Label-free imaging of human cells: algorithms for image reconstruction of Raman hyperspectral datasets

Raman microspectroscopy-based, label-free imaging methods for human cells at sub-micrometre spatial resolution are presented. Since no dyes or labels are used in this imaging modality, the pixel-to-pixel spectral variations are small and multivariate methods of analysis need to be employed to convert the hyperspectral datasets to spectral images. Thus, the main emphasis of this paper is the introduction and comparison of a number of multivariate image reconstruction methods. The resulting Raman spectral imaging methodology directly utilizes the spectral contrast provided by small (bio)chemical compositional changes over the spatial dimension of the sample to construct images that can rival fluorescence images in terms of spatial information, yet without the use of any external dye or label. Algorithms for label-free image reconstruction of biological entities via confocal Raman micro-spectroscopy are reported. The image depicts a photomicrograph and a pseudo-color Raman image of a stem-cell cluster.