Biomedical Raman Spectroscopy
Jason T. Motz
Harvard Medical School and The Wellman Center for Photomedicine
Massachusetts General Hospital
11 April 2006
A New Type of Secondary Radiation
C. V. Raman and K. S. Krishnan, Nature, 121(3048): 501-502, March 31, 1928
If we assume that the X-ray scattering of the 'unmodified' type observed by Prof. Compton corresponds to the normal
or average state of the atoms and molecules, while the 'modified' scattering of altered wave-length corresponds to
their fluctuations from that state, it would follow that we should expect also in the case of ordinary light two types of
scattering, one determined by the normal optical properties of the atoms or molecules, and another representing the
effect of their fluctuations from their normal state. It accordingly becomes necessary to test whether this is actually
the case. The experiments we have made have confirmed this anticipation, and shown that in every case in which light
is scattered by the molecules in dust-free liquids or gases, the diffuse radiation of the ordinary kind, having the same
wave-length as the incident beam, is accompanied by a modified scattered radiation of degraded frequency.
The new type of light scattering discovered by us naturally requires very powerful illumination for its observation. In
our experiments, a beam of sunlight was converged successively by a telescope objective of 18 cm. aperture and 230
cm. focal length, and by a second lens was placed the scattering material, which is either a liquid (carefully purified
by repeated distillation in vacuo) or its dust-free vapour. To detect the presence of a modified scattered radiation, the
method of complementary light-filters was used. A blue-violet filter, when coupled with a yellow-green filter and
placed in the incident light, completely extinguished the track of the light through the liquid or vapour. The
reappearance of the track when the yellow filter is transferred to a place between it and the observer's eye is proof of
the existence of a modified scattered radiation. Spectroscopic confirmation is also available.
Some sixty different common liquids have been examined in this way, and every one of them showed the effect in
greater or less degree. That the effect is a true scattering, and secondly by its polarisation, which is in many cases
quire strong and comparable with the polarisation of the ordinary scattering. The investigation is naturally much more
difficult in the case of gases and vapours, owing to the excessive feebleness of the effect. Nevertheless, when the
vapour is of sufficient density, for example with ether or amylene, the modified scattering is readily demonstrable.
2
The Nobel Prize in Physics 1930
"for his work on the scattering of light and for the discovery of the effect
named after him"
Professor Sir C.V. Raman
1888-1970
3
First photographed Raman spectra
Bangalore, India
The Raman Effect: Inelastic Scattering
hni
h(ni-nR)
3
hni
2
1
0
Inelastic Scattering
• Emitted light has
decreased energy (i<R)
Virtual Level
Energy
• Energy transferred from
incident light to
molecular vibrations
3
2
1
0
Rayleigh
(elastic) Scattering
4
S1
Raman (inelastic)
Scattering
S0
difference in energy
Some Vibrations in Benzene
Intensity (CCD Counts)
Kekule
5
5
Chubby Checker
x 10
Breathing
4
4
3
2
1
0
400
600
800
1000
1200
1400
-1
Raman Shift (cm )
1600
1800
Evolution of Raman Spectroscopy
• 1928~1960
– Minor experimental advances
• 1960
– Invention of laser expands scope experiments
• 1980s: rapid technological advances
–
–
–
–
Fourier Transform spectroscopy
Charge Coupled Device (CCD) detectors
Holographic and dielectric filters
Near-Infrared (NIR) lasers
• Late 1980s1990s
– Biomedical investigations
– Advanced dispersive spectrometers
• 2000 
– In vivo application
– Optical fiber probes
– Non-linear spectroscopy
6
Outline
• The Raman Effect
– Theory
– Techniques & Applications
• Biomedical Raman Spectroscopy
– History
– Excitation wavelength selection
– Advantages of Raman spectroscopy
• Instrumentation
– Laboratory
– Clinical: optical fiber probes
• Case Study: Atherosclerosis
– Disease background
– Impact of Raman spectroscopy
• Frontiers
7
Classical Raman Physics
• Interaction between electric field of incident
photon and molecule
– Electric field oscillating with incident frequency vi:
Ei  E0 cos(2n i t )
– Induces molecular electric dipole (p):
p E
• Proportional to molecular polarizability, 
– ease with which the electron cloud around a molecule
can be distorted
– Polarization results in nuclear displacement
q  q0 cos  2n R t 
8
Classical Raman Physics
• For small distortions, polarizability is
linearly proportional to the displacement
  
  0  
 q0  ...
