URINARY LEVEL OF NUCLEAR MATRIX PROTEIN 22 IN THE DIAGNOSIS OF BLADDER CANCER: : EXPERIENCE WITH 130 PATIENTS WITH BIOPSY CONFIRMED TUMOR

Abstract

PURPOSE: We prospectively evaluated the value of nuclear matrix protein 22 (NMP22dagger) and cytology in the diagnosis of bladder cancer. MATERIALS AND METHODS: We analyzed NMP22 in voided urine from 235 patients before cystoscopy. Of the patients 130 had transitional cell carcinoma of the bladder and subsequently underwent surgery. In a subset of 200 patients bladder washout samples for cytology were collected during cystoscopy. The cutoff for NMP22 was 10.0 units per ml. For cytology only high grade atypia was considered positive. RESULTS: Histology showed 77 superficial (pTa, pTis) and 53 invasive (pT1 or greater) tumors. Sensitivity of NMP22 was 51% and specificity was 83%. NMP22 sensitivity was 36% for superficial tumors and 73% for invasive transitional cell carcinoma. Overall sensitivity of cytology was 52% and specificity was 89%. Cytology sensitivity was 38% for superficial tumors and 83% for invasive transitional cell carcinoma. NMP22 sensitivity for grades 1, 2 and 3 tumors was 30%, 56% and 68%, respectively. Cytology sensitivity for grades 1, 2 and 3 tumors was 30%, 50% and 91%, respectively. Combined NMP22 and cytology had a sensitivity of 70%. CONCLUSIONS: NMP22 has sensitivity and specificity similar to those of cytology from bladder washout samples. Particularly in low stage and low grade tumors both tests show the same disappointing sensitivity. Because of a false-negative rate of 49%, NMP22 cannot replace cystoscopy in clinical practice, as the danger of missing NMP22 negative tumors is too high to rely on its results in an individual patient.