U.S. patent application number 12/894528 was filed with the patent office on 2011-04-28 for method and an arrangement for the determination of the optical properties of a multi-layered tissue.
This patent application is currently assigned to BALTER AS. Invention is credited to Jakob J. Stamnes, Knut Stamnes.
Application Number | 20110098575 12/894528 |
Document ID | / |
Family ID | 19912221 |
Filed Date | 2011-04-28 |
United States Patent
Application |
20110098575 |
Kind Code |
A1 |
Stamnes; Jakob J. ; et
al. |
April 28, 2011 |
METHOD AND AN ARRANGEMENT FOR THE DETERMINATION OF THE OPTICAL
PROPERTIES OF A MULTI-LAYERED TISSUE
Abstract
The present invention relates to a method and an arrangement for
the determination of the optical properties of a multi-layered
tissue. More specifically, the invention relates to a method for
the detection and characterization of tumors in a tissue.
Inventors: |
Stamnes; Jakob J.; (Oslo,
NO) ; Stamnes; Knut; (Maplewood, NJ) |
Assignee: |
BALTER AS
Bergen
NO
|
Family ID: |
19912221 |
Appl. No.: |
12/894528 |
Filed: |
September 30, 2010 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
10471111 |
Oct 23, 2003 |
7822468 |
|
|
12894528 |
|
|
|
|
Current U.S.
Class: |
600/476 |
Current CPC
Class: |
A61B 5/0062 20130101;
A61B 5/0064 20130101; A61B 5/446 20130101; A61B 5/444 20130101;
G01N 21/4795 20130101 |
Class at
Publication: |
600/476 |
International
Class: |
A61B 6/00 20060101
A61B006/00 |
Foreign Application Data
Date |
Code |
Application Number |
Mar 6, 2001 |
NO |
2001 1131 |
Claims
1-8. (canceled)
9. A method for determining optical properties of a multi-layered
tissue, the method comprising the steps of: illuminating the
multi-layered tissue, which includes a plurality of air/tissue
interfaces, with a continuous-wave collimated electromagnetic beam
successively from each of a plurality of directions at each of a
plurality of wavelengths; independently detecting, in one dimension
only for each successive illumination, reflectance and/or
transmittance data received from a column of the multi-layered
tissue that is substantially perpendicular to each of the plurality
of air/tissue interfaces; and comparing the detected reflectance
and/or transmittance data with simulated reflectance and/or
transmittance data generated for the multi-layered tissue to
determine optical properties of the multi-layered tissue.
10. A method for determining optical properties of a multi-layered
tissue as recited in claim 9, further including the step of
determining a tissue configuration of the multi-layered tissue
based on the determined optical properties and characterizing the
tissue configuration for diagnosis.
11. A method for determining optical properties of a multi-layered
tissue as recited in claim 9, further including the step of
generating an image of the optical properties of the multi-layered
tissue, the image including a plurality of pixels whereby each
pixel represents optical properties determined for a particular
column of the multi-layered tissue that is substantially
perpendicular to each of the plurality of air/tissue
interfaces.
12. A method for determining optical properties of a multi-layered
tissue as recited in claim 9, wherein the step of comparing
includes identifying the simulated reflectance and/or transmittance
data that approximate the detected reflectance and/or transmittance
data to determine the optical properties of the multi-layered
tissue.
13. A method for determining optical properties of a multi-layered
tissue as recited in claim 9, wherein the step of comparing
includes a synthetic database including the simulated reflectance
and/or transmittance data, the simulated reflectance and/or
transmittance data being based on a plurality of different tissue
configurations for the multi-layered tissue, each tissue
configuration consisting of a combination of optical properties
corresponding to a particular set of values of simulated
reflectance and/or transmittance data.
14. A system for determining optical properties of a multi-layered
tissue, the system comprising: an apparatus including: a
continuous-wave collimated light source configured to successively
illuminate the multi-layered tissue, which includes a plurality of
air/tissue interfaces, from each of a plurality of directions at
each of a plurality of wavelengths, and a detector configured to
detect reflectance and/or transmittance data received, in one
dimension only, from a column of the multi-layered tissue that is
substantially perpendicular to each of the plurality of air/tissue
interfaces; a synthetic database including simulated reflectance
and/or transmittance data, the simulated reflectance and/or
transmittance data being based on a plurality of different tissue
configurations for the multi-layered tissue, each tissue
configuration consisting of a combination of optical properties
corresponding to a particular set of values of simulated
reflectance and/or transmittance data; wherein the synthetic
database is configured to compare the detected reflectance and/or
transmittance data with the simulated reflectance and/or
transmittance data to identify the simulated reflectance and/or
transmittance data that approximates the detected reflectance
and/or transmittance data to determine optical properties of the
multi-layered tissue; and an image generator configured to generate
an image of the optical properties of the multi-layered tissue, the
image including a plurality of pixels whereby each pixel represents
optical properties determined for each column of the multi-layered
tissue that is substantially perpendicular to each of the plurality
of air/tissue interfaces.
Description
CROSS-REFERENCE TO RELATED APPLICATIONS
[0001] This patent application is a continuation of U.S.
application Ser. No. 10/471,111, filed in the U.S. Patent and
Trademark Office on Oct. 23, 2003 based on PCT/NO02/00095, filed on
Mar. 6, 2002 by Stamnes et al., which claims priority to NO 2001
1131 filed on Mar. 6, 2001, the entire contents of each of these
applications being incorporated herein by reference in its
entirety.
BACKGROUND OF THE INVENTION
[0002] 1. Field of the Invention
[0003] The present invention relates to a method and an arrangement
for the determination of the optical properties of a multi-layered
tissue. More specifically, the invention relates to a method for
the detection and characterization of tumours in a tissue.
[0004] 2. Description of Related Art
[0005] New developments in imaging by means of visible and
near-infrared light show promise in medical diagnosis, because
light can offer three, key benefits over traditional diagnostic
tools based on e.g. X-rays, ultrasound, or nuclear magnetic
resonance. First, light at different wavelengths interacts with
tissue in distinctive ways and forms the basis for spectroscopy,
which allows one to optimize the wavelength for a specific
application. Second, image-processing methods are becoming powerful
enough to make it possible to use just a few photons, and thus
allow imaging based on low-level or noisy signals. Third, light
offers a good compromise between the lower-resolution radio
frequency or ultrasound imaging and the shorter wavelength, higher
resolution, but harmful ionizing radiation of X-rays. Also, optical
methods are usually non-invasive and non-toxic, and have the
potential to be realized in terms of compact and inexpensive
devices.
