Objectives: Prostate-specific antigen (PSA) kinetics have failed to predict for the presence of prostate cancer in screening populations in which many patients harbor subclinical disease. We hypothesized that the prebiopsy PSA doubling time (PSADT) and PSA velocity (PSAV) could predict for cancer detection in a referral population with a suspicion of prostate cancer. Methods: Data were collected from 1699 consecutive veterans with a PSA level of 10 ng/mL or less who underwent prostate biopsy. Logistic regression analysis was performed on the following: age, race, family history, digital rectal examination findings, PSA, PSA density, PSADT, PSAV, prostate volume, and ultrasound lesions. Model building was accomplished with 70% of the data, and validation was done using the remaining 30%. These data were also analyzed using classification and regression tree analysis. Results: Using logistic regression analysis (P <0.05) on the model building set, prostate cancer was associated with age (older than 70 years), PSA level (greater than 2.9 ng/mL), PSA density (more than 0.12 ng/mL/cm3), digital rectal examination findings, and the presence of a lesion on ultrasonography. A PSADT of 2 to 5 years was marginally associated with prostate cancer detection (odds ratio 1.6, 95% confidence interval 1.1 to 2.3), and a PSADT of less than 2 years or longer than 5 years and PSAV were not predictive. On classification and regression tree analysis, PSADT was not selected as a predictive factor. Furthermore, neither PSADT nor PSAV was predictive of Gleason score 7 or worse cancer. Conclusions: In contrast to its prognostic value after the diagnosis of prostate cancer has been established, PSA kinetics offer little to clinical decision making as predictors of cancer or high-grade cancer in men with a PSA level of 10 ng/mL or less.
ASJC Scopus subject areas