 q 0
Rayleigh Scattering
• Resultant dipole:
p   E   0 E0 cos  2n i t  
  
1
E0 q0 
 cos  2 n i  n R  t   cos 2 n i n R  t 
2
 q 0

Anti-Stokes Raman
9

Stokes Raman
Photo-Molecular Interactions
Scattering
100
80
Intensity
3
2
1
0
Rayleigh
n1
60
40
20 Anti-Stokes
0
-2000
Stokes
-1000
0
1000
Raman Shift (cm-1)
2000
2
Energy
Virtual Levels
1
0
n’1
E=hnR
2
1
0
AutoFluorescence
10
IR
Absorption
Rayleigh
Scattering
Stokes Anti-Stokes
Raman Scattering
NIR
Fluorescence
n0
Classical Raman Physics
• Raman scattering occurs only when the molecule is ‘polarizable’

0
dq
• Raman intensity  n4
– Classical dipole radiation
– Stokes shifted Raman is more intense than anti-Stokes by Boltzmann factor:
hn
I A  n i  n R   kTR

 e
I S  n i n R 
4
• Consistent with other scattering phenomena, often reported in terms
of cross-section ( [cm2]), or probability of scattering:
I  I 0 dz
– : density of molecules
– dz: pathlength
11
Characteristics of Raman Scattering
• Very weak effect
–
–
–
–
–
Only 1 in 107 photons is Raman scattered
NIR elastic scattering in tissue: 1/ s'  1mm
NIR absorption in tissue: 1/ a  10cm
Red absorption in tissue or water: 1/ a  5m
Raman scattering in tissue or water: 1/  R  3km
• True scattering process
– Virtual state is a short-lived distortion of the electron cloud
which creates molecular vibrations
–  < 10-14 s
• Strong Raman scatterers have distributed electron clouds
– C=C
– -bonds
12
Quantum Mechanics: Normal Modes
• Only certain vibrational frequencies and atomic
displacements allowed
– Linear molecules: 3N-5
– Non-linear molecules: 3N-6
• Examples
–
–
–
–
Stretching between 2 atoms
Symmetric and asymmetric stretching with 3 atoms
Bending amongst 3 atoms
Out of plane deformations
• Vibrational energies are sensitive to
–
–
–
–
–
–
13
Atomic mass
Molecular structure and geometry
Bond strength
Bond order
Environment
Hydrogen bonding
Units & Dimensional Analysis
• Spectroscopic frequencies reported in
wavenumbers [cm-1], proportional to
transition energy :
E n 1
n  
hc c 
E  hn
c  n
• Raman frequencies are independent of
excitation wavelength and reported as shifts
– Wavenumbers relative to excitation frequency:
nR 
14
1
i

1
R
Units & Dimensional Analysis
• Example
– NIR excitation at 830 nm: 12,048 cm-1
– Typical Raman shift: n R ~1000 cm-1
• R = 905 nm
– Sharp biological Raman linewidths ~10 cm-1 FWHM
• R= 0.69 nm
15
Raman Spectrum of Cholesterol
16
Hanlon et al. “Prospects for in vivo Raman spectroscopy,” Phys Med Biol 45: R1 (2000)
Specific Raman Spectroscopic Techniques
• Non-resonant Raman spectroscopy
– Visible
– Near-infrared
•
•
•
•
•
•
•
(UV) Resonance Raman spectroscopy
Raman microscopy/imaging
Fiber optic sampling
Time resolved (pulsed) Raman spectroscopy
High-wavenumber Raman spectroscopy
SERS: Surfaced Enhanced Raman Spectroscopy
Non-linear Raman spectroscopy
– CARS: Coherent Anti-Stokes Raman Spectroscopy
17
General Applications of Raman Spectroscopy
•
•
•
•
•
•
•
•
18
Structural chemistry
Solid state
Analytical chemistry
Applied materials analysis
Process control
Microspectroscopy/imaging
Environmental monitoring
Biomedical
Outline
• The Raman Effect
– Theory
– Techniques & Applications
• Biomedical Raman Spectroscopy
– History
– Excitation wavelength selection
– Advantages of Raman spectroscopy
• Instrumentation
– Laboratory
– Clinical: optical fiber probes
• Case Study: Atherosclerosis
– Disease background
– Impact of Raman spectroscopy
