Download Light Scattering Spectroscopy

University of Cyprus
Biomedical Imaging and Applied Optics
Light Scattering Spectroscopy
Introduction
LSS – How did it start?
• What is the problem?
• Gastroenterologist: “Problem:
need real-time in situ cancer
detection”
• Want compatibility with
endoscopes
• Prefer non-invasive technique
(optical detection of cancer
with “magic” laser?)
• Must be affordable, user
friendly
• Optical Biopsy:
• Noninvasive Detection of
Cancer with light
Optical “biopsy”
biopsy
Surgical biopsy
2
Introduction
• Optical Biopsy
• Noninvasive Æ Optical
detection
• Endoscopic compatibility Æ
Fib
Fiber-optic
ti mediation
di ti
• Diagnosis Æ Optical
spectroscopy
• Various spectroscopies
can be used for optical
tissue diagnostics
•
•
•
•
•
Auto-fluorescence
Exogenous-drug fluorescence
Raman
Absorption & FTIR
Elastic scattering
The internist's dream: smart colonoscope
Smart
Spectrometer
S
t
t
Illumination
source
(magic laser)
Optical “biopsy”
biopsy
3
But what does a pathologist look for?
• But what does a pathologist look for?
• Architectural changes
• Cancer Characteristics
Shape of cell
Shape
p of nucleus
Ratio of nucleus to total cell volume
Chromatin distribution
Structure of organelles
PLEOMORPHISM (variations in
nuclear size and DNA density)
• Cell density and distribution
Normal cells
•
•
•
•
•
•
Cancer cells
• This information is available from
elastic optical-transport data
• Elastic scattering
• Reflection / transmission
• Absorption
Ab
ti
• Elastic Scattering
• Spectral and angular patterns of
li ht scattering
light
tt i d
depend
d on th
the size
i
and structure of the scattering
particle
4
Mie Theory
• Gustav Mie, 1908
• Particles
P ti l comparable
bl or larger
l
than
th the
th wavelength
l
th
• Why does wavelength dependence of scattering
change with variations in microscopic tissue
morphology?
⎡ E&s ⎤ e − jk ( r − z )
⎢ E ⎥ = − jjkr
⎣ ⊥s ⎦
⎡ S2
⎢
⎣ S4
S3 ⎤ ⎡ E&i ⎤
S1 ⎦⎥ ⎢⎣ E⊥i ⎥⎦
At some distance away from the particle:
⎡ S2 2
⎡ I&s ⎤
⎢ I ⎥ = constant ⎢
⎢⎣ 0
⎣ ⊥s ⎦
S1 = ∑
q
2q + 1
(a π + b τ )
q(q + 1) q q q q
S2 = ∑
q
0 ⎤ ⎡ I &i ⎤
⎥
⎥
2 ⎢
S1 ⎥⎦ ⎣ I ⊥i ⎦
2q + 1
(aq τ q + bq π q )
q(q + 1)
aq and bq are complex
p
expressions
p
invoking
g
spherical Bessel functions
πq and τq are defined as
Pn1
πn =
sinθ
Steven Jacques, Oregon Graduate Center
and
dPn1
τn =
dθ
where Pn1 is the Legendre polynomial
5
Mie Theory
• If you really love the math …
jn = spherical Bessel functions
x = α = 2πn
2 noa/λ the “size
“si e parameter”
hn(1) = spherical Hankel functions
m = Ni/No = ni/no if no absorp.
[F(x)]’ means differentiation with respect to the argument
Bohren & Huffman, Absorption and Scattering lf Light by Small Particles, Wiley Science
6
Light Scattering Spectroscopy
• Wavelength Dependence
• Th
The scattering
tt i cross section
ti off the
th nuclei
l i
exhibits a periodicity with wavelength
2
⎛ sin
⎞
i
2
δ
/
λ
sin
i
δ
/
λ
⎛
⎞
(
)
(
)
2
σ s ( λ , l ) = l ⎜1 −
+⎜
⎟ ⎟
2 ⎜
δ /λ
δ
/
λ
⎝
⎠ ⎟⎠
⎝
π
Scattering on
nuclei
δ = π l (nn − 1)nc
Epithelium
where
nc is the refractive index of cytoplasm
n is the refractive index of the nuclei relative
to that of cytoplasm.
