Task-based imaging of colon cancer in the Apc(Min/+) mouse model

Article (PDF Available)inApplied Optics 45(13):3049-62 · June 2006with63 Reads
DOI: 10.1364/AO.45.003049 · Source: PubMed
Abstract
Optical coherence tomography (OCT), laser-induced fluorescence (LIF), and laser-scanning confocal microscopy (LSCM) were used for the task of multimodal study of healthy and adenomatous mouse colon. The results from each modality were compared with histology, which served as the gold standard. The Apc(Min/+) genetic mouse model of colon cancer was compared with wild-type mice. In addition, a special diet was used for the task of studying the origins of a 680 nm autofluorescent signal that was previously observed in colon. The study found close agreement among each of the modalities and with histology. All four modalities were capable of identifying diseased tissue accurately. The OCT and LSCM images provided complementary structural information about the tissue, while the autofluorescence signal measured by LIF and LSCM provided biochemical information. OCT and LIF were performed in vivo and nondestructively, while the LSCM and histology required extraction of the tissue. The magnitude of the 680 nm signal correlates with chlorophyll content in the mouse diet, suggesting that the autofluorescent compound is a dietary metabolite.

Figures

Task-based imaging of colon cancer in the Apc
Min
mouse model
James B. McNally, Nathaniel D. Kirkpatrick, Lida P. Hariri, Alexandre R. Tumlinson,
David G. Besselsen, Eugene W. Gerner, Urs Utzinger, and Jennifer K. Barton
Optical coherence tomography (OCT), laser-induced fluorescence (LIF), and laser-scanning confocal
microscopy (LSCM) were used for the task of multimodal study of healthy and adenomatous mouse colon.
The results from each modality were compared with histology, which served as the gold standard. The
Apc
Min
genetic mouse model of colon cancer was compared with wild-type mice. In addition, a special
diet was used for the task of studying the origins of a 680 nm autofluorescent signal that was previously
observed in colon. The study found close agreement among each of the modalities and with histology. All
four modalities were capable of identifying diseased tissue accurately. The OCT and LSCM images
provided complementary structural information about the tissue, while the autofluorescence signal
measured by LIF and LSCM provided biochemical information. OCT and LIF were performed in vivo and
nondestructively, while the LSCM and histology required extraction of the tissue. The magnitude of the
680 nm signal correlates with chlorophyll content in the mouse diet, suggesting that the autofluorescent
compound is a dietary metabolite. © 2006 Optical Society of America
OCIS codes: 170.1790, 170.2150, 170.2680, 170.3890, 170.4500, 170.6510.
1. Introduction
Colorectal cancer (CRC) is the third most common
form of cancer in both men and women in the United
States. CRC accounted for an estimated 11% of new
cancer cases and 10% of cancer fatalities in 2004.
Early detection is essential, as CRC has a five-year
survival rate of 90% when it is detected at an early
stage, but the survival rate drops to less than 10%
after metastasis. Unfortunately, only 38% of CRC
cases are discovered in the early stage.
1
New tools
that minimize the invasiveness of screening, aid
early detection, and enable the serial monitoring of
disease progression and therapies for CRC are
needed. There are multiple schemes under which
these tools must function. In vivo, minimally invasive
imaging tools are needed for both human patients
and the mouse models that are used to develop and
test therapies. Ex vivo, nondestructive imaging is
needed to determine tissue function and structure
with minimal artifacts. Ex vivo, destructive analysis
(imaging of histological sections) provides the gold
standard for diagnosis and tissue evaluation.
There are a number of system considerations that
drive the design of imaging-based screening and diag-
nostic tools. The first important factor is resolution.
Cellular-level resolution enables the identification of
cell morphologies associated with early-stage dyspla-
sia, while tissue-structure resolution can be used to
identify gross tissue changes associated with disease.
A second design consideration is the utility of surface
imaging capabilities compared with depth-resolved
diagnostic tools. Surface imaging tools, such as tra-
ditional video colonoscopy, only provide data on super-
ficial tissue structure. For depth-resolved techniques,
the depth of penetration in highly scattering biological
tissue is an important consideration in system design.
For use in the diagnosis of early-stage CRC, a depth of
penetration of the order of 1–2 mm is sufficient for the
relatively thin colonic mucosa and submucosa. A final
design factor is the type of information the diagnostic
instrument provides. While structural information is
The authors are all with the University of Arizona, 1230 East
Speedway Boulevard, Tucson, Arizona 85721-0104. J. B. McNally
(james.mcnally@gmail.com), U. Utzinger, and J. K. Barton are
with the College of Optical Sciences. N. D. Kirkpatrick, L. P.
Hariri, A. R. Tumlinson, U. Utzinger, and J. K. Barton are with the
Division of Biomedical Engineering. D. G. Besselsen is with Uni-
versity Animal Care as well as Veterinary Science and Microbi-
ology. E. W. Gerner is with the Department of Cell Biology and
Anatomy and the Department of Biochemistry and Molecular Bio-
physics.
Received 16 May 2005; revised 4 December 2005; accepted 9
December 2005; posted 3 January 2006 (Doc. ID 62092).
0003-6935/06/133049-14$15.00/0
© 2006 Optical Society of America
1 May 2006 Vol. 45, No. 13 APPLIED OPTICS 3049
very useful in biological diagnosis, the addition of bio-
chemical clues to tissue function may increase the sen-
sitivity and specificity of a given system to disease.
Optical imaging modalities have resolutions that
range from subcellular to tissue level, can be depth
resolved, and are capable of providing functional and
structural information. For these reasons, they may
have an advantage over competing technologies. Con-
ventional ultrasound imaging is limited to resolu-
tions of approximately 50–100 m, 1 to 2 orders of
magnitude poorer than that of optical imaging mo-
dalities. Clinical magnetic resonance imaging (MRI)
and computed tomography (CT) are limited in reso-
lution to 1 mm and provide primarily anatomical
information.
However, optical techniques face some challenges
in the clinical and laboratory environments, primary
among them being limited depth of penetration and
low endogenous signal levels. Ex vivo imaging greatly
reduces system design constraints by eliminating
the need for miniaturization, biocompatibility, and
tissue-access considerations, but extraction of the tis-
sue for ex vivo imaging is not compatible with real-
time tissue evaluation or time-serial evaluation of
disease progression and therapies. In vivo imaging
enables nondestructive diagnosis, but access to the
target tissue site and low signal levels become major
design concerns. The design of optical tools is also
driven by whether the system will be used in human
or animal model studies. While imaging of human
subjects is ideal for the study of human disease patho-
genesis, such experiments are often very expensive,
time consuming, and limited in scope owing to safety
constraints. Animal models, particularly rodent mod-
els, are a useful alternative. Murine models can
greatly reduce the time and expense of early-stage
studies while still offering a sufficiently accurate
model of human physiology. However, small-rodent
models place additional design constraints, such as
extreme miniaturization, on optical instruments that
can realistically be used for in vivo experimentation.
The optimal choice of optical modality thus differs
depending on the imaging environment and the tar-
get imaging tasks. In this paper we utilize four opti-
cal modalities that are suitable for in vivo and ex vivo
imaging in a mouse model of colon cancer. These four
modalities include optical coherence tomography
(OCT), laser-induced fluorescence (LIF) spectroscopy,
laser-scanning confocal microscopy (LSCM), and his-
tology. We employ these four modalities in the pur-
suit of two specific imaging tasks. The first task is
multimodal identification and verification of healthy
and diseased tissue. The different modalities are used
to collect structural and biochemical information
about the tissue sample with both macroscopic and
microscopic resolution and to then arrive at a diag-
nosis of the disease state. The second task is studying
the origin of a 680 nm autofluorescent signal in an
attempt to determine whether the signal is correlated
with increased chlorophyll content in the mouse diet
and whether the autofluorescent signal is the result
of superficial contamination or cellular uptake.
