J.S. Maier1, A.J. Drauch1 and S.D. Stewart1 ... - ChemImage

bolivialodgeInternet and Web Development

Dec 14, 2013 (3 years and 7 months ago)

64 views

Raman Molecular Imaging: a Novel Technique for
Histopathological Evaluation of Tissue Specimens
J.S. Maier
1
, A.J. Drauch
1
and S.D. Stewart
1
.
1
Applications Science, ChemImage Corporation, Pittsburgh, PA, United States.
Introduction
We describe the use of Raman Molecular Imaging (RMI) as an
objective tool for the histopathological assessment of tissue. RMI
enables direct measurement of molecules in intact thin tissue
sections without the use of antibody-, or gene-based complex
biological reagents.
RMI builds on the demonstrated success of Raman spectroscopy
in the analysis of biological samples by integrating full field-of-view
digital imaging with high precision spectrally sensitive optics. Work to
date has shown feasibility for using this approach on tissue to deliver
information from imaging spectroscopy into the digital pathology
imaging workflow. In particular, preliminary work demonstrating
clinically significant distinction between progressive and indolent
prostate cancer [1] is currently in press for publication.
RMI brings a particular established tool from analytical chemistry,
Raman spectroscopy, into a digital imaging format to deliver images
where the contrast comes from the different molecules present in
the sample.
Raman spectroscopy is a nondestructive, analytical technique that
exploits the interaction of monochromatic light with the molecules
in a sample under investigation. When monochromatic light, usually
emitted from a laser, interacts with a sample, most of the light is
reflected back at the same wavelength (elastic scattering); however,
in a small percentage of the time, light interacts with molecules
by changing vibrational states which leads to a change in energy
causing a shift in the wavelength of light that is reflected back from
the sample (inelastic scattering). This change in energy, called a
Raman shift, is due to specific molecular vibrations, and results in
a specific fingerprint of the chemical structure within a molecule.
Figure 1
. RGB image of tissue sample indicating region of interest
for Raman molecular imaging.
Figure 2.
RGB image of region of interest for Raman imaging (left),
Grayscale Raman image at Raman shift of 1450 wavenumbers (right).
Design
Thin sections from formalin fixed, paraffin-embedded samples
were obtained in standard fashion and applied to aluminum coated
microscope slides. Paraffin was removed and digital Raman images
were acquired using a FALCON II
TM
Raman imaging system.
After Raman imaging measurements were made, samples were stained
to facilitate histological correlation between traditional H and E and
Raman imaging modalities. Figure 1 shows the H and E stained tissue.
Figure 2 shows the H and E stained tissue in a specific region of interest
along with the RMI signal in grayscale. Digital images were created
by process of constructing a reference library of spectra for selected
histological types (stroma, epithelium), and subsequent segmenting
using spectral mixture resolution. Analysis included subjective
comparison of Raman images to imagery of the stained sample.
Result
To demonstrate the molecular basis for the contrast in these images
the same Raman data set was masked using a segmented image
from the stained sample. Masking was performed to isolate pixels
associated with nuclei and pixels associated with tissue, but not
nuclei. The mean spectra from the two masked images are shown
below in Figure 3 along with the image segmented for nuclei and non-
nuclear material.
The difference spectrum, seen in Figure 4, has clear peaks which are
associated with nucleic acids in the published literature.
Nuclei
Not Nuclei
0.00
0.10
0.20
0.30
0.40
0.50
0.60
0.70
0.80
0.90
1.00
800
1000
1200
1400
1600
1800
Wavenumbers
Figure 3.
Segmented image distinguishing nuclei. Plot of mean
spectra from nuclei and non-nuclei.
Figure 4.
Difference between nuclei and non-nuclei image spectra
after baseline.
Table 1.
Reference table for band assignments.
[2]
Figure 6.
RGB image of region of interest with overlay of stroma and
epithelium as determined from spectral mixture resolution.
Data Analysis
Raman image data were processed using established methods for
spectral imaging datasets including noise reduction through spectral
smoothing, spectral truncation and baseline subtraction.
The data processed in this fashion were analyzed by applying a
method called spectral mixture resolution. In this method reference
spectra are used to determine a relative contribution of each reference
spectral trace to the measured spectrum at each pixel of the image.
Result
The result of this process is an image which depicts the relative
proportion of signal from each of the reference members. Figure 5
shows the results of this process mapped onto a gray scale. Figure
6 shows these images mapped to a brown color scheme and overlaid
with the H and E stained tissue image.
Figure 5.
Signal proportion images for stroma and epithelium
signatures.
Conclusion
RMI is a novel technique for digital image based analysis of
unstained thin sections of tissue. RMI carries information about
the molecular environment at the cellular and subcellular level in
tissues as manifested by accurate histological sectioning based
on comparison to existing, stain-based approaches. The molecular
information at the basis of the contrast in RMI may have correlation
with pathophysiology that is of clinical interest. Further study of RMI
in application to clinical questions of interest is warranted.
References
M. Tollefson et al, Raman spectral imaging of prostate cancer:
can Raman molecular imaging be used to augment standard
histopathology?,
BJUI
in press April 2010.
Z. Movasaghi et al, Raman spectroscopy of biological tissues,

