Early undetectable serum PSA is an independent predictive factor of progression-free survival (PFS) and bone metastases in patients with localized prostate cancer treated by androgen deprivation and radiotherapy

Abstract

4684 Background: Patients with locally-advanced prostate cancer are managed by androgen deprivation (AD) and radiation (RT). We attempted to identify early predictors of outcome in a large retrospective series with a long follow-up. Methods: From 1991 to 1998, 92 patients with locally-advanced prostate cancer were treated by AD and RT. Median pretreatment PSA level was 19 ng/mL (1.6–148). Clinical stage was T1 (4%), T2 (60%), and T3 (36%). Gleason score was 4–6 (46%), 7 (32%) and 8–10 (22%). Prognosis was classified as high risk (T3 and/or PSA>20 ng/mL and/or Gleason score > 7) (n= 59) or intermediate risk (n= 33). AD consisted in a complete androgen blockade (GnRH agonist and anti-androgen) (55 pts) or an anti-androgen as a single agent (37 pts) and was given for a median duration of 6 months (3–18). AD was usually initiated 3 months before RT. The prostate was to receive 65 Gy (2.5 Gy/day) (74 pts) or 70 Gy (2 Gy/day) (18 pts). The median follow-up was 70 months (14–126). Results: The 5-year PSA-PFS rate was 41% (95% CI : 30–51) and 57% (95% CI: 47–68) according to the ASTRO definition and the MDACC definition, respectively. The 5-year metastases-free survival and the 5-year cause-specific survival rates were respectively 82% (95% CI: 73–89) and 94% (95% CI: 88–99). The median PSA level assessed 3 to 6 months after the start of AD and before RT was 1.3 ng/mL (0.08–31). In multivariate analysis, significant predictors for PSA-PFS included high-risk disease (p<0.01) and an undetectable (<0,2 ng/mL) serum PSA after 3 months of neo-adjuvant AD (p<0,01). These two factors were also independent predictors of distant metastases (p=0.04). Conclusions: An undetectable serum PSA 3 months after AD is an independent predictor of both PSA-PFS and distant metastases in locally-advanced prostate cancer. Therefore, this factor may be a useful early surrogate endpoint in clinical trials using neo-adjuvant treatments. No significant financial relationships to disclose.