PURPOSE To assess the prognostic significance of serum prostate-specific antigen (PSA) in the monitoring of patients with localized prostate cancer treated with primary radiation therapy, we analyzed the data from 179 patients treated at our institution between 1987 and 1990. PATIENTS AND METHODS One hundred seventy-nine previously untreated patients received radiation at 69 Gy to the prostate with curative intent for prostate adenocarcinoma. The median follow-up duration is now 41 months. PSA levels were measured before radiotherapy and then evaluated periodically. RESULTS Baseline levels were greater than 4 ng/mL in 83% of cases and were significantly correlated with clinical tumor stage (P = .002). Six months after completion of therapy, PSA values had returned to normal in 53% of the patients with initially elevated values. At the time of analysis, 32 patients have relapsed, including three of 30 patients (10%) with normal initial and 6-month values, five of 79 patients (6%) with initially elevated but normal 6-month values, and 24 of 69 patients (35%) with persistently elevated PSA levels at 6 months. Actuarial 4-year relapse-free survival was significantly correlated with initial and 6-month PSA values (84% in patients with normal 6-month values v 60% in patients with persistently elevated levels). Furthermore, when the relative decline between initial and 6-month PSA values exceeded 50%, the crude rate of recurrence was 14% as opposed to 34% when it failed to exceed 50%. The 4-year relapse-free survival rates were 77% and 59%, respectively (P = .008). By multivariate analysis restricted to the patients with elevated baseline PSA levels, the rate of decline of PSA values reached the highest prognostic significance (P < .0001). Age at diagnosis, clinical tumor stage, and Gleason score only reached statistical significance in univariate analysis. CONCLUSION PSA values are of major prognostic significance in assessing the 4-year results of radical radiation therapy for localized prostate cancer. The rate of decline of PSA values is the strongest predictor of outcome and might help to identify a subset of patients with poorer prognosis who may benefit from early hormonal therapy.