• Frontiers
21
History of Biological Raman Spectroscopy
• 1970: Lord and Yu record 1st protein spectrum from lysozyme using
HeNe excitation
• Evolution to NIR excitation
– Decreased fluorescence
– Increased penetration (mm)
• 1980s:
– FT Raman with Nd:YAG and cooled InGaAs detectors
• long collection times (30 min)
– Clarke (1987-1988): visible excitation of arterial calcium hydroxyapatite
and carotenoids
• 1990s, advances in:
–
–
–
–
–
22
Lasers
Detectors
Dispersive spectrometers
Filters
Chemometrics
UV, Visible, and NIR Excitation
Raman signals have a constant shift  can vary excitation wavelength
• UV: resonance enhanced, R<F, photo damage, low penetration
• Visible: Raman  -4, fluorescence overlaps with Raman signal
• NIR: low fluorescence, deep penetration, Raman  -4
23
UV, Visible, and NIR Applications
• UVRR
– Biological macromolecules: nucleic acids, proteins, lipids
– Organelles, cells, micro-organisms, bacteria, phytoplankton
neurotoxins, viruses
– Clinically limited: photomutagenicity
• Visible
– Cells (minimal fluorescence)
– DNA in chromosomes, pigment in granulocytes and
lymphocytes, RBCs, hepatocytes
– First artery studies: hydroxyapatite and carotenoids
(Clarke 1987, 1988)
• NIR
– Hirschfeld & Chase, 1986: FT-Raman
– Tissue: artery, cervix, skin, breast, blood, GI, esophagus,
brain tumor, Alzheimer’s, prostate, bone
24
Raman Spectra: Fingerprinting a Molecule
• Raman spectra are
molecule specific
• Spectra contain information
about vibrational modes of
the molecule
• Spectra have sharp
features, allowing
identification of the
molecule by its spectrum
25
Examples of analytes found in blood
which are quantifiable with Raman spectroscopy
Spectroscopic Advantages of NIR Raman
• Narrow vibrational bands are chemical
specific and rich in information
• Freedom
to choose excitation
wavelength
1.0
CH bend
– minimize
unwanted
0.8
Phosphate
stretch tissue fluorescence
Carotenoid3
2
– optimize
sampling depth
0.6
1
106
Amide III
C=C
– utilize
0.4 H2OCCD technology Visible
0
104
n1
0.2
102 0.0
1 -0.2
Melanin
Sterol ring
-0.4
10-2
100
26
Energy
Intensity (a.u.)
Molar extinction coefficient
 (10-3M-1cm-1); H20 (cm-1)
2
800
350 mW
NIR
VirtualExcitation
Level
1 min
3
HbO
Phenylalanine
Ester linkage
2
HbO
1
1000
1200
1400
1600
1800
200
500
1000
0
-1
n0
Fluorescence
Raman
(inelastic)
Raman Shift
(cm ) (nm)
Wavelength
Scattering
2 000
Diagnostic Advantages of Raman Spectroscopy
• Wavelength selection
• No biopsy required
• Directly measures molecules
– Small concentrations
– Chemical composition
– Morphological analysis
• Quantitative analysis
• In vivo diagnosis
27
Outline
• The Raman Effect
– Theory
– Techniques & Applications
• Biomedical Raman Spectroscopy
– History
– Excitation wavelength selection
– Advantages of Raman spectroscopy
• Instrumentation
– Laboratory
– Clinical: optical fiber probes
• Case Study: Atherosclerosis
– Disease background
– Impact of Raman spectroscopy
• Frontiers
28
Laser Sources for Raman Spectroscopy
Source
29
Wavelength (nm)
Ar+
488.0, 514.5
Kr+
530.9, 647.1
He: Ne
632.8
Ti: Al2O3 (cw)
720-1000
Diode (InGaAs)
785, 830
Nd:YAG
1064
In Vitro Experimental Raman System
Argon ion
pump laser
Dichroic
beam-splitter
Ti: sapphire
laser
NIR
excitation
(830 nm)
CCD
f/4
Spectrograph
Confocal
pinhole
Band-pass filter
Notch filter
CCD Camera
30
Motorized translation stage
Collimating lenses
In Vitro Turbid Liquid Analysis
ppm accuracy for precise quantitative measurements