[0006] Any imaging or spectroscopic method must deal with both the
absorption by the primary component of tissue, i.e. the aquatic
solution, and the absorption and scattering by various types of
tissue "particles". Spectroscopy is used for measuring
time-dependent total variations in the absorption and scattering in
large volumes of tissue. For example, brain oxymetry (haemoglobin
spectroscopy) can reveal internal bleeding caused by head injury.
Imaging is important when it is of interest to detect a localized
heterogeneity of the tissue, such as an early breast or brain tumor
or a small amount of bleeding in the brain. Then imaging enables
one to identify the site of the trauma and differentiate it from
the surrounding tissue. Tumors represent a structural anomaly that
one desires to detect, localize, and classify. Tumor growth is
associated with (i) a larger blood volume resulting from a
relatively larger number density and volume fraction of blood
vessels in the tumor, (ii) increased concentrations of the
intracellular organelles required for the energy production
associated with rapid growth, and (iii) accumulation of highly
scattering calcium precipitates. Some of these properties are
expected to be helpful in classifying tumors as benign, malignant,
and so on.
[0007] Light that enters tissue is absorbed and scattered by
compounds in the tissue called chromophores. The major chromophores
are melanin, haemoglobin, and cytochromes. Melanin is a pigment
that colors our skin and protects us from sunburn. It attenuates UV
light strongly by acting as a Rayleigh scatterer. Haemoglobin (Hb)
is a colored pigment found in red blood cells. It is a large
molecule that can bind oxygen molecules to form HbO.sub.2.
Cytochromes consist of a series of enzymes found in the membrane of
the mitochondria. They have absorption spectra that depend on
whether the enzyme is in its oxidized or reduced state. Cytochromes
can be monitored by optical means. For example, NADH is an
important compound that absorbs strongly in the UV (310-375 nm). If
NADH is exposed to UV light, it will fluoresce with a broadband
emission spectrum around 460 nm. The absorption spectrum of Hb is
different from that of HbO.sub.2. Thus, highly oxygenated arterial
blood appears bright red, while venous blood containing more
deoxygenated haemoglobin appears bluish red. The absorption
coefficient varies with wavelength, but is typically 0.02-0.1 per
millimeter in the visible and NIR parts of the spectrum. It also
depends on chromophore content (especially the amount of blood). By
tissue particle is meant a small volume of tissue with a complex
refractive index that is different from that of the surrounding
medium. The absorption by the aquatic component is known or can be
measured for healthy tissue. A small volume surrounding a tumor is
characterized by increased blood concentration and thus enhanced or
anomalous absorption. Particles much smaller than the wavelength of
light consisting of cell nuclei or mitochondria are called Rayleigh
scatterers. Particles much larger than the wavelength of light
consisting of cells or groups of cells are called Mie scatterers.
The scattering coefficient varies with wavelength but is typically
in the range 20-100 per millimeter.
[0008] Knowledge of the optical properties of biological tissue is
the critical basis for carrying out studies in biomedical imaging
as well as for developing instruments for medical diagnosis.
[0009] The physiological state of a biological tissue can be
obtained from its absorption and scattering coefficients. By
physiological state is meant the relative concentrations of aquatic
and non-aquatic components as well as the chemical composition of
non-aquatic tissue components including blood vessels, organelles,
etc. Thus, it is desirable to develop accurate and reliable methods
for determining optical properties of tissue.
[0010] Optical coherence tomography (OCT) has, been successfully
applied in biomedical imaging. This method relies on the use of
coherent light, and it can be used to image the architectural
morphology or glandular organization of tissues. However, its
penetration depth is limited to 2-3 mm, and this technique, as
currently practiced, does not provide the information needed to
determine the optical properties of a biological system.
[0011] Since coherent light provides limited depth information,
many current optical imaging techniques rely on the use of diffuse
light, which carries information about deeper layers, i.e. about
4-8 mm within the tissue.
[0012] These methods include; [0013] (i) time-resolved techniques
for detection of so-called "snake" or "ballistic" photons that have
propagated along nearly-straight paths [0014] (ii) diffuse optical
tomography, and [0015] (iii) tomographic imaging using diffuse
photon density waves created by intensity-modulating the incident
light energy.
[0016] In imaging with diffuse light, approaches to the solution of
the radiative transfer problem are frequently based on the
diffusion approximation to the radiative transfer equation.
[0017] By the term "radiative transfer problem" is meant transport
of light in a multiple scattering medium such as tissue.
[0018] A review of the current literature is given in table 1.
TABLE-US-00001 Physical Equipment Commercial Method Principle
Strength Limitation Need Feasibility Comments OCT based on high
spatial penetration depth is optical gating available performs
architectural coherent light resolution limited & correlation
morphology, but is unable to devices image tissue optical
properties Laser CT (1) based on CT can determine penetration depth
is heterodyne under based on transmitted signal, can theory
extinction limited detection investigation only be applied to a
very thin layer coefficient device of tissue Time-resolved based on
ballistic can determine penetration depth is high time- promising
high time resolution is netessary, diffuse photons optical limited
resolution but is difficult to do imaging properties device CW
diffuse based on diffuse real-time signal to noise CW laser under
performs quantitative functional optical photons functional problem
sources investigation brain imaging, but does not tomography (2)
imaging determine tissue optical properties TOAST (3) based on
finite real-time based on time- pulsed laser under contains the
limitation inherent in element model functional resolved & high
time- investigation time-resolved techniques imaging measurements
resolution device Diffuse based on simple & fast determines
only CW laser under it is insufficient to use only optical
diffusion theory absorption coefficient source investigation
absorption coefficient to describe a reflection biological tissue
tomography (4) DPDW based on diffuse analytical transmitted DPDW
light source under cannot determine optical tomography (5) photon
clarity solution for signal is weak modulation investigation
properties of the background tissue waves DPDW Phased array based
on diffuse fast imaging of resolution is CW light under additional
work needed to retrieve imager (6) photon density brain function
questionable source & investigation optical properties from
measured waves modulation signal Frequency- based on determine
resolution & accuracy diode laser under cannot retrieve
scattering domain optical diffusion theory optical is questionable
and intensity investigation coefficient properly; unable to
tomography (7) properties modulator determine optical properties of
the background (1) Watanabe et al. (1998). (2) Siegel et al.
(1999). (3) Schweiger et al. (4) Cheng and Boas (1998). (5) Boas et
al. (1997). Chen et al. (1998). Li et al. (2000). (6) Chance at al.
