Background: Because cystoscopy is invasive and expensive and urine cytology has low sensitivity, alternative methods for detecting bladder cancer are sought. Minichromosome maintenance (Mcm) proteins have been used as diagnostic markers for cervical cancer. We investigated whether one Mcm protein, Mcm5, can be used to detect urothelial cancer cells in urine sediments.
Methods: We used two monoclonal antibodies against His-tagged human Mcm5 (amino acids 367-582) in an immunofluorometric assay to measure Mcm5 levels in cells in the urine of 353 patients who presented with hematuria or lower urinary tract symptoms or who were undergoing follow-up cystoscopy for urothelial neoplasia. Urine samples were also subjected to routine cytologic analysis. Patients underwent upper urinary tract imaging and cystoscopy within 12 hours of producing the urine sample. Data were analyzed by comparing areas under a nonparametric receiver operating characteristics (ROC) curve and by McNemar's test and Fisher's exact test. All statistical tests were two-sided.
Results: At the assay cut point where the false-negative and false-positive rates were the same, the Mcm5 test detected primary and recurrent bladder cancers with 87% (95% confidence interval [CI] = 77% to 94%) sensitivity and 87% (95% CI = 83% to 91%) specificity. At the cut point where the specificities of urine cytology and the Mcm5 test were equal (97%, 95% CI = 95% to 99%), the Mcm5 test was statistically significantly (P<.001) more sensitive than urine cytology, 73% (95% CI = 61% to 83%) versus 48% (95% CI = 35% to 60%). At the lower detection limit of the Mcm5 test, sensitivity was highest, 92% (95% CI = 83% to 97%) and specificity was 78% (95% CI = 72% to 83%). Patients with prostate cancer had higher levels of Mcm5 in their urine sediments than did men without malignancy (P<.001).
Conclusions: Elevated levels of Mcm5 in urine sediments are highly predictive of bladder cancer.