Evaluation of superficial bladder transitional-cell carcinoma by optical coherence tomography.

Abstract

Optical coherence tomography (OCT) is a new modality that allows noninvasive examination of the internal structure of biological tissue in vivo with a spatial resolution of 10 to 15 microm. This study evaluated the clinical application of OCT to determine epithelial and subepithelial anatomic structure and invasiveness of bladder epithelial lesions.
The OCT examination was performed with a 980-nm 10 mW superluminescent diode using a 2.7-mm-diameter optical fiber positioned cystoscopically. A total of 261 scans of 1.5 seconds' duration, which generated 200 x 200-pixel images, were performed on 87 areas in 24 patients at high risk of having transitional-cell carcinoma (TCC). Lesions, visually suspect, and normal areas were photographed, scanned, and biopsied. The scans were evaluated independently before comparison with histopathology findings.
Of the 87 areas, 29 of 36 visually suspect areas and 35 of 35 normal areas, were correctly diagnosed with OCT. Of the 16 areas with papillary TCC, all 16 were diagnosed correctly as tumor, and 9 of 10 were diagnosed correctly as invasive, including 6 with lamina propria invasion only. Papillary and flat tumors, carcinoma in situ, inflammation, chronic cystitis, and von Brunn's nests were scanned. Overall, OCT had a sensitivity of 100%, overall specificity of 89%, positive predictive value of 75%, and negative predictive value of 100%. The accuracy was 92%. The positive predictive value for invasion was 90%.
Optical coherence tomography is a simple, portable, promising modality for evaluation of bladder lesions and depth of tumor penetration. Further refinement of this technology may lead to the development of an optical surrogate for biopsy.

Full text

Page 1

JOURNAL OF ENDOUROLOGY
Volume 19, Number 5, June 2005
© Mary Ann Liebert, Inc.
Evaluation of Superficial Bladder Transitional-Cell Carcinoma
by Optical Coherence Tomography
MICHAEL J. MANYAK, M.D.,1NATALIA D. GLADKOVA, M.D.,4,5JOHN H. MAKARI, M.D.,1
ARNOLD M. SCHWARTZ, M.D., Ph.D,2ELENA V. ZAGAYNOVA, M.D.,4,5LADAN ZOLFAGHARI, M.D.,1
JASON M. ZARA, Ph.D.,3RASHID IKSANOV, M.S.E.E.,4and FELIX I. FELDCHTEIN, Ph.D.4,6
ABSTRACT
Background and Purpose: Optical coherence tomography (OCT) is a new modality that allows noninvasive
examination of the internal structure of biological tissue in vivo with a spatial resolution of 10 to 15 ?m. This
study evaluated the clinical application of OCT to determine epithelial and subepithelial anatomic structure
and invasiveness of bladder epithelial lesions.
Materials and Methods: The OCT examination was performed with a 980-nm 10 mW superluminescent
diode using a 2.7-mm-diameter optical fiber positioned cystoscopically. A total of 261 scans of 1.5 seconds’
duration, which generated 200 ? 200-pixel images, were performed on 87 areas in 24 patients at high risk of
having transitional-cell carcinoma (TCC). Lesions, visually suspect, and normal areas were photographed,
scanned, and biopsied. The scans were evaluated independently before comparison with histopathology find-
ings.
Results: Of the 87 areas, 29 of 36 visually suspect areas and 35 of 35 normal areas, were correctly diag-
nosed with OCT. Of the 16 areas with papillary TCC, all 16 were diagnosed correctly as tumor, and 9 of 10
were diagnosed correctly as invasive, including 6 with lamina propria invasion only. Papillary and flat tu-
mors, carcinoma in situ, inflammation, chronic cystitis, and von Brunn’s nests were scanned. Overall, OCT
had a sensitivity of 100%, overall specificity of 89%, positive predictive value of 75%, and negative predic-
tive value of 100%. The accuracy was 92%. The positive predictive value for invasion was 90%.
Conclusion: Optical coherence tomography is a simple, portable, promising modality for evaluation of blad-
der lesions and depth of tumor penetration. Further refinement of this technology may lead to the develop-
ment of an optical surrogate for biopsy.
INTRODUCTION
C
early stage but has a high mortality rate in advanced stages.
Most cancers are low grade and superficial when diagnosed,
but the first presentation can be with multiple high-grade and
diffuse lesions that are highly aggressive. Imaging has a defi-
nite role in staging bladder cancer, but conventional ultra-
sonography, CT, and MRI have significant limitations because
their spatial resolution is not sufficient to provide the equiva-
ANCER OF THE URINARY BLADDER is an epithelial
malignancy that is usually well controlled if found at an
lent of histopathologic detail to determine the aggressive su-
perficial lesions or to detect invasion unless it is deep or wide-
spread.
The invention of lasers 40 years ago and the convergence of
technology in the fields of material science, the defense and
communications industries, and biology have made possible the
use of light for diagnostic imaging. Precision in control of the
spatial and temporal properties of light, along with under-
standing of light and its heterogeneous scattering properties
within tissue, have extended our capability for imaging. Diag-
nosis using light differs in one fundamental respect from its
Departments of 1Urology, 2Pathology, and 3Electrical and Computer Engineering, The George Washington University, Washington, DC.
4Institute of Applied Physics of Russian Academy of Sciences and 5Nizhny Novgorod Academy of Medicine, Nizhny Novgorod, Russian Fed-
eration.
6Imalux Corporation, Cleveland, Ohio.
570