(1998). (7) Pogue at al. (1997). CT: Computed Tomography TOAST:
Time-resolved Optical Absorption and Scatter Tomography DPDW:
Diffuse Photon Density Wave
[0019] 1. Even though time-resolved techniques offer the potential
for determining the optical properties of tissues, their reliance
on high time-resolution measurements makes it a very challenging
task to carry out experimental studies for validating the
methodology, and to develop suitable bedside instrumentation.
[0020] 2. Quite a few of the available techniques have been used
only to study the differences between the absorption coefficients
of the object and its surrounding medium (e.g. Cheng, x. and D. A.
Boas, 1998: Diffuse optical reflection tomography with
continuous-wave illumination. Opt. Express 3, No. 3, 118-123). It
is insufficient to use solely the absorption coefficient to
describe a biological tissue or an object embedded in such a
tissue, since scattering can usually not be ignored in such a
medium.
[0021] 3. Most of the available tomographic optical imaging methods
are focused on studying the optical properties (the absorption and
scattering coefficients and the asymmetry factor or phase function)
of an object that is embedded in a turbid medium, e. g. a tumor in
healthy tissue, assuming that the optical properties of the
background medium (healthy tissue) are known. These techniques
cannot easily be used to study the optical properties of the turbid
background medium (healthy tissue). However, accurate knowledge of
the optical properties of the background medium is critical for
success in biomedical imaging.
[0022] The above-described limitations of existing approaches
clearly show that there is an urgent need to develop and provide
reliable methods to determine both; [0023] (1) the optical
properties of healthy biological tissue, and (2) the location and
optical properties of an object (such as a tumor) that is embedded
in the healthy tissue.
BRIEF SUMMARY OF THE INVENTION
[0024] The overall objective of the present invention is thus to
provide a new method for determining the optical properties of a
multi-layered tissue. In this context the term "optical properties"
includes the determination of the absorption and scattering
coefficients, and also the asymmetry factor. By the term
"multi-layered tissue" is meant a tissue with a stratified
structure such that the optical properties may vary from one layer
to the next.
[0025] A further object of the present invention is to use this
method for the diagnosis and localization of tumors embedded in
such a layered tissue.
[0026] Our approach is based on rigorous radiative transfer theory
(described in more detail below). A sophisticated state-of-the-art
radiative transfer model for the coupled air/tissue system is used
to provide both forward and inverse algorithms for characterizing
the optical properties of such a multi-layered tissue.
[0027] Further, the algorithms are used to provide a method to
determine the location and optical properties of an object (such as
a tumor) that may be embedded in such a layered tissue.
[0028] Also provided is an experimental arrangement (FIG. 3), and
validation studies will be carried out.
[0029] The theoretical foundation for the method and arrangement in
accordance with the present invention is detailed below, and
investigations in progress are as follows:
[0030] (1) Forward simulations in which a multi-layered tissue is
illuminated by a collimated electromagnetic beam, successively from
several directions. For each illumination direction a comprehensive
radiative transfer model for the coupled air/tissue system will be
used to solve for the reflected diffuse light at several viewing
angles.
[0031] (2) Development of a method for retrieving the optical
properties and layer thickness of each of the top two layers of a
multi-layered tissue based on reflected radiances obtained from
forward simulations.
[0032] (3) Design of an experimental set-up for testing and
validating the theoretical simulations described in (1) and (2)
above.
[0033] (4) Development of an algorithm to determine the location
and optical properties of an object that is embedded in a layered
tissue based on the results obtained in the studies described
above.
[0034] Optical diagnostics can be classified as either imaging or
spectroscopy, or a combination of both. Optical coherence
tomography (OCT), as a high-resolution, shallow imaging tool, has
been developed in several research groups around the world (Huang,
D., E. A. Swanson, C. P. Lin, J. S. Schuman, W. G. Stinson, W.
Chang, M. R. Hee, T. Flotte, K. Gregory, C. A. Puliafato, and J. G.
Fujimoto, 1991: Optical coherence tomography. Science 254,
1178-1181; Izatt, J. A., M. Kulkarni, K. Kobayashi, M. V. Sivak, J.
K. Barton, and A. J. Welsh, 1997: Optical coherence tomography for
biodiagnostics. Optics & Photonics News 8, 41-47; Schmitt, J.
M., 1998: OCT elastography: imaging microscopic deformation and
strain of tissue. Opt. Express 3, No. 6, 199-211). OCT can be used
to perform high-resolution, cross-sectional tomographic imaging of
the internal microstructure in materials and biological systems by
measuring echo time delays and magnitudes of backscattered light
(Huang et al., 1991; Fujimoto et al., 2000). With continuous wave
illumination a transverse resolution of 10 .mu.m can be obtained
with OCT, but the penetration depth is limited to 2-3 millimetres
at most (Carts-Powell, Y., 1999: Optical tools offer minimally
invasive medical diagnostics. Optics & Phonotics News 10, No.
6, 33-37; Fujimoto, J. G., W. Drexler, U. Morgner, F. Kartner, and
E. Ippen, 2000: Optical coherence tomography using echoes of light.
Optics & Photonics News 11, No. 1, 24-31).
[0035] OCT can be used to image the architectural morphology or
glandular organization of tissues, but it cannot be used to
determine the optical properties of a biological system. On the
other hand, OCT measurements of ultra-fast echo time delays require
optical gating and correlation techniques, since direct electronic
detection is not possible.
[0036] Coherent imaging is useful, but most of the light that
enters the tissue is either absorbed or scattered several times.
Much ongoing research, e.g., Boas, D. A., M. A. O'Leary, B. Chance,
and A. G. Yodh, 1997: Detection and characterization of optical
inhomogeneities with diffuse photon density waves: a
signal-to-noise analysis. Appl. Opt. 36, 75-92; Chance, B., E.
Anday, S. Nikoa, S. Zhou, L. Hong, K. Worden, C. Li, T. Murray, Y.
Ovetsky, D. Pidikiti, and R. Thomas, 1998: A novel method for fast
imaging of brain function, non-invasively, with light. Opt. Express
2, No. 10, 411-423; Chen, B., J. J. Stamnes, and K. Stamnes, 1998:
Reconstruction algorithm for diffraction tomography of diffuse
photon density waves in a random medium. Pure Appl. Opt. 7,
1161-1180; Cheng, x. and D. A. Boas, 1998: Diffuse optical
reflection tomography with continuous-wave illumination. Opt.
Express 3, No. 3, 118-123; Durduran, T., J. P. Culver, M. J.
Holboke, X. D. Li, L. Zubkov, B. Chance, D. N. Pattanayak, and A.
G. Yodh, 1999: Algorithms for 3D localization and imaging using
near-field diffraction tomography with diffuse light. Opt. Express
4, No. 8, 247-262; Jacques, S. L., I. S. Saidi, A. Ladner, D. G.