31
Hanlon et al. “Prospects for in vivo Raman spectroscopy,” Phys Med Biol 45: R1 (2000)
Clinical Raman Systems
830 nm
shutter
bandpass
filter
Diode Laser
holographic grating
notch filter
CCD
32
Current Raman Instrumentation
• Laser diodes
– Compact
– Stable narrow line
– NIR
• High throughput spectrographs (f/1.8)
• Holographic elements
– Bandpass filters (eliminates spontaneous emission of lasing medium)
– Notch filters (106 rejection of Rayleigh scattered laser line)
– Large area, highly efficient transmission gratings
• CCD detectors
– High QE (back-thinned, deep-depletion)
– Low noise (LN2 cooled)
– Multichannel detection
• High throughput, filtered fiber optics probes
• NIR FT and scanning PMT systems no longer useful
33
Early in vivo Data
• Simple 6-around-1 optical fiber probe
• 100 mW excitation, 3 second collection
1500
a
500
5000
b
0
-500
-5000
-1500
34
*
-2500
-10000
-3500
2000 1600 1200 800
Raman shift
-15000
2000 1600 1200
Raman shift
800
Optical Fiber Probes
Problems
Fiber background  NA2
1. Fiber background
•
•
Distorts signal
Adds shot-noise
2. Low signal collection
•
•
35
Raman effect is weak
Tissue is highly diffusive
Solution #1: Reduce Fiber Background
Fiber background produced equally in excitation and collection fibers
• Excitation laser of power Po generates Raman scattered light from tissue
• Posxlx: fraction of laser light Raman scattered and transmitted by excitation fiber*
• Bx=Posxlxes: Raman background detected from excitation fiber
– es: fraction of light elastically scattered (and collected) from sample
• Poes: intensity of scattered excitation light gathered by collection fibers
• Bc=Poessclc: intensity of background generated and transmitted by collection
fibers*
• BT=Bx+Bc=Poes(sxlx+sclc)
Excitation laser
*
NA2
Tissue Raman
x
c
x
Fiber background
From McCreery RL
“Raman Spectroscopy for Chemical Analysis,” 2000.
36
Tissue Sample
Tissue Sample
c
Filter Transmission
100
Transmission (%)
80
Region of Interest
60
Collection Filter
Excitation Filter
40
20
0
0
37
500
1000
-1
Raman Shift (cm )
1500
Solution #1: Filtering
Solutions
Problems
1. Fiber background
•
•
Distorts signal
Adds shot-noise
2. Low signal collection
•
•
Raman effect is weak
Tissue is highly diffusive
1. Micro-optical filters
•
•
2. Optimize optical design
•
•
•
38
Short-pass excitation filter
Long-pass collection filter
Characterize distribution
of Raman light in tissue
Define optimal geometry
Design collection optics
Excitation Light Diffusing Through Tissue
Experimental
1 mm
Monte Carlo
39
Solution #2: Optical Design
Problems
1. Fiber background
•
•
Distorts signal
Adds shot-noise
2. Low signal collection
•
•
Raman effect is weak
Tissue is highly diffusive
Solutions
1. Micro-optical filters
•
•
2. Optimize optical design
•
•
•
40
Short-pass excitation filter
Long-pass collection filter
Characterize distribution
of Raman light in tissue
Define optimal geometry
Design collection optics
Raman Probe Design Goals
• Restricted geometry for clinical use
– Total diameter ~2mm for access to coronary arteries
– Flexible
– Able to withstand sterilization
• Designed to work with 830 nm excitation
• High throughput
–
–
–
–
41
Data accumulation in 1 or 2 seconds
Safe power levels
SNR similar to open-air optics laboratory system
Accurate application of models
Raman Probe Design
collection fibers
aluminum jacket
excitation fiber
1 mm
long-pass
filter tube
metal
sleeve
