The utility of prostate-specific antigen velocity thresholds in clinical practice: a population-based analysis.

Abstract

To investigate the ability of prostate-specific antigen velocity (PSAV) to predict prostate cancer, and assess the test characteristics of several PSAV thresholds for identifying prostate cancer and high-grade cancers. From a population-based database of PSA results, men with an initial PSA level of <10.0 ng/mL, taken between I January 1994 and 31 December 2003, were identified. Those with three or more PSA tests before diagnosis, taken over > or =18 months, were included. Men were followed for a diagnosis of prostate cancer or histologically confirmed benign disease until 31 December 2003. In all, 24 709 men were included, with 716 (2.9%) diagnosed with prostate cancer and 1488 (6.0%) with benign histology. The mean (10.38 vs 0.43 ng/mL/year) and median (1.47 vs 0.03 ng/mL/year) PSAV were considerably higher in men with prostate cancer than in those with no cancer (P < 0.001). There was no PSAV threshold that could reliably identify prostate cancer or high-grade cancers without requiring many men to proceed to prostate biopsy. In this population, PSAV had additional value over one PSA value in identifying men with prostate cancer. Many men with prostate cancer might have a 'normal' (<0.75 ng/mL/year) PSAV. As with total PSA level, there was no PSAV threshold that could reliably predict prostate cancer, but rather a continuum of risk of cancer associated with PSAV level.