New Urinary Markers for Screening Bladder Cancer
Craig D. Zippe, M.D. and Ashok Agarwal, Ph.D.

Hematuria either gross or microscopic, is the most common sign of bladder cancer and traditionally has warranted a thorough investigation of the entire urinary tract. A complete workup has included an upper tract study, an office systoscopy and a urinary cytology. The yield of detecting a urothelial malignancy from this workup in our clinic, however, has been less than 10% with microscopic hematuria, and roughly 30% in the cases of gross hematuria. Historically, less than 10% of office cystoscopies detect bladder cancer, and the sensitivity of urinary cytology is on the average less than 50% and as low as 20% in low-stage, low-grade disease. A voided urine cytology costs nearly $100, which compromises its widespread use as a screening test.

For the past two years, we have been using several new urinary tumor markers, which were introduced to monitor patients for recurrent bladder cancer or to screen patients at risk for bladder cancer. We have prospectively examined three different urinary tumor markers on all patients presenting with hematuria or who were at risk for having bladder cancer. As an investigational study,we have studied the role of a quantitative test, NMP22 (Matritech) and a qualitative test, BTA Stat (Bard), in screening patients for bladder cancer. All patients at risk were screened with an upper tract study, a urinary cytology and an office cystoscopy. All positive cystoscopic findings were confirmed histologically with a biopsy. Urinary cytologies were recorded as positive, negative or atypical. For, calculation of sensitivity and specificity, all atypical cytologies were considered as negative. An NMP22 value of more than 10 U/ml. was considered positive for this screening population. A positive Bard BTA Stat test was scored visually in the office as per manufacturer's standards.

Study I: Role of NMP22 as a Screening Test
This screening study compared the use of NMP22 wilh urinary cytology and office cystoscopy in patients at risk for bladder cancer. Of the 330 screened patients, a total of 18 bladder cancers were detected. When using an NMP cutoff value of more than l0 U/ml., we detected all 18 cancers, with a sensitivity of 100%. In contrast urinary cytology detected only six cancers, for a sensitivity of 33%. The receiver operating curve (ROG) analysis of the data showed that a cutoff value of 10 U/ml., as compared to 6 U/ml., provided optimum sensitivity and specificity. The negative predictive value (NPV) for NMP22 was 100% (Table 1). The clinical value of this exceptionally high NPV is that we missed no bladder tumors in this cohort of patients when the NMP value was less than 10 U/ml.. The evolving implication of this data is that office cystoscopy may possibly be eliminated for the majority of our patients who are routinely screened for hematuria. Additional prospective studies are needed before further recommendations can be made.


The shortcoming of this study (as seen hy other investigators) is the low specificity. Clinically, this is manifested as a high false-positive rate or a low positive predictive value (PPV). In the current study, the specificity and PPV were 86% and 29% respectively. As we have gained familiarity with the NMP22 test, we have analyzed our false-positive data retrospectively and found that more than 90% of the false-positive results fall into five main categories : 1) any inflammatory or infectious condition such as recurrent cystitis or prostatitis (more than 5WBC/HPF), 2) presence of renal or bladder calculi, 3) any current or recent history of a foreign body in the urinary tract (double J stent, percutaneous nephrostomy tube), 4) any bowel interposition (ileal conduit. continent diversion), and 5) any concomitant urinary tract cancer (prostate cancer, renal cell Ca) (Table 2). Exclusion of "known causes" of false-positive results greatly increases the specificity and PPV to values nearly identical to cytology. We intend to validate these known causes in future prospective studies.


Study 2: Comparison of NMP22 and BARD BTA Stat as a Screening Test and in Monitoring Patients with Known Bladder Cancer
In this study, we screened 199 patients (incident cases) with NMP22 and the Bard BTA Stat test. The advantage of the BTA Stat test is that it is an office-based, one-step visual test and can be performed on individual samples by ancillary personnel. Conversely, the NMP22 requires at least 10 samples per run to be cost-effective, but it has the advantage of being quantitative. The clinical utility of having a quantitative test is still unclear. Although the quantitative value correlates with advanced stage and grade, its clinical value in screening is still being investigated. In this cohort of patients, we recently missed our first two cancers in a screening population using a cutoff value of 10 U/mL for NMP22 (Table 3). The NMP22 detected four of the six cancers, with the two missed cancers being solitary Ta, Grade 1 to 2 cancers. The Bard BTA Stat test results were similar as it detected four of the six cancers, missing the same two low-grade, low-stage cancers as NMP22.

In the cohort of patients with known bladder cancers (prevalent cases), the NMP22 detected a larger number of recurrences (24/28) than the BTA Stat (19/28) (P = 0.414) (Table 3). Cytology only detected nine of 28 (32%) cancers. In the prevalent cases, we used a cutoff NMP22 value of 6.0 U/mL to determine a positive result. This cutoff value has been reported previously by other investigators. Our results for sensitivity and specificity were similar to others in detecting recurrent cancers. NMP22 in our study showed a trend toward better sensitivity as compared to BTA Stat in detecting recurrent cases of bladder cancers, however, the differences were not significant. Comparision of the "known causes" of false-positive results between the BTA Stat and the NMP22 showed nearly identical results. Contrary to previous reports on the BTA Stat, we found no evidence that microscopic or gross hematuria by itself invalidated the test.

The role of these new markers in screening for bladder cancer is an exciting and potentially significant contribution. We all can agree that earlier detection of bladder cancer (at any stage} in any one individual patient will yield more promising results from our local therapies. Unlike annual prostate cancer screening, a significant difference in outcome may occur with detection of localized bladder cancer a year earlier. Similar to the evolution of PSA as a screening marker, greater utilization of these new urinary tumor markers will continue as we understand how they work.



Menu Page  |   CRM Home  |   Cleveland Clinic Home   |  Contact Us   |  Privacy Statement   |   Disclaimer   |   Cleveland Clinic 2008
Last Update : December 27, 2008
EmailTo A Friend
American Center for Reproductive Medicine
The Cleveland Clinic
10681 Carnegie Avenue, X11
Cleveland, OH 44195 USA
Tel: (216) 444-9485
Fax: (216) 445-6049