Gastrointest Endoscopy Clin N Am

15 (2005) 703 – 714

New Imaging Techniques at Colonoscopy:

Tissue Spectroscopy and Narrow Band Imaging
Evelien Dekker, MD, PhDT, Paul Fockens, MD, PhD
Department of Gastroenterology & Hepatology, Academic Medical Center, University of Amsterdam,
PO Box 22700, 1100 DE Amsterdam, The Netherlands

Colorectal cancer is one of the most common cancers in the Western world.
Most malignant colonic lesions arise in preexisting adenomatous polyps. It would
be ideal to detect all precursor lesions so that they could be removed during the
same procedure, resulting in prevention of colorectal carcinoma. Indeed, there is
a significant reduction in the incidence of colon cancer in screened populations
by white-light colonoscopy plus polypectomy [1]. Not all cancers are prevented,
however. There are three possible explanations: (1) missed polyps during
colonoscopy; (2) not all cancers follow the adenoma-carcinoma sequence; and
(3) flat lesions easily can be overlooked during colonoscopy. In back-to-back
colonoscopy studies, it is determined that 15% to 24% of small polyps (b 1 cm)
are missed by experienced endoscopists and 0% to 6% of the larger polyps [2,3].
Furthermore, not all adenomatous lesions are polypoid. Flat and depressed le-
sions may account for 12% to 40% of all adenomas and early colorectal carci-
nomas [4,5]. These lesions have a surprisingly high rate of submucosal invasion
even when they are small and, needless to say, are more difficult to detect
by colonoscopy.
Therefore, although colonoscopy has undergone an impressive development
in the past 30 years, there remains a need for better endoscopic visualization of
precursor lesions of colorectal cancer. The ideal colonoscopy should be able to
detect all adenomas, including flat and depressed ones. Furthermore, it should
be able to differentiate immediately between neoplastic and nonneoplastic le-
sions, because neoplastic lesions should be removed during the same colonos-
copy session and other lesions can be left in situ, reducing time, costs, and risks

T Corresponding author.
E-mail address: [email protected] (E. Dekker).

1052-5157/05/$ – see front matter D 2005 Elsevier Inc. All rights reserved.
doi:10.1016/j.giec.2005.08.006 giendo.theclinics.com
704 dekker & fockens

of the procedure. Also, ideally, it should be able to detect and identify dyplasia
within fields of transformed mucosa, such as longstanding ulcerative colitis or
Crohn’s disease.
Chromoendoscopy in combination with magnification endoscopy is a tech-
nique of proved success. Rembacken and colleagues and others report this tech-
nique as helpful for the detection and detailed morphologic assessment of
colorectal lesions [5]. Kudo and coworkers [6] performed pioneering studies,
combining high-magnification colonoscopy with chromoendoscopy for in vivo
prediction of histology using a pit-pattern classification. Pancolonic chromo-
endoscopy shows a significant increase in diminutive adenoma detection [7,8]
and detection rates of dysplasia in patients who have ulcerative colitis [9].
Many more, novel techniques currently are under investigation and offer prom-
ising means of improving endoscopic detection of precursor lesions of colorec-
tal cancer. This article focuses on different forms of tissue spectroscopy and
narrow band imaging (NBI) as applied in the colon; chromoendoscopy and
confocal leaser microscopy are discussed elsewhere in this issue.

Tissue spectroscopy

Tissue spectroscopy is based on the evaluation of light-tissue interactions,

not limiting itself to reflection as used for standard white-light endoscopy.
This technology detects relative changes in the way light interacts with tissue
along the disease transformation pathway (ie, from dysplasia to invasive cancer).
This article makes an arbitrary split between those spectroscopic methods that
use point measurement and still images and those that offer real-time endo-
scopic imaging.

Point measurement and still image techniques

Point spectroscopy involves a thin optical probe that touches the mucosa
gently, offering detailed information of the inspected tissue. It is not an appro-
priate method, however, for screening for mucosal abnormalities in the large
colon because of its small sampling volume. It typically acts as a virtual biop-
sying method. Whenever a lesion is seen or suspected with white-light endos-
copy, point spectroscopy is performed to get more information of the nature of
such a lesion.

