TRAP indicates telomeric repeat amplification protocol. The target condition
in this study was bladder cancer.
For the overall series, the area under curve is 0.951 (95% confidence
interval [CI], 0.925-0.976) and for individuals aged ≤75 years, 0.968
(95% CI, 0.942-0.993). Points are marked to demonstrate the sensitivity
and 1−specificity of urine cytology and of the telomeric repeat amplification
protocol (TRAP) test at cutoff points of 40, 50, 60, and 70 arbitrary enzymatic
Customize your JAMA Network experience by selecting one or more topics from the list below.
Sanchini MA, Gunelli R, Nanni O, et al. Relevance of Urine Telomerase in the Diagnosis of Bladder Cancer. JAMA. 2005;294(16):2052–2056. doi:10.1001/jama.294.16.2052
Context The identification of new molecular markers is one of the most challenging
goals for the early detection of bladder cancer because available noninvasive
methods have neither sufficient sensitivity nor specificity to be acceptable
for routine use.
Objective To develop a relatively simple, inexpensive, and accurate test that
measures telomerase activity in voided urine to apply to large-scale screening
programs for bladder cancer detection.
Design, Setting, and Participants Case-control study conducted in 218 men (84 healthy individuals and
134 patients at first diagnosis of histologically confirmed bladder cancer),
frequency matched by age and recruited between March 2003 and November 2004
in Italy. Urine telomerase activity was determined using a highly sensitive
telomeric repeat amplification protocol (TRAP) assay. Urine samples were processed
for cytological diagnosis and TRAP assay. The diagnosis of bladder cancer
was based on bioptic and cystoscopic examinations. The performance of the
TRAP assay to detect urine telomerase activity was compared with urine cytology
as an aid to early cancer detection. Quantification of urine telomerase activity
was conducted in a blinded manner.
Main Outcome Measure Sensitivity and specificity of TRAP to detect bladder cancer.
Results Using a 50 arbitrary enzymatic unit cutoff value, we validated the results
obtained in the pilot study. In the overall series, sensitivity was 90% (95%
confidence interval [CI], 83%-94%) and specificity was 88% (95% CI, 79%-93%).
Specificity increased to 94% (95% CI, 85%-98%) for individuals aged 75 years
or younger. The same predictive capacity of telomerase activity levels was
observed for patients with low-grade tumors or with negative cytology results.
Conclusions The present validation study demonstrated the ability of urine telomerase
activity levels to accurately detect the presence of bladder tumors in men.
This test represents a potentially useful noninvasive diagnostic innovation
for bladder cancer detection in high-risk groups such as habitual smokers
or in symptomatic patients.
The incidence of human bladder cancer has greatly increased over the
last few decades, with more than 60 000 new cases diagnosed each year
in the United States alone,1 and now represents
the 4th most common malignancy in men and the 10th most common in women.1,2 According to the latest reports from
the National Cancer Institute,3 the incidence
of this pathology is higher in industrialized than in developing countries.
Bladder cancer is 3 times more common among men than women, and the
incidence increases with age. Approximately 80% of newly diagnosed individuals
are aged 60 years or older.1 At present, about
20% of patients die each year, but when the disease is diagnosed and treated
in the early stage, the chances of survival are good, thus highlighting the
importance of a timely and accurate diagnosis.