short-pass
filter rod
retaining
sleeve
42
2 mm
ball lens
0.55
0.70
1.75 mm
Single Ring Probe has 15 Fibers
Motz et al. Appl Opt 43: 52 (2004)
Calcified Aorta
800
Single Ring, sapphire lens
Lab system
Intensity (CCD counts/mW/s)
700
600
500
400
300
200
100
0
0
43
500
1000
Raman Shift (cm-1)
1500
Outline
• The Raman Effect
– Theory
– Techniques & Applications
• Biomedical Raman Spectroscopy
– History
– Excitation wavelength selection
– Advantages of Raman spectroscopy
• Instrumentation
– Laboratory
– Clinical: optical fiber probes
• Case Study: Atherosclerosis
– Disease background
– Impact of Raman spectroscopy
• Frontiers
44
The Burden of Cardiovascular Disease†
• 71,300,000 people in United States afflicted
• 910,600 deaths per year
– 1 out of every 2.7 deaths
• Coronary artery disease claims 653,000 lives annually
– 1 out of every 5 deaths
– Economic cost: greater than $142.5 billion
45
†
American Heart Association, Heart and Stroke Statistics-2006 Update
Arterial Anatomy
Normal
Mildly Atherosclerotic Plaque
Ruptured Plaque
Media
Lumen
Fibrous
Cap
Atheroma
Intima
T
NC
Adventitia
• Intima: innermost layer of arterial wall
• composed of a single layer of endothelia cells in normal artery
• region of artery involved in atherosclerotic disease
• Media: arterial layer composed primarily of smooth muscle cells
• constricts and dilates to control blood flow
• in large arteries (e.g. aorta) this layer is largely composed of elastin
• Adventitia: outermost layer of arterial wall
46
• connective tissue and fat
T: thrombus
NC: necrotic core
Some Current Challenges in Cardiology
• Evaluation and development of therapeutics
• Etiology of atherosclerosis
• Mechanisms of re-stenosis
– Post-angioplasty
– Transplant vasculopathy
• Detection of vulnerable atherosclerotic plaques
– Prediction/prevention of cardiac events
47
Vulnerable Plaques
• Account for majority of sudden cardiac death
• Frequently occur in clinically silent vessels
– <50% stenosis
• Effective treatments unknown
• Characterized by:
–
–
–
–
–
Biochemical changes
Foam cells
Lipid pool
Inflammatory cells
Thin fibrous cap (<65 m)
• Currently undetectable
48
Standard Diagnostic Techniques
• Angiography
– Severity of stenosis, thrombosis, dense calcifications
– Provides no biochemical information
• Angioscopy
– Surface features of plaque, including color
– No information of sub-surface features
• Histopathology
– Biochemical and morphological information
– Requires excision of tissue
49
Emerging Diagnostic Techniques
• Magnetic resonance imaging
• External ultrasound
• Positron emission tomography
• Electron beam computed tomography
• Thermography
• Elastography
• Intravascular ultrasound
• Optical coherence tomography
50
Non-Invasive
Spectroscopic Diagnostic Techniques
• NIR Absorption spectroscopy
– Inhibited by water absorption
– Broad spectral features
• Fluorescence spectroscopy
– Limited chemical information
– Broad spectral features
• Raman Spectroscopy
– Quantitative biochemical information
– Morphological analysis
51
In Vitro Experimental Methods
• Macroscopic Raman Spectroscopy
– 1 mm3 volumes of excised tissue are examined
• 100-350 mW excitation with 830 nm laser light
• 10 - 100 s collection times
– Comparison with histopathology for disease classification
– Principal Component Analysis
52
Raman Spectral Pathology of Atherosclerosis
1.0
0.8
0.6
• Ca hydroxyapatite
• proteins
0.4
0.2
0.0
-0.2
1.2
Intensity (a.u.)