A. Optical Coherence Tomography
OCT is a depth-resolved, minimally invasive imaging
modality that achieves high axial resolution by using
low-coherence interferometry. Depth scanning is
achieved by varying the optical path length in the
reference arm of a Michelson interferometer and
measuring the envelope of the resulting interferomet-
ric signal. High axial resolution is achieved by using
a low-coherence source, as interference will occur
only when the optical paths of the light in both inter-
ferometer arms match to within the coherence length
of the source. OCT offers a number of advantages for
use in in vivo animal imaging. The use of near-
infrared (NIR) light allows for a relatively large depth
of penetration, typically up to 2 mm, in highly scat-
tering biological tissue. OCT is also compatible with
fiber-based technologies, enabling the fabrication of
miniature OCT probes.
2,3
OCT axial resolution is
generally unaffected by the imaging performance of
the beam-shaping optics because the coherence
length of the source determines the axial resolution.
OCT has been used in vivo to produce high-resolution
images of rabbit colon that resolve upper colonic mu-
cosa, muscular mucosa, and submucosa in addition to
crypt structures within the colonic mucosa.
4
Visual-
ization of these structures is important in the diag-
nosis of inflammatory bowel disease (IBD) and colon
cancer. However, the resolution of OCT systems still
falls short of that achievable with a high-power opti-
cal microscope. OCT systems are capable of real-time
imaging, which enables rapid, large-area screening.
OCT has effectively detected specialized intestinal
metaplasia,
5
adenoma, and carcinoma
6
in vivo in the
human gastrointestinal (GI) tract. We have used en-
doscopic OCT to image the development of adenoma
in vivo in mouse colon over time, demonstrating
OCT’s potential for early-stage diagnosis and serial
monitoring of CRC.
7
B. Laser-Induced Fluorescence
In LIF spectroscopy a sample is illuminated with
ultraviolet or visible radiation, exciting endogenous
fluorophores, such as NADH and collagen in the sam-
ple. The emission from the autofluorescence of the
tissue provides a spectral signature that differs be-
tween normal and diseased tissue. The depth of pen-
etration of LIF excitation light is typically of the order
of a few hundred micrometers in biological tissues.
The field of view and resolution of LIF spectroscopy
can vary based on system design; in some cases dif-
fuse excitation illumination is used, while in other
cases the excitation is localized by an objective lens.
LIF is well suited for miniaturization using fiber-
based implementations and is a robust analytical tool
even at the low levels typical of autofluorescence sig-
nals, making LIF advantageous for use in in vivo
studies of disease. Previous studies have shown that
LIF can accurately identify dysplasia in human colon
tissue both ex vivo and in vivo.
8,9
LIF has also been
used in vivo to distinguish adenomatous tissue in
3050 APPLIED OPTICS Vol. 45, No. 13 1 May 2006
mouse colon
7
and to monitor the progression of colo-
rectal tumors in rats.
10
C. Laser-Scanning Confocal Microscopy
LSCM is a high-resolution, spectrally resolved imag-
ing technique that combines the high-power magni-
fication and depth-resolving capabilities of a confocal
microscope with the autofluorescence techniques of
LIF spectroscopy. LSCM images give an en face view
of the tissue at different depths. The field of view in
these images depends on the magnification of the
objective being used but typically ranges from tens of
micrometers to several hundred micrometers in each
dimension. The depth of penetration in LSCM is de-
termined, much like LIF, by the attenuation of the
short-wavelength excitation light in the tissue and is
normally limited to a few hundred micrometers.
LSCM has been used to quantify and differentiate the
autofluorescence of normal, premalignant, and ma-
lignant human colon ex vivo.
11
LSCM also facilitates
the localization of autofluorescence signals in the tis-
sue by combining the spatial-filtering capabilities of a
confocal microscope with focused fluorescence excita-
tion. Ex vivo confocal microscopy has been used to
determine the spatial origin of colonic autofluores-
cence in an effort to identify the differences among
healthy tissue and adenomatous and hyperplastic
polyps.
12
The subcellular-level resolution capabilities
of LSCM have been exploited to study intracellular
processes and structures in mouse models.
13,14
Recent efforts have been made to develop miniature
endoscopes capable of LSCM that would pave the way
toward minimally invasive, in vivo imaging.
15–18
En-
doscope designs include both fiber-bundle imaging and
fiber-scanning configurations. In the fiber-bundle con-
figuration, lateral resolution is typically limited to
3 m by the fiber-to-fiber spacing. The fiber-scanning
designs, which include a commercial device, have re-
ported lateral resolutions of 0.7 m. While each of
these systems was designed with the goal of integrat-
ing the LSCM endoscope into a standard colonoscope,
none of these current systems are sufficiently small to
be used in nondestructive imaging of mouse colon.
Additionally, each of these studies used exogenous
fluorescent dyes, instead of the weaker autofluores-
cent signal, to increase the sensitivity of imaging.
D. Histology
Histology is considered the gold standard of patho-
logical evaluation. This technique enables high-
resolution, subcellular images of the tissue sample.
Fluorescence measurements can be made on the his-
tologic sections by using a fluorescence microscope to
identify biochemical tissue constituents, and exoge-
nous dyes can be employed to increase contrast with-
out concern for toxicity effects. However, histologic
examination requires the extraction, fixation, and
processing of a tissue sample. These steps perma-
nently alter the tissue before examination and may
introduce processing artifacts that can be difficult to
interpret. Additionally, histological tissue prepara-
tion is time consuming and labor intensive.
E. Animal Model
Mouse models can be used to study the development
and treatment of human GI disease. The C57BL
6J-Apc
Min
strain of mice is an attractive model for
studying neoplasia in the GI tract. A mutation in the
Apc gene causes these mice to spontaneously develop
benign tumors (adenomas) throughout the GI tract.
19
The homolog of this gene in humans (APC) is mutated
in familial adenomatous polyposis, an inherited CRC
condition. The dysplasia tends to occur more promi-
nently in the upper GI tract; however, dietary sup-
plements, such as arginine-enriched water, can
increase the tumor load in the distal colon of mice.
F. Special Diets
Our previous studies of mouse colon indicated the
presence of three distinct peaks in the LIF signal of
adenomatous tissue.
7
Peaks at 390 and 450 nm are
associated with collagen and NADH fluorescence
combined with hemoglobin absorption at 420 nm.
The origin of a third peak at 680 nm is controversial.
Others have suggested the 680 nm signal is associ-
ated with the porphyrin derivatives of hemoglobin
breakdown.
20–22
However, these porphyrin deriva-
tives are associated with dual emission peaks at
630 and 690 nm,
23,24
which is inconsistent with our
experimental data. A 680 nm peak is consistent with
the spectral profile of chlorophyll and its metabolites.
To investigate the origin of this peak, we fed the mice
in our study varying quantities of chlorophyll.
G. Multimodality Biological Imaging
OCT, LIF, and LSCM may be used in combination to
improve the diagnosis or interpretation of tissue
structure. Kuranov et al. combined OCT and LIF to
study the boundaries of carcinoma in the cervix.
25
By
combining modalities and employing an exogenous
fluorescent dye, they were able to more accurately
identify the tumor boundary. McNichols et al. re-
ported the use of a fluorescence-guided OCT endo-
scope for use in the detection of oral cancer.
26
They
employed fluorescence for rapid screening of a large
area to guide the use of OCT in areas identified as
suspect by the autofluorescence signal. OCT and
LSCM have been used in combination to improve the
interpretation of tissue structure. Neerken et al. used
OCT and LSCM in the study of age-related effects in
human skin.
27
They found that combined OCT and
LSCM helped identify the sources of contrast seen in
each imaging modality and gave a better understand-
ing of the layered structure of human skin.
Because OCT and LIF are both highly compatible
with fiber-based technology, we have developed en-
doscopes for in vivo applications in mouse models
that combine the two modalities.
28
The miniaturiza-
tion of OCT–LIF technologies enables minimally in-
vasive and nondestructive measurements, thereby
facilitating time-serial monitoring of therapies and
disease development.
7
The goal of the present study
was to compare in vivo combined OCT–LIF data to ex
vivo LSCM images and histology at corresponding
1 May 2006 Vol. 45, No. 13 APPLIED OPTICS 3051
locations. A qualitative, visual comparison of OCT
and LSCM images to histology was performed, and
the sensitivity and specificity of the LIF spectra to
disease was computed. Relative magnitudes of 390,
450, and 680 nm emission among disease, diets, and
mouse strain were analyzed in an effort to under-
stand the origins of the autofluorescent signals.
