Rate of change in serum prostate specific antigen levels as a method for prostate cancer detection.

To examine prospectively the usefulness of measurement of rate of change in serum prostate specific antigen levels (PSA slope) in detecting prostate cancer in a PSA-based prostate cancer screening study, we evaluated 982 serially screened men whose initial screening was negative for cancer. All men had at least 1 PSA value greater than 4.0 ng./ml. and all ultimately underwent prostatic biopsy. For those who entered the study with normal PSA levels, a PSA slope cutoff point of 0.75 ng./ml. per year or more maximized sensitivity and specificity for predicting cancer (odds ratio 7.20, 95% confidence interval 4.52 to 11.47). This cutoff point was most predictive for men 70 years old or younger. For men who entered the study with elevated PSA levels (greater than 4.0 ng./ml.) a lower PSA slope cutoff point (0.4 ng./ml. per year or more) maximized sensitivity and specificity for predicting cancer (odds ratio 2.73, 95% confidence interval 1.82 to 4.07). We conclude that PSA slope is useful for serial prostate cancer screening, although its predictive value varies with patient age and initial PSA level.