Urine markers for bladder cancer surveillance: A systematic review

Introduction: The follow-up of patients with urothelial cell carcinoma (UCC) of the bladder is done by cystoscopy and, in most cases, cytology. The last decade, many urine-based tests for UCC have been developed and tested in different populations. For the urological practice, considering the amount of follow-up cystoscopies, especially urine markers for recurrent disease would be useful. Therefore, we reviewed the literature on these markers for recurrent UCC and compared our findings with recent review-articles. Methods: We performed a PubMed search. In case of primary and recurrent disease, the study was included if the patients under surveillance were reported separately. Patients with no evidence of disease at surveillance cystoscopy were considered to determine specificity. A marker was included if at least 2 studies from 2 different institutions/ authors were available. Results: The literature review yielded 64 articles. We found 18 markers (BTAstat, BTAtrak, NMP22, FDP, ImmunoCyt, Cytometry, Quanticyt, Hb-dipstick, LewisX, FISH, Telomerase, Microsatellite, CYFRA21-1, UBC, Cytokeratin20, BTA, TPS, Cytology) that met our criteria. BTAstat, NMP22, ImmunoCyt and cytology were evaluated in more than 750 patients. Telomerase, Cytokeratin20 and Hb-dipstick were tested in less than 250 patients. The highest median sensitivities were reported for CYFRA21-1 (85%), Cytokeratin20 (85%) and Microsatellite analysis (82%). The highest specificities were reported for Cytology (94%), BTA (92%) and Microsatellite analysis (89%). In comparison with recent reviews, median sensitivity was >= 5% lower for the surveillance group in 13/18 urine-based tests while specificity remained relatively constant between different patient groups. Conclusions: To our knowledge, this is the first review that assesses sensitivity and specificity of urine markers solely for UCC surveillance. In our view, Microsatellite analysis, ImmunoCyt, NMP22, CYFRA21-1, LewisX and FISH are the most promising markers for surveillance at this time. Nevertheless, clinical evidence is insufficient to warrant the substitution of the cystoscopic follow-up scheme by any of the currently available urine marker tests. Future studies may test some of the most sensitive and specific assays to reduce the cystoscopy frequency. However, our results show that initiators of these studies should anticipate a lower sensitivity than reported in the current literature. (c) 2005 Elsevier B.V. All rights reserved.