2. Materials and Methods
A. Animals and Diets
Two groups of mice (Jackson Labs, Bar Harbor,
Maine) were used in this experiment. The first group
was fed a custom, chlorophyll-enriched diet 20%
spinach) (Harlan Teklad, Indianapolis, Indiana),
while the second group was fed the standard NIH-31
Modified MouseRat Sterilizable Diet (Harlan Tek-
lad, Indianapolis, Indiana). The special diet group
consisted of 10 mice: 6 cancer model C57BL
6J-Apc
Min
mice and 4 control C57BL6J mice. The
normal diet group consisted of 16 mice: 10 cancer
model C57BL6J-Apc
Min
and 6 control C57BL6J
mice. All mice were housed in microisolators at the
University Animal Care facility on a 12:12 light:dark
cycle. All mice had free access to food and a 2%
arginine-enriched drink. All animal protocols were
approved by the University of Arizona Insti-
tutional Animal Care and Use Committee. In addi-
tion, LIF data from a previous study (unpublished)
using a group of four C57BL6J and three C57BL
6J-Apc
Min
mice that were fed a synthetic,
chlorophyll-free diet (AIN-93G, Harlan Teklad, Indi-
anapolis, Indiana) were reviewed.
B. Combined OCT–LIF System
A combined OCT–LIF system was used to image all
mice in this study. Data for both modalities were
collected simultaneously through the use of a dual-
modality endoscope. We have previously described a
similar endoscope-based OCT–LIF system in detail.
28
Briefly, the OCT subsystem consisted of a superlu-
minescent diode (SLD) source (Superlum Diodes,
Moscow, Russia) with a center wavelength of
1310 nm and a bandwidth of 60 nm. The measured
axial and lateral resolutions were 15 m (in air) and
20 m, respectively. Assuming an average index of
refraction for tissue of 1.4, the OCT image depth was
approximately 1.4 mm. The LIF subsystem used the
325 nm line of a He–Cd laser (Kimmon Electric,
Englewood, Colorado) as the excitation source. The
excitation light was coupled into a multimode, alu-
minized, fused-silica fiber with a 200 m core and a
0.22 N.A. (Fiberguide Industries, Stirling, New Jer-
sey). The autofluorescence emission was collected by
an identical pair of multimode, aluminized, fused-
silica fibers and was focused onto the entrance slit of
a modified Czerny–Turner grating spectrometer
(Triax 180, Jobin Yvon, Edison, New Jersey). Spec-
tral data were collected over the wavelength region
from 300 to 750 nm with a resolution of approxi-
mately 5 nm. Fluorescence spectra were collected
continuously throughout the scans witha1sinte-
gration time per spectrum.
The OCT–LIF endoscope combined the two sepa-
rate subsystems into a miniature 2 mm outer
diameter package. The OCT optics consisted of a
single-mode fiber, a centering ferrule, a gradient-
index (GRIN) lens to focus the beam, and a rod prism
to bend the beam through the endoscope envelope
into the tissue. The LIF optics consisted of three fi-
bers and the same prism. The LIF fibers were secured
to the top of the OCT ferrule–GRIN lens assembly,
and the fiber tips were coupled directly to the front
surface of the rod prism. A schematic of the inte-
grated OCT and LIF subsystems is presented in Fig.
1, and a solid model of the OCT–LIF endoscope is
shown in Fig. 2.
While the basic optical design of the endoscope
used in this study was similar to that previously
described,
28
several modifications were made to facil-
itate more rapid and accurate mouse colon imaging.
The limited scan length and lack of rotation capabil-
ities of the original endoscope necessitated several
different insertions of the catheter to scan the entire
length of the distal colon and manual rotation of the
catheter to achieve scans at several angles. The cath-
eter mounts, drive motors, and optical packaging
were modified to enable collection of OCT and LIF
data over the entire 35 mm length of the distal mouse
colon in a single longitudinal scan, with automated
rotation of the catheter tip optics enabling collection
of OCT and LIF data at different angles within the
colon. Custom LabVIEW software was also developed
to control the longitudinal scans and angular rota-
tions without any experimenter intervention.
C. OCT–LIF Imaging
In vivo simultaneous OCT–LIF imaging was per-
formed on each 18-week-old mouse in the study. Ap-
proximately 12 h before the imaging procedure,
access to food was removed while free access to water
was maintained in an attempt to minimize the pres-
ence of feces in the lower GI tract. Immediately before
imaging, the mice were anesthetized with a 2.5%
Avertin solution administered intraperitoneally at a
dose of 0.015 mL Avertin per gram of mouse body
mass. The endoscope was lubricated (KY Liquid,
McNeil-PPC, Skillman, New Jersey) and then in-
serted into the anus. The imaging sequence consisted
of eight longitudinal scans, each 35 mm in length.
After each longitudinal scan, the catheter tip optics
were rotated 45° before the commencement of the
subsequent scan. The mouse remained in a prone
position throughout the scan. The entire imaging se-
quence took approximately 40 min to complete.
D. Tissue Harvesting
After the OCT–LIF imaging sequence was com-
pleted, the subjects were euthanized with either an
overdose of the 2.5% Avertin solution or by CO
2
as-
phyxiation. The abdominal cavity of the mouse was
opened, and the distal portions of the colon and rec-
tum were excised. The colon was then sliced longitu-
3052 APPLIED OPTICS Vol. 45, No. 13 1 May 2006
dinally, rinsed with a saline solution, and laid flat
with the luminal side outward. Tissue samples that
were imaged on the LSCM were stored in a physio-
logical phosphate-buffered saline solution and were
refrigerated until imaging. All the tissue samples
were ultimately placed in a Histochoice (AMRESCO,
Solon, Ohio) fixative solution, no later than 6 h after
harvesting.
E. Laser-Scanning Confocal Microscope Imaging
A LSM 510 META (Zeiss, Jena, Germany) laser-
scanning microscope was used to perform the LSCM
imaging. An excitation wavelength of 458 nm from a
30 mW, Ar
laser was used. Sample fluorescence was
collected in an epifluorescence geometry. Two water-
dipping objectives from Olympus 10 (0.3 N.A.) and
40 (0.8 N.A.) were used to achieve varying levels of
resolution and fields of view. The detector was a 32
element, spectral photomultiplier tube (PMT) array
(META detector). The META detector was configured
in a two channel setup, with one channel for the
460–600 nm spectral region and the second chan-
nel for the 600–700 nm region to isolate any
680 nm signal. Zeiss LSM5 software was used to ac-
quire the confocal images. Based on the appearance
of the OCT–LIF data, 16 colons (several each of
healthy and abnormal) were selected for imaging on
the confocal microscope.
F. Histology
Histology was used as the gold standard for deter-
mining the tissue disease state. After being fixed in
the Histochoice solution, the tissue samples were pro-
cessed, embedded in paraffin, and sectioned longitu-
dinally. Sections were obtained that corresponded to
the angular locations of each OCT–LIF scan. The
sections were stained with hematoxylin and eosin
and were analyzed by an experienced pathologist.
G. Data Analysis
The OCT and LSCM images were compared visually
with each other and with the corresponding histology
sections to establish a correlation between the struc-
Fig. 1. Schematic of the OCT–LIF system. The OCT superluminescent diode (SLD) source is passed to a 50:50 fiber beam splitter. The
reference arm path travels to a collimator, through a neutral-density (ND) filter, and then retraces its path back to the beam splitter after
reflection from a retroreflector. The sample arm path is coupled into the dual-modality endoscope, where the light is focused into and
collected from the tissue. The light from the two paths is recombined at the beam splitter and travels to a detector. The detector signal
is demodulated by a lock-in amplifier, and the data are collected and stored by the computer. The LIF He–Cd laser source is coupled into
a multimode fiber and into the dual-modality endoscope. The excitation light illuminates the tissue, and two adjacent collection fibers in
the endoscope collect the emission light and carry it to the CCD spectrometer, where the data are recorded by a computer.
Fig. 2. Rendering of the OCT–LIF endoscope tip optics. The OCT
channel consists of a single-mode fiber (F) centered on a GRIN lens
(G) by a silica ferrule (SF). The GRIN lens focuses the light into a
rod prism (R) that bends the beam at 90° through the silica window
(W) and into the tissue. The LIF channel consists of three multi-
mode, aluminized, fused-silica fibers. The He–Cd laser is coupled
into the excitation channel (Ex). The rod prism (R) reflects the light
through the silica window (W) onto the tissue. Autofluorescence
from the tissue is collected by the two emission channel fibers (Em)
that carry the signal to a spectrometer.
