H.S. and M.B. are joint first authors, as they contributed equally to the manuscript. A.B. and B.S. are joint senior authors, as they contributed equally to the manuscript
Early results of bladder-cancer screening in a high-risk population of heavy smokers
Article first published online: 11 MAR 2008
© 2008 THE AUTHORS. JOURNAL COMPILATION © 2008 BJU INTERNATIONAL
Volume 102, Issue 3, pages 291–296, August 2008
How to Cite
Steiner, H., Bergmeister, M., Verdorfer, I., Granig, T., Mikuz, G., Bartsch, G., Stoehr, B. and Brunner, A. (2008), Early results of bladder-cancer screening in a high-risk population of heavy smokers. BJU International, 102: 291–296. doi: 10.1111/j.1464-410X.2008.07596.x
- Issue published online: 17 JUL 2008
- Article first published online: 11 MAR 2008
- Accepted for publication 7 January 2008
- bladder cancer;
- urine dipstick;
To report first results of an early bladder-cancer detection programme, and to evaluate the detection rate and the diagnostic value of the tests used.
SUBJECTS AND METHODS
Urine samples of 183 screened subjects with a history of smoking of ≥40 pack-years were collected for analysis with a urinary dipstick test for haematuria, the nuclear matrix protein-22 test (BladderChek®, Matritech, Inc., Newton, MA, USA), voided urine cytology and a molecular cytology test (UroVysion, Abbott Molecular Inc., Des Plaines, IL, USA). Participants with at least one positive test result had a further evaluation including cystoscopy and radiological imaging. The subjects’ risk factors, test results and histological findings were analysed.
In all, 75 subjects had at least one positive test result and were evaluated further; abnormal histological findings were detected in 18 (24% of those who had cystoscopy, 9.8% of the original 183), 15 of those in the urinary bladder, with pTaG1 (one), carcinoma in situ (two), dysplastic lesions (11) and one an inverted papilloma. In the upper urinary tract, two urothelial tumours (pTaG1 and pTxN2G3) and one renal cell carcinoma (pT1G2) were detected by computed tomography. In summary, six of 183 subjects (3.3%) had a histologically confirmed malignant tumour and another 12 (6.6%) were identified with a possible pre-cancerous lesion of the urinary tract. The urinary dipstick, BladderChek, cytology and UroVysion detected (i.e. were true-positive in) nine (50%), one (6%), seven (39%) and 11 (61%) of the 18 tumours found, while they failed to detect nine (50%), 17 (94%), 11 (61%) and seven (39%) of these lesions, respectively. Omitting the urine dipstick test, the BladderChek, cytology or UroVysion from the test setting could have spared 40, five, two or one subjects(s) from unnecessary invasive interventions; however, three, none, two or six lesions, would have been missed. More positive screening tests per subject was associated with a higher probability of a (pre)-malignant lesion.
Screening a high-risk group with a history of smoking of ≥40 pack-years showed a significant proportion (3.3%) with malignancy. These first results are encouraging and warrant continuation of the screening programme. In this series the most efficient screening tool was the combination of UroVysion, cytology and urinary dipstick testing. Of special scientific interest will be the follow-up of those patients with a possible pre-cancerous lesion.
carcinoma in situ
fluorescence in situ hybridization
nuclear matrix protein
(upper) urinary tract
positive predictive value.
Bladder cancer is the fourth most frequent cancer diagnosed in men, and the 11th most frequent cancer diagnosed in women. In the USA >60 000 people will be diagnosed with bladder cancer every year, and >13 000 are predicted to die from the disease . Although the incidence is lower than for other malignancies, bladder cancer accounts for the second most prevalent malignancy in middle-aged and elderly men, because of high recurrence rates . While the incidence has been quite constant, the mortality is decreasing slightly, probably due to improvements in diagnosis and treatment. However, there are still considerably many patients with a tumour that has already become invasive at the time of diagnosis.
