Abstract

Optical spectroscopy/imaging has been one of the techniques, which has a major impact in the field of medical diagnostics. The degree of multiplexing (e.g. detecting multiple analytes through a single spectral acquisition) and versatility of spectral techniques in molecular sensing domain has been well established in the literature. Raman spectroscopy detects individual molecules and is independent of excitation wavelength, allowing selection of application specific wavelengths while using the same data library measured at a fixed wavelength, making RS a prime candidate for label-free sensing and imaging. Raman spectra are very sensitive biochemical markers due to the unique vibrational fingerprint spectra of the tissue. Raman spectroscopy is also very sensitive towards small molecular/chemical changes, such as an increased nucleus-to-cytoplasm ratio, disordered chromatin, higher metabolic activity, and changes in lipid and protein levels. We have collected Raman spectral libraries for obtaining operator independent diagnostics under ex-vivo and in-vivo conditions and we have tested this approach using multivariate analysis on known solutions with Raman active components. The ultimate successful classification architecture is a blend of multivariate analysis of the spectral data as well as their transformation into feature space with prominent features defined as peak location, peak widths and relative peak-peak ratios.

© 2015 Optical Society of America