Page 2

therapeutic application for destruction. Therapeutic uses of light
are determined by the effect of light on tissue, while diagnos-
tic uses generally detect changes resulting from the effect of tis-
sue on light.1The rapid, noninvasive determination of tissue
characteristics such as scattering and absorption is essential to
process information received from the light interaction in tis-
sue during a diagnostic procedure. A prime example of light–tis-
sue interaction already in daily use is diffuse reflectance spec-
troscopy, the mechanism by which pulse oximetry determines
tissue oxygen levels during surgery.
Optical coherence tomography (OCT) is another intriguing
diagnostic application of light. This technique was originally
used to image human tissue in 1991 and employs light (instead
of sound waves) to obtain images in a manner analogous to B-
mode ultrasonography. Reflected light, generated in the near-
infrared spectrum by a superluminescent diode, is registered by
interferometry to produce two-dimensional images. The method
images tissue in situ and in real time, providing resolution on
the order of 5 to 20 ?m, which is comparable to that traditional
microscopic analysis. Cellular features detected by ultra-high
spatial-resolution OCT include mitotic activity, nuclear-to-cy-
toplasmic ratios, and migration.2Optical coherence tomogra-
phy does not require a conducting medium and can therefore
image through air or water with far greater resolution than ul-
trasound.2–4In addition, OCT is relatively inexpensive and
portable and can be used with existing endoscopic instrumen-
tation.5Improvements in the ability to obtain OCT images
prompted an evaluation of this modality for its clinical applic-
ability for superficial bladder carcinoma.
MATERIALS AND METHODS
After approval by the institutional review board and obtain-
ing of informed consent from each patient, 24 patients at high
risk of having transitional-cell carcinoma (TCC) of the urinary
bladder underwent cystoscopic examination with the OCT pro-
tocol. Scanning was performed with a 980-nm 10-mW super-
luminescent diode using a 2.7-mm (OD) optical fiber positioned
through a cystoscope sheath. Patients underwent a standard cys-
toscopic examination. Visually suspect lesions, as well as nor-
mal-appearing urothelial tissue, were photographed, scanned
with OCT, and biopsied. A total of 261 scans of 1.5 seconds’
duration, which generated 200 ? 200-pixel images, were per-
EVALUATION OF SUPERFICIAL BLADDER TRANSITIONAL-CELL CARCINOMA
571
FIG. 1.
ithelial layer; LP ? lamina propria; M ? muscularis.
Benign OCT scan: normal bladder mucosa. EL ? ep-
FIG. 2.
Abnormal but not invasive OCT scan type: CIS.
FIG. 3.
sive into muscularis.
Abnormal and invasive OCT scan type: TCC inva-
FIG. 4.
row)
Benign OCT scan: chronic cystitis with edema (ar-