Initially bladder cancer is usually asymptomatic and it can take years until symptoms develop. The most common (but not necessarily early) symptom is painless haematuria in ≈85% of patients . Cystoscopy is the standard for detecting bladder cancer but it is invasive, and urine cytology has an unfavourably poor sensitivity for low-grade tumours, and is interpreter-dependent. Therefore, there has been an intense search for urine-based markers as an appropriate screening tool in the past few years. Apart from urine dipstick testing for haematuria, other methods have been used, but the ideal marker for detecting bladder cancer has not yet been identified. However, it was proposed that a combination of markers would probably be efficient in a high-risk population .
Several risk factors and carcinogenic agents are known in the development of bladder cancer, but latency periods are often >10–20 years. The most important risk factors are cigarette smoking and occupational exposure to aromatic amines. By producing lesions in the DNA of target cells these and other carcinogens can initiate and augment the process of tumorigenesis . Brennan et al.[6,7] showed in two gender-based case-control studies that 30% of bladder cancer cases among women and 66% among men are due to smoking. Also, there was a direct correlation between the amount of smoking (numbers of cigarettes smoked per day and duration of smoking career) and the risk of bladder cancer.
The survival rate strongly depends on tumour grade and stage; the 5-year overall survival rate after radical cystectomy for pT2 tumours is 80%, for pT3 50% and for pT4 only 10%. If metastases occur, most of the patients die within a year after diagnosis. In the last few years some data were published suggesting that screening for bladder cancer can have positive effects on patients’ morbidity and mortality, and would help to reduce the economic costs, especially in high-risk groups [8,9].
Therefore, the Urology Department of the Medical University Innsbruck launched an early bladder-cancer detection programme in February 2007 for a high-risk population, defined as those with a history of smoking of ≥40 pack-years. According to statistical estimates the number of heavy smokers meeting this inclusion criterion currently amounts to ≈20 000 people in the province of Tyrol. In this report we present the screening test results, histological findings and first analyses, to assess the efficiency of this screening programme.
SUBJECTS AND METHODS
Since February 2007 subjects with a history of smoking ≥40 pack-years were invited by mass media to participate in this early-detection programme for bladder cancer. The protocol was approved by the institutional review board. A previous history of urothelial malignancy or stopping smoking >10 years earlier were exclusion criteria. After a personal conversation with a urologist, written informed consent was given and a questionnaire about smoking habits, medical history and possible environmental carcinogens was completed. Subsequently, a midstream urine sample (≈150 mL) of the second void of the day was collected. This was tested with a urinary dipstick (Combur-9®, Roche, Basel, Switzerland), following the instructions of the manufacturer; the result was either ‘negative’ or ‘positive’ for microhaematuria. If there was a current UTI, antibiotic treatment was administered and the subject screened 3 weeks later. The nuclear matrix protein (NMP)-22 test (BladderChek®, Matritech, Inc., Newton, MA, USA) was also used according to the directions of the manufacturer, and an additional 100 mL of urine was immediately fixed and sent to the pathology department for cytology (haematoxylin and eosin staining); the results were categorized as ‘negative for malignancy’, ‘suspicious or positive for malignancy’ (= positive) or as ‘atypia of uncertain significance’ by the pathologist (A.B. or G.M.). A fluorescence in situ hybridization (FISH) test, UroVysionTM (Abbott Molecular Inc., Des Plaines, IL, USA) was then used according to recommendations of the manufacturer; after hybridization and visualization by fluorescence microscopy all images were interpreted at the department of pathology (by I.V.) as ‘positive’ or ‘negative’.
Subjects with at least one positive test result were further evaluated, including CT of the upper urinary tract (UUT). The lower urinary tract was investigated by cystoscopy if there was a positive result from the dipstick or BladderChek, but negative cytology and UroVysion test. If there was a suspicious result from cytology or UroVysion subjects were advised to undergo cystoscopy, including photodynamic diagnosis after instillation of 5-aminolaevulinic acid under anaesthesia, and tumour was resected or random biopsies were taken. All histological samples were interpreted by A.B. or G.M. All endoscopic procedures were performed by one urologist (H.S.).
