Evaluation of nuclear matrix protein 22 as a tumour marker in the detection of transitional cell carcinoma of the bladder

Objectives  To evaluate the sensitivity and specificity of urinary nuclear matrix protein-22 (NMP22) in detecting bladder cancer, to compare the diagnostic performance of NMP22 alone and when corrected by urinary creatinine level, and to correlate NMP22 level with the histological and clinical characteristics of bladder cancer. Patients, subjects and methods  The study included 267 patients classified into five groups: group 1 comprised 111 patients with active transitional cell carcinoma (TCC) of the bladder; group 2 included 76 patients who had had bladder TCC but were being followed and were free of disease, as confirmed by cystoscopy; group 3 comprised 25 patients with benign urological diseases; group 4 included 25 patients with other malignant pathological conditions; group 5 constituted a control group of 30 healthy subjects free of urological diseases. Urinary NMP22 was measured using an enzyme-linked immunosorbent assay. Receiver operating characteristic (ROC) curves were constructed to obtain the thresholds which gave optimal sensitivity and specificity for combinations of NMP22 alone and when corrected by urinary creatinine level. Stage, grade, tumour size, pattern of growth, focality and the presence of recurrence were recorded and their associations with NMP22 evaluated. Results  The mean levels of NMP22 were 122.8, 5.1, 3.7, 2.3 and 0.3 U/mL for groups 1–5, respectively; overall, these values were significantly different (P<0.001). The mean (95% CI) optimal combination of 78.2% (69.3–85.5) sensitivity and 95.5% (87.3–99.0) specificity was obtained from the ROC analysis with a threshold value of 13.7 U/mL NMP22. When values were corrected by urinary creatinine levels, the threshold given by the best combination of sensitivity and specificity, at 73.2% (63.2–81.7) and 97.0% (89.6–99.5), respectively, was 3.0 U/mg creatinine. NMP22 level was statistically associated with stage, grade, tumour size and focality. Conclusions  Urinary NMP22 appeared to be a potential tumour marker for detecting TCC of the bladder; when corrected by urinary creatinine level, it might provide a better interpretation than when used alone. NMP22 correlated with the most relevant variables in bladder cancer. As a noninvasive adjunct, NMP22 might have a role in guiding urologists about the need for cystoscopy in such patients.