Page 251 - Vibrational Spectroscopic Imaging for Biomedical Applications

Raman Micr oscopy for Biomedical Applications 227

complex signatures, features in Raman spectra from tissue, cells, and
bacteria are distributed over a broad region and multivariate algo-
rithms are required for analysis and classification. The principle of
supervised classification is that a model is trained by spectra of
known class membership and subsequently spectra of unknown ori-
gin can be assigned to one of these classes.
This contribution summarizes selected research results about
fiber-optic Raman spectroscopy of tissue, Raman imaging of tissue
and cells, and Raman spectroscopy of bacteria. Recent reviews
described more comprehensively biomedical applications of Raman
2
and infrared spectroscopy to diagnose tissues, metabolic finger-
3
printing in disease diagnosis, disease recognition by Raman and infra-
4
5,6
red spectroscopy, noninvasive studies on individual cells, and Raman
and coherent anti-Stokes Raman spectroscopy of cells and tissues. 7
8.2 Raman Imaging of Tissue
In principle, diseases and other pathological anomalies lead to chem-
ical and structural changes on the molecular level which also change
the Raman spectra and which can be used as sensitive, phenotypic
markers of the disease. As these spectral changes are very specific
and unique, they can be considered as a fingerprint. Early reports in
the literature regarding the utility of Raman spectroscopy to biomedical
problems were based on macroscopic acquisition of spectral data
only at single points which required an a priori knowledge of the
location or a preselection of the probed position. Since the inhomoge-
neous nature of tissue on the microscopic level was not considered in
these early studies, an accurate correlation between the histopathology
of the sampled area and the corresponding spectra was only possible
for homogenous tissue on the macroscopic level. Considerable progress
was made when high throughput and more sensitive instruments
became available for Raman microspectroscopic imaging. They
enable to microscopically collect larger number of spectra from larger
sample populations in less time, improving statistical significance
and spatial specificity.
Besides microscopic imaging, another promising medical applica-
tion of Raman spectroscopy is the combination with endoscopy. The
basic components consist of a light source, a light guide, and an endo-
scope which can be rigid or flexible. The imaging contrast is usually pro-
vided by white light illumination and registration of the scattered light
by an optical system. To validate a medical diagnosis micromechanical
instruments are inserted into the working channel and a tissue biopsy is
collected which is subsequently analyzed using other methods such as
light microscopy. However, the selection of a biopsy and the localization
of early or small cancer lesions based on visual inspection are often dif-
ficult. Tissue changes should be detected as early as possible because the
prospects for cure are at maximum in an early stage of tumor genesis.