Oelberg, 1997: Developing an optical fiber reflectance spectrometer
to monitor bilirubinemia in neonates. SPIE Proceedings of
Laser-Tissue Interaction VIII, edited by S. L. Jacques, 2975,
115-124; Kienle, A., M. S. Patterson, N. Dognitz, R. Bays, G.
Wagnieres, and H. van den Bergh, 1998: Noninvasive determination of
the optical properties of two-layers turbid media. Appl. Opt. 37,
779-791; Klose, A. D. and A. H. Hielscher, 1999: Iterative
reconstruction scheme for optical tomography based on the equation
of radiative transfer. Medical Phys. 26, 1698-1707; Patterson, M.
S. B. Chance, and B. C. Wilson, 1989: Time resolved reflectance and
transmittance for noninvasive measurement of tissue optical
properties. Appl. Opt. 28, 2331-2336; Siegel, A. M., J. J. A.
Marota, and D. A. Boas, 1999: Design and evaluation of a
continuous-wave diffuse optical tomography system. Opt. Express 4,
No. 8, 287-298; Wang, R. K. and Y. A. Wickramasinghe, 1998. Fast
algorithm to determine optical properties of a turbid medium from
time-resolved measurements. Appl. Opt. 37, 7342-7351, Yodh, A. and
B. Chance, 1995, Spectroscopy and imaging with diffusing light.
Physics Today 3, 34-40) is aimed at developing methods for
interpreting diffuse images. One approach is based on a
time-resolved technique to measure the diffuse photons from pulsed
light sources that have propagated along nearly straight paths
("ballistic" photons) (Wang and Wickramasinghe, 1998). Techniques
that rely on time-resolved measurements provide the opportunity, at
least in principle, to determine the optical properties of tissue
(Patterson, M. S. B. Chance, and B. C. Wilson, 1989: Time resolved
reflectance and transmittance for noninvasive measurement of tissue
optical properties. Appl. Opt. 28, 2331-2336; Wang, R. K. and Y. A.
Wickramasinghe, 1998. Fast algorithm to determine optical
properties of a turbid medium from time-resolved measurements.
Appl. Opt. 37, 7342-7351).
[0037] In practice, however, it is very difficult to perform such
measurements, not only because scattering rapidly reduces the
number of "ballistic" photons along the path, but also because an
extremely high time resolution is necessary.
[0038] Research on diffuse optical tomography has relied on several
approaches. Siegel et al. (Siegel, A. M., J. J. A. Marota, and D.
A. Boas, 1999: Design and evaluation of a continuous-wave diffuse
optical tomography system. Opt. Express 4, No. 8, 287-29) built a
portable continuous-wave diffuse optical tomography system
consisting of 18 laser diode sources (9 at 780 nm and 9 at 830 nm)
and 16 silicon detectors, but they did not use their method and
results to obtain the optical properties of the biological tissue.
The time-resolved optical absorption and scattering tomography
technique developed by the group at the University College London
(Schweiger, M., L. Zhukov, S. R. Arridge, and C. Johnson, 1999:
Optical tomography using the SCIRun problem solving environment:
Preliminary results for three-dimensional geometries and parallel
processing. Opt. Express 4, No. 8, 263-269) relies on a finite
element forward model to simulate light transport in scattering
media in order to reconstruct the internal distribution of optical
parameters from time-of-flight data. Since this technique is based
on time-resolved measurements it suffers from the same limitations
as the methods discussed above. Another example is the diffuse
optical reflection tomography developed by Cheng and Boas (1998),
which can be used to determine the absorption coefficient of an
object embedded in a turbid medium. Note that, in general, it is
insufficient to use only the absorption coefficient to describe a
turbid medium such as a biological tissue or an object embedded in
such a turbid medium.
[0039] Diffuse photon density waves have been used for tomographic
imaging of objects embedded in a turbid medium (Boas, D. A., M. A.
O'Leary, B. Chance, and A. G. Yodh, 1997: Detection and
characterization of optical inhomogeneities with diffuse photon
density waves: a signal-to-noise analysis. Appl. Opt. 36, 75-92;
Chen, B., J. J. Stamnes, and K. Stamnes, 1998: Reconstruction
algorithm for diffraction tomography of diffuse photon density
waves in a random medium. Pure Appl. Opt. 7, 1161-1180; Li, X., D.
N. Pattanayak, T. Durduran, J. P. Culver, B. Chance, and A. G.
Yodh, 2000: Nearfield diffraction tomography with diffuse photon
density waves. Phys. Rev. E 61, 4295-4309). The advantage of using
diffuse photon density waves is that such a wave has a well-defined
phase front. Measurements of both the phase and the amplitude of
the transmitted diffuse photon density wave can be used to study
objects that are embedded in a turbid background medium with known
optical properties (scattering and absorption coefficients). Since
the signal of the transmitted diffuse photon density wave is
usually very weak, this technique cannot be applied to a thick
turbid medium. Also, tomographic imaging using diffuse photon
density waves cannot be employed for determining the optical
properties of the turbid background medium. Another approach relies
on the use of diffuse photon density waves in conjunction with
reflectance measurements of light from layered media to determine
the structure of a layered tissue (Svaasand, L. O., T. Spott, J. B.
Fishkin, T. Pham, B. J. Tromberg, and M. W. Berns, 1999:
Reflectance measurements of layered media with diffuse
photon-density-waves: a potential tool for evaluating deep burns
and subcutaneous lesions. Phys. Med. Biol. 44, 801-813). Since the
wavelength of the diffuse photon density wave is usually large, say
at least a few centimetres, measurements of the phase of the
diffuse photon density wave can only provide low-resolution
information about the internal structure of the tissue.
[0040] Fluorescence spectroscopy constitutes an established way of
applying medical diagnostics (e.g., Chance, B., 1996: Use of
intrinsic fluorescence signals for characterizing tissue metabolic
states in health and disease. SPIE Proc. 2679, 2-7; Gardner, C. M.,
S. L. Jacques, and A. J. Welch, 1996: Fluorescence spectroscopy of
tissue: recovery of intrinsic fluorescence from measured
fluorescence. Lasers surg. Med. 18, 129-138), either to the tissue
itself or to some agent such as a dye used in photodynamic therapy.
By analysing the "colours" of the emitted light following optical
excitation of a tissue sample, one may determine if the tissue is
normal, benign, or cancerous. The underlying physical basis for
other spectroscopic techniques such as Raman scattering,
phosphorescence, and elastic scattering, is that the light-tissue
interaction is strongly influenced by the chemical composition and
the cellular structure of the tissue. These methods require that
either optical biopsies are performed, or that healthy tissues is
cut out or exposed to harmful ionizing radiation (Carts-Powell, Y.,
1999: Optical tools offer minimally invasive medical diagnostics.