Connective Tissue
Perelman et al. Phys Rev Let 80: 627-630 (1998); Backman et al. Nature 406: 35-36 (2000)
7
Light Scattering Spectroscopy
• Wavelength dependence
• With diffuse
diff
reflectance,
fl t
allll particle
ti l sizes
i
look
l k white
hit
0.033μm 0.3μm 0.5μm 2.9μm 4.5μm 9.6μm 9.9μm
8
Light Scattering Spectroscopy
• Wavelength Dependence
• As d ↑
d=1 μm
• more rapid oscillations
• spacing of peaks Æsize of
scatterer
• depth of modulation
Ænumber of such
scatterers
• More often Æ mixture of
scatterers
Normalized
d Qbacksc
• Characteristics
d=2 μm
d=4 μm
• Superposition of spectra
400
500
600
700
800
Wavelength (nm)
9
Light Scattering Spectroscopy
• Wavelength Dependence
broadband
polarized
illumination
normal colon cells
p
polarization-resolved
detection
Perelman et al. Phys Rev Let 80: 627-630 (1998); Backman et al. Nature 406: 35-36 (2000)
cancerous cells
10
Light Scattering Spectroscopy
• Polarization Contrast
0.25
• Separate single vs. multiple scattering
Sq. metaplasia
0.20
parallel
0 15
0.15
perpendicular
HSIL
0.10
Analyzer
Polarizer
0.05
400
500
600
700
wavelength (nm)
0.15
Cells
0.10
0.05
Diffusive Layer
0.00
-0.05
0.05
400
500
600
700
wavelength (nm)
11
Principles of LSS Imaging
Polarizer
Mirror
Source
Filters
Beam
Splitter
CCD
Analyzer
Tissue
12
Principles of LSS Imaging
λn
λn
λ3
λ2
λn
λ3
λ3
λ2
λ1
λ2
λ1
I||
I⊥
Backscattering Signal
pixel 25 x 25 μm
λ1
Δ
I
0.06
0.04
d, σ, n
0.02
0
400
500
600
700
Wavelegth, nm
13
Principles of LSS Imaging
• LSS Imaging of Colon Adenoma: Nuclear enlargement
Adenoma
0
0.5
Enlarged
Nuclei, %
mm
1
1.5
0
0.5
1
1.5
mm
d, normal
σ, normal
d, adenoma
σ, adenoma
Morphometry
5.6 ± 0.20
1.01
7.44 ± 0.23
1.59
40-50
30-40
20-30
20
30
10-20
Normal
LSS
5.7 ± 0.13
0.82
7.67 ± 0.40
2.19
Backman et al. Nature Medicine 7: 1245-1248, 2001
14
Light Scattering Spectroscopy
λ =0.630 μm
• Angularly-resolved scattering
• Angular
A
l distribution
di t ib ti has
h
interferometric (oscillatory)
behavior as well
• As d ↑
• more rapid oscillations
d
n2
n1
Normaliz
zed Magnitud
de of Scatterin
ng
d =0.6
0 6 μm
d =1.2 μm
d =2.4 μm
0
45
90
135
180
Angle (deg)
15
Angular LSS Imaging
Source
Filter
Polarizer
Beam
Splitter
CCD
f
f
Polarizer
Tissue
16
Angular LSS Imaging
• Angular LSS Studies: Experiment with T84 Cells
Normal Mesothelial Cells
0.0045
Mean diameter
diameter, μm
Standard deviation, μm
Morphometry
LSS
10 4
10.4
10 5
10.5
1.35
1.52
Backman et al. IEEE J Sel Top Quant Electron, 7: 887:893, 2001
LSS signal,, a.u.
Cell Nuclei
LSS signal,, a.u.
0.0045
N(d) ~ d - β, β = 2.7 vs. 2.2
0 003
0.003
0.0015
0
450
Tumor T84 Cells
550
650
wavelength, nm
Fitting parameter β = 2.7
0 003
0.003
0.0015
0
450
550
650
wavelength, nm
Fitting parameter β = 2.2
17
Summary of LSS
• LSS contrasts:
• Polarization: single vs. multiple scattering
• Angle: small vs. large particles
• Spectrum: size and refractive index
• Advantages:
• Strong signal - allows use of lower cost components
components.
• Sensitive to important chromophores that are not
fluorescent: e.g., hemoglobin.
• Sensitive to both tissue structure and biochemistry.
• can distinguish different normal tissues.
• Disadvantages:
Di d
t
• May not be sensitive to subtle biochemical changes
preceding structural changes.
18
Schematic of simple LSS system
spectrometer
Light
source
19
Optical fiber probes
20
Optical biopsy for breast cancer
• Application #1:
• Improve reliability and
capability of transdermal
needle diagnosis (increase
th volume
the
l
off sensitivity
iti it
compared to FNA).
• Application #2:
• Provide a real-time in-situ
diagnostic tool for tumor
margins during breastbreast
conserving surgery.
• Application #3:
• Provide a real-time in-situ
assessment of the
“sentinel” lymph node
d i surgery.
during
21
Optical biopsy for breast cancer
• Breast Anatomy
yArchitecture Complex
structures
•
•
•
•
Heterogeneous
Hormonal control
Changes with age
Varietyy of cell types
yp
22
Optical biopsy for breast cancer
• Malignancies
g
of the Breast
Ductal
Lobular
Medullary
Mucinous
Comedo
Paget s
Paget’s
disease
Metaplastic
Tubular
23
Optical biopsy for breast cancer
• Breast Tissue Spectra
200
150
100
Normal
Tumour
Fibroadenoma
50
0
320
420
520
620
720
24
Optical biopsy for breast cancer
• Sentinel Lymph Node
• SLN biopsy aims to
determine the necessityy
of extensive surgery to
the armpit
• Patients requiring further
surgery may have to
undergo
d
separate
t
surgery
• Current techniques to
examine SLN during
operation
p
are impractical
p
25
Optical biopsy for breast cancer
• SNL Spectra
160
140
120
100
80
60
+ve Node
-ve Node
40
20
0
320
420
520
620
720
26
Primary Care
• Primary
y care p
physicians
y
have p
poor accuracy
y in
determining when to refer a patient to the specialist
27
An engineer’s chicken
28