PSA kinetics after prostate brachytherapy: PSA bounce phenomenon and its implications for PSA doubling time

PURPOSE: To analyze prostate-specific antigen (PSA) kinetics in patients treated with prostate brachytherapy (PI) with a minimum of 5 years of PSA follow-up. METHODS AND MATERIALS: The records of 162 patients treated with PI for localized prostate cancer with a minimum of 5 years of PSA follow-up were reviewed. A variety of pretreatment and posttreatment variables were examined. Patients were coded as having a PSA bounce if their PSA achieved a nadir, elevated at least 0.2 ng/mL greater than that nadir, and decreased to, or below, the initial nadir. Two definitions of biochemical failure (bF) or biochemical relapse-free survival (bRFS) were used: the classic American Society for Therapeutic Radiology and Oncology consensus definition of three consecutive rises (bF3) and the nadir plus 2 ng/mL definition (bFn+2). Associations between a PSA bounce and the various pre- and posttreatment factors were assessed with logistic regression analysis, and the association between a PSA bounce and bF was examined with the log-rank test. The Mann-Whitney U test was applied to test for differences in the PSA doubling time (PSADT) and the time to a PSA rise between the PSA bounce patients and the bF patients. PSADT was calculated from the nadir to the time of the first PSA rise, because this point is known first in the clinical setting. RESULTS: The 5-year overall bRFS rate was 87% for the bF3 definition and 96% for the bFn+2 definition. A PSA bounce was experienced by 75 patients (46.3%). Patients who experienced a PSA bounce were less likely to have a bF, regardless of the bRFS definition used (bF3: p=0.0015; bFn+2: p=0.0040). Among the pre- and posttreatment factors, only younger age predicted for a PSA bounce on multivariate analysis (p=0.0018). The use of androgen deprivation had no effect on PSA bounce. No difference was found in the PSADT between patients who had a PSA bounce and those with bF. The median PSADT for those with a PSA bounce was 8.3 months vs. 10.3 months using the bF3 definition and 8.8 months using the bFn+2 definition. However, a significant difference was found in the time to the first rise in PSA after PI for patients with a PSA bounce vs. patients with bF. The median time to the first rise in PSA after nadir for those with a PSA bounce was 15.1 months vs. 30.0 months using the bF3 definition (p=0.001) and 22.3 months using the bFn+2 definition (p=0.013). CONCLUSION: Patients experiencing a PSA bounce are more likely to be younger and will have a better bRFS. The PSADT cannot differentiate a PSA bounce from bF. The time to the initial PSA rise after nadir is an excellent discriminator of bF from PSA bounce. The time of the PSA rise after nadir occurs far sooner for a PSA bounce than for bF. This factor should be considered when assessing a patient with a rising PSA level after PI before a patient is administered salvage therapy.