Predictive Value of PSA Trend During and After Postprostatectomy Salvage Radiation Therapy

Abstract

There is a high rate of biochemical recurrence in post prostatectomy patients who are treated with salvage radiation therapy (S-RT). Early initiation of androgen deprivation therapy has been shown to improve outcome in this cohort. In this IRB approved study we evaluate various parameters for their predicative value in defining a cohort of patients at higher risk of biochemical recurrence (BR). Eighty-seven patients with BR defined by two consecutive increases of PSA > 0.2 ng/mL or a single measurement of PSA>0.4ng/mL were enrolled in this study. Patients were excluded if they were receiving ADT or if they failed within three months of prostatectomy. Patients were treated by a single radiation oncologist using IMRT to a dose of 66.6 Gy in 37 fractions. PSA was measured prior to the initiation of S-RT and during S-RT at two week intervals. BR after S-RT was defined as PSA value >= 0.1 ng/mL above the post-S-RT PSA nadir confirmed by a second measurement, no response in PSA, or the initiation of ADT. Kaplan-Meier (KM) analysis with log rank test, univariate and multivariate Cox regression hazard ratio (HR) was used to analyze the probability of BR for Gleason (G) score (6 vs. >6), pre-treatment PSA (=10), seminal vesicle invasion (SVI), peri-neural invasion (PNI), margin status, extra-capsular extension (ECE), pre-S-RT PSA nadir (0.2), pre-S-RT PSA velocity (=0.075/m), pre-S-RT PSA (=1), PSA trend during S-RT (>2 consecutive decreases in PSA), PSA response at 40 Gy, 60 Gy, within three months and within six months of S-RT. The BRFS, overall survival and PCa specific survival in this group at a median follow up of 45 months (m) is 70.1%, 94.3% and 97.7% respectively with a median time to BR of 17.5m. A total of 79.7% of the patients with a decreasing PSA trend during S-RT were free of BR compared to only 14.3% of patients with an increasing PSA trend during S-RT (p < 0.001). Additionally in univariate analysis, Gleason grade, pre-S-RT PSA velocity, PSA response at 40 Gy, 60 Gy, within three months and within six months of S-RT were predictors of BR. In multivariate Cox regression modeling, PSA trend during S-RT (p < 0.001), PSA response at 40 Gy (p < 0.001), 60 Gy (p < 0.001), within three months (p < 0.001) and within six months (p < 0.001) of S-RT were independent predictors of BR. PSA trend during S-RT and an early PSA response has significant predictive value for BR after S-RT and has the potential to strengthen existing nomograms for early identification of patients at higher risk of BR. These parameters should be evaluated in additional prospective studies before clinical guidelines can be established.