The consecutive procedures were as described in the flowchart (Fig. 1); if no signs of malignancy were detected during normal cystoscopy and bladder-wash cytology, subjects were advised to be re-tested after a year. A control cystoscopy was done after 6 months if the bladder-wash cytology showed atypia of uncertain malignancy. If there were suspicious findings at cystoscopy or atypia in bladder wash-cytology, another cystoscopy was done (this time under anaesthesia) with photodynamic diagnosis and bladder biopsy. If histology and cytology was normal, subjects were advised to be re-tested after a year. If there was dysplasia on histology or atypia in cytology, control cystoscopies were scheduled at 3, 9, 18 and 30 months. Malignant lesions were treated as proposed by commonly accepted guidelines .
In all, 183 subjects were included in the screened population up to July 2007; the demographic characteristics and results of the screening tests are shown in Table 1. The mean (range) number of pack years was 58 (40–152.5). In all, 108 subjects (59.0%) had no positive test result and were advised to be re-tested after a year; 54 (29.5%) had a positive result on urine dipstick testing and/or BladderChek only and had cystoscopy with no anaesthesia. However, in seven of these 54 subjects a second cystoscopy was required (under anaesthesia) because of abnormal findings. Overall, 21 subjects (11.5%) had a positive cytology and/or UroVysion test, and had subsequent cystoscopy combined with acquisition of samples for histology.
|All negative||Positive||No. of positive tests|
|Mean (sd)||59.1 (8.08)||61.4 (8.60)||65.9 (13.62)||63.6 (9.01)||62.6 (9.47)||60.1 (8.70)||60.38 (8.34)||65.70 (12.80)||58.17 (7.08)|
|Gender, n (% of all M or F)|
|M||72 (58.5)||35 (28.5)||9 (7.3)||10 (8.1)||13 (10.6)||123 (100)||39||9||3|
|F||36 (60)||22 (37)||2 (3)||0||2 (3)||60 (100)||22||2||0|
|Pack-years, n (%)|
|40–59.9||66 (57.9)||39 (34.2)||5 (4.4)||5 (4.4)||8 (7.0)||114 (100)||41||6||1|
|60–79.9||28 (62)||14 (31)||1 (2)||5 (11)||4 (9)||45 (100)||11||4||2|
|≥80||14 (58)||4 (17)||1 (4)||0||3 (13)||24 (100)||9||1||0|
Of the 75 subjects with at least one abnormal urine-based test, 18 (24% of the 75, 9.8% of the original 183) had abnormal histological findings detected during the evaluation. Fifteen (of 18) were detected by cystoscopy in the urinary bladder and the histological evaluation showed the following lesions: urothelial cancer pTaG1 (one), carcinoma in situ (CIS, two), dysplastic lesions (11) and inverted papilloma (one; Table 2). Two upper urinary tract (UUT) TCCs (pTaG1 and pTxN2G3) and one RCC (pT1G2) were identified by CT. In summary, six of 183 subjects (3.3%) had histologically confirmed malignant disease and another 12 (6.5%) were identified with a possible pre-cancerous lesion of the UT.
|Histological finding||Dipstick||BladderChek||Cytology||UroVysion||Detected by|
|CIS + dysplasia||neg||neg||neg||pos||Cystoscopy|
|UUT TCC pTaG1||pos||neg||pos||pos||CT|
|UUT TCC pTxN2G3||pos||pos||unc||np||CT|
One subject with no positive test result (negative urine dipstick, negative BladderChek, invalid cytology and UroVysion, because of too few cells in the urine sample) was advised to be re-tested after a year. At 7 weeks after screening this subject had macrohaematuria, and a papillary urothelial carcinoma of low malignant potential was diagnosed by transurethral resection.
There was no statistically significant difference in risk factors for bladder cancer (occupational exposure, history of radiation and family history) between subjects with or with no histological findings (data not shown). Interestingly, there was no statistical difference in numbers of pack-years between subjects with a histological finding (mean 55.6 pack-years) and those with no suspicion of tumour (mean 56.6 pack-years).