Page 3

formed on 87 areas. Scans were performed at 1-mm intervals
on the lesions and at their junctions with the bladder epithe-
lium. All scans were obtained by placing the end-firing OCT
probe on the desired area perpendicular to the wall of the blad-
der. Each patient had at least one apparently normal area that
was photographed, scanned with OCT, and biopsied. Biopsy
specimens were preserved in Formalin for standard histopatho-
logic analysis.
The OCT scans were reviewed at a later date and graded as
either benign, abnormal but not invasive, or abnormal and in-
vasive. Benign scans are characterized by sharp, stratified, high-
contrast images consisting of three visible layers. The upper-
most epithelial layer is sharply demarcated and well organized,
with relatively uniform thickness. Beneath the epithelium, the
normal lamina propria is a bright, thin, continuous layer be-
neath which the muscularis appears as a darker layer (Fig. 1).
In some cases, the epithelium was atrophic and could not be re-
solved by OCT, but the lamina propria and muscularis were
normal in appearance.
Scans with stratification but low contrast between epithelium
and poorly defined but present and intact lamina propria and
organized muscularis are considered positive for intraepithelial
pathology (Fig. 2). In these cases, the epithelial layer is irreg-
ular and not sharply demarcated. Scans with intraepithelial
focal areas of disorganization are also considered positive. Im-
ages with disorganized epithelium, disrupted lamina propria,
and homogeneous muscularis with poor differentiation between
layers are considered abnormal and invasive (Fig. 3).
After evaluation, OCT scans were compared with the
histopathologic diagnosis previously determined independently
by another investigator (AMS).
RESULTS
The 87 areas of interest were 16 diverse papillary lesions,
36 areas termed “visually suspicious for malignancy,” and 35
areas relatively normal in appearance. Of the 87 areas, 29 of
36 suspect areas and all 35 of the normal areas were correctly
diagnosed with OCT, giving an accuracy of 92% for this study.
All 21 lesions thought to be carcinoma (16 papillary and 5 flat
lesions) had abnormal scans and were interpreted as malignant,
providing a sensitivity of 100%. In 31 suspect lesions, 24 were
considered benign by scan and confirmed, with the other 7 eval-
uated as scan positive but without evidence of malignancy on
biopsy, thereby yielding a specificity of 77% for suspect le-
sions. The overall OCT specificity (incorporating the 35 true-
normal areas) was 89%. The positive predictive value (PPV)
was 75%, while the negative predictive value (NPV) was 100%
(Table 1).
Other microscopic findings included inflammation and
chronic cystitis (Fig. 4), as well as normal tissue structures such
as von Brunn’s nests and several instances of subepithelial
blood vessels (Fig. 5). In these instances, the tissue structures
were organized, and the tissue layers were not violated. In-
flammatory conditions contributed to the false-positive findings
in the suspect lesions judged to have carcinoma but no inva-
sion (Figure 6).
Ten tumors were diagnosed by OCT as being invasive, with
nine of them being confirmed as such by histopathology ex-
amination. There was one false-positive scan for invasion. In-
terestingly, six of the nine lesions diagnosed by OCT as inva-
sive had invasion of the lamina propria only. Therefore, the
PPV for invasion was 90%.
DISCUSSION
The natural history of superficial urinary bladder cancer is dis-
tinctly different from that of lesions that present with either in-
vasion of the muscle wall or metastases. It is well established
that the prognosis following these latter presentations is signifi-
cantly worse than that of patients with superficial disease. Ap-
propriate management requires proper diagnosis and staging
because superficial tumors generally pose only a local manage-
ment problem, whereas progression to invasion necessitates more
extensive treatment. The need to identify patients with refractory
disease or those with tumors in the earliest stages of invasion is
underscored by the results of cystectomy in patients with inva-
sive bladder carcinoma.6
Papillary lesions remain the most common presentation of
bladder carcinoma, with biopsy including the superficial muscu-
laris required to rule out invasive disease. Normally, the bladder
is lined by transitional epithelium that ranges in thickness from
five to seven cell layers in the collapsed state to two to three cell
layers when the bladder is distended.7The lamina propria is a
thin, loose, fibrovascular sheet of tissue that separates the deeper
muscular layers from the superficial epithelial layer. Invasion
through the lamina propria has prognostic significance, and its
identification is important for proper tumor staging. Papillary le-
sions are usually visibly prominent but carcinoma in situ (CIS)
is a diffuse, high-grade, intraepithelial malignancy with a rather
innocuous appearance but an aggressive natural history. Although
the exact incidence of this entity is not clear, 10% of patients in
a large series of newly diagnosed bladder carcinomas had CIS.8
Frequently, CIS appears as a velvety or erythematous patch of
mucosa either alone or in association with either low-grade su-
perficial or high-grade invasive disease. Appearance alone can
be deceptive, however, because dysplasia or frank CIS has been
found in biopsies of normal-appearing bladder epithelium in 18%
of patients with a papillary tumor.9In any event, the clinician
must rely on visual appearance to guide biopsies and should per-
form systematic biopsies even of normal-appearing epithelium in
cases where CIS is suspected. Multiple lesions, frequent recur-
rences, and CIS are associated with a higher risk of progression
to invasion and mandate periodic surveillance, often with biopsy.
No imaging technology to date reliably assists in the diagnosis
of occult malignancy or invasion.
MANYAK ET AL.
572
TABLE 1. ACCURACY OF OCT FOR 87 SCANNED AREASa
True positive
False positive
Sensitivity: TP/(TP ? FN)
Specificity: TN/(TN ? FP)
PPV: TP/(TP ? FN)
NPV: TN/(TN ? FN)
Accuracy: (TP ? TN)/total
21 True negative
False negative
59
70
100%
89%
75%
100%
92%
aAreas were 16 papillary tumors, 36 visually suspect sites,
and 35 apparently normal regions.