Optics & Phonotics News 10, No. 6, 33-37).
[0041] Since many parts of the body such as skin, esophagus,
stomach, intestine, bladder, and head have a layered tissue
structure, it is increasingly recognized that for biomedical
imaging a layered tissue model is more realistic than a homogeneous
model. Several researchers have investigated the solution of the
diffusion equation for layered turbid media. Takatani and Graham
(Takatani, S. and M. D. Graham, 1979: Theoretical analysis of
diffuse reflectance from a two-layer tissue model. IEEE Trans.
Biomed. Eng. BME-26, 656-664) and Schmitt et al. (Schmitt, J. M. G.
X. Zhou, E. C. Walker, and R. T. Wall, 1990: Multilayer model of
photon diffusion in skin. J. Opt. Soc. Am. A 7, 2141-2153) derived
analytical formulas for the steady-state reflectance by use of
Green's functions to solve the diffusion equation, while Dayan et
al. (Dayan, I., S. Havlin, and G. H. Weiss, 1992: Photon migration
in a two-layer turbid medium, A diffusion analysis. J. Mod. Opt.
39, 1567-1582) applied Fourier and Laplace transforms to obtain
expressions for the steady-state and time-resolved reflectance.
Keijzer et al. (Keijzer, M., W. M. Star, and P. R. M. Storchi,
1988: Optical diffusion in layered media. Appl. Opt. 27. 1820-1824)
and Schweiger et al. (Schweiger, M., S. R. Arridge, M. Hiraoka, and
D. T. Deply, 1995: The finite element model for the propagation of
light in scattering media: Boundary and source conditions. Med.
Phys. 22, 1779-1792) employed a finite element method, and Cui and
Ostrander (Cui, W. and L. E. Ostrander, 1992: The relationship of
surface reflectance measurements to optical properties of layered
biological media. IEEE Trans. Biomed. Eng. 39, 194-201) used a
finite difference approach.
[0042] However, these researchers did not compare their results to
solutions of the transport equation, and the feasibility of
deriving the optical properties of the two layers adopted in their
models has not been studied. Kienle et al. (Kienle, A., M. S.
Patterson, N. Dognitz, R. Bays, G. Wagnieres, and H. van den Bergh,
1998: Noninvasive determination of the optical properties of
two-layers turbid media. Appl. Opt. 37, 779-791) solved the
diffusion equation using a Fourier transform approach for a
two-layer turbid medium having a semi-infinite second layer. It is
of interest to note the following statement of these authors:
[0043] "although the solutions of the diffusion equation for the
reflectance from a two-layered medium are quite close to the
results of the transport theory, the errors in determining the
optical properties caused by this approximation are greater than in
the semi-infinite case. Therefore, it would be advantageous to have
a solution of the transport equation for a two-layered medium fast
enough to be used for determination of the optical properties."
[0044] As described above, existing methods for imaging of tissue
are inadequate because they do not provide us with the information
we need to obtain, namely the three-dimensional spatial variation
of the optical properties of said tissue, including the absorption
coefficient, the scattering coefficient, and the asymmetry
factor.
BRIEF DESCRIPTION OF THE SEVERAL VIEWS OF THE DRAWING(S)
[0045] FIG. 1 shows a schematic diagram of a multi-layered
tissue.
[0046] FIG. 2 shows simulations of reflected radiances for a
3-layer tissue model with an abnormal ("malignant") middle layer,
and normal ("healthy") upper and bottom layers. The absorption and
scattering coefficients for the "healthy" upper and bottom layers
are taken to be .alpha..sub.20=.alpha..sub.3=0.1 mm.sup.-1 and
.sigma..sub.1=.sigma..sub.3 10.0 mm.sup.-1, while the asymmetry
factor is fixed at g.sub.1=g.sub.3=0.95. Upper left panel: The
scattering coefficient of the middle layer is fixed at
.sigma..sub.2=10.0 mm.sup.-1 and the asymmetry factor is
g.sub.2=0.95 (the same as for the "healthy" upper and bottom
layers), but the absorption coefficient is allowed to vary. Upper
curve: .alpha..sub.2=0.10 mm.sup.-1 ("healthy" value); Middle
curve: .alpha..sub.2=0.15 mm.sup.-1 ("malignant" value); Lower
curve: .alpha..sub.2=0.20 mm.sup.-1 ("malignant" value). Upper
right panel: The absorption coefficient of the middle layer is
.alpha..sub.2=0.1 mm.sup.-1 and the asymmetry factor is
g.sub.2=0.95 (the same as for the "healthy" upper and bottom
layers), but the scattering coefficient is allowed to vary. Lower
curve: .sigma..sub.2=0.10 mm.sup.-1 ("healthy" value); Middle
curve: .sigma..sub.2=15 mm.sup.-1 ("malignant" value); Upper curve:
.sigma..sub.2=20 mm.sup.-1 ("malignant" value). Lower panel: The
absorption coefficient of the middle layer is .alpha..sub.2=0.1
mm.sup.-1 and the scattering coefficient is .sigma..sub.2=10.00
mm.sup.-1 (the same as for the "healthy" upper and bottom layers),
but the asymmetry factor is allowed to vary. Lower curve:
g.sub.2=0.95 ("healthy" value); Middle curve: g.sub.2=0.80
mm.sup.-1 ("malignant" value); Upper curve: g.sub.2=0.65 mm.sup.-1
("malignant" value).
[0047] FIG. 3 shows a schematic diagram of the experimental
setup.
BRIEF DESCRIPTION OF THE INVENTION
[0048] Our novel concept for characterizing biological tissues and
retrieving their optical properties is based on a comprehensive
radiative transfer model together with multi-angle illumination and
viewing at several wavelengths. Thus we propose to illuminate the
tissue successively from several different directions by an
extended continuous-wave collimated beam. For each illumination
direction we will measure the backscattered light successively in
different viewing directions through directional scanning with a
CCD camera.
[0049] The equipment required for these measurements consists of
standard optical components and a CCD camera with suitable
resolution and sensitivity. Neither temporal modulation of the
illuminating beam to create a photon density wave nor fast
electronics to do time gating will be required.