Raman spectroscopy
Raman spectroscopy is a form of image enhancement based on the principle
that incident light (with wavelengths in the near-infrared region of the spectrum)
can cause molecules within a tissue to vibrate and rotate [10]. The charged
molecules can resonate emitting energy that can be measured spectrally and the
resulting resonance spectrum represents the tissue’s content of specific com-
ponents making up a molecular profile or ‘‘fingerprint.’’
 new imaging techniques at colonoscopy 705

Recently, Shim and Wilson designed and built a fiberoptic device for in vivo
Raman spectroscopy measurements [11]. The first clinical study appeared in
2000 and demonstrated feasibility of Raman spectroscopy during endoscopy
[10]. Molckovsky and coworkers [12] report that in vivo, ten adenomas were
distinguished from nine hyperplastic polyps, with 89% specificity. In this study,
Raman spectra were subjected to a variety of spectral analysis algorithms to
determine the optimal diagnostic values. The exact mechanisms involved in
which Raman scattering differentiates normal from abnormal gastrointestinal
tissues are not known. More clinical studies are in progress. Although the speci-
ficity of this system makes it promising, inherent difficulties of the technique
itself and considerable technical problems should be solved before it becomes
applicable in clinical practice.

Light scattering spectroscopy

Light scattering spectroscopy technique is based on white-light reflectance,
whereby photons incident on tissue are back scattered without a change in their
wavelength, providing structural information about the tissue. Analysis of the
intensity and wavelength of light reflected from the surface gives an estimate of
the size and degree of crowding of surface epithelial nuclei [13]. Initial studies
published in 1995 by Mourant and colleagues suggest that colorectal cancer
can be diagnosed with a sensitivity of 100% and a specificity of 98% [14]. Since
then, only one clinical study was reported as an abstract at the Digestive Disease
Week 2004. Dhar and colleagues performed a study in 45 patients undergoing
colonoscopy [15]. Data were obtained from 138 sites and matched biopsies were
taken for histology. The sensitivity and specificity varied between 75% and 84%
in differentiating the different types of polyps, cancer, colitis, and normal mucosa.
More clinical studies should be performed before judging the clinical use of
this method in the colon.

Fluorescence spectroscopy
Depending on the wavelength of the light used in tissue spectroscopy and
depending on the tissue characteristics, different spectra of light can be identi-
fied. One of the possible light-tissue interactions is fluorescence. Characteristic
of this phenomenon is that when light hits fluorophores in tissue, the wavelength
of the emitted light is longer than that of the light entering the tissue. Normal
and dysplastic tissues may have different fluorescence spectra because malignant
transformation might be associated with alteration of the types, concentrations,
and microdistribution of the constituent fluorophores and chromophores, leading
to differences in the emission and absorption of the fluorescent light, respectively,
resulting in altered surface-measured fluorescence spectra.
Tissue fluorescence can be detected in a small region via an optical probe
(point spectroscopy) in near real time, requiring close contact between the
probe and the polyp. The disadvantage of this technique is that endoscopic sur-
veillance of a large surface area of mucosa with no visible clues as to the presence
706 dekker & fockens

of dysplasia is problematic. As in surveillance through random histologic biop-

sies, many point sites have to be sampled; this takes a lot of time and still harbors
the risk of missing foci of dysplasia.
The initial studies of light-induced fluorescence in the gastrointestinal tract
used laser-induced point fluorescence spectroscopy in an effort to differentiate
colon adenoma and cancer from normal colon and hyperplastic polyps. After
assessment of the optimal excitation wavelength and spectral differences, the
accuracy of the system approximated that of histopathologic assessment, with
sensitivity ranging from 80% to 100% and specificity from 77% to 95% [16–18].
Although there is a high degree of accuracy in differentiating nondysplastic
from dysplastic lesions, this technique does not seem promising, because it car-
ries the drawbacks of a point measurement (ie, small volume sampling) and also
lacks, unlike imaging, contextual information often needed to distinguish abnor-
mal from normal tissue.

Optical coherence tomography

Optical coherence tomography (OCT) involves the use of a special probe
through the channel of an endoscope, which is brought into contact with the
tissue and takes (again) so-called ‘‘optical biopsies’’ (two-dimensional cross-
sectional images). Like endoscopic ultrasonography, OCT provides true ana-
tomic images corresponding to the layers of the gastrointestinal tract, but it uses
light instead of ultrasound waves. The resolution approaches that of conventional
histopathology (resulting in 10-fold greater resolution but less depth compared
with endoscopic ultrasonography) [13].
In the colon, OCT can visualize the mucosa, muscularis mucosae, and (parts
of the) submucosa; visualization of single colonic crypts is beyond its current
resolution. Only a few studies using this method in the colon are published
[19–21]. In the first studies, feasibility was proved [19,20]. In the only pro-
spective clinical study performed in the colon, in a total of 44 polyps, real time
endoscopic OCT differentiated adenomatous from hyperplastic polyps and nor-
mal tissue in degree of tissue organization and light scattering [21].
Although in its infancy as a clinical tool, OCT provides high-resolution
images over the same depth as conventional mucosal biopsy and may prove a
useful and minimally invasive technique for evaluating early neoplastic changes.
These polyps, however, should be detected first by other techniques.