Page 4

Optical coherence tomography was initially used to image
the transparent ocular structures in vitro. Expansion of this tech-
nique included imaging of the retina and macula in vivo, with
resolution of 10 ?m.10,11After initial studies demonstrated the
ability to image the nearly transparent ocular structures with
OCT, efforts focused on imaging tissue with complex scatter-
ing properties.12In vivo imaging of dermatologic structures was
performed successfully, followed by endoluminal imaging of
vascular tissue components.13Compared with high-frequency
intravascular ultrasonography, OCT produced resolution more
than six times that of the highest frequency imaging modality
available.14
Endoscopic application of OCT also has been demonstrated
in the human gastrointestinal tract. Compared with histologic
specimens, OCT has demonstrated promise in discriminating
inflammatory, neoplastic, metaplastic, and cancerous processes
of the esophagus and colon.15,16Additionally, OCT has been
used in other epithelial areas to provide high-resolution imag-
ing of both normal tissue and neoplastic gynecologic lesions.17
Imaging of human genitourinary tissue by OCT was first
demonstrated in 1997. In vitro OCT images from biopsies of
prostate, bladder, and ureter were compared with standard
preparations with hematoxylin and eosin staining. The OCT im-
ages demonstrated differentiation between the prostatic urethra
and prostate, depicted the neurovascular bundle and the
prostate–adipose border and the prostatic capsule, and distin-
guished the anatomic layers in the bladder and ureter with an
axial resolution of 16 ?m.18
The use of OCT for endoscopic evaluation of bladder pathol-
ogy also has been demonstrated.19,20The present study used an
improved imaging system with greater resolution to further
evaluate the utility of OCT for delineation of subepithelial an-
atomic structure and presence and depth of tumor invasion. In
this study, which encompassed 261 scans, the PPV (75%) is
encouraging, but the sensitivity (100%) and specificity (77%
for suspect lesions, 89% overall) suggest that this modality may
be quite useful for assessment of superficial lesions. This claim
is bolstered by the NPV (100%), which suggests that there is a
low likelihood of missing an aggressive lesion, at least as dem-
onstrated in this preliminary study. Very encouraging is the PPV
for invasion (90%), especially as two thirds of these patients
had lesions invading the lamina propria only. These findings
warrant an expansion of this pilot study to determine the true
utility of OCT for superficial bladder carcinoma.
Another intriguing issue raised by this study is the possibil-
ity of elimination of multiple bladder biopsies. The sensitivity
and NPV for OCT in this setting suggested that a significant
reduction in biopsies could be achieved. Because of the rela-
tively high risk of bladder carcinoma, areas of probably normal
tissue and suspect but not clearly cancerous areas were biop-
sied in addition to all papillary lesions and suspect flat lesions.
Because all 35 normal areas and 24 of 31 suspect areas were
correctly read as negative, 56 of the 87 biopsies (64%) could
have been avoided. All papillary lesions and all areas of posi-
tive scans would still be biopsied, because definitive diagnosis
and grade remain unknown with OCT at this resolution. There-
fore, properly targeted biopsies could become a more tangible
realization.
A current limitation of OCT in this setting is the scanning
dimensions, which are limited to a maximum of 1 mm in depth
and 2 mm in width. Such scanning fields limit the ability of
OCT to provide complete staging information on a large scale.
However, variations in the wavelength of the light source to en-
hance tissue penetration and allow greater image depth are un-
der evaluation. The addition of contrast agents such as chro-
mophores also holds great promise for enhancement of OCT
images. Clearly, OCT is an intriguing component of sensor fu-
sion technology in the quest to obtain a true optical biopsy.21
CONCLUSION
Optical coherence tomography is a simple, portable, promis-
ing modality for the evaluation of bladder lesions and hte depth
of tumor penetration. Further refinement of this technology may
lead to a practical office-based surveillance system that could
serve as an optical surrogate for biopsy.
ACKNOWLEDGMENTS
The authors gratefully acknowledge the technical assistance
of the Institute of Applied Physics (Nizhny Novgorod, Russia)
and Imalux Corporation (Cleveland, OH), which supplied the
EVALUATION OF SUPERFICIAL BLADDER TRANSITIONAL-CELL CARCINOMA
573
FIG. 5.
blood vessels.
Benign OCT scan: chronic cystitis with subepithelial
FIG. 6.
reaction.
False-positive OCT scan: cystitis with granulomatous