[0050] The multi-angle, multi-wavelength approach described above
is very important in order to make the solution of the inverse
problem unique. When using only one wavelength, one illumination
direction, and one viewing direction, there may be many different
combinations of optical properties of the various layers that give
essentially the same backscattered light. To avoid this problem we
constrain our solution of the inverse problem by forcing it to be
consistent with many different sets of scattering data acquired
from several different combinations of wavelengths as well as
illumination and viewing directions. The solution of the inverse
problem contains two main ingredients, which are described
below.
[0051] One essential ingredient in our approach to the solution of
the inverse problem is the creation of a synthetic database of
simulated measurements. This is accomplished as follows. (1) For a
given measurement configuration, we use rigorous forward modeling
for the coupled air/tissue system to compute the field that would
be measured for a given tissue configuration. Here the term
"measurement configuration" means a particular combination of
illuminating wavelength, illumination direction, and viewing
direction, and the term "tissue configuration" means a particular
combination of optical parameters in the various layers of the
tissue. (2) We repeat the computations in (1) for a number of
different tissue configurations so as to obtain look-up tables or a
partial synthetic database of simulated measurements associated
with the given measurement configuration. (3) We repeat the
computations in (1) and (2) for many different measurement
configurations in order to create a complete set of look-up tables
or a complete synthetic database of simulated measurements that
covers all desired combinations of measurement configurations and
tissue configurations. The rigorous forward modeling for the
coupled air/tissue system is described below under "Rigorous
forward modeling".
[0052] The second essential ingredient of our inversion approach
consists in comparing measured data for many different combinations
of measurement configurations with simulated data contained in our
synthetic database in order to determine that particular tissue
configuration which provides best agreement between measured and
synthetic data. To accomplish this in a cost-effective manner we
employ the method of "global optimisation" that is described below
under "Inverse modeling to obtain tissue configuration".
[0053] To explain the novel concept of the present invention, we
will start by considering a stratified multi-layered turbid medium.
Thus we will assume that the tissue consists of plane parallel
layers, each having different optical properties. To be specific we
will first consider a tissue consisting of three layers, and each
layer will be described by its optical properties, i.e. its
[0054] i) absorption and scattering coefficients a and a
[0055] ii) asymmetry parameter g, and
[0056] iii) thickness z.
[0057] Note that the absorption and scattering coefficients and the
asymmetry parameter are wavelength-dependent.
[0058] We employ the discrete-ordinate method (to be described
below) to solve the radiative transfer equation pertaining to a
slab of biological tissue that is stratified into a number of
layers (FIG. 1). The radiative transfer equation provides a
rigorous description of the transport of light in an absorbing and
multiple scattering material, such as tissue. However, the change
of the refractive index between the air and the tissue affects both
the reflected field and the field inside the tissue significantly.
Therefore, one must take the reflection and refraction of the
incident light at the air/tissue interface into account to obtain a
rigorous solution for the coupled air/tissue system (Z. Jin and K.
Stamnes, Radiative transfer in non-uniformly refracting layered
media, Appl. Opt. 33, 431-442, 1994; G. E Thomas and K. Stamnes,
Radiative Transfer in the Atmosphere and Ocean, Cambridge
University Press, 1999, section 6.6; A. R. Degheidy and M. S. Abdel
Krim, Effects of Fresnel and diffuse reflectivities on light
transport in a half-space medium, J. quant. Spectrosc. Radat.
Transfer, 61, 751-757, 1999).
[0059] The assumption of a stratified tissue implies that the
optical properties depend only on the depth. Radiative transfer in
this stratified coupled air/tissue system can be described by the
Radiative Transfer Equation (RTE), given as the first equation in
section 6.6.1 in G. E. Thomas and K. Stamnes, Radiative Transfer in
the Atmosphere and Ocean, Cambridge University Press, 1999 as
follows:
u I ( .tau. , u , .phi. ) .tau. = I ( .tau. , u , .phi. ) - a (
.tau. ) 4 .pi. .intg. 0 2 .pi. .phi. ' .intg. - 1 1 u ' p ( .tau. ,
u ' , .phi. ' ; u , .phi. ) I ( .tau. , u ' , .phi. ' ) - S * (
.tau. , u , .phi. ) ##EQU00001##
[0060] where I(.tau.,u,.phi.) is the diffuse radiance, and
S * ( .tau. , u , .phi. ) = a ( .tau. ) I 0 4 .pi. .mu. 0 .mu. t b
( - .mu. 0 ; m rel ) p ( .tau. , - .mu. t , .phi. 0 ; u , .phi. ) -
.tau. / .mu. t . ##EQU00002##
[0061] Here m.sub.rel is the index of refraction in the tissue
relative to air, Io is the incident beam irradiance,
T.sub.b(-.mu..sub.0;m.sub.rel).ident.T.sub.b(-.mu..sub.0,.phi..sub.0i;-.m-
u..sub.t,.phi..sub.0;m.sub.rel) is the beam transmittance through
the interface, .mu..sub.0 is the cosine of the incident beam angle,
and .mu..sub.t is the cosine of the refracted beam angle in the
tissue, related to .mu..sub.0 by Snell's Law
.mu..sub.t.ident..mu..sub.t(.mu..sub.0,m.sub.rel)= {square root
over (1-(1-.mu..sub.0.sup.2)/m.sub.rel.sup.2)}.
[0062] In this RTE u' is cosine of the polar angle prior to
scattering, u is the cosine of the polar angle after scattering,
.phi.' is the azimuth angle prior to scattering, .phi. is the
azimuth angle after scattering, .tau. is the optical depth,
a(.tau.)=.sigma.(.tau.)/[.sigma.(.tau.)+.alpha.(.tau.)] is the
single-scattering albedo, .alpha.(.tau.) is the absorption
coefficient, .sigma.(.tau.) is the scattering coefficient, and
p(.tau., u', .phi.', u, .phi.,) is the scattering phase function,
which gives the probability of scattering from an incident
direction (u',.phi.') into an new direction (u, .phi.). Note that
the internal source S*(.tau.,u,.phi.) depends on the direction
(.theta..sub.0, .phi..sub.0) of the incident light (see FIG. 1),
the collimated beam intensity, and the refractive index in the
tissue relative to that in air. A numerical code that solves the
multiple scattering problem for such an air/tissue system is
described in Z. Jin and K. Stamnes, Radiative transfer in
non-uniformly refracting layered media, Appl. Opt. 33, 431-442,
1994, and in section 6.6.1 in G. E Thomas and K. Stamnes, Radiative
Transfer in the Atmosphere and Ocean, Cambridge University Press,
1999. The solution is obtained by employing the so-called
discrete-ordinate method, which amounts to replacing the integral
on the right-hand side of the RTE by a sum. Thereby, the original
RTE, which is an integro-differential equation, is replaced by a
set of ordinary differential equations, which are solved using
standard techniques of linear algebra. Thus, one obtains a fast and
rigorous forward modeling scheme for the coupled air/tissue system.