Applied Spectroscopy Reviews
, 42: 493-541, 2007.
Acknowledgements

Supported by: Western Pennsylvania Prostate Foundation

Tissue provided by Allegheny Dept of Pathology Drs. Silverman
and Uihlein
Observed peak
Assignment Reference
784
780 cm
-1
Uracil based ring breathing mode
781 cm
-1
Cytosine/uracil ring breathing (nucleotide)
782 cm
-1
DNA
Thymine, cytosine, uracil

RNA

U,T,C (ring breathing modes in the DNA/RNA bases)
784/5 cm
-1
Phosphodiester; cytosine
785 cm
-1
U,T,C (ring breathing modes in the DNA/RNA bases)
Backbone O-P-O
786 cm
-1
DNA; O-P-O, cytosine, uracil,thymine
Pyrimidine ring breathing mode
787 cm
-1
Can be taken as a measurement for the relative quantity of nucleic acids present
Phosphatidylserine
788 cm
-1
C’
5
-O-P-O-C’
3
phosphodiester bands in DNA
DNA

O-P-O stretching DNA
1332
1325-30 cm
-1
CH
3
CH
2
wagging mode in purine bases of nucleic acids
1330 cm
-1
Typical phospholipids
Region associated with DNA & phospholipids
Collagen
Nucleic acids and phosphates
1332 cm
-1
-C stretch of Phenyl (1) and C
3
-C
3
stretch and C
5
-O
5
stretch CH
a
in-plane bend
1333 cm
-1
Guanine

1364
1355/7 cm
-1
Guanine (N
7
, B, Z-marker)
1359 cm
-1
Tryptophan
1360 cm
-1
Tryptophan
1361/2/3/5 cm
-1
Guanine (N
7
, B, Z-marker)
1365 cm
-1
Tryptophan
1367 cm
-1
n
s (CH
3
) (phospholipids)
1369 cm
-1
Guanine, TRP (protein), porphyrins,lipids
1485
1480 - 575 cm
-1
Amide II
(largely due to a coupling of CN stretching & in-plane bending of the N-H group, is not often used for
structural studies per se because it is less sensitive & is subject to interference from absorption bounds
of amino acid side chain vibrations)
1485 cm
-1
G, A (ring breathing modes in the DNA bases)
Nucleotide acid purine bases (guanine and adenine)
1485 cm
-1
NH
3
+
1487/8 cm
-1
Guanine (N
7
)
1488 cm
-1
Collagen
1490 cm
-1
DNA
Formalin peak appearing in the fixed normal and tumor tissues
1572
1573 cm
-1
Guanine, adenine, TRP (protein)
1575 cm
-1
Ring breathing modes in the DNA bases
G, A (ring breathing modes of the DNA/RNA bases)
1576 cm
-1
Nucleic acid mode
Nucleic acid modes indicating the nucleic acid content in tissues
1576/7 cm
-1
Guanine (N
3
)
Difference
0.00
0.10
0.20
0.30
0.40
0.50
0.60
0.70
0.80
0.90
1.00
sca
ledd
iffe
rence
800
1000
1200
1400
1600
1800
Wavenumbers
1572
1364
784
1332
1485
1.
2.