PSA doubling time as a predictor of clinical progression after biochemical failure following radical prostatectomy for prostate cancer

OBJECTIVES: To characterize the clinical progression of disease in men who have undergone prostatectomy for clinically localized prostate cancer and have postoperative biochemical failure (elevated prostate-specific antigen [PSA] level) and to identify predictors of clinical disease progression, including the possible effect of PSA doubling time (PSADT). PATIENTS AND METHODS: Between 1987 and 1993, 2809 patients underwent radical retropubic prostatectomy for clinically localized (< or =T2) disease. In our database, all patients with postoperative biochemical failure (PSA level > or =0.4 ng/mL) were identified. The PSADT was estimated using log linear regression on all PSA values (excluding those values determined after administration of hormonal therapy) within 15 months after biochemical failure. All patients had regular PSA measurements from the time of surgery through the follow-up period. Systemic progression (SP) was defined as evidence of metastatic disease on a bone scan. Local recurrence (LR) was defined on the basis of digital rectal examination, transrectal ultrasonography, and biopsy. The SP-free survival and LR/SP-free survival (survival free of both LR and SP) after biochemical failure was estimated with use of the Kaplan-Meier method. Patients with prostate cancer treatment after biochemical failure had their follow-up censored from this study at the time of treatment. RESULTS: Postoperative biochemical failure occurred in 879 men (31%). The mean follow-up from time of biochemical failure was 4.7 years (range, 0.5-11 years). The mean time to biochemical failure was 2.9 years (median, 2.4 years). The overall mean SP-free survival from time of biochemical failure was 94% and 91% at 5 and 10 years, respectively. The mean LR/SP-free survival was 64% and 53% at 5 and 10 years, respectively. By using univariate analysis on the 587 patients with PSADT data, significant risk factors for SP were PSADT (P<.001) and pathologic Gleason score (P=.005); for LR/SP, significant risk factors included PSADT (P<.001) and pathologic Gleason score (P<.001). In multivariate Cox models analysis, only PSADT remained a significant risk factor for both SP and LR/SP (P<.001). Mean 5-year SP-free survival was 99%, 95%, 93%, and 64% for patients with PSADT of 10 years or longer, 1.0 to 9.9 years, 0.5 to 0.9 year, and less than 0.5 year, respectively; the respective mean LR/SP-free survivals were 87%, 62%, 46%, and 38%. The percentage of patients with PSADT of less than 0.5 year was considerably higher if the type of first clinical event was SP (48%) compared with LR (18%) (P<.001). CONCLUSIONS: For patients who have undergone radical prostatectomy, a rising PSA level suggests evidence of residual or recurrent prostate cancer. Many men remain free of clinical disease for an extended time after biochemical failure following radical prostatectomy for clinically localized prostate cancer. The PSADT appears to be an important predictor of SP and also of any clinical progression (local or systemic). These data may be useful when counseling men regarding the timing of adjuvant therapies.