Real-time imaging techniques

In contrast to point measurements and still images, real-time imaging enables

immediate in vivo imaging during colonoscopy. The main advantage of this
technique is that an entire field of view can be investigated with this method. This
opens the possibility to scan large areas of, for instance, colonic mucosa. Es-
pecially for conditions, such as dysplasia in ulcerative colitis, it seems preferable
to have such a real-time imaging technique. Depending on the resolution, the
time it takes to image an area of colonic mucosa differs. One of the highest
 new imaging techniques at colonoscopy 707

resolutions can be obtained with OCT, which at the same time makes it un-
practical or even impossible to be used to scan a large area of colonic mucosa in a
reasonable amount of time.

Fluorescence imaging
Computer generated endoscopic image. A real-time, false-color system for
light-induced fluorescence imaging was developed first by Xillix Technologies
(Richmond, Canada), using only selected emission wavelength bands using
special optical filters and detected by two high-sensitivity cameras to form the
final displayed fluorescence image. Haringsma and Tytgat and Duval and col-
leagues investigated the feasibility of this system in the colon [22,23]. Haringsma
and Tytgat examined 109 colonic lesions in an uncontrolled study and found that
the system differentiated dysplastic from nondysplastic lesions with a sensitivity
of 96% but a disappointing specificity of 70% [22]. Duvall and colleagues were
able to identify all lesions evident by standard endoscopy but also one additional
high-grade dysplastic lesion in the colon not seen with standard endoscopy [23].
Despite these positive results, several problems were encountered that make this
system less applicable to the clinical situation. For example, there seems to be a
high rate of false-positive findings, especially in inflamed tissue, making its use
in ulcerative colitis patients less valuable. Also, the third generation of this
system is based on a fiberoptic colonoscope, whereas videoendoscopy has taken
over most of the market because of its ease of use and higher resolution.

Spectral videocolonoscope. Endoscopic images also can be generated by pro-

cessing fluorescence information (computer composed image) or directly ob-
tained through incorporation of spectroscopic technology into the imaging system
of a flexible endoscope (‘‘spectral endoscope’’). The aim of these systems is to
convert the fluorescent light into a real-time image of the whole endoscopic view
leading to the possibility of screening large areas of the colon in a way
comparable to standard endoscopic examination. Wang and coworkers developed
a system consisting of a videoendoscope, a source of UV light, and an optical
fiber to deliver the excitation light [24,25]. The charge-coupled device in the
endoscope is insensitive to UV light, and if the white-light illumination of
the videoendoscope is removed, the endoscopic images obtained essentially
are fluorescence images. The technique seemed feasible, but several problems
have become clear, especially the motion and the geometry of the colon, which
make a fixed relationship between the light source and the colon impossible,
obscuring interpretation.
Olympus Medical Systems (Tokyo, Japan) developed a new prototype
videoautofluorescence endoscopy system that may overcome the most important
limitations of earlier prototypes (Fig. 1). Recently, the first clinical study using
this system in 60 patients who had Barrett’s esophagus was published [26]. In
this study, the Barrett’s esophagus was screened first with white-light endoscopy
and then examined with autofluorescence to detect additional areas with high-
grade dysplasia or early carcinoma. Videoautofluorescence imaging increased the
708 dekker & fockens

Fig. 1. Identification of a colonic adenoma with a new prototype autofluorescence videoendoscopic

imaging system. (A) Conventional image. (B) Autofluorescence image.

detection rate of high-grade dysplasia and early cancer from 23% to 33%.
Because of the high false positivity rate of autofluorescence (positive predictive
value of 49%), it seems attractive to combine this technique with a point detec-
tion method, such as NBI. Olympus recently developed such a system combining
videoautofluorescence with NBI and a prototype is available. Studies using this
combination of videoautofluorescence and NBI in the esophagus and the colon
are being performed by the authors’ group.

