An ideal tumour marker should be present in all patients with that tumour, absent in people without the tumour and its concentration should be proportional to the tumour burden. The marker would be easy and cheap to measure, and its estimation would be accurately reproducible. Such a marker could be used for screening, for monitoring the response to treatment, or both.
The natural history of superficial bladder cancer (SBC) is unique in its propensity for local recurrence over the course of a patient's life. Although the statistical probability of developing invasive disease can be assessed from the histology and pattern of recurrence, it can never be completely discounted in an individual patient even in the lowest risk group. For this reason, patients with SBC are subjected to a long-term endoscopic follow-up.
‘Check cystoscopy lists’, whether under local or general anaesthetic and whether performed by nurses or doctors, represent a considerable part of the workload of any urological unit. If a tumour marker could identify those patients with recurrent tumours, many patients would be spared unnecessary invasive procedures and the NHS saved a considerable sum of money. Several markers have been identified and evaluated. Some are readily available and some are still research tools. How closely do currently available urinary markers meet these needs?
Voided urine cytology (VUC) has been used both in the diagnosis and follow-up of SBC since its first description by Papanicolou and Marshall in 1945. There have been many commercial tests assaying urinary constituents that have not stood the test of time. The best of the currently available tests are BTA-TRAK, BTAstat, nuclear matrix protein-22 (NMP22) and ImmunoCyst. Other tests under development but not yet commercially available include telomerase, hyaluronic acid/hyaluronidase (HA/HAase) and fibrinogen/fibrin degradation product (F/FDP) assays.
BTAstat is a qualitative test and BTA-TRAK a quantitative test for human complement factor H-related protein. The BTAstat test is simple and can be used as a ‘near-patient’ test. BTA-TRAK is a more complex procedure requiring trained personnel .
Several NMPs have been described; NMP22 is found in human epithelial cells, is present in voided urine and higher levels are associated with TCCs. The test for NMP22 is a multistage process undertaken by trained technicians in a hospital or reference laboratory setting .
ImmunoCyst uses three monoclonal antibodies to stain exfoliated malignant cells in voided urine and thus improve the sensitivity of cytology. It requires skilled laboratory technicians for the staining and to assess the specimen with fluorescent microscopy .
Telomerase is a ribonucleoprotein that enables the synthesis of telomere endings; it can be measured in urine by a PCR technique. The process is complicated and requires a reference laboratory .
HA and HAase can be measured using an ELISA-like assay. Levels of HA are elevated in certain tumours and it appears to be a promoter of metastasis. HAase breaks down HA and has been detected at increased levels in the urine of patients with bladder tumours. The test is not yet commercially available, and will require a reference laboratory .
F/FDP have long been known to be increased in the urine of patients with bladder cancer . The assay was available as a point-of-care test, but has now been withdrawn.
The results of comparative studies of all these tests are summarized in Table 1, which is a composite of several different studies, none of which were a direct comparison of all of the tests listed. Moreover, each study had a different proportion of patients with no bladder cancer, and the grades of cancer also varied among the studies. However, the three readily available tests were compared directly in several studies [2,7,8] and the findings are relatively consistent:
|Test||N||Sensitivity, %||Specificity, %||Comment|
|BTAstat||1482||67.7||65.8||False positives with infection/haematuria|
|Telomerase||(187)||74||89||Complex test*– not commercially available|
|HA/HAase||504||91||86||Complex test*– not yet commercially available|
|F/FDP||192||68||86||No longer available as commercial test|
- (i) VUC has poor specificity but high sensitivity; its sensitivity increases with the grade of tumour, so that in carcinoma in situ or grade 3 tumours the sensitivity is ≈ 90%.
- (ii) The BTA and NMP22 tests have a higher sensitivity as they are better able to detect lower grade tumours. They both have a lower specificity than VUC. False-positives may occur in patients with stone disease, haematuria or UTI.
- (iii) Some of the newer tests have high sensitivity and specificity but are complex, require reference laboratory analysis and may be quite expensive.
How can these tests be integrated into clinical practice? As yet, none of the available tests is sufficiently accurate to replace cystoscopy in the investigation of a patient with a possible bladder tumour. VUC if positive indicates a high-risk tumour and may be useful in prioritising patients who have been found to have a tumour by a haematuria clinic, if there is any delay in arranging definitive treatment. VUC may also be useful in the follow-up of patients with known high-risk disease. Positive cytology with negative cystoscopy should prompt a careful investigation for extra-vesical disease (prostate or upper tract).
Could any of the tests be used to direct the frequency of cystoscopic follow-up of patients with bladder cancer? There is already good evidence as a guide; low-risk (G1pTa disease) with no recurrence at 3 months requires no more than annual cystoscopy. High-risk disease requires vigilant follow-up because of the risk of invasion, and 3-monthly cystoscopy plus VUC remains the safest option.
Could the intermediate risk group be managed differently? If review cystoscopy after appropriate intravesical therapy is clear, might 3-monthly cystoscopy be replaced with noninvasive urine testing? The evidence suggests that VUC would detect high-grade recurrence, and BTAstat or NMP22 would detect the less threatening lower grade disease. This would seem to be the most promising area for noninvasive urine testing to be integrated into the care of patients with SBC. So far, this hypothesis has not been tested in a large-scale prospective study, but the time may be right for this to be evaluated .
One factor that should not be discounted is patient preference. In an unpublished study from Nijmegen, Witjes interviewed patients who had been on regular cystoscopic surveillance for> 12 months. Patients were asked whether they would prefer to continue with flexible cystoscopy or have a noninvasive test. When the sensitivity of the test was set at < 90%, over 90% of the patients chose cystoscopy.