Ancillary methods for the detection of recurrent urothelial neoplasia

Abstract

The diagnosis of tumors of the urinary bladder and urinary tract account for 6% of new cancer cases in men and 3% of new cancer cases in women in the U.S. each year.1, 2 Two distinct clinical forms of the disease are recognized: noninvasive or superficially invasive, typically low grade, “papillary” tumors with a favorable prognosis and “flat” tumors, which often are of high histologic grade and detected either as noninvasive in situ lesions or with bladder wall penetration and the potential for an adverse clinical outcome. Both papillary and flat lesions have a strong tendency to recur either as a regrowth of an incompletely excised original lesion (rare) or as a new primary lesion (more common). Of the 54,000 new diagnoses of urinary bladder carcinoma in the U.S. each year, 75% are at an early stage at the time of presentation with no invasion of the smooth muscle wall of the bladder.3, 4 Of these superficial bladder tumors, 50–70% will recur and 10–20% will progress to muscle invasive disease.5 The relatively high new case rate coupled with the high recurrence rate results in an extremely high overall disease prevalence and a significant cost to the health care system for the continued surveillance, follow-up, and management of these patients. Moreover, because not all superficial lesions will progress to muscle invasive disease, it may be neither clinically necessary nor cost-effective to detect them when they recur. Thus, there is considerable interest in the development of new ancillary techniques that could detect clinically significant recurrent urothelial neoplasia with higher sensitivity and specificity, lengthen the time intervals and thereby potentially reduce the overall cost of cystoscopic surveillance, and allow for curative surgical intervention prior to the development of life-threatening tumor invasion of the bladder wall. Cytologic examination of urothelial cells from voided urine, urinary bladder washing, and upper urinary tract brushing specimens in combination with cystoscopic examination has been the gold standard for the detection of recurrent urothelial neoplasia.6 The sensitivity for the detection of recurrent urothelial neoplasms (papilloma, urothelial neoplasm of low malignant potential, urothelial carcinoma, and urothelial carcinoma in situ) has shown a wide variation in the published literature.7, 8 For patients with higher grades of urothelial carcinoma, cytologic examination can achieve a high sensitivity and specificity (Table 1).9 However, for patients with lower grade tumors, false-positive and false-negative rates can be >10%.9 The detection of low grade urothelial neoplasia, a continuing challenge for cytopathologists, can be improved when experts use vigorous cytomorphologic criteria to separate low grade lesions from benign urothelial cells.9, 10 Yet, when used by nonexperts in a general pathology practice setting, exfoliative cytology has been criticized for having a high false-negative rate in patients with low grade lesions.11 In addition, the atypical cytomorphology of the urothelium associated with the use of intravesical chemotherapeutic agents can complicate the interpretation of urothelial cytologic specimens.12 Moreover, despite a published range of 0–73% sensitivity for the detection of low grade urothelial neoplasia by cytologic methods, substantial increases can be achieved when key cytologic criteria for detection are identified and effectively applied.12 The multifocal nature of urothelial neoplasia also must be considered as well as the impact of specimen type; a voided urine sample has a major impact on the cytologic detection of significant lesions compared with a bladder washing specimen.13 However, in clinical practice, a negative cytology report is significant in that although a low grade lesion might be missed, a clinically significant high grade lesion generally can be excluded. In summary, despite the high sensitivity and specificity for the detection of both low grade and high grade urothelial neoplasms achievable in those laboratories directed by expert cytopathologists, the general consensus is that cytologic techniques are not capable of identifying all cases of recurrent disease. The development of ancillary techniques to compliment the relatively inexpensive and convenient cytologic methods is relevant to both clinical patient outcome and financial outcome in an era of health care cost containment. A variety of wet laboratory immunoassays, on-slide immunoassays, and in situ hybridization procedures and postnucleic acid extraction molecular techniques have been designed to compliment cytology and improve the overall sensitivity and specificity for the detection of recurrent urothelial neoplasia (Table 2).2, 3, 14 Although the current consensus maintains that no new individual marker can eliminate the need for follow-up cystoscopy, there is significant agreement that, with the use of the ancillary tests, the sensitivity and specificity of cytologic diagnosis can be increased and the intervals