420
K. malek et al.
cillary diagnostics (e.g., IhC, flow cytometry, PCR) to further characterize their cell
origin, clonality and/ or malignant risk, and these tests are often time-consuming,
laborious, and costly. Cancer research has focused intently on identifying mutations
and other biomarkers to help detect neoplasias earlier, or to develop treatments [2–
5]. however, disease phenotypes are typically extremely complex, and the examination of a single gene mutation for instance, does not adequately characterize the
disease [6]. metabolomics, in comparison to proteomics or genomics, can quantify
many metabolites in complex biological systems [7]. metabolism is more closely
associated with an organism’s phenotype [8] and will almost certainly be affected
by disease, so metabolic biomarkers are likely to be accurate indicators of disease.
Fourier transform infrared spectroscopy and Raman spectroscopy are examples of
newer applications used in metabolomics. FtIR has proven to be a sensitive analytical tool to examine the structure and conformation of lipids, proteins and nucleic
acids and also of complex biological materials such as body fluids and cell cultures
in pursuit of biomarkers. We define an “infrared biomarker” as a region within a
complex mid-infrared spectrum that is diagnostically useful. the ability of FtIR
spectroscopy to detect changes in the chemical composition and morphology of
cells enables its use in distinguishing differences among cell populations, such as
normal cells versus neoplastic cervical tissue. given the ease with which infrared
spectroscopy can be performed on small tissue or cell samples with high precision
and minimal preparation, an infrared parameter that relates to cellular biosynthetic
activity may be valuable as a prognostic indicator, particularly in cases where the
current “gold standards” are wanting.
the adaptation of multi-channel infrared array detectors from military hardware
to FtIR microscopes in the early 1990s resulted in new methodologies to investigate the macromolecular architecture of cells in tissue sections [9]. the new generation of focal plane array (FPA) and linear array detectors are capable of recording
thousands of spectra in rapid time. Each FtIR image is essentially a digital hyperspectral data cube containing absorbance, wavenumber and x, y spatial coordinates.
multivariate imaging techniques including uhCA and K-means clustering, Principal Components Analysis (PCA), Linear discriminant Analysis (LdA), Fuzzy Cmeans clustering, and neural networks have proven to be invaluable in the identification and correlation of spectral groups or “clusters”, which can be directly compared to stained cells for traditional light microscopic evaluation by pathologists.
A number of exciting new developments have also emerged in the field of FtIR
spectral processing and image reconstruction. these include algorithms to correct
for light scattering and determine end member spectra in complex data sets. over
the past few years, we have seen developments in the understanding of sample preparation and the choice of sample substrate has also become a hot topic especially in
regard to trans-reflection based substrates, because an electric-field standing wave
is produced in a transflection measurement, which causes a non-linear change in absorbance with sample thickness [10]. New insights into the interpretation of spectral
bands and the mystery as to why the dNA bands appear so weak in the spectra of
fixed cells and tissues have now been resolved [11, 12].
K. malek et al.
cillary diagnostics (e.g., IhC, flow cytometry, PCR) to further characterize their cell
origin, clonality and/ or malignant risk, and these tests are often time-consuming,
laborious, and costly. Cancer research has focused intently on identifying mutations
and other biomarkers to help detect neoplasias earlier, or to develop treatments [2–
5]. however, disease phenotypes are typically extremely complex, and the examination of a single gene mutation for instance, does not adequately characterize the
disease [6]. metabolomics, in comparison to proteomics or genomics, can quantify
many metabolites in complex biological systems [7]. metabolism is more closely
associated with an organism’s phenotype [8] and will almost certainly be affected
by disease, so metabolic biomarkers are likely to be accurate indicators of disease.
Fourier transform infrared spectroscopy and Raman spectroscopy are examples of
newer applications used in metabolomics. FtIR has proven to be a sensitive analytical tool to examine the structure and conformation of lipids, proteins and nucleic
acids and also of complex biological materials such as body fluids and cell cultures
in pursuit of biomarkers. We define an “infrared biomarker” as a region within a
complex mid-infrared spectrum that is diagnostically useful. the ability of FtIR
spectroscopy to detect changes in the chemical composition and morphology of
cells enables its use in distinguishing differences among cell populations, such as
normal cells versus neoplastic cervical tissue. given the ease with which infrared
spectroscopy can be performed on small tissue or cell samples with high precision
and minimal preparation, an infrared parameter that relates to cellular biosynthetic
activity may be valuable as a prognostic indicator, particularly in cases where the
current “gold standards” are wanting.
the adaptation of multi-channel infrared array detectors from military hardware
to FtIR microscopes in the early 1990s resulted in new methodologies to investigate the macromolecular architecture of cells in tissue sections [9]. the new generation of focal plane array (FPA) and linear array detectors are capable of recording
thousands of spectra in rapid time. Each FtIR image is essentially a digital hyperspectral data cube containing absorbance, wavenumber and x, y spatial coordinates.
multivariate imaging techniques including uhCA and K-means clustering, Principal Components Analysis (PCA), Linear discriminant Analysis (LdA), Fuzzy Cmeans clustering, and neural networks have proven to be invaluable in the identification and correlation of spectral groups or “clusters”, which can be directly compared to stained cells for traditional light microscopic evaluation by pathologists.
A number of exciting new developments have also emerged in the field of FtIR
spectral processing and image reconstruction. these include algorithms to correct
for light scattering and determine end member spectra in complex data sets. over
the past few years, we have seen developments in the understanding of sample preparation and the choice of sample substrate has also become a hot topic especially in
regard to trans-reflection based substrates, because an electric-field standing wave
is produced in a transflection measurement, which causes a non-linear change in absorbance with sample thickness [10]. New insights into the interpretation of spectral
bands and the mystery as to why the dNA bands appear so weak in the spectra of
fixed cells and tissues have now been resolved [11, 12].