The importance of serum prostate-specific antigen testing frequency in assessing biochemical and clinical failure after prostate cancer treatment.

OBJECTIVES

To assess the relationship between prostate-specific antigen (PSA) testing frequency and biochemical failure (bF) and clinical failure (cF).

METHODS

The records of 5616 patients with low-, intermediate-, or high-risk prostate cancer treated (brachytherapy, external beam radiotherapy, or surgery) between 1996 and 2007 were reviewed. Factors influencing bF and cF were recorded including age, initial PSA, androgen deprivation, race, clinical stage, biopsy Gleason score, and the frequency of follow-up PSA testing. Univariate and multivariate analyses were performed to assess the effect of these factors on bF and cF. Sensitivity and specificity were calculated to determine the optimal frequency of PSA testing.

RESULTS

The median follow-up is 45 months. The median number of PSA tests per year before the occurrence of bF and cF is 1.9 for both endpoints. The multivariate analysis of factors significantly associated with bF and cF demonstrate that PSA frequency, initial PSA, clinical stage, and biopsy Gleason score are independently predictive of outcome. PSA testing achieves the best sensitivity and specificity when 2 PSA tests are drawn per year for both bF (sensitivity = 66.3%, specificity = 58.0%) and cF (sensitivity = 75.1%, specificity = 60.3%).

CONCLUSIONS

The frequency of PSA testing is strongly associated with the detection of bF and cF. Because it is a variable that can be controlled, PSA testing frequency should be standardized to minimize spurious conclusions from studies with bF and cF endpoints. The sensitivity and specificity can be optimized by obtaining 2 PSA tests per year.