1.0
0.8
0.6
0.4
• cholesterol
• -carotene
• proteins
0.2
0.0
-0.2
-0.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0
-0.2
-0.4
-0.6
53
800 1000 1200 1400 1600 1800
800 1000 1200 1400 1600 1800
• collagen
• elastin
• actin
800 1000 1200 1400 1600 1800
-1
Image from medstat.med.utah.edu/WebPath/webpath.html
Raman Shift (cm )
Raman Spectral Modeling
• Goal: Diagnose disease by analyzing the complex macroscopic
spectra (R) obtained from biopsy samples
• Strategy: Develop a library of microscopic or chemical basis
spectra (B) that compose the macroscopic data
• Implementation: Use ordinary least squares fitting to determine a
weighted linear combination of basis spectra to evaluate
the biopsies
R()artery = wcollagenB()collagen+
wcholesterolB()cholesterol+
wcalcificationB()calcification+…
54
Potential Features for Spectral Identification
56
Collagen
Elastin
Actin
Adventitial fat
-carotene
Foam cells
Cholesterol
Necrotic core
Calcification
Hemoglobin
Fibrin
In Vitro Experimental Methods
• Macroscopic Raman Spectroscopy
– 1 mm3 volumes of excised tissue are examined
• 100-350 mW excitation with 830 nm laser light
• 10 - 100 s collection times
– Comparison with histopathology for disease classification
– Principal Component Analysis
• Confocal Microscopic Raman Spectroscopy
– ~(2x2x2) m3 sampling volume of microscopic structures
• 100 mW excitation of 6 m thick sections with 830 nm laser light
• 10-360 s collection times
– Development of morphological model
– Spectroscopic mapping of tissue sections
57
Atherosclerosis In Vitro: Confocal Microscopy
Foam Cell
0.6
0.6
0.4
0.4
0.2
0.0
0.2
0.0
-0.2
-0.2
-0.4
-0.4
0.1 mm
800
1000
1200
Elastic Lamina
0.8
Intensity (a.u.)
Intensity (a.u.)
0.8
1400
1600
1800
800
-1
Raman Shift (cm )
1000
1200
1400
1600
1800
-1
Raman Shift (cm )
0.8
Smooth Muscle Cell
Intensity (a.u.)
0.6
0.4
0.2
0.0
-0.2
-0.4
800
1000
1200
1400
1600
1800
-1
58
Intima
Media
Adventitia
Raman Shift (cm )
~(2x2x2) m3 Sampling Volume
Coronary Artery Morphological Structures
Smooth Muscle Cell
Intensity (a.u.)
-Carotene Crystal
Foam Cell/Core
Adventitial Fat
Elastic Lamina
Calcification
Cholesterol Crystal
Collagen
800
1000
1200
1400
1600
1800
-1
Raman Shift (cm )
59
Buschman HPJ, et al. Cardiovascular Pathology 10(2), 69-82 (2001)
Morphological Model of Coronary Arteries
Normal Coronary Artery
Intensity (a.u.)
Intensity (a.u.)
Non-Calcified Plaque
Calcified Plaque
800
1200
1600
800
Raman shift (cm-1)
1600
Raman shift (cm-1)
Intensity (a.u.)
Macroscopic Data
Microscopic
Model Fit
800
1200
Raman shift (cm-1)
60
1200
1600
Residual
Buschman HPJ, Motz JT, et al. Cardiovascular Pathology 10(2), 59-68 (2001)
Morphological Assay of Coronary Arteries
Intensity (a.u.)
Normal Coronary Artery
1.2
1.0
0.8
0.6
0.4
0.2
0.0
-0.2
-0.4
-0.6
-0.8
-1.0
-1.2
800
1000
1200
1400
1600
1800
Structure
Collagen
Cholesterol
Calcification
Elastic Lamina
Fat
Foam Cell / Core
-Carotene
Smooth Muscle
Contribution
20%
6%
0%
6%
38%
0%
3%
27%
Structure
Collagen
Cholesterol
Calcification
Elastic Lamina
Fat
Foam Cell / Core
-Carotene
Smooth Muscle
Contribution
39%
10%
11%
0%
28%
0%
0%
12%
-1
Raman Shift (cm )
Intensity (a.u.)