For details, see Z. Jin and K. Stamnes, Radiative transfer in
nonuniformly refracting layered media, Appl. Opt. 33, 431-442,
1994.
[0063] Rigorous Forward Modeling
[0064] Based on the three-layer model described above we will solve
the Radiative Transfer Equation (RTE) to obtain the backscattered
light for various angles of illumination and viewing. In this
endeavor we will use the solution of the RTE for a coupled
air/tissue system as described above, where the change in the
refractive index between the air and tissue is taken into account.
This is very important in order to get a correct description of the
backscattered light.
[0065] As explained in detail previously, these simulations are
carried out for many different tissue configurations (i.e.
combinations of the optical properties in the three layers) in
order to create look-up tables (i.e. a synthetic database of
simulated measurements) for the backscattered intensity for a
variety of measurement configurations (i.e. combinations of
wavelengths and illumination and viewing angles). Such look-up
tables are essential in order to obtain a reasonably fast solution
of the inverse problem. The third or bottom layer will be assumed
to be so thick that no radiation is scattered back from its bottom
surface. We noted above that the optical properties depend on
wavelength. Thus, by systematically varying the optical properties
within their expected ranges for the wavelength interval of
interest, we automatically include their wavelength dependence.
[0066] In our forward simulations we assume that the tissue layers
in FIG. 1 have thicknesses of z.sub.1, z.sub.2, and z.sub.3,
respectively. The refractive indices of the air and the tissue are
taken to be n.sub.1 and n.sub.2, respectively, and the absorption
and scattering coefficients of layers 1, 2, and 3 are denoted by
.alpha..sub.1, .alpha..sub.2 and .alpha..sub.3 and .sigma..sub.1,
.sigma..sub.2 and .sigma..sub.3, respectively. The asymmetry
factors of layers 1, 2, and 3 are denoted by g.sub.1, g.sub.2, and
g.sub.3, respectively. As explained previously, the scattering
phase function gives the probability that a photon that is incident
in the direction (u',.phi.') is scattered into the direction (u,
.phi.). One step in the solution of the RTE involves expanding the
scattering phase function in a series of Legendre polynomials, and
the asymmetry factor is the first expansion coefficient in this
series. It has the values g=0 for isotropic scattering or
scattering that is symmetric about the forward direction, g=-1 for
complete backscattering, and g=1 for complete forward scattering.
Lacking better information about the scattering phase function, we
will use the synthetic Henyey-Greenstein scattering phase function
to describe the angular light scattering pattern due to particles
in the tissue. This scattering phase function is given in equation
(6.49) in G. E Thomas and K. Stamnes, Radiative Transfer in the
Atmosphere and Ocean, Cambridge University Press, 1999, and it
depends only on one parameter, namely the asymmetry factor. To
provide an example of the capability of the radiative transfer
model when applied to the coupled air/tissue system we consider the
following cases below:
[0067] (1) We keep the optical properties of the upper and lower
layer fixed, while allowing those of the middle layer (i.e. the
asymmetry factor g.sub.2 and the absorption and scattering
coefficients .alpha..sub.2 and .sigma..sub.2) to vary. This is a
simple way of mimicking a "malignant" layer located between two
healthy layers. Our goal is to show that the reflected radiances
are sensitive to changes in the optical properties of the
"malignant" middle layer when the optical properties of layers 1
and 3 are kept fixed. Thus, FIG. 2 shows that the reflected
radiances are sensitive to changes in the absorption and scattering
coefficients (.alpha..sub.2 and .sigma..sub.2) and the asymmetry
factor (g.sub.2) of layer 2. This illustrates that the measured
reflected radiances carry information that can be used to retrieve
the optical properties of this layer.
[0068] (2) When either the absorption or the scattering coefficient
is large, the photon penetration depth is small. If the photon
penetration depth becomes less than the thickness of layer 1,
reflected radiances are no longer sensitive to changes in the
optical properties of layer 2. Also, the transmitted radiance then
becomes too small to be measured by conventional detection
techniques. For a specified direction of illumination the photon
penetration depth can easily be determined by increasing the upper
layer thickness until the reflected intensity does not change.
Knowledge of the photon penetration depth is important, because the
reflected photons do not carry any information about the tissue
located beneath that depth.
[0069] (3) FIG. 2 displays results only for light at normal
incidence. By keeping the middle layer thickness so large that no
photons are expected to reach the lowest layer, we can vary the
optical properties of the two upper layers in a systematic fashion
to create look-up tables for the reflected light. Entries in these
look-up tables will depend on the optical properties of each of the
two upper layers as well as on the direction of the incident light
and the viewing angle of the detector. The look-up tables will be
designed to allow for retrieval of the optical properties of the
two upper layers, the photon penetration depth, and the thickness
of the upper layer.
[0070] Inverse Modeling to Obtain Tissue Configuration
[0071] The solution of the inverse problem to retrieve tissue
optical properties from measurements of reflected radiances, will
be based on forward modeling combined with a suitable optimisation
method (see, e.g. O. Frette, J. J. Stamnes, and K. Stamnes, Optical
remote sensing of marine constituents in coastal water: A
feasibility study, Appl. Opt. 37, 8318-8326, 1998; O. Frette, S. R.
Erga, J. J. Stamnes, and K. Stamnes, Optical remote sensing of
waters with vertical structure, Appl. Opt. 40, 1478-1487 (2001)).
Our radiative transfer model for the air/tissue system will be used
for simultaneous retrieval of the tissue optical properties (the
absorption and scattering coefficients, and the asymmetry factor)
and the layer thickness using the method described by O. Frette, J.
J. Stamnes, and K. Stamnes, Optical remote sensing of marine
constituents in coastal water: A feasibility study, Appl. Opt. 37,
8318-8326, 1998. To retrieve the optical properties of the three
layers we will use inverse modeling based on the look-up tables or
synthetic database described above. Thus we will compare the
measured backscattered data with simulated backscattered data from
the look-up tables for a variety of combinations of optical
parameters (.alpha..sub.j, .sigma..sub.j, g.sub.j, and z.sub.j;
j=1, 2, 3) for the three layers. In this comparison we will use a
"global optimisation" method such as simulated annealing to find
that combination of optical properties which minimises the
difference between measured data and simulated measured data stored
in the look-up tables (see, e.g. O. Frette, J. J. Stamnes, and K.
