From the Andrology-Oncology Research Laboratory, and Department of Urology, Biostatistics and Epidemiology, The Cleveland Clinic Foundation, Cleveland Ohio

ABSTRACT

Purpose: The limitation of current urinary tumor markers is the low specificity and positive predictive value, which clinically manifests as a high false-positive rate. We analyzed the false-positive data of 2 urinary tumor markers, NMP22 and the BTA stat tests. We examined the clinical categories of the false-positive results, established relative exclusion criteria, and recalculated the specificity and positive predictive value after using the exclusion criteria.

Materials and Methods: A total of 278 symptomatic patients who presented to a urology clinic were asked to submit a single voided urine sample. Each sample was divided into 3 aliquots of which 1 was stabilized with the NMP22 test kit stabilizer and assayed for NMP22, 1 was tested for BTA stat and 1 was sent for cytological examination. All patients subsequently underwent office cystoscopy and bladder biopsy if indicated.

Results: Of the 278 symptomatic patients, 112 presented with microscopic hematuria, 77 gross hematuria and 89 chronic symptoms of urinary frequency or dysuria. Of 34 cases (12%) of histologically confirmed bladder cancer NMP22 detected 28 (82.4%), BTA stat 23 (67.7%) and cytology only 10 (29.4%). When atypical cytologies were considered positive, cytology then detected 19 cases (55.9%). Elevated NMP22 values were positive in 28 cases and false-positive in 44 for a specificity of 82% and a positive predictive value of 38.9%. Similarly, BTA stat test was positive in 23 cases and false-positive in 43 for a specificity of 82.4% and a positive predictive value of 34.9%. When atypical cytologies were considered positive, the specificity and positive predictive value were 93% and 55.9%. Greater than 80% of the false-positive results were clinically categorized as benign inflammatory or infectious conditions, renal or bladder calculi, genitourinary cancer or an instrumented urinary sample, A category of "no known pathology" was included in analysis as a control. History of ureteral stents or any bowel interposition segment has a 100% false-positive rate. Exclusion of all 6 clinical categories improved the specificity and positive predictive value of NMP22 (95.6%, 87.5%) and BTA stat (91.5%, 69.7%), and was similar to urinary cytology.

Conclusions: Awareness and exclusion of the categories of false-positive results can increase the specificity and enhance the clinical usefulness of NMP22 and BTA stat results. Similarly, treating an atypical cytology as positive can enhance the sensitivity and usefulness of urinary cytology.

KEY WORDS: bladder neoplasms; carcinoma, transitional cell; tumor markers, biological; nuclear matrix; antigens, neoplasm

Accepted for publication January 29, 1999.

* Matritech, Cambridge, Massachusetts

‡ Requests for reprints: Department of Urology, Cleveland Clinic Foundation, 9500 Euclid

   Ave., Cleveland, Ohio, 44195

† Bard Diagnostics, Redmond, Washington

§ Financial interest and/or other relationship with Matritech and Pfizer