Mildly Calcified Plaque
1.2
1.0
0.8
0.6
0.4
0.2
0.0
-0.2
-0.4
-0.6
-0.8
-1.0
-1.2
800
61
1000
1200
1400
1600
-1
Raman Shift (cm )
1800
Diagnostic Database
• 165 intact samples from explanted hearts
– Biopsies snap frozen until examination
– 2 data sets
• Calibration (n=97)
• Prospective (n=68)
• Three tissue categories determined by histopathology
– Non-atherosclerotic
– Non-calcified plaque
– Calcified plaque
62
Coronary Artery Disease Classification:
A Prospective Study
1.0
Normal Artery
Non-Calcified Plaque
Calcified Plaque
Punctate Calcification
Calcification
0.8
0.6
0.4
3  error zone
0.2
0.0
0.0
0.1
0.2
0.3
0.4
0.5
0.6
CholesterolNCR + Lipid CoreNCR
63
Buschman HPJ, Motz JT, et al. Cardiovascular Pathology 10(2), 59-68 (2001)
0.7
0.6
0.5
0.3
0.2
0.1
ch
ol
es
te
ro
l
ca
lci
f ic
at
io
n
el
as
tic
la
m
in
a
ad
fo
ve
am
nt
ita
ce
lf
ll /
at
ne
cr
ot
ic
sm
co
oo
re
th
m
us
be
cle
ta
-c
ce
ar
ll
ot
en
e
cr
ys
ta
l
0
co
64
Normal Coronary Artery
Non-Calcified Plaque
Calcified Plaque
0.4
lla
ge
n
Relative Fit Coefficient (Mean +/- SEM)
Raman Morphometry of Coronary Artery
Buschman HPJ, et al. Cardiovascular Pathology 10(2), 59-68 (2001)
Clinical Data: Methods
• Peripheral vascular surgery
– Femoral bypass
– Carotid endarterectomy
• Laser power calibration set with Teflon
– ~100 mW (82-132mW)
• OR room lights turned off as during angioscopy
• Spectra collected for a total of 5 seconds
– 20 accumulations of 0.25s each
– Probe held normal to arterial wall
• Analysis of 1s and 5s data
– Additional model components: sapphire, epoxy, water, HbO2
All data presented are integrated for only 1 second
65
Clinical Data: Intimal Fibroplasia (Normal)
0.4 mm
1
Data
Fit
Residual
Intensity (a.u.)
0.5
0
-0.5
-1
800
66
1000
1200
1400
Raman Shift (cm-1)
1600
1800
Motz JT et al., J Biomed Opt 11(2): 021003
Clinical Data: Atheromatous Plaque
1
Data
Fit
Residual
1.8 mm deep
calcification
Intensity (a.u.)
0.5
0
0.4 mm
0.1 mm
-0.5
800
67
1000
1200
1400
Raman Shift (cm-1)
1600
Motz JT et al., J Biomed Opt 11(2): 021003
1800
Clinical Data: Calcified Plaque
50 m
1
Data
Fit
Residual
Intensity (a.u.)
0.5
0
-0.5
800
68
1000
1200
1400
Raman Shift (cm-1)
1600
1800
Motz JT et al., J Biomed Opt 11(2): 021003
Clinical Data: Ruptured Plaque
1
Data
Fit
Residual
Intensity (a.u.)
0.5
0
-0.5
-1
-1.5
800
1000
1200
1400
Raman Shift (cm-1)
1600
1800
0.1 mm
0.4 mm
69
Motz JT et al., J Biomed Opt 11(2): 021003
Clinical Data: Thrombotic Plaque
0.4 mm
1
Data
Fit
Residual
Intensity (a.u.)
0.5
0
-0.5
-1
0.1 mm
-1.5
800
70
1000
1200
1400
Raman Shift (cm-1)
1600
1800
Motz JT et al., J Biomed Opt 11(2): 021003
Clinical Data: Representative Analysis
Model Component
Intimal
Atheromatous Calcified Ruptured Thrombotic
Plaque
Plaque
Plaque
Plaque
Fibroplasia
Collagen (%)
9
0
7
0
0
Cholesterol (%)
0
44
2
27
14
Calcification (%)
0
16
71
1
12
Elastic Lamina (%)
0
4
3
0
0
Adventitial Fat (%)
50
13
0
1
0
Lipid Core (%)
13
16
0
0
0
-Carotene (%)
0
7
4
23
13
28
0
12
47
61
3
0
0
13
27
Smooth Muscle (%)
Hemoglobin (a.u.)
