Utility of PSA doubling time (PSADT) in the detection of prostate cancer in men with a serum PSA ≤ 10 ng/ml

Abstract

4645 Background: PSADT is a strong predictor of adverse outcomes in men with recurrent prostate cancer. Its utility has not been examined, however, for the most common diagnostic challenge: a man referred for suspected prostate cancer and a modest PSA elevation (PSA <=10 ng/ml). The optimal diagnostic approach to such patients is not known and multivariable models have been constructed to facilitate the selection of patients who require a prostate biopsy. Diurnal variation of PSA is common in such patients (JAMA 289:2695, 2003) and may reduce the utility of PSADT. We sought to determine if PSADT is an independent predictor of the presence of prostate cancer in men with a PSA < 10 ng/ml. Methods: Data were collected on 1,699 veterans seen at the Portland VA with a serum PSA of ≤ 10 ng/ml who underwent an initial prostate biopsy. Variables analyzed included: age, race, family history, digital rectal exam (DRE), prostate-specific antigen (PSA), PSA density (PSAD), PSA doubling time (PSADT), prostate volume, and ultrasound (US) findings. Data were analyzed using both logistic regression (LR) and Classification and Regression Tree (CART) analysis. Model building was carried out on 70% of the data and study validation was done on the remaining 30% of the randomly selected dataset. Results: PSADT was < 2 yrs in 256 patients (15.2%), 2 to 5 yrs in 297 patients (17.7%) and ≥ 5 yrs in 1128 patients (67.1%). Prostate cancer was histologically detected in 448 (26.4%) patients. Prostate cancer diagnosis was associated with age, PSA > 2.9 ng/ml, PSAD > 0.12 ng/ml/cc, positive DRE, and an US-detected lesion by LR (p < .05). PSADT of 2 to 5 years was weakly associated with the diagnosis of prostate cancer (RR 1.6, 95% CI 1.1- 2.3) while a PSADT < 2 years or > 5 years were not predictive. The LR model predicted the presence of prostate cancer with an area under the ROC curve of 0.76 in the validation set. Similarly, CART failed to select PSADT in the formation of a decision tree. Using the validation set, CART sensitivity was 96.2% and specificity was 15.8%. Conclusion: While useful in several clinical settings once the diagnosis of prostate cancer is established, PSADT is of limited utility in the initial evaluation of a patient suspected of harboring prostate cancer. No significant financial relationships to disclose.