Application of photosensitizers. Exogenously applied sensitizers should be

chosen for their preference to accumulate selectively in malignant and prema-
lignant lesions and to induce fluorescence when illuminated with light of the
appropriate wavelength. 5-Aminolevulinic acid (5-ALA) is of interest because
it produces minimal side effects (cutaneous sensitivity) and better tumor selec-
tivity compared with other sensitizers [27]. 5-ALA is not a sensitizer but is con-
verted intracellularly into the photoactive compound, protoporphyrin IX, which
accumulates preferentially in malignant tissue. Indeed, sensitivity of fluorescence
colonoscopy after local sensitization with 5-ALA for detection of dysplasia in
ulcerative colitis was high (87%–100%) [28]. Specificity was only 62% after
local application, however. The hydrophilic properties of 5-ALA limit its uptake
and penetration into the tissue, particularly after topical administration and the
long instillation times (1–2 hours after topical and 4–6 hours after systemic
administration) and high concentrations needed are drawbacks for its clinical use.
Hexaminolevulinate can induce higher protoporphyrin IX fluorescence
intensities with lower instillation times and lower concentrations. In a pilot-
study, hexaminolevulinate-based fluorescence endoscopy in patients who had
rectal cancer and adenoma was well tolerated and produced selective fluo-
rescence of all rectal adenomas with intraepithelial neoplasia [29]. For rectal
cancer, however, there was only weak fluorescence or none at all. Thus, photo-
sensitizing agents currently in use have a relatively poor capacity to accumulate
preferentially in malignant tissue.
 new imaging techniques at colonoscopy 709

Immunoscopy
Another attractive option may be the coupling of fluorescent dyes to tumor-
related antigens, as they are in everyday use in most pathology labs. Im-
munoscopy is performed with the use of injected probes consisting of small
molecules (enzyme substrates, receptor ligands, monoclonal antibodies, and
peptides) tagged to a fluorescent dye to achieve high-affinity binding to specific
biochemical and molecular markers of disease [30]. If such a fluorophore could
be coupled to a tumor specific monoclonal antibody, this could help discriminate
normal tissue from tumor. These fluorophores typically emit light in the near-
infrared zone and usually produce powerful fluorescence that is detectable even
through millimeters of tissue. Labeling of an anti–carcinoembryonic antigen
monoclonal antibody to fluorescein isothiocyanate was reported by a Japanese
group in 1989 [31]. They studied 42 gastric lesions, including 20 early tumors in
vitro and applied the conjugated antibody topically. Sixty minutes after ap-
plication of the conjugated antibody, 90% of the tumors showed up positive with
no false positives. A similar technique was used by a German group in 27 pa-
tients who had colonic polypoid lesions [32]. During conventional colonoscopy,
a fluorescein-labeled monoclonal antibody against carcinoembryonic antigen
was applied directly onto the mucosal surface. Fluorescence in vivo was present
in 19 out of 25 carcinomas and in 3 of 8 adenomas. The technique failed in
the presence of mucosal ulceration or bleeding.
Recently, Kelly and colleagues developed an imaging agent for real-time en-
doscopic tumor detection in a murine model using a previously identified phage
library-derived, colon cancer–specific cyclic peptide and fluorescent moieties
[33]. It was shown that a modified near-infrared fluorescent agent containing a
cyclicpeptide motif could be used to detect invasive colorectal carcinoma by
near-infrared fluorescence endoscopy. Now that the mechanism is exploited in
a murine model, the development of optical tumor-specific imaging agents could
have far-reaching potential in real-time polyp and cancer detection in humans.
Possibilities seemingly are endless and, given the limitations in current fluo-
rescence imaging in the sensitive and specific detection of very early lesions in
the colon (also in the presence of confounding circumstances like chronic inflam-
mation), these developments will complement existing fluorescence endoscopy.
The field of immunoscopy is close to molecular imaging, and even may be con-
sidered a form of molecular imaging. In the future, other techniques, such as MRI,
may be more appropriate than colonoscopy and one could imagine a MR colo-
graphy, which not only depicts possible polyps but also differentiates their nature
through molecular imaging techniques. This would obviate optical cololonoscopy
as a diagnostic method. Unfortunately, the road to these techniques seems long.