between surveillance cystoscopies for the management of urothelial neoplasia can be lengthened.14 The Bard Bladder Tumor Antigen (BTA)™ test (Bard Corp., Murray Hill, NJ) is an FDA-approved latex agglutination test designed to detect bladder carcinoma in voided urine samples. Two new improved versions; the BTA Stat™ (performed as a point-of-care test at the bedside) and the BTA Track™ test (which is mailed to a reference laboratory), recently were introduced.3 Table 3 lists a series of studies of the BTA™ test for the detection of urothelial neoplasia and compares the results with other methods. The sensitivity for this technique has ranged from a low of 32% to a high of 100%, with a range in specificity of 40–96%.14-25 Although the majority of studies have indicated that the BTA™ test is superior in sensitivity to cytology,15-18 one study found low sensitivity and specificity rates for high grade urothelial carcinomas,14 another reported high false-positive and false-negative rates,21 and another concluded that the test was useful to monitor disease aggressiveness but unable to replace cystoscopy for the treatment of patients.23 In addition, the BTA™ test results have been compared with other ancillary methods and found to be less sensitive and specific than the Quanticyt karyometry assay20 and significantly less specific than telomerase assays.24 However, one study found the BTA™ test to have a higher sensitivity than the nuclear matrix protein NMP 22™ test.22 Recently, Nasuti et al. reported a high sensitivity but low positive predictive value for the one-step BTA Stat™ test compared with voided urine cytology in 100 patients and concluded that the high false-positive rate for the BTA Stat™ test limited the test to a role as an adjunct to cytology in a screening format.25 Sarosdy et al.15 D'Hallewin and Baert16 Ianari et al.17 Leyh and Mazeman18 Miyanaga et al.19 Wiener et al.14 Van der Poel et al.20 Zimmerman et al.21 Abbate et al.22 Schamhart et al.23 Ramakumar et al.24 Nasuti et al.25 In general, studies of the Bard BTA™ test were performed on voided urine samples rather than bladder washing specimens and the comparison of sensitivity and specificity used cytology reports from “nonexpert” cytopathologists. In summary, despite high false-positive results in the range of 4–34%3 and variable accuracy for the detection of low grade lesions,10 the BTA™ test shows significant promise as an adjunct to the detection of early urothelial neoplasia in bladder washing and voided urine samples. The impact of test improvements currently under clinical trial evaluation is not yet known. The nuclear matrix protein (NMP) 22 is a member of a family of nuclear matrix proteins involved in DNA configuration, structure, and function regulating fee replication and transcription of a variety of genes.26 The NMP 22™ detection method is an immunoassay currently performed in reference laboratory settings as a referred specimen. A summary of studies of the NMP 22™ assay for the early detection of urothelial neoplasia is shown in Table 4. NMP 22™ is a quantitative assay in which the setting of the cutoff point, ranging from 6–12 IU/mL in published series, is critical for optimizing the test performance. This test has been approved by the FDA as an adjunct to cystoscopy, but is not approved as a stand alone test capable of replacing cystoscopy. Soloway et al.27 Miyanaga et al.28 Witjes et al.29 Serretta et al.30 Abbate et al.22 Stampfer et al.31 Wiener et al.14 Landman et al.32 Ramakumar et al.24 Hughes et al.33 As seen in Table 4, previously published studies have reported a range in sensitivity from 48–81% and specificity from 60–86% for the detection of urothelial neoplasia using the NMP 22™ assay.14, 22, 24, 27-33 In general, the test has been listed as having a higher sensitivity than cytology and is considered useful for monitoring bladder carcinoma recurrence.14, 27-29 In one study the NMP 22™ assay had a higher sensitivity than the BTA™ test,32 and in another study the sensitivity was lower than the BTA™ test.22 Of two studies combining NMP 22™ and telomerase assays, both were found in one study to have a higher sensitivity than the BTA™ test and cytology,32 whereas in the other study telomerase assessment outperformed the NMP 22™ test, as well as the BTA™ test and conventional cytology.24 In a recent study of 107 patients, the positive predictive value for the NMP 22™ assay combined with conventional cytologic examination was 74% and the negative predictive value was 81%.33 In summary, when compared with cytology performed on voided samples by nonexpert cytopathologists, the NMP 22™ test appears to provide additional information regarding the risk assessment of patients for recurrence of carcinoma of the urinary bladder. However, given the relative high cost of the test and the necessity of mailing the specimen to a reference laboratory, it remains to be seen whether the marginal improvement in cytology can be justified by the inconvenience and cost of the procedure. Telomerase is a ribonucleoprotein enzyme with reverse transcriptase activity that