Contribution of the prostate limits the usefulness of survivin for the detection of bladder cancer.

Abstract

We determined if urinary survivin is influenced by the prostate and analyzed survivin levels in a healthy control population. This was a blinded, retrospective analysis of frozen urine samples. Three groups were designated. Group 1 patients were diagnosed with prostate cancer and samples were collected immediately prior to radical retropubic prostatectomy. Group 2 patients had already undergone radical retropubic prostatectomy at least 3 months prior to providing a urine sample. Group 3 comprised healthy controls. Groups 1 and 2 patients were recruited at an institutional practice. Group 3 patients were recruited from the local community. The frequency of high survivin scores was much lower in patients who did not have a prostate (p < 0.0002). The survivin score was 2 in 35% and 44% of the patients in groups 1 and 3, respectively, while only 9% of those in group 2 had a survivin score of 2 (p < 0.0002). Of healthy controls 50% had a false-positive survivin score. Urinary survivin predicted prostate cancer poorly with 52% sensitivity and 50% specificity, and it did not predict any biological features of prostate cancer. Urinary survivin markedly decreases after prostatectomy. There is a 50% false-positive rate when using urinary survivin in controls. These 2 features make urinary survivin a poor bladder cancer biomarker.