Narrow band imaging

The application of filters to the standard white-light endoscopy could be a

simple possibility to improve image quality in endoscopy. NBI (Olympus Medi-
710 dekker & fockens

cal Systems) is a new technique that uses optical filters for imaging the mucosal
morphology without the use of dye spaying (chromoendoscopy).
It currently uses a sequential (red-green-blue [RGB]) electronic endoscope
system with a standard xenon light source but special filters. In the NBI system,
the band-pass ranges of the different components of white light are narrowed and
there is a higher relative intensity of blue light. Because blue light does not
penetrate into deeper layers of the mucosa, NBI is expected to give improved
details of the surface structures and microvasculature [34]. It usually is com-
bined with high-resolution and magnification endoscopy. A new system based on
color chip technology has been released by Olympus in the fall of 2005.
All endoscopic experience is obtained by using a prototype NBI system,
consisting of an Olympus CV240 processor and a CLV-U40 light source. The
endoscopy system uses sequential RGB illumination, in which white light from a
xenon lamp passes first through RGB filters. The filters separate the white light
into the colors red, green, and blue that illuminate the mucosa sequentially. The
RGB reflected light is detected separately by a monochromatic charge-coupled
device placed at the tip of the endoscope, and the three images are integrated
into a single color image by the videoprocessor. In addition to the conventional
RGB filters for white-light endoscopy, the light source is equipped with special
RGB filters for NBI. The RGB band-pass ranges of the NBI filters are 400 to
430 nm, 530 to 550 nm, and 600 to 620 nm, respectively.
Most clinical studies using this technique are performed in the upper gas-
trointestinal tract and only three studies using NBI in the colon are published:
one as a full article and two as abstracts. Machida and coworkers recently
published a study using NBI in 43 lesions in 34 patients [35]. The NBI system
was useful for detecting lesions and observation of surface structure and capillary
vessels and the diagnostic accuracy rate of the endoscopic diagnosis compared
with histology was equivalent to chromoendoscopy. The authors’ group also
performed a clinical feasibility study of 33 colonic polyps (Fig. 2) [36]. Using
visual analog scales, NBI zoom mucosal images were found superior to stan-
dard high-resolution images and comparable to high-resolution chromoscopy.

Fig. 2. Use of NBI in colonic polyps: tubular adenoma. (A) Conventional image. (B) NBI image.
 new imaging techniques at colonoscopy 711

Fig. 3. Use of NBI for dysplasia screening in patients who have longstanding ulcerative colitis.
(A) Conventional image of a DALM. (B) NBI image of a DALM.

The authors also performed a study using the NBI system for dysplasia screening
in patients who had longstanding ulcerative colitis (Fig. 3) [37]. Preliminary
results of this study comparing NBI with only targeted biopsies to a standard
surveillance colonoscopy with random biopsies.
Although this method is under development, it seems promising. It has several
advantages compared with chromoendoscopy: one, the elimination of the need to
spray contrast-enhancing agents; a second, the absence of irregular or unequal
distribution of dye; and, finally, allowing the immediate back-and-forth switching
between the standard endoscopic image and the NBI image. Like chromoendos-
copy, it can be used as a screening method for adenomatous lesions and as a
point measurement evaluating the pit pattern of the detected lesion. Its use in the
colon awaits proof in randomized, controlled clinical trials, especially when
comparing this technique with standard endoscopy and chromoendoscopy.

Summary and future directions

The ideal colonoscopic technique should have a high diagnostic accuracy and
the ability to screen large mucosal areas. Most of the new techniques described in
this article need to be developed further and should be tested in randomized,
controlled clinical trials. Furthermore, they should be compared with chromo-
endoscopy, a technique with proved success in endoscopic practice [7,8]. Such
techniques are not yet available for routine use, but when the most promis-
ing ones have proved their clinical use, they may take their place in standard
patient care.
It is difficult to choose the most promising method in this range of interesting
and mostly promising techniques. From a clinical perspective, a technique that
is easy to use, relatively cheap, and quick is desired. Care should be taken,
however, not to focus on these issues early in the research, risking frustrating
researchers. More appropriately, the most sensitive and specific technique or
combination of techniques, which at first may be one of the most expensive
712 dekker & fockens