1 May 2006 Vol. 45, No. 13 APPLIED OPTICS 3053
ture visualized in each modality and the presence
and size of any adenomas. The OCT and LIF data
were also compared visually to identify fluorescence
changes over any abnormal tissue regions.
In addition, quantitative statistical analyses were
performed on the LIF data. Previous studies of the
LIF spectra of human and murine colonic adenomas
have found a reduction in signal at 390 and 450 nm
and an increase in signal at 680 nm.
7,9
Based on
these findings, we analyzed the relative intensities at
these wavelengths for all the mice in the study. The
mice were divided into six groups based on three
parameters: state of tissue health, diet, and breed.
Table 1 summarizes the groups and the number of
subjects in each group. The boundaries of the tissue
used for group 1 (adenoma) and group 2 (healthy
tissue in colon with adenoma) mice were determined
by visual inspection and correlation of the histologi-
cally confirmed adenoma in the OCT image with the
LIF spectra. The mean and standard deviation of the
680:390 nm, 680:480 nm, and 680:390 nm ratios
were calculated for each group.
The absolute signal intensities at 390, 450, and
680 nm were then used to analyze the sensitivity and
specificity of LIF in identifying adenomas, where sen-
sitivity is defined as
# true positives兾共# true positives
# false negatives),
and specificity is defined as
# true negatives兾共# true negatives
# false positives).
Based on data from previous studies, positive identi-
fication of adenoma was associated with 390 and
450 nm spectral intensities below a threshold value
and a 680 nm spectral intensity above a threshold.
7,9
Optimal threshold values for each wavelength (high-
est sensitivity and specificity for diseased and non-
diseased areas of tissue) were determined for mice
that developed adenomas (groups 1 and 2). These
same thresholds were then applied to nondiseased
mice of both breeds in the normal diet group to de-
termine the specificity in this group.
Finally, the fluorescent signal intensities at 390
and 450 nm were analyzed for the different mouse
strains in nondiseased mice. The absolute signal in-
tensities and their variation were compared to deter-
mine if the LIF signal varied between the different
strains.
3. Results
As expected, none of the C57BL6J control mice in
either diet group spontaneously developed disease in
the colon. Of the six C57BL6J-Apc
Min
mice on the
enriched-chlorophyll diet, none developed adenomas
that were identified by the OCT–LIF, LSCM imaging,
or pathological evaluation of the histology. Of the 10
C57BL6J-Apc
Min
mice on the normal diet, 1 mouse
died before imaging and 5 mice each developed one
histologically confirmed adenoma in the lower colon.
In one of these mice the adenoma was just beyond the
scan range of the endoscope and was not imaged with
OCT–LIF; however, it was imaged with LSCM. The
adenoma in a second mouse had significantly ob-
structed the bowel, and insertion of the catheter re-
sulted in a perforated colon distal to the adenoma.
This mouse was excluded from all data analysis. Con-
sequently, OCT and LIF data were acquired from
three of the five adenomas that developed in the dis-
tal colon, and LSCM images were taken from four of
the five samples.
A. Imaging Modality Performance Metrics
Table 2 summarizes the performance metrics, as
identified in Section 1, for the OCT, LIF, and LSCM
systems used in the study. The OCT and LSCM im-
ages provide structural information about the tissue
samples at different scales; the OCT endoscope has
only tissue layer resolution but relatively large depth
of penetration, whereas the LSCM obtains subcellu-
lar resolution but shows significant signal reduction
after depths of 100 m. The OCT and LSCM data
were also complementary in the sense that the OCT
images provided a depth-resolved image of the tissue,
while the LSCM images provided an en face perspec-
tive. Biochemical information about the tissue state
was provided by both the LIF and LSCM data sets.
Table 1. LIF Data Analysis Groups
a
Group Number Group Description
Number of
Mice in Group
1 Adenoma mice (Apc
Min
, normal diet) 3
2 Adjacent healthy tissue in mice with adenoma (Apc
Min
, normal diet) 3
3 Nonadenomatous Apc
Min
mice on normal diet 4
4 Nonadenomatous control mice on normal diet 5
5 Nonadenomatous Apc
Min
mice on high-chlorophyll diet 5
6 Nonadenomatous control mice on high-chlorophyll diet 4
a
The colon tissue samples were divided into six different groups for the LIF data analysis based on tissue state, mouse strain, and diet.
Groups 1 and 2 represent different areas (adenomatous versus healthy) of the same tissue sample. The mice in groups 3–6 were all healthy
and were grouped based on genetic strain and diet.
3054 APPLIED OPTICS Vol. 45, No. 13 1 May 2006
The LIF provided low-resolution, gross-tissue fluo-
rescent properties, while the LSCM images provided
biochemical information at the cellular and subcellu-
lar levels. The relevant performance metrics and
trade-offs of each modality are further highlighted in
the following sections with the comparison of healthy
and diseased tissue samples.
B. Appearance of Healthy Tissue
Figures 3– 6 show results from histologically con-
firmed healthy colon tissue. Figure 3 compares a rep-
resentative OCT image to the histology. OCT was
readily capable of imaging the layered structure of
the healthy tissue. The OCT image shows the mu-
cosa, the mucosa–submucosa, and the submucosa–
tunica media boundaries and the adventitia. In this
image, as well as in many others in our study, addi-
tional tissue structures outside the colon were visu-
alized, including pancreas and adipose tissues.
Figure 4 compares a LSCM image to an en face
histology image of healthy colon tissue from a mouse
on the regular diet. The LSCM images of healthy
colon showed a well-structured tissue characterized
by regular crypt structures. The circular crypt struc-
tures associated with the surface tissue layer gave
way to the cell-lined crypt walls at greater depths in
the LSCM images (data not shown). The structure
was consistent with the histology, and the diameter
of the crypt structures was in close agreement in both
modalities. At higher resolutions, the LSCM was ca-
pable of visualizing cellular- and subcellular-level de-
tails of the cells in the mucosa lining the crypt
structures. Figure 5 shows a high-magnification
LSCM image in which the nuclei and other cellular
organelles were resolved. The autofluorescence signal
from the healthy tissue was predominantly in the
wavelength region of 460–600 nm.
Figure 6 shows the LIF spectra of a healthy mouse
correlated to the OCT image. These spectra were typ-
ical of healthy colonic tissue and were consistent with
previous studies.
22
The two peaks at approximately
390 and 450 nm were attributed to the autofluores-
cence of collagen and NADH, respectively. The spec-
tral dip at approximately 420 nm was consistent with
hemoglobin absorption. This set of LIF spectra were
taken from a mouse on the enriched-chlorophyll diet.
The distinct fluorescence peak seen at 680 nm was
present in all mice on the high-chlorophyll diet.
C. Chlorophyll Study
To determine if the 680 nm LIF emission peak was
the result of chlorophyll metabolites, we compared
Fig. 3. OCT and histology images of healthy colon. (a) In the OCT image the top reflective surface is the glass window (W) of the endoscope.
The mucosa (M), mucosa–submucosa (MSM) boundary, and the submucosa–tunica media (SMTM) boundary are all clearly visible. Other
tissue outside the colon [pancreas (P) and adipose tissue (Ad)] is visible. (b) Structure of the histology closely correlates with the OCT image.
The pancreas and adipose tissue in the OCT image can also be seen in the histology.
Table 2. Modality Performance Comparison
a
Modality
Lateral
Resolution
Axial
Resolution
Depth of
Penetration
Structural
Data
Biochemical
Data
Image
Perspective
OCT 20 m15m 1.4 mm Yes No Cross sectional
LIF 1 mm NA 150 m No Yes NA
LSCM 1 m3m 100 m
b
Yes Yes En face
a
Comparison of important performance metrics, as identified in Section 1, of the optical modalities employed in this study.
b
Autofluorescent signal was detectable over approximately ten 10 m steps in the LSCM system.
