20 Medical Applications of NIR Spectroscopy
461
marker substances that are providing differences in absorption and scattering characteristics. Hemoglobin concentrations and its redox status, described by oxy- and
deoxygenated forms for providing oxygenation saturation, water and lipid content
have been listed for discrimination from healthy tissue. Tumor tissues were found
with significantly high levels of water and of hemoglobin because of higher vascularization, but also with less lipids and lower oxygen saturation due to cell hypermetabolism. Commonly, in tumorous tissue of breast cancer, abnormal proliferation of
cells will result in an angiogenesis with an increased number of blood vessels, eventually increasing the local blood Hb concentration. A recent breast cancer study has
been reported by Mehnati et al., who differentiated between normal and tumorous
breast tissue by local hemoglobin concentrations [48].
Cancerous tissues from pancreatic and colorectal tumors have been classified by
studying, in particular, first and second overtone bands of the lipid CH 2 stretching
vibrations, which could be intensified by first spectral derivative calculations; for
the characteristic wavenumber intervals and band assignments, see also Fig. 20.2.
Besides diagnosis, also margin location of cancerous tissue, especially intraoperative
malign tumor margin assessment is important for supporting the surgeon’s decisions
during operations. Furthermore, NIR spectroscopy is useful for therapeutics monitoring, especially after chemotherapy, radiation treatment, neoadjuvant and photodynamic therapy. For more details, the reader is again referred to our review [46]. A
recent instrumental development of an LED-based NIR sensor for kidney tumor diagnostics has also been published [49]. Four LEDs with emission band maxima at 940,
1170, 1300 and 1440 nm were used in combination with a fiber-optic reflection probe
for an improved sensor prototype as compared with the established application of a
tungsten halogen lamp. A dispersive monochromator-based spectrometer equipped
with an InGaAs linear array detector was employed. Spectra of renal biopsies with
diameters of one to two cm were investigated, applying principle component analysis
(PCA) and partial least squares discriminant analysis (PLS-DA) for kidney tumor
detection.
Novel imaging instrumentation has been developed, e.g., on the basis of a broadband NIR source such as a tungsten halogen lamp and a liquid-crystal tunable filter
(LCTF) for wavelength dispersion, realizing bandpass filtering without mechanical
movement. A recent review on NIR imaging for biological tissues, including chemometric tools for spectral data analysis, has been published by Ozaki et al. [50], so
that the reader is referred to that chapter for more information.
A very recent application dealing with multispectral imaging has been reported
for the histopathology of skin biopsies. This technique combines spectral resolution of spectroscopy with spatial resolution of imaging to show several merits for
biomedical applications. As mentioned above, infrared and Raman microscopy have
been employed for spectral histopathology, also in combination with conventional
tissue staining for tumor diagnostics. The biofingerprint interval within the lowwavenumber region has been mentioned already, but instrumentation for imaging will
require different photonics allowing only shorter wavelengths. Spreinat et al. [51]
described an imaging set-up using a Xenon lamp for NIR radiation, a monochromator
461
marker substances that are providing differences in absorption and scattering characteristics. Hemoglobin concentrations and its redox status, described by oxy- and
deoxygenated forms for providing oxygenation saturation, water and lipid content
have been listed for discrimination from healthy tissue. Tumor tissues were found
with significantly high levels of water and of hemoglobin because of higher vascularization, but also with less lipids and lower oxygen saturation due to cell hypermetabolism. Commonly, in tumorous tissue of breast cancer, abnormal proliferation of
cells will result in an angiogenesis with an increased number of blood vessels, eventually increasing the local blood Hb concentration. A recent breast cancer study has
been reported by Mehnati et al., who differentiated between normal and tumorous
breast tissue by local hemoglobin concentrations [48].
Cancerous tissues from pancreatic and colorectal tumors have been classified by
studying, in particular, first and second overtone bands of the lipid CH 2 stretching
vibrations, which could be intensified by first spectral derivative calculations; for
the characteristic wavenumber intervals and band assignments, see also Fig. 20.2.
Besides diagnosis, also margin location of cancerous tissue, especially intraoperative
malign tumor margin assessment is important for supporting the surgeon’s decisions
during operations. Furthermore, NIR spectroscopy is useful for therapeutics monitoring, especially after chemotherapy, radiation treatment, neoadjuvant and photodynamic therapy. For more details, the reader is again referred to our review [46]. A
recent instrumental development of an LED-based NIR sensor for kidney tumor diagnostics has also been published [49]. Four LEDs with emission band maxima at 940,
1170, 1300 and 1440 nm were used in combination with a fiber-optic reflection probe
for an improved sensor prototype as compared with the established application of a
tungsten halogen lamp. A dispersive monochromator-based spectrometer equipped
with an InGaAs linear array detector was employed. Spectra of renal biopsies with
diameters of one to two cm were investigated, applying principle component analysis
(PCA) and partial least squares discriminant analysis (PLS-DA) for kidney tumor
detection.
Novel imaging instrumentation has been developed, e.g., on the basis of a broadband NIR source such as a tungsten halogen lamp and a liquid-crystal tunable filter
(LCTF) for wavelength dispersion, realizing bandpass filtering without mechanical
movement. A recent review on NIR imaging for biological tissues, including chemometric tools for spectral data analysis, has been published by Ozaki et al. [50], so
that the reader is referred to that chapter for more information.
A very recent application dealing with multispectral imaging has been reported
for the histopathology of skin biopsies. This technique combines spectral resolution of spectroscopy with spatial resolution of imaging to show several merits for
biomedical applications. As mentioned above, infrared and Raman microscopy have
been employed for spectral histopathology, also in combination with conventional
tissue staining for tumor diagnostics. The biofingerprint interval within the lowwavenumber region has been mentioned already, but instrumentation for imaging will
require different photonics allowing only shorter wavelengths. Spreinat et al. [51]
described an imaging set-up using a Xenon lamp for NIR radiation, a monochromator