More than 90% of newly diagnosed bladder cancers are transitional-cell
carcinomas. Approximately 75% of patients present with superficial cancer,
20% with invasive disease, and the remaining 5% with metastatic disease at
Established approaches for detecting bladder cancer include urine cytology
and cystoscopy, used singly or in sequence. However, the invasiveness and
relatively high cost of cystoscopic examination and the limited sensitivity
of urinary cytology, especially for low-grade superficial lesions, make it
of the utmost importance to develop a noninvasive, reliable, and simple test
to increase the rate of detection of bladder cancer. Among the markers investigated
for this purpose, an important role has been played by telomerase activity
in voided urine or bladder washings,6-10 determined
by the telomeric repeat amplification protocol (TRAP) assay.11 Initially,
studies dealt with qualitative determinations. To obtain a more accurate and
reliable estimate of telomerase activity levels, a quantitative TRAP assay
was developed, based on the exponential amplification of the primer-telomeric
repeats generated in the telomerase reaction.12-15 Using
this assay, telomerase activity has been detected in almost all superficial
urothelial cell carcinomas, but not in healthy urothelia.16 We
used the TRAP assay with the internal standard developed by Wright et al17 and added a reference curve to obtain more accurate
and reproducible results.18
The promising results from our pilot study18 prompted
us to carry out a case-control study, prospectively planned and performed
blindly on urine from male individuals to validate the 50 arbitrary enzymatic
units (AEUs) that emerged as the best cutoff and to define the diagnostic
accuracy of different telomerase activity cutoff values in terms of sensitivity
The study was conducted in 218 men (Figure
1), of whom 84 were healthy individuals and 134 were patients at
first diagnosis of bladder cancer, frequency matched by age (≤75 years
and >75 years). Median age was 62.4 years (range, 22-98 years) in healthy
individuals and 69.8 years (range, 33-88 years) in patients.
Healthy individuals were recruited from hospital laboratory staff and
geriatric wards, and none had been previously clinically diagnosed with any
type of cancer or with inflammatory pathologies of the urogenital tract.
Patients were prospectively enrolled from the Urology Departments of
Pierantoni-Morgagni Hospital (Forlì) and Infermi Hospital (Rimini)
between March 2003 and November 2004. All patients underwent cystoscopy as
a reference standard for bladder cancer detection, and all tumors or suspicious
lesions were resected. Patients who had undergone previous treatment were
The final diagnosis of bladder cancer was based on histologic examination.
Histologic type and tumor cell differentiation were determined according to
World Health Organization criteria. Fifteen (11%) tumors were well differentiated
(G1), 55 (41%) were moderately differentiated (G2), and 57 (42%) were poorly
differentiated (G3). There was 1 carcinoma in situ. Grading was not available
for 6 patients.
Demographic data and medical history were collected at study entry.
The local ethics committee reviewed and approved the study protocol for each
center, and all participants provided written informed consent.
Urine samples from both healthy individuals and patients were processed
for cytological diagnosis and TRAP assay. Each patient evaluated for bladder
cancer provided a voided urine sample immediately before cystoscopy.
Cytological examination was performed in all the urine samples from
healthy individuals (n = 84) and in 103 of the 134 bladder cancer
patients analyzed with TRAP assay. Forty-eight (46.6%) patients had positive
cytology, 40 (38.8%) had negative cytology, 8 (7.8%) patients with suspicious
cytology findings had evidence of bladder cancer at histologic examination,
and 7 (6.8%) had nonassessable cytology because of a lack of exfoliated cells.
The cytological examination was unavailable for 31 patients because they bypassed
this preliminary urine evaluation and directly underwent cystoscopy.
Cell extract preparation and TRAP assay were carried out as previously
described.18,19 Cells were pelleted
by centrifugation (850g for 10 minutes at 4°C)
within 1 to 3 hours of urine sample collection, washed once in phosphate-buffered
saline, resedimented by centrifugation (2300g for
5 minutes at 4°C), and stored at −80°C until use (a maximum
of 12 months). The pelleted cells were resuspended in 200 μL of lysis reagent19 and left on ice for 30 minutes.
Cell lysates were centrifuged (10 000g for
20 minutes at 4°C), and the supernatant extracts were stored at −80°C.
Aliquots of each urine sample containing 1 μg of protein lysate were used
for the TRAP assay. Telomerase products were evaluated on fluorescence electropherograms,
and the area underlying the different peaks was calculated. To obtain semiquantitative
levels of telomerase activity, an internal telomerase assay standard (ITAS;
25 attograms17), amplified by the same 2 primers
used for the telomerase activity assay, was included in the TRAP buffer. Protein
concentrations corresponding to 10, 30, 100, 300, 1000, and 3000 cells of
a human bladder cancer line (MCR)18 were analyzed
in each assay and used as the reference curve. To obtain quantitative evaluations
of telomerase activity, the areas of each sample were also normalized to the
150–base pair ITAS peak. The relative telomerase activity per cell for
each sample is presented as the percentage of the ratio of TRAP ladder/ITAS
per cell vs the value of MCR and expressed in AEUs. All experiments were performed
in duplicate, and when variations were greater than 15%, observed in about
10% of cases, a third analysis was performed. Telomerase activity was expressed
as a continuous variable in all analyses.
