COMPARISON OF THE MONOCLONAL UBC-ELISA TEST AND THE NMP22 ELISA TEST FOR THE DETECTION OF UROTHELIAL CELL CARCINOMA OF THE BLADDER CHRISTINE MIAN, MICHELE LODDE, ANDREA HAITEL, EDUARD EGARTER VIGL, MICHAEL MARBERGER, AND ARMIN PYCHA ABSTRACT Objectives. To compare the diagnostic value of two enzyme-linked immunosorbent assay (ELISA) tests, the nuclear matrix protein 22 (NMP22) test and a newly developed urinary bladder cancer (UBC) test, in patients having symptoms suggestive of urothelial cell carcinoma (UCC) and patients under follow-up after transure- thral resection (TUR). Methods. Two hundred forty patients with a mean age of 65.8 years (range 22 to 92) were included in this retrospective study. The tests were performed on previously frozen urine samples. Eighty-one patients had symptoms suggestive of bladder cancer and 159 patients were being followed up after complete TUR of UCC. Voided urine was evaluated by the NMP22 test and the monoclonal UBC-ELISA test, which traces cytokeratins 8 and 18. All patients underwent subsequent cystoscopy and biopsy evaluation of any suspi- cious lesion. The cutoff levels for bladder cancer positivity were 10 U/mL for the NMP22 test and 12 g/L for the UBC test. Results. In the 54 patients with histologically proved UCC, the NMP22 test had a sensitivity of 55.5% and the UBC test a sensitivity of 64.8%. According to the histologic stages, the sensitivity of NMP22 was 51.7% in pTa tumors, 46.1% in pT1, and 70% in pT2 or higher tumors; the sensitivity of UBC was 62.1% in pTa, 53.8% in pT1, and 80% in pT2 or higher tumors. For histologic grades 1 to 3, the sensitivity was 50%, 50%, and 68.7% for NMP22 and 66.6%, 60%, and 68.7% for UBC, respectively. The specificity was 79% and 92% for NMP22 and UBC, respectively. Conclusions. The monoclonal UBC-ELISA test is superior to the NMP22 test in both sensitivity and speci- ficity. Nevertheless, neither test can replace cystoscopy. UROLOGY 55: 223–226, 2000. © 2000, Elsevier Science Inc. C ystoscopy is the reference standard for the identification of primary or recurrent urothe- lial cell carcinoma (UCC) of the bladder, but it is invasive and causes significant discomfort to the patient. Furthermore, flat tumors or carcinoma in situ may be difficult to detect. 1 Urinary cytology is noninvasive and very effective in diagnosing high- grade lesions, but it has a sensitivity of 11% to 17% in grade 1 tumors, which are the most common type of UCC. 1–4 The limitations of cytology and cystoscopy, both for primary diagnosis and monitoring of patients after UCC has been removed, led to the develop- ment of new urine-bound tests for the early detec- tion of UCC. 1–7 Methods based on the immuno- logic detection of soluble antibodies in voided urine, such as the BTA stat test and the nuclear matrix protein 22 (NMP22) test, are of particular interest. These tests are more sensitive than cytol- ogy in low-grade tumors, but their specificity is so low that cystoscopy is still essential. Moreover, in grade 3 tumors, their sensitivity is about 25% lower than that of routine cytology, so that the latter is again not rendered unnecessary. 1 However, the de- velopment of effective urine-bound tests is of on- going interest. The aim of the present study was to compare the diagnostic value of two enzyme-linked immu- From the Department of Clinical Pathology, General Hospital of Bolzano, Bolzano, Italy; and Departments of Urology and Clini- cal Pathology, University of Vienna, Vienna, Austria Reprint requests: Armin Pycha, M.D., Department of Urology, University of Vienna, Wa ¨ hringer Gu ¨ rtel 18-20, 1090 Vienna, Austria Submitted: April 14, 1999, accepted (with revisions): August 4, 1999 ADULT UROLOGY © 2000, ELSEVIER SCIENCE INC. 0090-4295/00/$20.00 ALL RIGHTS RESERVED PII S0090-4295(99)00383-0 223