Page 312 - Vibrational Spectroscopic Imaging for Biomedical Applications

286 Cha pte r Ni ne

resolve, e.g., the detection and characterization of neoplastic changes,
discrimination of the cancer type and staging of the cancer, and the
effects of drug an/or radiation therapy on cancer development.
Understanding the chemical basis of the tissue spectra can be
very useful to extract diagnostic parameters from these spectra.
52
Shafer-Peltier et al. demonstrated that they could discriminate
between macroscopic spectra of different types of breast tissue on by
fitting them with a set of reference spectra derived from Raman imag-
ing. The reference set of nine independent basis spectra consisted of
various breast tissue morphological structures, such as the epithelial
cell cytoplasm, cell nucleus, fat, β-carotene, collagen, calcium
hydroxyapatite, calcium oxalate dehydrate, cholesterol-like lipid
deposits, and water. The resulting fitting coefficients yielded the con-
tribution of each basis spectrum to the macroscopic tissue spectrum,
thereby elucidating the chemical/morphological makeup of the tis-
58
sue. The fitting coefficients were shown to be different for the differ-
ent breast tissue pathologies, enabling tissue identification based on
a thorough understanding of the chemical composition.
Raman images of tissue are often compared with hematoxylin
and eosin (H&E)-stained images of the same or adjacent section that
are routinely used in pathology. Although similar tissue structures
can be observed, Raman images show more also other features that
are not present in the H&E-stained sections. This is illustrated by the
66
work of Koljenovic et al. on healthy human bronchial tissue. Using
−1
big maps, up to 1. 5 × 0.5 mm, with a high spectral (8 cm ) and spatial
resolution (1 to 12 μm), she identified the chemical composition of
different morphologic structures and provided the basis for further in
vitro and in vivo investigations of the biochemical changes that
accompany pathologic transformation of tissue Fig. 9.12. All maps
were made with a point mapping system using 847 nm, 100 mW for
excitation and 10 seconds signal collection time.
As mentioned above, much work has focused on the different
aspects of cancer research. Much of this work is fundamental research
on the differences between normal and (pre)cancerous tissue in order
to develop clinical applications. Biochemical chances that occur can
be spatially identified which is important for targeting the most dis-
criminative areas using fiber-optic Raman probes.
An example of the kind of information that Raman imaging can pro-
67
vide is given in Fig. 9.13, published in a study by Amareff et al. Using
785 nm, 160 mW laser light, focused to a spot of 3 to 4 μm, the sampled
2
large areas (20 to 30 mm ) of mouse glioma tissue with a relatively low
resolution of 50 μm and 5 seconds signal integration time per spectrum.
She showed that the cholesterol and phospholipid contents were highest
in the corpus callosum and decreased gradually toward the cortex
surface as well as in the tumor. Using a combination of subsequent
k-means and hierarchical cluster analysis techniques, she demon-
strated clear distinction between normal, tumor, two types of necrotic