contains an RNA component that provides a template for the synthesis of repeat telemeric sequences.34 Tests for telomerase activity (Fig. 1) using the polymerase chain reaction (PCR)-based telomere repeat amplification protocol (TRAP) assay have been positive in nearly all urothelial tumors.34 The published range for sensitivity for the detection of urothelial carcinomas by telomerase measurement is from 70–93% with a specificity range of 60–99% (Table 5).24, 35-39 Although the false-positive rate was high in one study,35 the telomerase assay generally has outperformed cytology in several additional studies.24, 35, 36, 38, 39 In a study of 196 patients, Ramakumar et al. found that telomerase had the highest combined sensitivity and specificity of the ancillary procedures tested and outperformed both the NMP 22™ and BTA™ assays, as well as conventional cytology.24 Muller et al.35 Yoshida et al.36 Lance et al.37 Ito et al.38 Lee et al.39 Ramakumar et al.24 In summary, although of significant potential, the routine use of telomerase assays in the management of patients with recurrent urothelial neoplasia currently is limited by the cost and cumbersome nature of the bioassay, which requires intact tissue and use of the PCR-based TRAP procedure. The development of an on-slide method, utilizing either in situ hybridization or immunohistochemistry, appears necessary to create an easy ideally automated method that can be used on a routine basis for clinical material. Microsatellites are inherited short tandem repeat DNA sequences with low mutation rates unique to individuals.40 Assays for abnormalities in microsatellites generally have included the detection of mutational copy errors and deletions of gene loci seen as loss of heterozygosity (LOH; Fig. 2).2 In addition, it recently has been recognized that low grade papillary urothelial neoplasms often are associated with chromosomal instability and loss in parts of chromosome 9 in and near the p16INK 4A (MTS1) tumor suppressor gene.41 Abnormalities of chromosomes 11, 13, 3, 4, 8, 17, and 18 also have been associated with the development of urothelial malignancy. Microsatellite instability assays have been identified most frequently as LOH in the 9P region for low stage tumors.42 Additional mutations and genetic divergence appear to accumulate after alterations in primary chromosome 9 in the majority of low grade, superficial urothelial neoplasias.43 Four studies of microsatellite instability (including LOH and mutational analyses) have demonstrated the apparent highest sensitivity (range, 83–95%) and specificity (100%) for the early identification of recurrence of urothelial neoplasia in bladder washing and urinary cytology specimens (Table 6).44-47 Microsatellite analysis has been performed successfully on stored, frozen human samples of urine specimens that lacked telomerase activity,47 has achieved detection rates of twice the sensitivity of urine cytology,45 and has been capable of detecting tumor recurrence months before cystoscopic examination findings became positive. In addition, the development of an automated, low cost method of microsatellite analysis appears feasible.48 Thus, microsatellite analysis shows significant potential for clinical use to monitor early recurrence in urothelial neoplasia and, if an inexpensive automated assay can be developed in the near-future, this procedure may well become a part of the routine care of patients with tumors of the urinary bladder and upper urinary tract. Mao et al.44 Steiner et al.45 Linn et al.46 Mourah et al.47 In the early 1980s flow cytometry was introduced to detect abnormal total DNA content in urothelial cells obtained from urothelial cytology specimens.49 Early hope that DNA aneuploidy would be a major adjunct to the cytologic detection of recurrent urothelial carcinoma ultimately gave way to the reality that the technique could improve on the sensitivity and specificity of cytology only to a marginal degree.50, 51 Improvements in flow cytometric techniques, including dual parameter immunoflow cytometry, have achieved an increased sensitivity for the detection of WHO system Grade 1 transitional cell carcinomas (urothelial neoplasms of low malignant potential) to 86%.52 The introduction of DNA content measurements by digital static image analysis (Fig 3) in the late 1980s also reported an increase in the sensitivity and specificity for the detection of recurrent urothelial malignancies.53 Subsequently, it was reported that for patients with urothelial neoplasms, approximately 33% of cases with diploid urothelial cytology samples ultimately are associated with progression to muscle invasive disease.54 In addition, although the combination of conventional cytology with DNA ploidy measurement by image analysis has been reported to be capable of detecting up to 85% of cases of recurrent urothelial neoplasia, it has not appeared to be able to identify all patients with disease recurrence.55, 56 The Quanticyt™ karyometric analysis was developed in the Netherlands as an automated system for urothelial cell grading using Feulgen-stained