A study comparing various noninvasive methods of detecting bladder cancer in urine

To compare the nuclear matrix protein (NMP)-22 assay, bladder tumour specific antigen (BTAstat) test, telomerase activity (using the telomeric repeat amplification protocol assay, TRAP) and a haemoglobin dipstick test for their ability to replace voided urine cytology (VUC) for detecting bladder cancer.

The study included 120 urological patients prospectively recruited and assessed before surgery. A single freshly voided urine sample (approximate 100 mL) was collected from each patient and aliquoted for each test. All assays were conducted according to the manufactures' guidelines; 79 patients were tested for telomerase activity. The results were then compared with VUC and the diagnosis confirmed by cystoscopy and histology.

Fifty-two patients had histologically confirmed transitional cell carcinoma. The overall sensitivity for BTAstat, NMP22, telomerase, VUC and dipstick testing was 63%, 81%, 84%, 48% and 50%, respectively. Combining the results for telomerase and NMP22 gave a sensitivity of 100%. For G1 tumours the respective sensitivities were 23%, 62%, 56%, 23% and 15%, for G2 tumours, 68%, 86%, 92%, 50% and 41% and for G3 tumours 88%, 88%, 100%, 71% and 82%. For pTa tumours the respective detection rates were 48%, 70%, 84%, 39% and 30%, for pT1 tumours 80%, 90%, 90%, 50% and 50%, for pT2/pTis tumours, 100/100%, 100/100%, 100/100%, 88/100% and 88/83%. The overall specificity for the respective tests was 82%, 87%, 93%, 87% and 54%; combining the results of NMP22 and telomerase activity increased the specificity to 96%.

There was significantly better detection than VUC when using the NMP22 and TRAP assay, especially for well-differentiated (P < 0.001 and 0.0027, respectively) and superficial tumours (P < 0.001 and 0.034, respectively). Combining the results of NMP22 and telomerase activity yielded values comparable with cystoscopy.