A multiple prognostic index predictive of disease outcome after irradiation for clinically localized prostate carcinoma

BACKGROUND. This investigation was conducted to identify independent pretherapy disease-related factors associated with disease outcome in patients with clinically localized carcinoma of the prostate (CaP) and to develop models that incorporated relevant covariates for estimating the risk of disease relapse after irradiation (RT). METHODS. The outcome of 500 patients treated only with RT between March 1987 and June 1993 for clinical Stages T1-4N0,XM0 CaP was evaluated. The risk of disease relapse as a function of individual prognostic variables, and combinations thereof, was determined using logistic regression. RESULTS. With a median follow-up of 43 months (range, 4-103 months), 69 patients (14%) had clinical evidence of local recurrence (27 patients), regional lymph node relapse (4 patients), or metastatic relapse (38 patients) within 5 years of RT. Forty additional patients (8%) had biochemical relapse based solely on the post-RT serum prostate specific antigen (PSA) profile. Clinical tumor stage (P = 0.0006), Gleason score (P = 0.001) of the diagnostic biopsy specimen, and pretherapy PSA (P < 0.0001) were associated with disease relapse. The risk of any relapse within 5 years of RT was determined and graphically displayed as risk estimate plots for combinations of these pretherapy prognostic variables. CONCLUSIONS. The combination of pretherapy clinical tumor (T) stage, Gleason score, and PSA level can be used to obtain improved estimates of the risk for disease relapse in patients treated solely with RT for clinically localized CaP. Risk estimate plots of this type may facilitate exchange of therapeutic outcome information, be instrumental in pretherapy decision-making for the new patient with this condition, and aid in the selection of patients for future studies.