The population size was defined on the basis of results from the previous
pilot study18 in which we obtained 93% sensitivity
and 90% specificity using the 50-AEU cutoff value for the subgroup of male
individuals. In fact, for the 84 healthy individuals and 134 bladder cancer
patients of the present study, we predicted the 95% confidence interval (CI)
to be ±5% with respect to the single estimated value for sensitivity
and specificity. To avoid bias in the clinical utility of the TRAP assay,
we analyzed all samples prospectively, without previous knowledge of the patient’s
The threshold value for optimal sensitivity and specificity was determined
using a receiver operating characteristic (ROC) curve,20 constructed
by calculating the true-positive (sensitivity) and false-positive (1−specificity)
rates at several cutoff values. Sensitivity, specificity, and relative 95%
CIs were calculated for the most discriminant cutoff values. The relationship
between urine telomerase activity and histological grading was analyzed using
the median test. For all tests, a 2-sided P=.05 was
regarded as significant. Data analyses were performed with SAS release 8.0
(SAS Institute Inc, Cary, NC).
All statistical analyses were performed at the Unit of Biostatistics
and Clinical Trials of Istituto Oncologico Romagnolo, Forlì, Italy.
The median telomerase activity value in urine was 27 AEUs (range, 0-88)
in healthy individuals and 112 AEUs (range, 30-382) in patients. We did not
observe any patients with a telomerase activity value lower than 30 AEUs or
any healthy individuals with a telomerase activity value higher than 90 AEUs.
Moreover, in patients with negative or positive cytology, the median telomerase
activity values in urine were 99 (range, 38-265) and 134 (range, 37-253) AEUs,
As primary end point, we validated the results obtained in the pilot
study using a 50-AEU cutoff value. In the overall series, 90% (95% CI, 83%-94%)
sensitivity and 88% (95% CI, 79%-93%) specificity were observed.
As secondary end point, the diagnostic relevance of urine telomerase
activity was analyzed for the overall series and for the subgroups of individuals
75 years or younger and older than 75 years. The ROC curve analysis provides
a graphic demonstration of the sensitivity and specificity of telomerase activity
in the overall series and the even higher specificity in the subgroup of individuals
75 years or younger (Figure 2).
In particular, sensitivity in the overall series ranged from 61% to
100% and specificity from 54% to 100% according to the different AEU cutoff
values (Table 1). As shown in Figure 2, a similar sensitivity and an even higher
specificity (94%) (95% CI, 85%-98%) was obtained in the subgroup of individuals
75 years or younger.
Although an increase in urine telomerase activity levels was observed
from histologic grades 1 to 3, it did not reach statistical significance (Table 2).
The sensitivity of urine telomerase activity in detecting bladder tumors
was similar in the subgroups of patients with different tumor grades at all
AEU cutoff values. In particular, at 50 AEUs the sensitivity was 93%, 87%,
and 89% for grades 1, 2, and 3, respectively (Table 3).
Telomerase has been investigated as a potentially useful biomarker for
early cancer detection8,21-23 and
prognosis24 and for monitoring residual disease.21 Elevated levels of telomerase expression, in particular
of the human telomerase reverse transcriptase catalytic subunit, have been
observed in almost all human tumor histotypes, including bladder cancer. In
contrast, telomerase activity has not been detected in most of the somatic
We confirmed the high sensitivity and specificity of urine telomerase
levels, in particular of the 50-AEU cutoff observed in our pilot study, in
detecting bladder cancer.18
Moreover, we observed that sensitivity is not age dependent, whereas
specificity is higher in individuals younger than 75 years.
The test we developed requires a small amount of urine; is noninvasive,
inexpensive, and easy to perform; and permits a quantitative evaluation of
telomerase activity in urine. Furthermore, it is objective, reproducible,
and specific and is not reliant on the expertise of the cytopathologist.9 Another important advantage of this test is its ability
to also identify low-grade tumors, which often escape detection during cytologic
However, notwithstanding the validated optimal diagnostic accuracy of
the test, it is not recommended for use in routine screening programs because
of the low incidence of bladder cancer, and should be aimed at high-risk subgroups.