1 May 2006 Vol. 45, No. 13 APPLIED OPTICS 3055
the absolute intensity of the signal at 680 nm in all
healthy mice on the enriched-chlorophyll and normal
lab-chow diets (Fig. 7). There were nine healthy mice
in each diet sample set. The intensity of the 680 nm
signal in the high-chlorophyll diet group appeared
significantly stronger than the signal in the normal-
diet mice, an observation confirmed via a Student’s
t-test p 0.01. The variation in the 680 nm signal,
represented by the standard deviation of the signal
intensity, was also slightly larger in the high-
chlorophyll-diet mice relative to the normal-diet
mice. In previous studies of mice on a chlorophyll-free
diet, the observed 680 nm intensity was signifi-
cantly smaller than the normal p 0.02 and high-
chlorophyll p 0.01 diet groups. The average
intensity of the 680 nm signal in the high-
chlorophyll-diet mice was 650% larger than the av-
erage 680 nm signal in the chlorophyll-free-diet mice.
In the normal-diet mice the average intensity of the
680 nm signal was 60% larger than that of the
chlorophyll-free-diet mice.
D. Appearance of Diseased Tissue
Of the three adenomas within the imaging range of
the OCT–LIF endoscope, all three were distinctly vi-
sualized in the OCT images. The adenomatous areas
showed a loss of the layered structure that charac-
terized the normal colon tissue. The layered structure
Fig. 4. (a) LSCM image (10) of healthy colon. (b) En face his-
tology image (10) of healthy colon. In both images the regular,
circular structure of the crypts is readily discernible. The diameter
of the crypt structures in both the LSCM and the histology is in
close agreement (of the order of 25–50 m).
Fig. 5. High-magnification (40) LSCM image of healthy mouse
colon. The cellular-level structure of the tissue is clearly visualized,
with the cellular boundary characterized by areas of continuous
heterogeneous fluorescence. The hypointense areas visible within
some of the cells are nuclei. The bright autofluorescent signal from
within the cells is an unknown fluorescent component, potentially
other cellular organelles.
Fig. 6. OCT image (top) and corresponding LIF spectra (bottom)
of healthy colon taken from a high-chlorophyll diet mouse. The LIF
emission peaks at approximately 390 and 450 nm, with an absorp-
tion dip at 420 nm that is typical of healthy tissue. The presence of
a distinct peak at approximately 680 nm is visible throughout the
colon as well.
3056 APPLIED OPTICS Vol. 45, No. 13 1 May 2006
of the healthy tissue immediately proximal and distal
to the adenomas was seen to drop sharply at the
boundary of the adenoma and was replaced by a dis-
organized mass, typically a few millimeters in length.
An OCT image of one of the adenomas and the cor-
responding histology is presented in Fig. 8. In this
OCT image several hypointense oval-shaped regions
can be seen within the adenoma that correlate to
mucin-filled dilated crypts seen in the histology. All
three adenomas were visualized over an angular
range of 45°–90°.
The LSCM images of adenomas also showed an
irregular structure [Fig. 9(a)]. LSCM images taken
over the adenomas did not show the same organized
crypt structure that characterized the normal colon
tissue. Instead, the colon tissue appeared smoother
on the surface, with deep folds. At higher resolutions,
the cellular-level structure and organization of the
adenomatous tissue differed notably from the healthy
tissue. While the healthy tissue was characterized by
tightly packed, consistently organized cells of similar
size and shape, the cellular-level organization of the
adenoma appeared more haphazard. LSCM was also
capable of identifying the gross boundary of the ade-
noma, where the mucosa thickness changed rapidly
over a depth of greater than 1 mm. This boundary is
seen in Fig. 9(b).
Figure 10 shows the LIF spectra from an adenoma
correlated to the OCT image. As expected, the spectra
show a decrease in the absolute 390 and 450 nm sig-
nals and an increase in the absolute 680 nm signal.
The mean and standard deviation of the 680:390 nm,
680:450 nm, and 450:390 nm ratios for each mouse
group are plotted in Fig. 11. The mean for all three
ratios was larger in regions identified as adenoma
compared with healthy tissue of mice on the same
(normal) diet. The mean ratio in the healthy areas of
colons with adenomas was similar to those in compa-
rable mice in which no adenoma had developed, al-
though the standard deviations were larger. The
680:450 nm and 680:390 nm mean ratios in the
enriched-chlorophyll-diet mice were higher than all
healthy, normal-diet mice; however, the means of the
450:390 nm ratios were comparable between all
healthy mice of different diets and strains.
Using the absolute signal intensity at 390, 450, and
680 nm, we calculated the sensitivity and specificity
of the LIF data to disease to be 81% and 80%, respec-
tively. The same test was then applied to nondiseased
mice in the normal diet group. A specificity of 87%
was achieved for all mice in this diet group, with 82%
Fig. 7. Average and standard deviation of absolute intensity of
the LIF signal at 680 nm in normal-diet mice (1–9) and high-
chlorophyll-diet mice (10–18).
Fig. 8. (a) OCT image of colonic adenoma (A). Note the healthy, structured tissue immediately proximal and distal to the adenoma. The
adenoma is characterized by a thickening of the mucosal layer and a loss of layered structure. Mucin-filled dilated crypts (DC) can be seen
in the adenoma in the OCT image. (b) Corresponding histology. The histology of the adenoma is consistent with the structure seen in the
OCT image, including the presence of the dilated crypts. Other abbreviations as in Fig. 3.
1 May 2006 Vol. 45, No. 13 APPLIED OPTICS 3057
specificity in the control group and 93% specificity in
the Apc
Min
group.
The fluorescent signal intensity at 390 and 450 nm
for nondiseased mice of the two different strains is
shown in Fig. 12. The variance and range of the sig-
nal at both 390 and 450 nm is notably larger in the
Apc
Min
mice compared with those of the C57 control
mice. The variance in the signal intensity was be-
tween 207% and 547% greater in the Apc
Min
group
relative to the control group on the same diet. An
F-test p 0.01 confirmed that the difference in
variances at both 390 and 450 nm across diet groups
was statistically significant.
4. Discussion
In this study we used four different optical modalities
to accomplish two specific imaging tasks in the inves-
tigation of mouse colon cancer.
Our first task was multimodal classification of both
healthy and diseased tissue. In each modality differ-
ences were seen between healthy and adenomatous
tissue. However, each modality provided different
resolution scales and complementary information
about the tissue structure and biochemistry that po-
tentially could be used to improve the diagnostic and
screening capabilities relative to any single modality
on its own.
In pursuit of this first task, OCT successfully
Fig. 9. (a) LSCM image of an adenoma. The consistent crypt structure that characterizes the healthy colon tissue is not visible. The
surface of the adenoma appears to have a number of folds. (b) The xy projection of a z stack taken at the edge of the adenoma. The adenoma
is denoted by (A), and healthy tissue adjacent to the adenoma is denoted by (H). This LSCM image was taken over a depth of 1.3 mm at
z steps of 50 m to demonstrate the depth scale of the adenoma. The large z steps required to take the image resulted in the staircase
appearance of the adenoma edge.
Fig. 10. OCT–LIF spectra of adenoma. The adenoma can be vi-
sualized in the OCT image (A). The corresponding LIF signal
shows a significant reduction in the 390 and 450 nm signals and a
peak in the 680 nm signal over the adenomatous region.
Fig. 11. Mean and standard deviation of 680:390 nm, 680:450
nm, and 450:390 nm ratios for each of six groups of mice.
3058 APPLIED OPTICS Vol. 45, No. 13 1 May 2006
showed tissue-layer resolution of the structure of the
colon in both healthy and diseased tissue. The layers
visible in the depth-resolved OCT images of healthy
colon closely correlate with the layered structure vis-
ible in histology (Fig. 3). The resolution of our OCT
system was not sufficient to identify the crypt struc-
tures in the mucosa or other cellular-level details.
However, recent development of high-resolution OCT
systems has enabled the visualization of structures
within the rabbit mucosa, including crypts.
4
We have
recently developed an ultrahigh-resolution catheter
with approximately 4 m axial and lateral resolution
that has imaged structures consistent with murine
colon crypts in situ.