Stamnes, Optical remote sensing of marine constituents in coastal
water: A feasibility study, Appl. Opt. 37, 8318-8326, 1998). Here
the term global optimisation refers to an optimisation method that
is used to search for a global minimum in a cost-effective manner
among several local minima. An example of such a global
optimisation method is simulated annealing, which is described in
H. Press, S. A. Teukolski, W. T. Wetterlig, and B. P. Flannery,
Numerical Recipes, Cambridge University Press, 1992. It is
important to emphasise that what we can retrieve using inverse
radiative transfer modeling are values for the optical parameters
.alpha..sub.j, .sigma..sub.j, g.sub.j, at different wavelengths and
z.sub.j (j=1, 2, 3). This knowledge can be used to determine the
physiological state of the tissue.
[0072] Testing of the Retrieval Algorithm Using Synthetic Data
[0073] First we will test the inverse modeling approach described
above by using synthetic data generated from the forward model.
Thus we will use synthetic data for a variety of configurations
other than those included in the look-up table as input to the
inverse-modeling algorithm. The retrieved optical parameters are
those that yield the minimum difference between the synthetic
radiances and the radiances retrieved from the look-up tables.
[0074] Testing of the Retrieval Algorithm Using Data from a
Controlled Experiment.
[0075] After having tested the retrieval algorithm thoroughly in
the way just described, we will carry out a controlled experimental
test. To that end we will use a three-layer suspension, and in each
layer we will have particles with known optical properties of known
concentrations. But the optical properties and concentrations will
vary from one layer to the next.
[0076] Horizontal Imaging Using the Independent Column or Pixel
Approach
[0077] The next step in the development of the novel concept of the
present invention for imaging of a tissue is to investigate under
what circumstances we can apply the one-dimensional approach
described above to a tissue with variation in the horizontal
direction parallel to the layer interfaces. The basic assumption
that we will make is that each pixel in the image receives
backscattered light only from the vertical column that lies
directly underneath the corresponding area on the air/tissue
interface. Using this assumption, we can apply the retrieval
algorithm described above independently to each pixel in the image
and thus provide information about the horizontal variation of the
optical properties of different vertical columns, one for each
pixel in the image. Thus within the range of validity of this
independent-pixel approximation we can obtain a three-dimensional
image of the optical properties of the tissue.
[0078] Testing of the Independent-Pixel Approach Using Synthetic
Data
[0079] First we will test the independent-pixel approach described
above by using synthetic data generated by use of forward modeling.
Again the physical model will consist of three layers, but now
there will be an abrupt change in its vertical physical and optical
properties at a certain horizontal position. To generate synthetic
data in this case we will use Monte Carlo simulations. In the
retrieval we examine how close to the horizontal position of
discontinuity we can place the area under investigation on the
air/tissue interface, before the corresponding pixel in the image
starts to deteriorate. In this manner a good estimate of the range
of validity of the independent-pixel approach is obtained.
[0080] Testing of the Independent-Pixel Approach Using Data from a
Controlled Experiment
[0081] In a similar manner as described above for the testing of
the one-dimensional retrieval algorithm, a controlled experimental
test of the independent-pixel algorithm is conducted. A three-layer
suspension is used, and in each layer we will have particles with
known optical properties of known concentrations. But now we will
arrange to have an abrupt change of the optical properties at a
given horizontal position.
[0082] Three-Dimensional Imaging Beyond the Independent-Pixel
Approximation
[0083] The experimental data acquired by the multi-angle
illumination and viewing approach described above can also be
applied to obtain an image that is not based on the
independent-pixel approximation. Dependent upon the outcome of the
testing of the independent-pixel approach, as described above, we
will decide whether it is worthwhile to develop inverse modeling
methods that are not based on this approximation.
[0084] Experimental Setup
[0085] FIG. 3 shows a sketch of the experimental arrangement in
accordance with the present invention. A collimated laser beam is
incident upon the air/tissue interface in the direction
(.theta..sub.0,.phi..sub.0). The incident beam illuminates an
extended area of the air/tissue interface, and some of the incident
light is refracted through the air/tissue interface and penetrates
into the tissue where it may be scattered or absorbed. A portion of
the scattered light reaches the air/tissue interface and is
refracted back through the interface in various directions. A CCD
camera is directionally scanned to detect the light leaving the
illuminated area of the interface in various directions
(.theta.,.phi.). By directional scanning is meant to detect the
light leaving the illuminated area of the interface from many
different viewing directions. For each combination of illumination
and viewing directions several different wavelengths will be used
successively.
[0086] The Kr--Ar multi-line laser used in this experiment emits
light at 5 wavelengths, i.e. at .lamda.=458 nm, .lamda.=488 nm,
.lamda.=514 nm, .lamda.=530 nm, and .lamda.=647 nm. The CCD camera
used to detect the backscattered light has a high spatial
resolution with 2048.times.2048 pixels and a dynamic range of 16
bits. In a preferred embodiment of the invention the CCD chip is
cooled by a Peltzier element down to -32.degree. C. in order to
minimize the dark current noise. Both the laser and the detector
are of high quality and are very well suited for the controlled
experiments.
[0087] The scattering medium consists of three plane parallel
layers with different optical properties. Each layer can be a
liquid with a given density and absorption containing a suspension
of scattering particles. Alternatively, each layer can be of a
gelatine material with given absorption and scattering
coefficients. The latter choice will be better for experiments over
a prolonged period, since particles in a liquid suspension will
tend to change their vertical position with time.
[0088] The feasibility of retrieving information about optical
properties for an absorbing and scattering medium like tissue from
experiments as described above, has recently been demonstrated for
the coupled atmosphere/ocean system, see e.g. O. Frette, J. J.
Stamnes, and K. Stamnes, Optical remote sensing of marine
constituents in coastal water: A feasibility study, Appl. Opt. 37,
8318-8326, 1998; O. Frette, S. R. Erga, J. J. Stamnes, and K.
Stamnes, Optical remote sensing of waters with vertical structure,
Appl. Opt. 40, 1478-1487 (2001); Stamnes, K., W. Li, B. Yan, A.
Barnard, W. S. Pegau and J. J. Stamnes, A new ocean color
algorithm: Simultaneous retrieval of aerosol optical properties and
chlorophyll concentrations, Appl. Opt., submitted. In this
connection it is important to emphasise that the coupled
atmosphere/ocean system is far more complicated than the air/tissue
system. In the former case a simultaneous retrieval of optical
properties in the atmosphere and ocean is required because 90% or
more of the measured signal is due to absorption and scattering in
the atmosphere. In the latter case it suffices to retrieve the
optical properties of the tissue because the air does not
significantly affect the measured signal.
* * * * *