BACKGROUND: The clinical and pathologic predictors of prostate cancer-specific mortality (PCSM) many years after radical prostatectomy (RP) remain to be fully elucidated. We explored the association between pre-operative prostate-specific antigen (PSA) and other pathologic predictors and PCSM in men who have undergone (RP). METHODS: We report on 459 patients with PCSM data after RP who were followed prospectively over a 23-year period between 1987 and 1997. Cox regression and Kaplan-Meier analysis were used to evaluate pre-operative PSA, pathologic Gleason sum, pathologic stage, and surgical margin status as predictors of PCSM. RESULTS: The median PSA was 6.6 ng/ml (± 9.9) and the median follow-up time was 9.4 (± 4.9) years. Fourteen patients (3.1%) died of PC. On multivariate analysis, only PSA (HR: 1.050; P = 0.001) and binary Gleason sum (HR: 3.402; P = 0.043) remained significant predictors of PCSM. The predicted 10-year PCSM was significantly worse in those patients in the highest PSA tertile compared to those in other tertiles [PSA > 9.9: 87% (82-92%) vs. PSA = 4-9.9: 95% (93.0-97.0%) vs. PSA = 0-3.9: 100.0% (100.0-100.0%)]. CONCLUSIONS: We have highlighted the importance of pre-operative PSA in predicting PCSM many years after RP. It is a more significant predictor than Gleason sum and pathologic stage. Thus, PSA may help identify patients with life-threatening PC at a time when their disease is curable with definitive therapy.
BACKGROUND: The clinical and pathologic predictors of prostate cancer-specific mortality (PCSM) many years after radical prostatectomy (RP) remain to be fully elucidated. We explored the association between pre-operative prostate-specific antigen (PSA) and other pathologic predictors and PCSM in men who have undergone (RP). METHODS: We report on 459 patients with PCSM data after RP who were followed prospectively over a 23-year period between 1987 and 1997. Cox regression and Kaplan-Meier analysis were used to evaluate pre-operative PSA, pathologic Gleason sum, pathologic stage, and surgical margin status as predictors of PCSM. RESULTS: The median PSA was 6.6 ng/ml (± 9.9) and the median follow-up time was 9.4 (± 4.9) years. Fourteen patients (3.1%) died of PC. On multivariate analysis, only PSA (HR: 1.050; P = 0.001) and binary Gleason sum (HR: 3.402; P = 0.043) remained significant predictors of PCSM. The predicted 10-year PCSM was significantly worse in those patients in the highest PSA tertile compared to those in other tertiles [PSA > 9.9: 87% (82-92%) vs. PSA = 4-9.9: 95% (93.0-97.0%) vs. PSA = 0-3.9: 100.0% (100.0-100.0%)]. CONCLUSIONS: We have highlighted the importance of pre-operative PSA in predicting PCSM many years after RP. It is a more significant predictor than Gleason sum and pathologic stage. Thus, PSA may help identify patients with life-threatening PC at a time when their disease is curable with definitive therapy.