cytospin preparations and a digital image analysis system.57 The system combines DNA content analysis with nuclear shape calculations and, according to the developers, provides significant additional prognostic information that could be used to reduce the number of cystoscopies needed to care for patients with superficial bladder carcinoma.57 A major application of DNA ploidy analysis in urothelial cytology samples has been to differentiate recurrent urothelial malignancy (diploid or aneuploid) from cytologic atypia associated with intravesical installation of chemotherapy and immunotherapy (uniformly diploid). More recently, the technique of fluorescent in situ hybridization (FISH) has been applied to urothelial cytology specimens to detect chromosomal aneusomy.58 The success of this technique is dependent on the number of chromosomal centromeric probes used to detect gains or losses59 and will have relatively low sensitivity if insufficient numbers of probes are utilized.59 In one study, routine ploidy analysis was combined with FISH-based assessment for chromosome 9 and resulted in an increased sensitivity compared with that of either technique alone.60 Chromosome 9 also has been evaluated by single strand conformational polymorphism analysis (SSCP) with LOH found to be a marker of recurrent urothelial carcinoma that was more sensitive than routine cytology.61 However, urothelial carcinomas are a heterogeneous group of neoplasms and multiple probes must be used to detect recurrent disease. Thus, the FISH and SSCP methods are too complex, time-consuming, and expensive for general use in most laboratories.62 The loss of cellular expression of the AB and O blood group antigens has long been recognized as markers of urothelial neoplasia.63, 64 More recently, antibodies designed to detect Lewis X blood group expression in urothelial cytology have been evaluated as indicators of recurrent disease. However, to our knowledge to date, blood group expression assays have not achieved significant levels of improvement in the detection rates compared with conventional cytology to warrant their widespread clinical usage. A variety of proteins, glycoproteins, and other antigens have been studied for their ability to improve on the detection rates for recurrent urothelial neoplasia.64 In one study, cytokeratin 20 staining achieved 91% sensitivity compared with 56% sensitivity for cytology alone.65 In a study of cytokeratin 19 using a quantitative assay, a sensitivity of 96% and specificity of 74% was achieved, which was more than double the 43% sensitivity for cytology alone.66 Studies of epithelial membrane antigen have yielded conflicting results in their ability to improve the detection of urothelial neoplasia.3 AM43 and BB639 are 2 monoclonal antibodies utilized in a radioimmunoassay, which, although relatively sensitive, was associated with a 20% false-positive staining rate.67 The 486 P3/12 antigen has achieved an 89% sensitivity for detecting both high and low grade tumors and thus shows significant potential as an adjunct for the detection of recurrent disease.68 The M344 and 19A211 antigens similarly show potential for use in the follow-up of patients due to their ability to detect low grade lesions.69 Psorisian is a squamous epithelial differentiation marker that has shown potential for the follow-up of patients with squamous cell carcinoma of the bladder.70 Finally, the DD23 antigen is up-regulated in approximately 80% of bladder neoplasms and may be of significant value for the early detection of recurrent bladder neoplasms.71 Recently, a panel of monoclonal antibodies (M344, LDQ10, and 19A211) comprising the Immunocyt™ test was found to be a highly sensitive detector of all grades of urothelial carcinomas that, when combined with cytologic examination, showed potential for reducing the need for cystoscopy in selected patients.72 Hyaluronidase and hyaluronic acid have been evaluated as predictors of urothelial neoplasia in urine samples. Although urinary hyaluronidase measurements have been associated with a sensitivity of 100% and a specificity of 89% for the detection of high grade bladder neoplasms, the ability to detect low grade lesions by hyaluronidase measurements has not, to our knowledge as yet been proven.73, 74 A variety of growth factors have been measured in urine samples as both adjuncts to the detection of urothelial neoplasia and as predictors of disease outcome. Acidic fibroblast growth factor (acidic FGF) is associated with proliferation, differentiation, angiogenesis, and cell motility.3 Urothelial tumors show a significant increase in staining for acidic FGF compared with normal urothelium,75 and urine samples from patients with bladder carcinoma contain significantly more acidic FGF than do control specimens.76, 77 This marker appears more suited as a prognostic test than as a detector of low grade lesions. Basic fibroblast growth factor (basic FGF) also has been evaluated as an adjunct to urothelial carcinoma detection.77 Urinary basic FGF levels are higher in patients with bladder carcinoma than control subjects