A phase II validation of in vivo Raman spectroscopy for bladder cancer diagnosis

Bladder Cancer Epidemiology
Bladder cancer is a significant public health problem responsible for more than
130,000 deaths annually worldwide. It represents the fourth most common cancer
in men and the 8th most common in women for a 3:1 male predominance. Bladder
cancer is a malignant neoplasm originating from the surface lining (mucosa) of
the bladder. The most common form is urothelial cell carcinoma (UCC) which
accounts for 90-95% of all bladder cancers. The remainders are squamous cell
carcinomas (3-7%), adenocarcinomas (1-2%).

Disease presentation
When cancer is limited to the mucosa of the bladder it is referred to as
"superficial", mainly defined by its appearance at surgical cystoscopic removal
rather than its intrinsic invasive potential. Approximately 70% to 80% of
patients with newly diagnosed bladder cancer will present with superficial
bladder tumors. Those who do present with superficial, noninvasive bladder
cancer are often curable. Patients in whom superficial tumors are less
differentiated, large, multiple, or associated with carcinoma in situ (CIS) in
other areas of the bladder mucosa are at greatest risk for recurrence and the
development of invasive cancer. Such patients may be considered to have the
entire urothelial surface at risk for the development of cancer. Early
diagnosis and complete resection of tumor lesions is essential to bring a
change in prognosis for those with CIS and high grade neoplasms in particular.

Current Diagnostic Methods
The gold standard in diagnosis for bladder cancer is a selection of the region
of interest by ordinary white light cystoscopy and pathological assessment of
biopsies from selected lesions of the bladder wall. In case of non-muscle
invasive cancer, the treatment strategy is to eradicate existing disease,
prevent tumor recurrence and avoid development of invasive disease. To prevent
tumor recurrence, adjuvant intravesical (immuno- or chemo-) therapy (IVT) with
use of various drugs is used to destroy viable tumor cells. The high recurrence
rate may be due to overlooked and or incompletely removed lesions related to
the diagnostic assessment by white light cystoscopy. Photodynamic diagnosis
(PDD) is a kind of fluorescence guided resection during cystoscopy. It was
introduced in urology as a modality that would enhance visual contrast of
normal versus tumor tissue of the bladder and enhances the diagnostic value of
cystoscopy. However despite the increase in sensitivity this contrast
enhancement has been received with skepticism because of its lack in
specificity due to more false positive lesions.

Raman spectroscopy
Raman spectroscopy is a molecular specific technique that can be used as a
biochemical tool to study different (biological) materials; in particular this
technique has the capability to provide differential diagnosis of pre-cancers
and cancers.
An in vitro study was performed by Crow et al. to determine the sensitivity and
specificity of Raman spectroscopy. Bladder samples collected during cystoscopic
procedures were snap-frozen and a section was taken for histological
examination. Samples were classified as normal, cystitis, carcinoma in situ
(CIS), urothelial cell carcinoma and squamous cell carcinoma (SCC). In 76
patients, 1685 spectra were recorded (590 benign and 1095 malignant spectra).
These spectra were analyzed using principal-component fed linear-discriminant
analysis (PCA/LDA), to construct a diagnostic algorithm. The algorithm was
tested for its accuracy in predicting the histological diagnosis. The accuracy
achieved by the algorithm for normal, cystitis, CIS, TCC and SCC, were
respectively (sensitivity), 91%, 79%, 86% and 84%, and 98% and (specificity)
96%, 92%, 97%, 96% and 100%.

Raman spectroscopy and Photodynamic Diagnosis
Raman spectroscopy is a highly specific technique, but suffers from lack of
screening possibilities with respect to the surface area assessed (order of
magnitude square cm). Photodynamic diagnosis (PDD) of bladder cancer enables
gros evaluation of the bladder wall surface with high sensitivity (97%), but
relatively low specificity (50%)16, 28. Therefore, a combination of PDD and
Raman spectroscopy, may enable optical diagnosis at the lesion of interest and
would enhance the specificity and efficacy for early bladder cancer diagnosis.
An in vitro study of Raman spectroscopy of bladder wall samples after
fluorescence image guided biopsy showed the feasibility of the combination of
these
techniques.
Our group investigated the combination of Raman spectroscopy and PDD in a
retrospective study. Patient groups were identified who are associated with an
increased number of false positives in fluorescence diagnosis and would
therefore benefit most from the addition of Raman spectroscopy. Multivariate
analysis showed that female patients and patients who have had a recent TURBT,
within 12 weeks before PDD, would benefit most form highly specific Raman
spectroscopy because recent TURBTs and female gender are significant
independent predictors of fals positive findings in PDD.