A comparison of the test results exclusively in the 75 subjects who had cystoscopy is shown in Table 3; urine dipstick, BladderChek, cytology and UroVysion detected (were true positive in) nine (50%), one (6%), seven (39%) and 11 (61%) of 18 lesions found on cystoscopy, while they failed to detect (were false negative in) nine (50%), 17 (94%), 11 (61%) and seven (39%) of these lesions, respectively. Of the 57 subjects with normal findings during the evaluation, the results were true-negative in nine (16%), 47 (83%), 54 (95%) and 53 (93%), and false-positive in 48 (84%), 10 (18%), three (5%) and four (7%), respectively.
From the data in Table 3, assumptions about the diagnostic value of the four screening tests can be made. Although overall sensitivity and specificity analyses are not possible because only 75 of the 183 subjects had cystoscopy, the positive predictive value (PPV) of each test can be used as utility criteria. Omitting urine dipstick, BladderChek, cytology or UroVysion from the test setting could have prevented 40, five, two or one subjects from unnecessary invasive interventions, respectively; however, three, none, two or six lesions, would have been missed. An increasing number of different positive screening tests per person was associated with a higher probability of a (pre) malignant lesion (Table 4).
|N positive results||N subjects||Positive findings, n/N (%)|
|1||61||8/61 (13)||3/61 (5)|
Cancer screening intends to detect malignancies in early or even pre-invasive stages when therapeutic measures are promising, leading to a reduction of morbidity or mortality and to reduced costs for the healthcare system. Current cancer screening is mainly focused on prostate, breast, colorectal and cervical cancer. Interestingly, a screening method was established for cervical cancer but not for bladder cancer, despite bladder cancer in women being diagnosed significantly more often than cervical cancer (estimated incidence of 17 120 for bladder than 11 150 for cervical cancer in 2007, USA) . Bladder cancer is not among the most common cancers but its diagnosis usually means lifelong monitoring and treatment because of high recurrence rates, and this significantly affects the patient’s quality of life, making bladder cancer the most costly malignancy per patient from diagnosis to death . Moreover, bladder cancer is hardly ever detected incidentally at autopsy, and this obviously has a short preclinical duration, contrary to prostate cancer. This could be interpreted to suggest that no one with a tumour detected is treated unnecessarily. Because of this short preclinical phase, Messing et al. proposed that bladder cancer screening should be repeated at least annually.
Most bladder cancers are diagnosed when superficial, but about a quarter of cases have already become invasive at the time of diagnosis. Goldstein and Messing  argued for screening for bladder cancer because superficial bladder tumours are easily treated and have a very good prognosis, even if they are high-grade tumours, but once the disease has invaded into deeper layers the prognosis deteriorates rapidly, despite aggressive therapy.
As haematuria is the most common symptom of TCC, testing the urine for occult blood using chemical reagent strips seems to be an effective method. Two screening studies using haematuria ‘home testing’ have been published. Messing et al. and Britton et al. requested middle-aged and elderly men to test their urine repeatedly for haemoglobin. Both studies found that 15–20% of the screened subjects had blood in their urine and 6–8% had urothelial cancer. Most important, Messing et al.[8,15] compared long-term outcomes of screening participants with those of an unscreened population and found a significantly higher proportion of muscle-invasive cancers in the latter. This information suggests that screening can indeed lead to a down-staging of the tumours. Recently, an update of this series was published that indicated a reduced mortality from bladder cancer in screened subjects with diagnosed bladder cancer, which underlines the possible clinical implication of screening for bladder cancer .