29
We plan a future study to con-
firm this finding. While not quantitatively examined
in this study, the highly resolved tissue-layer thick-
nesses in the OCT images can be measured and an-
alyzed as an indicator of tissue condition. Previous
studies have shown that layer thicknesses in OCT
images closely correlate with healthy and diseased
tissue-layer thicknesses measured in histology.
7
The LSCM images also provided structural data
that correlated closely with histology but with a sig-
nificantly higher resolution relative to the OCT
images. The circular structures visible in the
autofluorescence LSCM images closely correlate with
the crypts visible in histology (Fig. 4). Subcellular
structures were also identifiable in the higher-
magnification LSCM images that correlate with
structures, such as nuclei, visible in the histology
(Fig. 5). Further miniaturization and improvements
in the dynamic range of LSCM endoscopic-based sys-
tems may enable imaging of these subcellular struc-
tures in vivo and in real time, although complete
coverage of the lower colon may be unachievable in a
reasonable amount of time. With a weak colon
autofluorescence signal, collecting a single high-
resolution, multiple-depth image stack with accept-
able image quality can take several minutes with the
LSM 510 instrument. Because of the high magnifica-
tion of the microscope objective, the field of view of
LSCM images is limited to a few hundred microme-
ters. Using such a time-consuming technique to cover
the entire colon would not be practical for in vivo
applications in a clinical setting. However, because
high-speed, helical-scanning OCT systems could per-
mit a complete, high-resolution, depth-resolved scan
of the lower mouse colon in a few minutes, OCT could
be used to identify potentially abnormal tissue struc-
tures that could be further investigated by using
higher-resolution LSCM. In our study of the colon
tissue ex vivo on the LSCM, we used the OCT–LIF
endoscopic data to guide our imaging to particular
areas of interest, such as adenomas. Combining mo-
dalities to decrease measurement time and improve
tissue targeting is a clinically relevant benefit of the
multimodality approach.
The large depth of penetration of OCT imaging
enabled visualization and differentiation of tissue
outside the colon. In the OCT image of healthy colon
presented in Fig. 3, two different tissue types (pan-
creas and adipose) can be visualized external to the
colon. In this context one major advantage of OCT in
imaging biological tissue is its capability of probing
more than 1 mm into the tissue to determine the
thickness of layers and to identify subsurface
structures. LSCM has a limited penetration depth
共⬃100 m, making autofluorescent LSCM impracti-
cal for identifying layers below the mucosa. However,
LSCM can be useful in determining the thickness of
certain structures. The LSCM image [Fig. 9(b)] and
the OCT image [Fig. 8(a)] of the adenoma both show
Fig. 12. Variability of 390 and 450 nm signals in
healthy Apc
Min
and C57 control mice. The C57
mice exhibit less variability than do the Apc
Min
mice across all diets.
1 May 2006 Vol. 45, No. 13 APPLIED OPTICS 3059
the steep slope of the tissue at the edge of the ade-
noma. In the LSCM image a change in tissue height
of more than 1 mm occurs in less than 200 m later-
ally. In the OCT image this edge is seen as the sharp
falloff in the healthy, structured tissue adjacent to
the adenoma. The insertion of the catheter into the
colon pushes against the adenoma and deeper colon
tissue, forcing this steep drop-off at the edge of the
adenoma. OCT is not capable of imaging through the
entire depth of the adenoma, making it difficult to
determine its thickness. However, a single-axis pro-
jection of the en face LSCM image over a full depth
scan at the edge of the adenoma enables estimation of
its height [Fig. 9(b)]. The measurement of approxi-
mately 1 mm in the LSCM image is consistent with
the histology of this adenoma, seen in Fig. 8(b).
The LIF modality was also capable of aiding the
task of multimodal classification and differentiation
of healthy and diseased colon tissue. LIF does not
provide structural information and in our system was
capable of providing only macroscale fluorescence
data. The sensitivity and specificity computed from
our LIF data are not as high as others who have
reported LIF sensitivities and specificities in the
range of 90%–100%.
8,9
Our sensitivity and specificity
might be improved by the use of more complex algo-
rithms. In the present study we simply used constant
threshold values for each of the three wavelengths
examined, and our results might also have been neg-
atively affected by the limited number of adenomas.
The sensitivity and specificity of the LIF modality
might also be further increased with improved design
of the LIF excitation and emission collection channels
in the dual-modality endoscope.
An advantage of our dual-modality endoscope is
that we can spatially and temporally correlate the
LIF spectra with the structure of the tissue seen in
the OCT images. By combining the two modalities,
we do not have to rely solely on the sensitivity and
specificity of the LIF spectra. Suspect LIF signal can
be confirmed with visual inspection of the tissue
structure in the OCT image and vice versa. This ben-
efit can be important from a practical perspective for
in vivo, endoscopic analysis of the colon. The insertion
of the endoscope can cause folds, stretching, or dips in
the tissue that can make visualizing the tissue struc-
ture with OCT difficult in places. Also, fecal contam-
ination can result in OCT images that are difficult to
interpret. Consistent with our task of multimodal
imaging, LIF was successfully used to identify an
ambiguous, abnormal OCT region as adenoma.
LSCM and histology confirmed that this region was
in fact adenomatous.
Our second specific task was the investigation of
the 680 nm autofluorescent signal. LIF was the most
useful modality for the task of studying the correla-
tion of the 680 nm autofluorescent signal and the
enriched-chlorophyll diet. Both the LIF and LSCM
data contributed to the goal of attempting to isolate
the location of the fluorophore. Because OCT provides
only structural data, it did not provide useful data on
the biochemical origins of the 680 nm signal.
In the LIF data we saw a consistent and distinct
peak at 680 nm throughout the colon of all mice on
the enriched-chlorophyll diet. The average 680 nm
intensity was significantly larger in the high-
chlorophyll diet group than in the normal diet group.
While there was some 680 nm signal in the normal
diet group, this signal was less frequently seen as a
distinct peak, and more commonly it was simply the
low-intensity tail of the emission signal from NADH
and collagen. For mice on the chlorophyll-free diet,
the 680 nm emission was even smaller and consisted
solely of this low-intensity tail. The ratio analysis
performed on the LIF data for different mouse
groups further demonstrates the effect of the high-
chlorophyll diet. As we would expect, both the
680:450 nm and 680:390 nm ratios were higher in
both strains of mice on the high-chlorophyll diet com-
pared with the normal diet groups. However, the ra-
tio for 450:390 nm was very similar in both strains
across the diet groups. This evidence strongly sug-
gests that the distinct, single 680 nm peak seen in the
LIF spectra is a by-product of the mouse diet with
chlorophyll metabolites, such as pheophorbide-a and
pyropheophorbide-a, as the autofluorescent source.
30
Both of these compounds are known to have a distinct
fluorescence peak at approximately 680 nm, and nei-
ther has a second peak at approximately 630 nm that
is characteristic of other porphyrin derivatives, such
as protoporphyrin IX and hematoporphyrin deriva-
tive.
24,31
The LIF data, however, was less useful for the task
of determining the location of the 680 nm fluores-
cence. While LIF indicated the macroscale presence
of a fluorophore at 680 nm emission, it could not de-
termine where in the tissue this autofluorescent sig-
nal originated on account of the limited resolution of
our catheter design. The LSCM data complemented
the LIF data and was helpful in providing clues as to
the origins of the fluorescence signal. For this specific
task, LSCM offered a major advantage because it
combined fluorescence data with high-resolution im-
ages of the tissue structure that enabled spatial cor-
relation of image data with biochemical information.
The LSCM data showed the red signal seen in the
high-chlorophyll-diet mice was often from a large par-
ticle on top of the tissue, clearly resulting from fecal
contamination. In some cases, however, the red fluo-
rescence was more diffuse across and within the crypt
structures of the colon. Further studies are needed to
determine whether this diffuse signal is indeed inter-
cellular or simply due to extremely small particle-size
contamination. Interestingly, the 680 nm fluores-
cence in the adenomas was significantly higher than
in the adjacent tissue. It is still unclear whether this
signal increase resulted from an increase in fecal con-
tamination due to the convoluted surface of the
adenoma or the selective uptake of chlorophyll
metabolites in the diseased cells. If this signal is
determined to be intercellular, chlorophyll might po-
tentially be used as a natural exogenous fluorescent
contrast agent.
