Objectives. This prospective study was undertaken to evaluate the diagnostic efficacy of the BTAstat test and nuclear matrix protein (NMP22) compared with voided urine cytology (VUC) in the detection of primary and recurrent bladder cancer.

Methods. A total of 147 patients provided a single voided urine sample for the BTAstat test, NMP22, and cytology prior to cystoscopy. Eighty-five of them had no bladder cancer history, whereas the remaining 62 were monitored for superficial bladder cancer. A group of 21 healthy age-matched volunteers were also enrolled in the study.

Results. Bladder cancer was confirmed histologically in 99 patients, of which 62 had primary tumors and 37 had recurrent ones. The overall sensitivity and specificity were 71.7% and 56.5% for the BTAstat test, 62.6% and 73.9% for NMP22, and 38.4% and 94.2% for VUC. The optimal threshold value for NMP22 calculated with receiver operating characteristics curve, was 8 U/mL. BTAstat test was significantly more sensitive than VUC in detecting bladder cancer in all stage and grade subgroups, except GIII. On the contrary, NMP22 was significantly more sensitive than VUC only in stage Ta, grade I and II patients. BTAstat test had higher but not significantly different sensitivity than NMP22.

Conclusions. Our data indicate a superiority of both BTAstat test and NMP22 over VUC in the detection of bladder cancer. Comparing BTAstat test with NMP22, the former proved to be more sensitive, whereas the latter was more specific. Ruling out diseases with potential interference can increase the overall specificity of both tests. False-positive results of either test in patients followed up for bladder cancer seem to correspond to future recurrences.