Besides haematuria screening, several other instruments have been considered for detecting bladder cancer. The cytological analysis of exfoliated cells found in the urine, FISH to detect genetic abnormalities, urinary marker antigens, growth factors and other biomarkers involved in urothelial carcinogenesis have been discussed as possible useful tools for screening. For the present early-detection programme, a combination of commercially available tools was used. Cystoscopy, the reference standard for detecting bladder cancer, but not appropriate for screening because it is invasive, served as a further diagnostic investigation if any of the screening test results was positive. Cancer was diagnosed in six of 183 subjects screened, 11 having dysplastic lesions and one an inverted papilloma. Obviously, these latter findings are not obligate pre-cancerous lesions, but there are several lines of evidence that urothelial dysplasia can indeed be a precursor lesion for invasive bladder cancer . An inverted urothelial papilloma has a very low risk of recurrence after complete excision . However, some cases of inverted papilloma have been associated with urothelial carcinoma, occurring either concurrently or subsequently, and even cases of urothelial carcinoma arising within inverted urothelial papilloma have been reported . Based on these facts it might be concluded that 18 subjects (i.e. 9.8%) were identified either with bladder cancer or being at potential risk of developing the disease in the defined population of heavy smokers. According to Lotan et al., who evaluated the cost per life-years saved associated with bladder cancer screening using an urinary based marker, screening is the most cost-effective strategy if the cancer incidence is >1.6% in the screened population, as long as screening costs, marker sensitivity and the rate of downstaging stay within certain ranges. In the present population malignant lesions were detected in 3.3% of subjects.
Comparing the screening tests used, the UroVysion and urinary cytology clearly outperformed both BladderChek and the urinary dipstick test. Although the sensitivity and specificity could not be calculated because fewer than half of all subjects were further evaluated by cystoscopy, the present results indicate that the BladderChek added no useful information. This is in contrast to other studies where the BladderChek was considered as a promising marker to detect bladder cancer, even in the early stages. Both, Grossman et al., who investigated the use of the BladderChek as an adjunct to cystoscopy, and Ponsky et al., who refined the use of the BladderChek in screening and monitoring for bladder cancer considered the BladderChek as a reliable and tumour-specific marker for detecting bladder cancer. It is difficult to interpret these differing results compared with the present study, but this issue will be addressed during the further course of this screening programme. Urine dipstick testing and the BladderChek are cheap and easy to use but have a low specificity. This was also reported by Britton et al., as besides four cases of bladder cancer, prostate cancer, cystitis, bladder stones and other pathologies were also found. Equally, the BladderChek test can be positive in patients many kinds of abnormalities of the UT.
UroVysion identified 11 of 18 subjects with dysplasia or tumour, and was false-positive in only four subjects. Sarosdy et al. evaluated the UroVysion FISH assay for diagnosing bladder cancer in subjects with haematuria, showing that the PPV of a positive UroVysion result was significantly higher in smokers with a history of >40 pack-years (65%) than in non-smokers (20.4%) or those smoking <40 pack-years (13.6–24.2%). It was concluded that heavy smokers might be a target group for molecular screening even in the absence of haematuria. In the present data, urine cytology showed seven true-positive and three false-positive results, leading to a PPV of seven of 10. Urine cytology and UroVysion seem to complement each other, because in 14 of 18 subjects with abnormal histological findings at least one of these tests was positive. The disadvantages of UroVysion and cytology are the high costs and the time lag to obtain results compared with the two ‘point-of-care’ tests. Furthermore, a considerable number of tests. In all, 64 UroVysion (35%) and 48 cytology (26%) results were not interpretable because there were too few urothelial cells in the voided urine samples, and were considered as not positive. Because of this, one tumour was missed and later diagnosed as bladder cancer because of symptoms (see above).
In conclusion, screening a high-risk group with a smoking history of ≥40 pack-years showed a significant proportion of patients with malignancy (3.3%). These early results are encouraging and warrant the continuation of the screening programme. Of specific scientific interest will be the follow-up of those subjects with a possible pre-cancerous lesion. The most efficient screening tool was the combination of UroVysion, cytology and urine dipstick; the value of the BladderChek in this setting will be further investigated.
CONFLICT OF INTEREST
- 5Campbell’s Urology. Philadelphia: Saunders, 2002.
- 17Urothelial dysplasia. In EbleJN, SauterG, EpsteinJI, SesterhennIA eds, World Health Organisation Classification of Tumours. Pathology and Genetics of Tumours of the Urinary System and Male Genital Organs. Lyon: IARC Press, 2004: 111–2.
- 18Inverted papilloma. In EbleJN, SauterG, EpsteinJI, SesterhennIA eds, World Health Organisation Classification of Tumours. Pathology and Genetics of Tumours of the Urinary System and Male Genital Organs. Lyon: IARC Press, 2004: 114–5.