Page 25 - Vibrational Spectroscopic Imaging for Biomedical Applications

Towar d Automated Br east Histopathology 3

attractive, since visual evidence readily relates to the knowledge
base of pathology and provides information in a compact form that
can be universally comprehended. Simple structural imaging (e.g.,
optical microscopy of hematoxylin and eosin (H&E)-stained tissue)
and manual recognition is already practiced in the clinic. Hence,
efforts to improve this process are the logical first attempts at
improving practice. More recently, molecular imaging has provided
some understanding of specific epitopes’ roles in cancer progres-
sion. Hence, it provides an alternative to add more information to
classical images. Molecular bases for disease diagnoses are not uni-
versal, however, and there are significant numbers of patients in
every cancer category, for whom the approach fails to provide any
useful information. Another alternative of chemical imaging is
emerging in which the contrast arises from endogenous chemical
constitution of the tissue.
FT-IR spectroscopic imaging, the imaging analogue of molecu-
lar infrared spectroscopy, provides an alternative platform for histo-
−1
18
pathologic imaging. Near-IR (NIR) light (14000 to 4000 cm ) is
most commonly employed for biomedical imaging as it encom-
passes a region of low absorption and scattering within the body.
NIR light can penetrate deep into samples and has been applied to
develop noninvasive medical diagnostics. The primary NIR con-
trast mechanism is scattering as the region only consists of broad
and significantly overlapped molecular vibrational overtones.
Therefore, the NIR spectral region has limited utility in distinguish-
ing biochemical features in complex tissue. The far-IR spectral
−1
region (below 400 cm ) contains absorption frequencies for atoms
with a high mass. This is a common feature for metals and metal
19
complexes with organic molecules. In contrast, the frequencies in
−1
the mid-IR region (4000 to 400 cm ) correspond directly to funda-
mental vibrational modes of organic chemical species. Hence, the
spectral response of any material is a chemical fingerprint that can
uniquely identify chemical species, their local environment and
their macromolecular conformation. Therefore the mid-IR spectral
region is most appropriate for distinguishing biochemical features
in tissues and is especially attractive for cancer pathology due to its
ability to detect subtle transformations.
By measuring the intrinsic chemical composition of tissue, FT-IR
imaging can provide significant biochemical information without the
application of contrast agents or chemical stains. The use of nonper-
turbing radiation and compatibility of developed approaches with
clinical practice are additional advantages that can lead to translation
of this technology to pathology laboratories. In addition, a knowl-
edge base for spectral changes corresponding to disease states exists
and significant understanding of tissue spectroscopy is available. 20
Last, instrumentation is well developed and has a large user base.