and, in one study, this test was more sensitive than cytology for the detection of urothelial neoplasia.78 Concerns regarding specificity and the ability to detect low grade lesions have limited the adoption of basic FGF measurements into clinical practice. Urinary autocrine motility factor has been detected in 100% of patients with muscle invasive urothelial carcinomas and in 80% of patients with superficial lesions.79 Autocrine motility factor levels also have been found to correlate with urothelial carcinoma stage and disease recurrence.80 However, the 25% false-positive rate identified for control specimens80 has created concern that the low specificity of this test would preclude its widespread clinical use. Epidermal growth factor (EGF) concentrations are lower in patients with urothelial neoplasia than in control patients, possibly due to EGF binding to urothelial cells during carcinogenesis.81 To our knowledge EGF measurements have not been used successfully for the early detection of recurrent urothelial carcinoma. Transforming growth factor-beta (TGF-β) also has been studied in patients with urothelial neoplasia. TGF-β expression is higher in urothelial neoplasia than normal urothelium and may be a significant marker of progression of bladder carcinoma.82 The use of TGF-β measurements for the detection of recurrent bladder carcinoma in urine samples to our knowledge has not been reported to date. Telomerase detection in cytologic specimens using the telomeric repeat amplification protocol (TRAP) method. Lanes 1, 2, 3, and 4 contain fluids that were positive for malignant cells and showed a ladder of base pairs (bps) indicating the presence of telomerase activity. Lanes 5, 6, 7, and 8 are fluids that were negative for malignant cells and showed no bp ladder. Lane 9 is the control cell line known to be positive for telomerase activity and Lane 10 is the telomerase negative control fluid. Microsatellite instability assays in urine samples. The six blots shown were obtained from urinary sediments. N: normal; U1: urinary sediment prior to treatment; U2: urinary sediment obtained 3–6 months after the initiation of treatment; T: the original first tumor tissue sample obtained at cystoscopy. In Specimen B919, the arrows indicate a loss of heterozygosity (LOH) at the same locus in both the original sample and the sample obtained at the time of recurrence. In Specimens B604 and B884, the initial urinary and tumor samples show similar LOHs (arrows), but the subsequent samples (U2) show restoration of both alleles and indicate that the patient was free of disease at the time of follow-up. In Specimen B816, continued LOH is observed in the tumor, original, and follow-up urine samples from 3 separate loci that were associated with confirmed disease recurrence at 6 months. Modified with permission from Steiner et al., Nature Med 1997;3:621–4.45 DNA ploidy in urine cytology of chemotherapy-associated atypia. Although cytologic atypia of urothelial cells was present, the image analysis-based DNA ploidy histogram revealed a diploid pattern with a DNA index (D.I.) of 1.02. Although diploid histograms are associated both with therapy atypia and low grade urothelial carcinomas, aneuploidy is characteristic of high grade tumors only. E-cadherin cell adhesion molecule in urine cytology showing overexpression in papillary low grade urothelial neoplasms. Benign squamous cells surround the central, darkly stained cluster of cells obtained from a patient with a papillary low grade urothelial carcinoma. E-cadherin staining is increased in the cells from a low grade papillary tumor, but progressively down-regulated and completely lost in flat high grade tumors with progressive invasion and subsequent metastasis (avidin-biotin immunoperoxidase, ×400). The E-cadherin/catenin system has been studied extensively in patients with bladder carcinoma. Loss of expression of E-cadherin has been associated with high histologic grade, advanced pathologic stage, and adverse outcome in patients with urothelial neoplasia (Fig. 4).83, 84 Elevated urinary E-cadherin measurements have been associated with the presence of papillary urothelial neoplasms that have featured up-regulation of E-cadherin (which subsequently is down-regulated in patients in whom the lower grade papillary tumors dedifferentiate, become flat lesions, and invade the bladder wall.85 Widespread measurements of soluble E-cadherin by liquid immunoassay and cellular E-cadherin by immunocytochemistry to our knowledge have not been studied widely for the detection of recurrent urothelial neoplasia. The CD44 cell adhesion molecule also has been studied in bladder carcinoma. Similarly, loss of expression of the standard form of this molecule by immunohistochemistry has been associated with high grade, high stage disease.86 However, attempts to use CD44 measurement in lysates of urine sediments were not successful in adding significant information for the detection of urothelial carcinoma.87 Finally, integrins have been studied in urine samples and also have shown a relation between loss of cell adhesion expression