Short-term androgen deprivation and PSA doubling time: their association and relationship to disease progression after radiation therapy for prostate cancer.

PURPOSE: The goal of this study was to investigate the relationship between PSA doubling time (PSADT) and initial management of prostate cancer with short-term androgen deprivation (STAD) and the impact of these factors on disease progression after radiation therapy.

METHODS AND MATERIALS: Between May 1989 and October 1998, 284 patients treated with 3D-CRT experienced biochemical failure (BF) as defined under the ASTRO consensus statement. All patients had sufficient follow-up data for PSADT calculations. Linear regression was used to assess predictors of PSADT among STAD, time to biochemical failure (TTBF), Gleason Score, tumor stage, dose, posttreatment PSA nadir, pretreatment PSA, and age. A composite covariate was created from the various combinations of factors found to be predictive of PSADT. The composite covariate was then included, along with PSADT and the factors previously mentioned, in proportional hazards modeling of freedom from distant metastasis (FDM), cause-specific survival (CSS), and overall survival (OS).

RESULTS: Fifty-four (19%) patients developed distant metastasis, 20 (7%) died of prostate cancer, and 53 (19%) died of any cause. The median PSADT was 12 months. Predictors of a longer PSADT were TTBF >12 months, Gleason Score 2-6, and STAD. An ordinal composite covariate was created with eight levels on the basis of the magnitude of observed mean PSADT within the eight possible combinations of the three dichotomized predictors. The most significant predictor of higher FDM rates in Cox modeling was the composite covariate, followed by longer PSADT, STAD, lower PSA nadir, higher RT dose, and Gleason Score 2-6. Predictors of higher CSS rates were lower nadir, longer PSADT, T1/T2ab tumors, the composite covariate, and STAD. The most significant predictor of a higher OS rate was STAD, followed by longer PSADT, younger age at diagnosis, the composite covariate, lower nadir, and T1/T2ab tumors.

CONCLUSIONS: Longer TTBF, Gleason Score 2-6 tumors, and STAD were predictive of longer PSADT. Even after adjusting for these factors in the capacity of their predictive properties for PSADT, STAD and observed PSADT continued to be significant independent predictors of FDM, CSS, and OS. STAD appears to have a pronounced impact on disease progression, probably the result partly of the prolongation of PSADT.