The additional value of free prostate specific antigen to the battery of age-dependent prostate-specific antigen, prostate-specific antigen density and velocity.

Abstract

This study describes the value of using the fraction of free prostate-specific antigen as a further marker in the early detection of prostate cancer. This newly introduced marker is compared to the usual battery of age-dependent total prostate-specific antigen, prostate-specific antigen density (microg/l x g tissue) and prostate-specific antigen velocity (microg/l x year). Determination of total prostate-specific antigen and free prostate-specific antigen was performed on fresh serum samples obtained from 3470 symptomatic patients aged 45-80 attending the Urology Clinics, or their General Practitioners. Among them, 310 patients had total prostate-specific antigen above the age-dependent cut-off, and/or free/total prostate-specific antigen under 11%, with different prostate-specific antigen densities and velocities. Only 147 patients complied to undergo biopsy: in 72 of those patients, benign prostatic disease was histologically confirmed, while in 75 patients primary prostate cancer was histologically confirmed. Total and free prostate-specific antigen levels were determined using the third generation DPCs prostate-specific antigen assay performed on the Immulite automated immunoassay instrument. Total prostate-specific antigen age reference values were adopted from Oesterling et al. (J Am Med Ass 1993; 270:860-4); the prostate-specific antigen density was considered suspicious of prostate cancer if it was greater than 0.15 microg/l prostate-specific antigen per gram tissue (Seaman et al. Urol Clin N Am 1993; 20:653); prostate-specific antigen velocity greater than 0.75 microg/l x year (Carter et al., J Am Med Ass 1992; 267:215) was considered suspicious for prostate cancer. Of the 147 patients, 75 had prostate cancer and 72 had benign prostatic hypertrophy. The difference between prostate cancer and benign prostatic hypertrophy was significantly reflected only by free/total prostate-specific antigen and prostate-specific antigen velocity. These parameters also provided the best sensitivity and specificity. Only these parameters proved to be significant when using a backwards logistic regression model (prostate-specific antigen velocity, p = 0.007 odds ratio 2.782; free/total prostate-specific antigen %, p = 0.016 odds ratio 2.678). Combinations of various parameters became significant when including free/total prostate-specific antigen, increasing prostate cancer detection to 88%. We conclude that free/total prostate-specific antigen is the most significant among prostate-specific antigen quantities (total age-dependent prostate-specific antigen, prostate-specific antigen density and prostate-specific antigen velocity). Adding this parameter to other prostate-specific antigen parameters improves the discrimination between prostate cancer and benign prostatic hypertrophy for the population at risk.