ones, should be determined. Once its accuracy is demonstrated, efforts should

be moved to developing a way in which it can be clinically useful and cost
effective. This process has a foreseeable need of a form of artificial intelligence
or decision support software. Maybe with the best technique, each and every cell
of the colon can be studied in vivo. The next part of the research then will have
to be computer assistance and making this technique ready for clinical practice in
a reasonable timeframe.
Of all the techniques available, NBI seems attractive, as it seems a sim-
plification of chromoendoscopy and is close to commercial release. Fluores-
cence techniques, including immunoscopy, may help endoscopists of the future
by waving a red flag whenever a clinically important lesion has been scanned
through the field of view. A combination of such a red-flag technique and a
microscopic technique carries enormous potential, especially in the screening
setting. Not one technique but different techniques for different indications may
be the end result.
Thus, in the authors’ opinion, the most successful clinical method probably
will be a combination of techniques that can provide a wide-area surveillance
needed in dysplasia screening in ulcerative colitis and adenoma detection during
screening colonoscopy (such as fluorescence endoscopy or immunoscopy), and,
once a target lesion is identified, point detection methods (such as NBI or
confocal microendoscopy) that can be used to identify the lesion further.

References

[1] Winawer SJ, Zauber AG, Ho MN, et al. Prevention of colorectal cancer by colonoscopic
polypectomy. The National Polyp Study. N Engl J Med 1993;329:1977 – 81.
[2] Hixson LJ, Fennerty MB, Sampliner RE, et al. Prospective blinded trial of the colonoscopic
miss-rate of large colorectal polyps. Gastrointest Endosc 1991;37:125 – 7.
[3] Rex DK, Cutler CS, Lemmel GT, et al. Colonoscopic miss rates of adenomas determined by
back/to/back colonoscopies. Gastroenterology 1997;112:24 – 8.
[4] Kudo S, Tamura S, Hirota Y, et al. The problem of de novo colorectal carcinoma. Eur J Cancer
1995;31:1118 – 20.
[5] Rembacken BJ, Fujii T, Cairns A, et al. Flat and depressed colonic neoplasms; a prospec-
tive study of 1000 colonoscopies in the UK. Lancet 2000;355:1211 – 4.
[6] Kudo S, Rubio CA, Teixeira CR, et al. Pit pattern in colorectal neoplasia: endoscopic magnify-
ing view. Endoscopy 2001;33:367 – 73.
[7] Brooker JC, Saunders BP, Shah SG, et al. Total colonic dye-spray increases the detection of
diminutive adenomas during routine colonoscopy: a randomized controlled trail. Gastrointest
Endosc 2002;56:333 – 8.
[8] Hurlstone DP, Cross SS, Slater R, et al. Detecting diminutive colorectal lesions at colonoscopy:
a randomized trial of pan-colonic versus targeted chromoscopy. Gut 2004;53:376 – 80.
[9] Kiesslich R, Fritsch J, Holtmann M, et al. Methylene blue-aided chromoendoscopy for the
detection of intraepithelial neoplasia and colon cancer in ulcerative colitis. Gastroenterology
2003;124:880 – 8.
[10] Shim MG, Wong Kee Song LM, Marcon NE, et al. In vivo near-infrared Raman spectroscopy:
demonstration of feasibility during clinical gastrointestinal endoscopy. Photochem Photobiol
2000;72:146 – 50.
 new imaging techniques at colonoscopy 713

