Association of free PSA percent, total PSA, age, and gland volume in the detection of prostate cancer.

Abstract

Measurement of the free fraction of total prostate-specific antigen (fPSA%) has been proposed as a useful addition to total PSA for the detection of prostate cancer. We assessed the performance of fPSA% in differentiating men with prostate cancer from men without cancer in a group of 1,709 subjects studied in five institutions. On the basis of PSA testing, digital rectal examination, and ultrasound examination conducted at one or more visits, 229 cancers were diagnosed. The mean fPSA% in men with cancer was 9.1% compared to 18.9% in men without cancer. The fPSA% varied by age, with men under 60 having a mean fPSA of 13.9% compared to 17.5% in men 60-69 years old and 19.8% in men over age 70. Significant associations of fPSA% with gland volume and PSA level were also observed. The sensitivity, specificity, and positive predictive value of different fPSA% cutoff levels were assessed in 513 men who underwent sextant biopsy. Sensitivity of 85.4%, 32.1% specificity, and a 39.2% positive predictive value were observed using an fPSA cutoff of 15% in men with PSA in the 4.0-9.9 ng/ml range. Sensitivity of 96.9%, 12.3% specificity, and a 36.2% positive predictive value were observed using an fPSA cutoff of 20% in the same men. If 15% fPSA had been used as a biopsy criterion in men with PSA of 4.0-9.9 ng/ml, the number of biopsies performed could have been reduced by 21.2%, with a concomitant reduction in cancer detection of 14.6%. Using a 20% fPSA criterion, biopsies would have been reduced 12.7%, with a 3.1% reduction in cancer detection. These results provide some evidence that the detection of prostate cancer is enhanced by measuring fPSA% in addition to the established measure of total PSA level. Further research is needed to identify other markers that have better sensitivity and specificity.