Rationale for differentiation between benign and malignant, grade and stage
Application of Raman spectroscopy in pre-clinical and clinical diagnosis in
preliminary research, has shown that algorithms are not one-to-one
interchangeable in translation from an in vitro to an in vivo situation. This
is caused on one hand, by the sensitivity of this technique for subtle
biochemical changes and on the other hand, the biochemical difference of
biopsies of the same tissue in vivo. Nevertheless the result of patient
measurements can be compared between the patients diagnosed with use of PDD and
white light, with some adjustments.The preliminary study phase I "Determination
of instrument parameters for the in vivo application of Raman spectroscopy for
Bladder Cancer Diagnosis" showed that spectra with enough signal/noise ratio
can be obtained in a clinical setting in vivo. Furthermore a correlation of the
Raman signal and the invasion was determined. Further differentiaton of
subgroups requires more research.

Primary Objective:
to clinically use and validate the optimized Raman probe in a patient database
in order to differentiate between benign and malignant bladder lesions in vivo.

Secondary objective:
to develop an algorithm to predict tumor grade in exophytic and flat lesions

Tertiary Objective:
to develop an algorithm to predict tumor stage in exophytic or solid lesions by
accurate prediction of the depth of invasion.

This study will be performed in the UMC Utrecht and Sint Antonius Hospital in
Nieuwegein. 285 patients will be included, who are sceduled for a TURBT by
standard medical care.

Intervention study: This concerns an experimental study to determine a
differentiation between benign and malignant bladder lesions, using in vivo
Raman spectroscopy in bladder cancer detection.

All subjects are measured at the time of the procedure; after selection of
regions as determined by standard medical
diagnostics. In each patient the acquisition time is 3 seconds per measurement.
A database of patient measurements
will be produced. Using data-analysis, spectral differentiation between benign
and malignant is conducted as well as
between different grades and stages.

First objective:
- benign
- malignant

Second objective:
To grade :
- flat lesions
o normal
o inflammation
o pre-malignant (hyperplasia, atypia and dysplasia)
o Carcinoma In Situ (CIS)
- exphytic lesions
o grade 1
o grade 2
o grade 3

Third objective:
To stage:
- non-invasive Ta (NMI-BC)
- superficial invasion T1 (NMI-BC)
- muscle invasion T >= 2 (MI-BC)

The Raman spectroscopic diagnoses are compared to the pathologic diagnoses of
285 patients from the UMC Utrecht and the Sint Antonius Hospital.

The extent of the burden for the patient in this study is limited to an
extension of the surgical procedure time in the operating room. The standard
assessment of the disease by visual evaluation of the bladder wall is carried
out regularly. Suspicious regions selected for biopsy then are recorded on a
form designed for this purpose. When present, lesions that are expected to be
normal, deviant benign and/or malignant are selected for Raman spectral
measurements. At each location multiple measurements are taken with acquisition
times of 3 seconds and use of two different probes, to determine
differentiation between benign and malignant, grade and stage using in vivo
Raman spectroscopy in the bladder. The extension in time to the procedure will
not exceed 15 minutes. A minimal increase in risk for damage of the bladder
wall considering the increase in instrument manipulation time in the urinary
bladder is the additional risk.
Another additional risk in this procedure might be that an extra biopsy is
being taken from healthy mucosa which might lead to bleeding in the bladder.
Nevertheless, during a TURBT procedure, the bleeding can be stopped by
electrical coagulation which is used regularly during a TURBT.

Risks of a surgical procedure: pain, infection, scarring, bleeding.
Risks of anesthesiology: hypersensitivity of medicins, respiratory arrest,
cardiac arrest. The risks during local anesthesiology are minimal.