Page 5

OCT system and probes. The authors also acknowledge the as-
sistance of Dr. Sherry Sedberry in tissue acquisition.
REFERENCES
1. Parrish JA, Wilson BC. Current and future trends in laser medi-
cine. Photochem Photobiol 1991;53:731.
2. Boppart SA, Bouma BE, Pitris C, et al. In vivo cellular optical co-
herence tomography imaging. Nature Med 1998;4: 861.
3. Huang D, Swanson EA, Lin CP, et al. Optical coherence tomog-
raphy. Science 1991;254:1178.
4. Manyak MJ, Warner JW. An update on laser use in urology. Con-
temp Urol 2003;15(10):13.
5. Feldchtein FI, Gelikonov GV, Gelikonov VM, et al. Endoscopic
applications of optical coherence tomography. Optics Express
1998;3:257.
6. Madersbacher S, Hochreiter W, Burkhard F, et al. Radical cystec-
tomy for bladder cancer today—A homogeneous series without
neoadjuvant therapy. J Clin Oncol 2003;21:690.
7. Mostofi FK, Davis CJ Jr, Sesterhenn IA. Pathology of tumors of
the urinary tract. In: Skinner DG, Lieskovsky G (eds): Diagnosis
and Management of Genitourinary Cancer. Philadelphia: WB Saun-
ders, 1988, pp 83–117.
8. Utz DC, Farrow DM. Carcinoma in situ of the urinary tract. Urol
Clin North Am 1984;11:735.
9. Wallace DMA, Hindmarsh JR, Webb JN, et al. The role of multi-
ple biopsies in the management of patients with bladder cancer. Br
J Urol 1979;51:535.
10. Hee MR, Izatt JA, Swanson EA, et al. Optical coherence tomog-
raphy of the human retina. Arch Ophthalmol 1995;113:325.
11. Puliafito CA, Hee MR, Lin CP, et al. Imaging of macular diseases
with optical coherence tomography. Ophthalmology 1995;102:217.
12. Schmitt JM, Knüttel A, Yadlowsky, et al. Optical-coherence to-
mography of a dense tissue: Statistics of attenuation and backscat-
tering. Phys Med Biol 1994;39:1705.
13. Brezinski ME, Tearney GJ, Bouma BE, et al. Optical coherence
tomography for optical biopsy. Circulation 1996;93:1206.
14. Brezinski ME, Tearney GJ, Weissman NJ, et al. Assessing ather-
osclerotic plaque morphology: Comparison of optical coherence
tomography and high frequency intravascular ultrasound. Heart
1997;77:397.
15. Jäckle A, Gladkova N, Feldchtein F, et al. In vivo endoscopic op-
tical coherence tomography of the human gastrointestinal tract—
Toward optical biopsy. Endoscopy 2000;32:743.
16. Jäckle A, Gladkova N, Feldchtein F, et al. In vivo endoscopic op-
tical coherence tomography of esophagitis, Barrett’s esophagus,
and adenocarcinoma of the esophagus. Endoscopy 2000;32:750.
17. Pitris C, Goodman A, Boppart SA, Libus JJ, Fujimoto JG, Brezin-
ski ME. High-resolution imaging of gynecologic neoplasms using
optical coherence tomography. Obstet Gynecol 1999;93:135.
18. Tearney GJ, Brezinski ME, Southern JF, et al. Optical biopsy in
human urologic tissue using optical coherence tomography. J Urol
1997;157:1915.
19. Jesser CJ, Boppart SA, Pitris C, et al. High resolution imaging of
transitional cell carcinoma with optical coherence tomography:
Feasibility for the evaluation of bladder pathology. Br J Radiol
1999;72:1170.
20. Zagaynova EV, Streltsova OS, Gladkova ND, et al. In vivo opti-
cal coherence tomography feasibility for bladder disease. J Urol
2002;167:1492.
21. Tearney GJ, Brezinski ME, Bouma BE, et al. In vivo endoscopic
optical biopsy with optical coherence tomography. Science
1997;276:2037.
Address reprint requests to:
Michael J. Manyak, M.D.
Dept. of Urology
George Washington University
2322 Blaine Drive
Chevy Chase, MD 20815
E-mail: uromjm@gwumc.edu
MANYAK ET AL.
574