[11] Shim MG, Wilson BC. Development of an in vivo Raman spectroscopic system for diagnostic
applications. J Raman Spetrosc 1997;28:131 – 42.
[12] Molckovsky A, Wong Kee Song LM, Shim MG, et al. Diagnostic potential of near-infrared
Raman spectroscopy in the colon: differentiating adenomatous from hyperplastic polyps.
Gastrointest Endosc 2003;57:396 – 402.
[13] Van Dam J. Novel methods of enhanced endoscopic imaging. Gut 2003;52(Suppl IV):iv12 – 6.
[14] Mourant JR, Boyer JD, Johnson TM. Detection of gastrointestinal cancer by elastic scattering
and absorption spectroscopy with the Los Alamos Optical Biopsy system. SPIE 1995;2387:
210 – 7.
[15] Dhar A, Johnson K, Novelli M, et al. Optical biopsy of colonic lesions: comparison of elastic
scattering spectroscopy with histology. Gastroenterology 2004;126(Suppl 2):A76.
[16] Cothren RM, Richards-Kortum R, Sivak Jr MV, et al. Gastrointestinal tissue diagnosis by laser-
induced fluorescence spectroscopy at endoscopy. Gastrointest Endosc 1990;36:105 – 11.
[17] Schomacker KT, Frisoli JK, Compton CC, et al. Ultraviolet laser-induced fluorescence of colonic
polyps. Gastroenterology 1992;102:1155 – 60.
[18] Cothren RM, Sivak Jr MV, Van Dam J, et al. Detection of dysplasia at colonoscopy using laser-
induced fluorescence: a blinded study. Gastrointest Endosc 1996;44:168 – 76.
[19] Sivak Jr MV, Kobayashi K, Izatt JA, et al. High-resolution imaging of the gastrointestinal tract
using optical coherence tomography. Gastrointest Endosc 2000;51:474 – 9.
[20] J7ckle S, Gladkova N, Feldchtein F, et al. In vivo endoscopic optical coherence tomography
of the human gastrointestinal tract—toward optical biopsy. Endoscopy 2000;32:743 – 9.
[21] Pfau PR, Sivak Jr MV, Chak A, et al. Criteria for the diagnosis of dysplasia by endoscopic topical
coherence tomography. Gastrointest Endosc 2003;58:196 – 202.
[22] Haringsma J, Tytgat GN. The value of fluorescence techniques in gastrointestinal endoscopy.
Better than the endoscopist’s eye? The European experience. Endoscopy 1998;30:416 – 8.
[23] Duvall GA, Saidi R, Kost J, et al. Real-time light induced fluorescence endoscopy in the
gastrointestinal tract. Gastrointest Endosc 1997;45:AB28.
[24] Wang TD, Van Dam J, Crawford JM, et al. Fluorescence endoscopic imaging of human co-
lonic adenomas. Gastroenterology 1996;111:1182 – 91.
[25] Wang TD, Crawford JM, Feld MS, et al. In vivo identification of colonic dysplasia using
fluorescence endoscopic imaging. Gastrointest Endosc 1999;49:447 – 55.
[26] Kara MA, Peters FP, ten Kate FJW, et al. Endoscopic video autofluorescence imaging may
improve the detection of early neoplasia in patients with Barrett’s esophagus. Gastrointest
Endosc 2005;61:679 – 85.
[27] Messmann H. 5-aminolevulinic acid-induced protoporphyrin IX for the detection of gastro-
intestinal dysplasia. Gastrointest Endosc Clin North Am 2000;10:497 – 512.
[28] Messmann H, Endlicher E, Freunek G, et al. Fluorescence endoscopy for the detection of
low and high grade dysplasia in ulcerative colitis using systemic or local 5-aminolevulinic
acid sensitisation. Gut 2003;52:1003 – 7.
[29] Endlicher E, Gelbmann CM, Knuchel R, et al. Hexaminolevulinate-induced fluorescence
endoscopy in patients with rectal adenoma and cancer: a pilot study. Gastrointest Endosc
2004;60:449 – 54.
[30] Massoud TF, Gambhir SS. Molecular imaging in living subjects F seeing fundamental biologi-
cal processes in a new light. Genes Dev 2003;17:545 – 80.
[31] Tatsuta M, Iishi H, Ichii M, et al. Diagnosis of gastric cancers with flourescein-labelled
monoclonal antibodies to carcinoembryonic antigen. Lasers Surg Med 1989;9:422 – 6.
[32] Keller R, Winde G, Terpe HJ, et al. Fluorescence endoscopy using a fluorescein-labeled
monoclonal antibody against carcinoembryonic antigen in patients with colorectal carcinoma
and adenoma. Endoscopy 2002;34:801 – 7.
[33] Kelly K, Alencar H, Funovics M, et al. Detection of invasive colon cancer using a novel,
targeted, library-derived fluorescent peptide. Cancer Res 2004;64:6247 – 51.
[34] Gono K, Obi T, Yamaguchi M, et al. Appearance of enhanced tissue features in narrow-band
endoscopic imaging. J Biomed Opt 2004;9:568 – 77.
714 dekker & fockens

[35] Machida H, Sano Y, Humamoto Y, et al. Narrow band imaging in the diagnosis of colorec-
tal mucosal lesions: a pilot study. Endoscopy 2004;36:1094 – 8.
[36] Dekker E, Ennahachi M, Kara M, et al. Narrow band imaging (NBI) for improved pit pat-
tern imaging in colonic polyps. Gastroenterology 2004;126(Suppl 2):A625.
[37] Dekker E, van Deventer SJH, Hardwick J, et al. The value of narrow band imaging for the
detection of dysplasia in longstanding ulcerative colitis. Gastroenterology 2004;126(Suppl 2):A77.