Specifically, smokers have about a 3-fold increased risk of developing bladder
cancer compared with nonsmokers. It might be even more advantageous in terms
of cost/benefit to use the TRAP assay in selected individuals who present
with hematuria. For this subgroup, the incidence of bladder cancer is about
10% to 15% and the sensitivity of urinary cytology is only 30% to 50%.31-33 Although cystoscopy
is the gold standard for the diagnosis of bladder cancer because of its 90%
sensitivity, the invasiveness and low specificity of the procedure in symptomatic
patients34 make it important to identify a
manageable and more accurate diagnostic tool.
In addition to telomerase, several new, alternative laboratory tests
based on the detection of different substances (eg, BTA tests, NMP22, fibrinogen
degradation products, hyaluronic acid, multicolor fluorescence in situ hybridization
well as novel research procedures (microsatellite analysis, DNA methylation,
RNA expression, real-time polymerase chain reaction analysis),44-46 have
become available in an attempt to improve the sensitivity of cytology for
the diagnosis of bladder cancer. However, many problems, such as low sensitivity,
unsatisfactory specificity levels, or technical difficulties for the application
of these tests in large population studies, have limited their clinical utility.
In conclusion, we believe that our telomerase activity urine assay,
with the reliability verified in pilot and confirmatory studies, represents
a promising and potentially important contribution to the early diagnosis
of bladder carcinoma, in particular for high-risk subgroups. Further prospective
studies on larger patient populations are needed to assess the diagnostic
role of urinary telomerase, to define the ability of this assay to detect
low-grade tumors, and to forecast clinical relapse.
Corresponding Author: Daniele Calistri,
PhD, Division of Oncology and Diagnostics, Morgagni-Pierantoni Hospital, Via
Forlanini 34, 47100 Forlì, Italy (firstname.lastname@example.org).
Author Contributions: Ms Sanchini had full
access to all of the data in the study and takes responsibility for the integrity
of the data and the accuracy of the data analysis.
Study concept and design: Sanchini, Nanni.
Acquisition of data: Sanchini, Gunelli, Fabbri,
Sermasi, Bercovich, Ravaioli, Amadori, Calistri.
Analysis and interpretation of data: Sanchini,
Drafting of the manuscript: Sanchini, Gunelli,
Bercovich, Ravaioli, Calistri.
Critical revision of the manuscript for important
intellectual content: Sanchini, Nanni, Bravaccini, Fabbri, Sermasi,
Statistical analysis: Sanchini, Nanni.
Obtained funding: Sanchini, Amadori, Calistri.
Administrative, technical, or material support:
Sanchini, Gunelli, Bravaccini, Fabbri, Sermasi, Bercovich, Ravaioli, Amadori,
Study supervision: Sanchini, Calistri.
Financial Disclosures: None reported.
Funding/Support: This work was supported by
Istituto Oncologico Romagnolo, Forlì and the National Research Council
(Consiglio Nazionale delle Ricerche [National Research Council]–Ministero
dell’Istruzione, dell’Universita e della Ricerca [Ministry of
Education, University and Research] “Oncologia”; grants 02.00082.ST97
Role of the Sponsor: Istituto Oncologico Romagnolo
and the National Research Council supplied the funding but did not participate
in the design and conduct of the study; in the collection, management, analysis,
and interpretation of the results; or in the preparation, review, or approval
of the manuscript.
Independent Statistical Analysis: Independent
statistical analysis was performed by Oriana Nanni, MSc, at the Unit of Biostatistics
and Clinical Trials of Istituto Oncologico Romagnolo, Forlì, Italy.
Acknowledgment: We thank Rosella Silvestrini,
PhD, Istituto Oncologico Romagnolo, for her invaluable scientific contribution
and Gráinne Tierney, BSc, Division of Oncology and Diagnostics, Morgagni-Pierantoni
Hospital, Forlì, for editing the manuscript. We also thank Giuliana
Amadori, RN, of the Department of Geriatrics, Morgagni-Pierantoni Hospital,
for assistance in sample collection. None of those acknowledged received compensation
from the study sponsors.