe pain and discomfort. So, non-invasive methods for early detection of corneal pathologies can have a large impact on the monitoring and management of patients and contribute to prevent disease progression to irreversible conditions.

The clinicians need an imaging tool capable of detecting the very early signs of cell dysfunction. In other words, a functional imaging modality with cell-level resolution could have a deep impact on understanding the mechanisms causing corneal diseases, on the detection of disease signs at very early stages and, ultimately, on the patients’ prognosis. Fluorescence lifetime metabolic imaging microscopy has the potential of fulfilling this need.

This research project comprised the development and evaluation of two different approaches for implementing fluorescence lifetime metabolic imaging microscopy, both having the cornea as final application:

 - Using multiphoton imaging techniques, namely two-photon excitation and second harmonic generation, with serial pixel imaging using laser scanning excitation and time-correlated single photon counting detection.

- Using one-photon excitation with parallel, wide-field imaging and time-gated detection through an intensified camera.

The first approach offered greater promise due to the inherent optical sectioning of multiphoton imaging, the use of safer infrared wavelengths and the possibility of exciting both NAD(P)H and FAD fluorescence with the same laser source. However, there were concerns on the safety of the high peak radiant powers required by multiphoton excitation and on the very high cost of femtosecond lasers. 

The second approach uses much less expensive picosecond diode lasers and was considered safe for in vivo ocular imaging. However, there were concerns on the possibility of achieving optical sectioning on a thick transparent tissue, on the suitability of Rapid Lifetime Determination algorithms for obtaining the decay parameters with adequate accuracy and precision and on obtaining an adequate signal using only FAD fluorescence. Relying on NAD(P)H fluorescence was not considered since one-photon excitation of this metabolic co-factor implies UV radiation that was deemed unsafe for in vivo ocular imaging, due to its phototoxicity.