71
Raman Spectroscopy in Cardiology
• Clinical contributions
– Detection of vulnerable plaque (Prediction and Prevention)
• Collagen content  fibrous cap thickness
• Chemical composition of plaques
• Identification of mechanical instabilities
– Evaluation and selection of interventional methods
• Drug therapy
• Restenosis of bypassed vessels
– Guidance for laser ablation therapy
• Basic science
– Etiology: monitoring of chemical and morphological
changes during disease progression
– Differences in diabetic atherosclerosis
72
Application To Other Diseases
Normal Breast Tissue
1
1
Data
Fit
Residual
0.5
Malignant Breast Tumor
Data
Fit
Residual
Intesnity (a.u.)
Intesnity (a.u.)
0.5
0
0
-0.5
-1
-0.5
800
1000
1200
1400
Raman Shift (cm-1)
1600
1800
-1.5
800
1000
1200
1400
Raman Shift (cm-1)
100 mW excitation, 1 second collection
73
1600
1800
Outline
• The Raman Effect
– Theory
– Techniques & Applications
• Biomedical Raman Spectroscopy
– History
– Excitation wavelength selection
– Advantages of Raman spectroscopy
• Instrumentation
– Laboratory
– Clinical: optical fiber probes
• Case Study: Atherosclerosis
– Disease background
– Impact of Raman spectroscopy
• Frontiers
74
Frontier Investigations: High-Wavenumber Raman
Cholesterol
Advantages
• Higher Raman signal
• Lower fluorescence
• No fiber background
• Distinguishes cholesterol esters
75
Disadvantages
• Broader lineshapes
• Smaller spectral region
• Mostly limited to lipids
• No calcification signalc
www.sigma.com
High-Wavenumber Raman Spectroscopy
Normal Bladder
DNA
Glycogen
Collagen
Actin
High-Wavenumber
Cholesteryl
palmitate
Cholesteryl linoleate
H&E
lp: lamina propria
u: urothelium
Triolein
76
Koljenovic S et al., J Biomed Opt 10(3): 031116 (2005)
Frontier Investigations: Pulsed Excitation
Remitted Intensity with Pulsed Excitation
1
0.9
0.8
0.7
Remitted Intensity with Pulsed Excitation
1
0.6
0.9
0.5
0.8
0.4
0.7
Normalized Power
Normalized Power
• 80 MHz repetition rate
• 2 ns fluorescence lifetime
• Rayleigh scattering ~t-3/2
• Raman scattering ~t-1/2
Laser Pulse
Raman
Rayleigh
Fluorescence
0.3
0.2
0.1
0.6
0.5
0.4
0.3
Laser Pulse
Raman
Rayleigh
Fluorescence
0.2
0
77
0
5
10
15
20
Time (ns)
250.1
0
0
30
35
10
40
20
30
Time (ps)
40
50
60
Decay kinetics based on work of Everall N et al., Appl Spectrosc 55(12): 1701 (2001)
Frontier Investigations: Pulsed Excitation
78
Martyshkin DV et al., Rev Sci Instr 75(3): 630 (2004)
Conclusions
• Raman spectroscopy ‘fingerprints’ molecules by
characterizing interactions between photons and
molecular vibrations
• Near-infrared excitation is preferred for
biomedical applications
• Recent optical fiber probe developments allow
accurate real-time analysis in vivo
• New areas of research are promising for
widespread clinical application
79
References
• McCreery RL. Raman Spectroscopy for Chemical Analysis.
Wiley-Interscience, New York, 2000.
• Ferraro JR, Nakamoto K, and Brown CW. Introductory
Raman Spectroscopy 2nd ed. Academic Press, Boston, 2003.
• Hanlon EB, et al. “Prospects for in vivo Raman
spectroscopy,” Phys Med Biol 45: R1-R59 (2000).
• Mahadevan-Jensen A and Richards-Kortum R. “Raman
spectroscopy for the detection of cancers and precancers,”
J Biomed Opt 1:31-70 (1996).
• Utzinger U and Richards-Kortum R. “Fiber optic probes for
biomedical spectroscopy,” J Biomed Opt 8: 121-147 (2003).
80