Recently, the phenomena known as the PSA bounce has been recognized as occurring in at least 30% of patients postimplant. Critz et al first reported on this event in 2000 where 273 of 779 men experienced a transient increase in PSA a median 18-months postimplant. The median bounce was 0.4 ng/mL and 92% of bounces occurred within 36 months.
The group from the Seattle Prostate Institute made similar observations later that year in 591 patients treated between 1988 and 1993. Thirty-five percent of their patients also experienced a temporary increase of at least 0.2 ng/mL. Seventy-five percent of these patients experienced a PSA rise between 0.3 and 3.4 ng/ mL. In this study, the mean time to rise was 24.8 months. Both of these studies and unpublished data from Fox Chase Cancer Center (E.M. Horwitz, personal communication, 2003) suggested that there was no clinical significance in the PSA bounce. Predictors of failure following treatment with an implant include both clinical and dosimetric parameters.
Clinical parameters include pretreatment PSA level, Gleason score, and T stage, where the ideal patient for treatment with an implant alone is one with a pretreatment PSA level of less than or equal to 10 ng/mL, Gleason score of less than or equal to 6, and stage T1c/ T2a. Multiple authors have reported that patients experience increased rates of biochemical failure when intermediate- or high-risk patients are treated with implant alone.
In addition to these known clinical parameters, several predictors of failure are based on the quality of the implant itself. Stock et al demonstrated that bNED control rates varied depending on the D90. Sixty-five patients with a D90 less than 140 Gy had 68% 4-year bNED control rates compared with 92% 4-year bNED control rates for the 69 patients with a D90 greater than or equal to 140 Gy (P = 0.02). Two-year posttreatment biopsies were negative in 70% of patients with a D90 less than 140 Gy versus 83% for patients with D90 greater than or equal to 140 Gy. Dose remained the most significant predictor of bNED control on multivariate analysis. Management of recurrence following brachytherapy Restaging and risk stratification of the patient with recurrent disease Despite improvements in clinical staging, patient selection, and implant technique, the management of disease recurrence following brachytherapy remains a clinical challenge. Considerable debate exists over the definition of failure following radiation treatment and the significance of the postbrachytherapy prostate biopsy.

In practical terms, the diagnosis of recurrence poses a number of difficult clinical issues. Clinicians must rely on clues supporting tumor recurrence—such as pretreatment PSA level, Gleason score, and clinical stage—and the definitions (and limitations) of biochemical failure, and must be aware of the ‘‘PSA bounce’’ phenomenon. Furthermore, they must have a good understanding of the adequacy of the initial implant as it relates to D90 and V100, because this may influence the probability of treatment failure. Once a patient is determined to have failed primary therapy, the pressing issue then becomes distinguishing local from systemic recurrences.
Local failure has been defined as histologically proven active adenocarcinoma on repeat prostate biopsy in the absence of radiographic evidence of disease. Unfortunately, multiple studies have demonstrated the relative lack of sensitivity and specificity of most radiographic tests including CT, MRI, bone scan, and more recently monoclonal antibody–labeled nuclear scans (Prosta- Scint, Cytogen Corp, Princeton, NJ) for the diagnosis of systemic disease in biochemically recurrent prostate cancer. Despite these shortcomings, these modalities should be applied if recurrence is suspected, because the presence of overt metastatic disease may obviate the patient being exposed to unnecessary local therapies. In this setting, pathologic confirmation via prostate biopsy of locally recurrent disease is warranted before consideration of invasive salvage therapies.

The PSA doubling time following external RT also appears to aid in predicting the eventual development of metastatic compared with local disease. In one study, faster doubling times were significantly associated with higher T stage, higher Gleason grade, and higher pretreatment PSA levels.