One further result of potential interest is the vari-
3060 APPLIED OPTICS Vol. 45, No. 13 1 May 2006
ation observed in both the 390 and the 450 nm LIF
signals between the Apc
Min
mice and the C57 con-
trol mice. In the OCT images and histological sections
of these different mice strains, no discernible differ-
ence in the structure of healthy tissue was noted.
However, we found that the variation of the LIF sig-
nal from these healthy mice was dependent on strain,
with the variance of the signal at both 390 and
450 nm significantly larger in the Apc
Min
group
than in the control group. This suggests that geno-
type differences between the mouse groups are ex-
pressed as phenotype changes evident in the LIF
signal. This genotype-dependent variation in LIF in-
tensity at 390 and 450 nm was present in both diet
groups.
5. Conclusion
The choice of modality and instrument design is
driven by the tissue being studied and whether the
diagnosis is to be performed in vivo or ex vivo. Optical
modalities offer several unique advantages in resolu-
tion and depth of penetration as well as in miniatur-
ization for in vivo applications. This study considered
four optical modalities that are suitable for in vivo
and ex vivo imaging in a mouse model of colon cancer.
OCT and LSCM offer the potential of acquiring
tissue-structure data that closely correlate with his-
tology in real time and nondestructively. Both modal-
ities were effectively capable of identifying and
differentiating healthy and diseased tissue states.
The LSCM and LIF data give further insight into the
biochemical properties of the colon tissue that can be
combined with structural information to help deter-
mine disease state.
The greatest promise of the OCT, LIF, and LSCM
modalities for clinical applications lies in their com-
patibility with optical fiber-based designs. Improving
endoscope technologies makes the goal of a device
that combines all three modalities to enable real
time, in vivo imaging a realistic, near-term prospect.
While histology will remain the gold standard for
pathological diagnosis, the capability of studying and
monitoring tissue in vivo and nondestructively with
optical imaging techniques is beneficial for studies of
the progression of disease over time and the efficacy
of therapies.
This research was supported in part by grants from
the National Institutes of Health (CA109385) and the
Specialized Program of Research Excellence in Gas-
trointestinal Cancers (CA095060).
References
1. American Cancer Society, “Cancer Facts and Figures 2004”
(American Cancer Society, 2004), http://www.cancer.org/
downloads/STT/CAFF_finalPWSecured.pdf.
2. G. J. Tearney, S. A. Boppart, B. E. Bouma, M. E. Brezinski,
N. J. Weissman, J. F. Southern, and J. G. Fujimoto, “Scanning
single-mode fiber optic catheter-endoscope for optical coher-
ence tomography,” Opt. Lett. 21, 543–545 (1996).
3. B. E. Bouma and G. J. Tearney, “Power-efficient nonreciprocal
interferometer and linear-scanning fiber-optic catheter for op-
tical coherence tomography,” Opt. Lett. 24, 531–533 (1999).
4. P. R. Herz, Y. Chen, A. D. Aguirre, J. G. Fujimoto, H. Mashimo,
J. Schmitt, A. Koski, J. Goodnow, and C. Petersen, “Ultrahigh
resolution optical biopsy with endoscopic optical coherence to-
mography,” Opt. Express 12, 3532–3542 (2004).
5. J. M. Poneros, S. Brand, B. E. Bouma, G. J. Tearney, C. C.
Compton, and N. S. Nishioka, “Diagnosis of specialized intes-
tinal metaplasia by optical coherence tomography,” Gastroen-
terology 120, 7–12 (2001).
6. S. Jäckle, N. Gladkova, F. Feldchtein, A. Terentieva, B. Brand,
G. Gelikonov, V. Gelikonov, A. Sergeev, A. Fritscher-Ravens, J.
Freund, U. Seitz, S. Schröder, and N. Soehendra, In vivo
endoscopic optical coherence tomography of the human gastro-
intestinal tract—toward optical biopsy,” Endoscopy 32, 743–
749 (2000).
7. L. P. Hariri, A. R. Tumlinson, N. H. Wade, D. G. Besselsen, U.
Utzinger, E. W. Gerner, and J. K. Barton, “Endoscopic optical
coherence tomography and laser induced fluorescence spec-
troscopy in murine colon cancer model,” Lasers Surg. Med. (to
be published).
8. R. Richards-Kortum, R. P. Rava, R. E. Petras, M. Fitzmaurice,
M. Sivak, and M. S. Feld, “Spectroscopic diagnosis of colonic
dysplasia,” Photochem. Photobiol. 53, 777–786 (1991).
9. R. M. Cothren, R. Richards-Kortum, M. V. Sivak, M. Fitzmau-
rice, R. P. Rava, G. A. Boyce, M. Doxtader, R. Blackman, T. B.
Ivanc, G. B. Hayes, M. S. Feld, and R. E. Petras, “Gastroin-
testinal tissue diagnosis by laser-induced fluorescence spec-
troscopy at endoscopy,” Gastrointest. Endosc. 36, 105–111
(1990).
10. S. Fu, C. T. Chia, C. L. Tang, C. H. Diong, and C. Seow,
“Changes in in-vivo autofluorescence spectra at different peri-
ods in rat colorectal tumor progression,” in Diagnostic Optical
Spectroscopy in Biomedicine, T. G. Papazoglou and G. A. Wag-
nières, eds., Proc. SPIE 4432, 118–123 (2001).
11. H. W. Wang, J. Willis, M. I. F. Canto, M. V. Sivak, and J. A.
Izatt, “Quantitative laser scanning confocal autofluorescence
microscopy of normal, premalignant, and malignant colonic
tissues,” IEEE Trans. Bio-Med. Eng. 46, 1246–1252 (1999).
12. G. S. Fiarman, M. H. Nathanson, A. B. West, L. I. Deckelbaum,
L. Kelly, and C. R. Kapadia, “Differences in laser-induced
autofluorescence between adenomatous and hyperplastic
polyps and normal colonic mucosa by confocal microscopy,”
Dig. Dis. Sci. 40, 1261–1268 (1995).
13. J. Chu, S. Chu, and M. H. Montrose, “Apical Na
H
exchange
near the base of mouse colonic crypts,” Am. J. Physiol. 283,
C358–C372 (2002).
14. K. Kataoka, E. Suzaki, and K. Komura, “The Golgi apparatus
of goblet cells in the mouse descending colon: three-
dimensional visualization using a confocal laser scanning mi-
croscope,” Histochem. Cell Biol. 116, 329–335 (2001).
15. J. Knittel, L. Schnieder, G. Buess, B. Messerschmidt, and T.
Possner, “Endoscope-compatible confocal microscope using a
gradient index-lens system,” Opt. Commun. 188, 267–273
(2001).
16. A. R. Rouse, A. Kano, J. A. Udovich, S. M. Kroto, and A. F.
Gmitro, “Design and demonstration of a miniature catheter for
a confocal microendoscope,” Appl. Opt. 43, 5763–5771 (2004).
17. W. J. McLaren, P. Anikijenko, S. G. Thomas, P. M. Delaney,
and R. G. King, In vivo detection of morphological and micro-
vasculature changes of the colon in association with colitis
using fiberoptic confocal imaging,” Dig. Dis. Sci. 47, 2424–2433
(2002).
18. R. Kiesslich, J. Burg, M. Vieth, J. Gnaendiger, M. Enders, P.
Delaney, A. Polglase, W. McLaren, D. Janell, S. Thomas, B.
Nafe, P. R. Galle, and M. F. Neurath, “Confocal laser endos-
copy for diagnosing intraepithelial neoplasias and colorectal
cancer in vivo, Gastroenterology 127, 706–713 (2004).
19. L. K. Su, K. W. Kinzler, B. Vogelstein, A. C. Preisinger, A. R.
Moser, C. Luongo, K. A. Gould, and W. F. Dove, “Multiple
1 May 2006 Vol. 45, No. 13 APPLIED OPTICS 3061
intestinal neoplasia caused by a mutation in the murine ho-
molog of the APC gene,” Science 256, 668 670 (1992).
20. G. I. Zonios, R. M. Cothren, J. T. Arendt, J. Wu, J. Van Dam,
J. M. Crawford, R. Manoharan, and M. S. Feld, “Morphological
model of human colon tissue fluorescence,” IEEE Trans. Bio-
Med. Eng. 43, 113–122 (1996).