and disease invasiveness.88 The shedding of fibrinogen degradation products (FDPs), reflecting focal disseminated intravascular coagulation associated with bladder neoplasia, has been used as a method of detection of the disease.3 The Aura Tek FDP Test™ (Perimmune, Inc., Rockville, MD) uses an enzyme-linked immunoadsorbent assay (ELISA) method to detect FDPs in urine. The sensitivity and specificity of this method have been reported to be 81% and 75%, respectively.89 Similar to other markers, the ability of this method to detect well differentiated, early stage tumors has not been established. However, in a comparison of the Aura Tek FDP Test™ with the BTA test, urine hemoglobin detection, and urothelial cytology, the Aura Tek FDP Test™ had the highest sensitivity for low grade lesions.89 High levels of the urokinase type I plasmin activator (uPA), a protease associated with tumor invasion, have been detected in bladder tumor tissues.2 Although elevated uPA levels have been associated with adverse outcome in bladder carcinoma independent of other prognostic factors,90 to our knowledge the use of uPA measurements for the early detection of recurrent urothelial neoplasia has not been reported. Glutathione-S-transferase π-1 has been tested in urine as a marker for bladder carcinoma using an ELISA method. The sensitivity was relatively low and the potential of this marker to detect low grade, early recurrent lesions to our knowledge remains unproven.91 Genes and proteins regulating the transition from the G1 to the S-phase of the cell cycle frequently have been implicated both in the development and detection of urothelial neoplasia.92 LOH in the 9p21 region has confirmed that the p16INK4A tumor suppressor gene is associated with the development of papillary and low grade urothelial neoplasms.93 Large clinical trials evaluating LOH at chromosome 9p to detect abnormalities of the p16 gene (point deletions) or loss of p16 protein production due to p16 promoter gene hypermethylation currently are in progress.92 As discussed earlier, microsatellite instability assays also are focused in the 9p region and show substantial potential for the early detection of recurrent urothelial neoplasia. In addition, chromosomal abnormalities in the 9q region also have been associated with the development of bladder carcinoma and may generate a new series of molecular detection strategies.93 Cell cycle regulators are well established as prognostic factors for bladder carcinoma,94 but to our knowledge their use as early detection markers is, to date, unproven. The p53 tumor suppressor gene has been studied extensively as a cell cycle inhibitor at the G1 to S checkpoint, an apoptosis regulator, and target for cancer therapy.95 Mutations of the p53 gene and the resulting production of mutant p53 protein have been associated with nonpapillary, deeply invasive, high stage disease.95-97 However, attempts to use p53 immunostaining on urine cytology samples to enhance the detection of bladder carcinoma have shown relatively low rates of sensitivity and specificity.98 Continued studies of p53 mutations for the management of bladder carcinoma appear likely to continue, but the use of this marker to aid in the detection of low grade tumors appears unlikely to achieve clinical success. Mutations of the retinoblastoma tumor suppressor gene (Rb) have been studied in bladder carcinoma and identified in both superficial and invasive urothelial lesions.99 However, to our knowledge the use of Rb protein detection or gene mutation analysis in urine samples to detect recurrent bladder carcinoma has not been reported. Cell proliferation markers have been studied as prognostic factors in bladder carcinoma, including the Ki-67 antigen and proliferating cell nuclear antigen.100 To our knowledge the use of cell proliferation for the detection of recurrent urothelial neoplasia has not been performed on a wide scale. Ki-67 labeling can supplement the routine use of histologic grade and stage in predicting bladder carcinoma outcome, but it appears unlikely to prove useful for the detection of early disease recurrence.100 Recently, a series of ancillary tests have been introduced with claims that they can improve the sensitivity and specificity of routine urinary cytology, facilitate the earlier detection of urothelial carcinoma recurrence, and reduce the number of cystectomies needed to monitor patients with this disease. Although the performance features of the ancillary tests have varied and the cytologic methods to which they have been compared have not been standardized, several of the procedures have received considerable support from urologists as assisting in patient management. Over the next several years, the ancillary tests will continue to be measured prospectively against cytology alone in clinical trials designed to determine whether the additional costs of the new procedures can be overcome by the added benefits of a reduction in costly cystectomies and the earlier specific detection of disease recurrence leading to earlier therapy and reduced morbidity and mortality.