Long-Term Prostate Specific Antigen Stability and Predictive Factors of Failure after Permanent Seed Prostate Brachytherapy

Purpose: Defining biochemical failure as nadir + 2 may overestimate cure after radiotherapy. We assessed long-term prostate specific antigen stability after low dose rate prostate brachytherapy and predictors of biochemical failure when prostate specific antigen was slowly rising below the nadir + 2 ng/ml threshold. Materials and Methods: A total of 2,339 patients with low or intermediate risk prostate cancer received 125 iodine brachytherapy from 1998 to 2010 with a minimum 3-year followup. In addition, 49.7% of the patients received 6 months of androgen deprivation. Clinical, dosimetric and prostate specific antigen data were retrieved from a prospective database. Biochemical results were classified as stable or rising prostate specific antigen (0.2 ng/ml or greater and increased 0.1 ng/ml or greater during the preceding 2 years), or biochemical failure (defined as nadir + 2). Multivariate analysis was done to identify predictors of failure used to create logistic regression models. Results: At a median followup of 89 months (range 37 to 199) prostate specific antigen was stable (nadir 0.03 ng/ml and at 60 months 0.04 ng/ml) in 2,004 patients (86%) and rising (nadir 0.16 ng/ml and at 60 months 0.29 ng/ml) in 145 (6%) while biochemical failure (nadir 0.51 ng/ml, p <0.001) was noted in 190 (8%). When there was no prior androgen deprivation therapy, the prostate specific antigen nadir and prostate specific antigen at 60 months were the strongest predictors of failure (OR 20.6 and 18.3, respectively, each p <0.0001). The logistic regression model had 85% sensitivity and 98% specificity, and predicted failure in 8 of 82 men (9.8%). A second model was created for the group with androgen deprivation therapy and rising prostate specific antigen using the predictive factors prostate specific antigen at 60 months (OR 53.9, p <0.0001) and T stage (OR 0.25, p = 0.0008). This model predicted biochemical failure in 30 of 56 men (54%) with 85% sensitivity and 93% specificity. The 2 predictive models yield an anticipated 90% cure rate in the entire cohort. Conclusions: Brachytherapy is highly curative with stable prostate specific antigen at a surgical ablation level in 86% of patients. Rising prostate specific antigen is rare at a 6% incidence and often innocuous.