21. Z. Huang, T. Chia, S. M. Krishnan, and C. Seow, “Study of
laser autofluorescence of human colon tissues,” in Proceedings
of the 20th Annual International Conference of the IEEE En-
gineering in Medicine and Biology Society New York (IEEE,
1998), pp. 2963–2966.
22. K. T. Schomacker, J. K. Frisoli, C. C. Compton, T. J. Flotte,
J. M. Richter, N. S. Nishioka, and T. F. Deutsch, “Ultraviolet
laser-induced fluorescence of colonic tissue: basic biology and
diagnostic potential,” Lasers Surg. Med. 12, 63–78 (1992).
23. C. Sun, E. Duzman, J. Mellot, L.-H. Liaw, and M. W. Berns,
“Spectroscopic, morphologic, and cytotoxic studies on major
fractions of hematoporphyrin derivative and photofrin II,” La-
sers Surg. Med. 7, 171–179 (1987).
24. R. S. DaCosta, H. Andersson, and B. C. Wilson, “Molecular
fluorescence excitation– emission matrices relevant to tissue
spectroscopy,” Photochem. Photobiol. 78, 384 –392 (2003), http:
//eemdb.uhnres.utoronto.ca/cgi-bin/WebObjects/WebFluor.
25. R. V. Kuranov, V. V. Sapozhnikova, N. M. Shakhova, V. M.
Gelikonov, E. V. Zagainova, and S. A. Petrova, “Combined
application of optical methods to increase the information con-
tent of optical coherent tomography in diagnostics of neoplastic
processes,” Quantum Electron. 32, 993–998 (2002).
26. R. J. McNichols, A. Gowda, B. A. Bell, R. M. Johnigan, K. H.
Calhoun, and M. Motamedi, “Development of an endoscopic
fluorescence image guided OCT probe for oral cancer detec-
tion,” in Biomedical Diagnostic, Guidance, and Surgical-Assist
Systems III, T. Vo-Dinh, W. S. Grundfest, and D. A. Benaron,
eds., Proc. SPIE 4254, 23–30 (2001).
27. S. Neerken, G. W. Lucassen, M. A. Bisschop, E. Lenderink, and
A. M. Nuijs, “Characterization of age-related effects in human
skin: a comparative study that applies confocal laser scanning
microscopy and optical coherence tomography,” J. Bio-Med.
Opt. 9, 274–281 (2004).
28. A. R. Tumlinson, L. P. Hariri, U. Utzinger, and J. K. Barton,
“Miniature endoscope for simultaneous optical coherence to-
mography and laser-induced fluorescence measurement,”
Appl. Opt. 43, 113–121 (2004).
29. A. R. Tumlinson, J. McNally, A. Unterhuber, B. Hermann,
H. Sattmann, W. Drexler, and J. K. Barton, “Endoscopic
ultrahigh-resolution OCT for in vivo imaging colon disease
model mice,” in Advanced Biomedical and Clinical Diagnostic
Systems III, T. Vo-Dinh, W. S. Grundfest, D. A. Benaron, and
G. E. Cohn, eds., Proc. SPIE 5692, 307–315 (2005).
30. L. Ma and D. Dolphin, “The metabolites of dietary chloro-
phylls,” Phytochem. 50, 195–202 (1999).
31. K. D. Ashby, J. Wen, P. Chowdhury, T. A. Casey, M. A. Ras-
mussen, and J. W. Petrich, “Fluorescence of dietary porphyrins
as a basis for real-time detection of fecal contamination on
meat,” J. Agric. Food Chem. 51, 3502–3507 (2003).
3062 APPLIED OPTICS Vol. 45, No. 13 1 May 2006
    • "The first dual-modal instrument based on combined fluorescence and OCT was developed by Jennifer K. Barton and her coworkers [2]. They have not only reported development of combined fluorescence-OCT systems with progressively improved features and finally its miniaturization [13,40– 42], but also carried out a series of studies demonstrating the use of these dual-modal systems in a variety of organ systems in different animal models [12,14,17,28,44] . Byeong H. Lee et al. reported a novel dualmodal fluorescence-OCT system using a double-cladding fiber [37,38] and employed it for the analysis of pearls and pearl treatments [16]. "
    Full-text · Article · Jul 2016 · BMC Systems Biology
    • "Autofluorescence of skin, fur, and tissue, due to several cellular components, including NADPH, flavin coenzymes, elastin, and collagen, can interfere significantly with signal from fluorescent reporters if emission wavelengths overlap (Figure 2) [6]. Additionally, chlorophyll present in standard mouse food autofluoresces thus interfering with many common reporters [7]. To compensate for autofluorescence, software has been developed with advanced mathematical modeling to separate the sources of different wavelengths. "
    [Show abstract] [Hide abstract] ABSTRACT: The detection and subsequent quantification of photons emitted from living tissues, using highly sensitive charged-couple device (CCD) cameras, have enabled investigators to noninvasively examine the intricate dynamics of molecular reactions in wide assortment of experimental animals under basal and pathophysiological conditions. Nevertheless, extrapolation of this in vivo optical imaging technology to the study of the mammalian brain and related neurodegenerative conditions is still in its infancy. In this review, we introduce the reader to the emerging use of in vivo optical imaging in the study of neurodegenerative diseases. We highlight the current instrumentation that is available and reporter molecules (fluorescent and bioluminescent) that are commonly used. Moreover, we examine how in vivo optical imaging using transgenic reporter mice has provided new insights into Alzheimer’s disease, amyotrophic lateral sclerosis (ALS), Prion disease, and neuronal damage arising from excitotoxicity and inflammation. Furthermore, we also touch upon studies that have utilized these technologies for the development of therapeutic strategies for neurodegenerative conditions that afflict humans.
    Full-text · Article · Jul 2014
    • "Specifically, studies have investigated the interaction of fat content in the diet and genetic susceptibility to colon cancer in the Apc Min/+ mouse model. Because mutations in Adenomatous Polyposis Coli (APC) was observed in over 80% of sporadic human colon cancer cases444546, Apc Min/+ mice carrying a dominant mutation in the Apc gene commonly serve as the mouse model of choice for the human Familial Adenomatous Polyposis (FAP) syndrome . Apc Min/+ mice spontaneously develop multiple intestinal neoplasia (Min) and numerous intestinal polyps, which increase in number and accelerate in development in response to a high fat diet [47]. "
    [Show abstract] [Hide abstract] ABSTRACT: BackgroundTo determine how diets high in saturated fat could increase polyp formation in the mouse model of intestinal neoplasia, Apc Min/+ , we conducted large-scale metabolome analysis and association study of colon and small intestine polyp formation from plasma and liver samples of Apc Min/+ vs. wild-type littermates, kept on low vs. high-fat diet. Label-free mass spectrometry was used to quantify untargeted plasma and acyl-CoA liver compounds, respectively. Differences in contrasts of interest were analyzed statistically by unsupervised and supervised modeling approaches, namely Principal Component Analysis and Linear Model of analysis of variance. Correlation between plasma metabolite concentrations and polyp numbers was analyzed with a zero-inflated Generalized Linear Model.ResultsPlasma metabolome in parallel to promotion of tumor development comprises a clearly distinct profile in Apc Min/+ mice vs. wild type littermates, which is further altered by high-fat diet. Further, functional metabolomics pathway and network analyses in Apc Min/+ mice on high-fat diet revealed associations between polyp formation and plasma metabolic compounds including those involved in amino-acids metabolism as well as nicotinamide and hippuric acid metabolic pathways. Finally, we also show changes in liver acyl-CoA profiles, which may result from a combination of Apc Min/+ -mediated tumor progression and high fat diet. The biological significance of these findings is discussed in the context of intestinal cancer progression.ConclusionsThese studies show that high-throughput metabolomics combined with appropriate statistical modeling and large scale functional approaches can be used to monitor and infer changes and interactions in the metabolome and genome of the host under controlled experimental conditions. Further these studies demonstrate the impact of diet on metabolic pathways and its relation to intestinal cancer progression. Based on our results, metabolic signatures and metabolic pathways of polyposis and intestinal carcinoma have been identified, which may serve as useful targets for the development of therapeutic interventions.
    Full-text · Article · Jun 2014
Show more