The optimal management of patients with clinically localized prostate carcinoma remains undefined due in part to the absence of well-designed, prospective, randomized trials. The current study was conducted to compare and contrast outcomes with different forms of therapy for patients with prostate carcinoma who were treated at several institutions using predefined prognostic categories.
A retrospective study of 6877 men with prostate carcinoma who were treated between 1989 and 1998 at 7 different institutions with 6 different types of therapy was conducted. Five-year actuarial rates of prostate specific antigen (PSA) failure were calculated based on predefined prognostic categories, which included combinations of pretreatment PSA level, tumor stage, and Gleason score. In addition, outcome was calculated using consistent biochemical failure definitions and a minimum, median length of follow-up.
Substantial differences in outcome were observed for the same type of treatment and at the same institution, depending on the number of prognostic variables used to define treatment groups. However, estimates of 5-year PSA outcomes after all forms of therapy for low-risk and intermediate-risk patient groups were remarkably similar (regardless of the type of treatment) when all three pretreatment variables were used to define prognostic categories. For patients in high-risk groups, the 5-year PSA outcomes were suboptimal, regardless of the treatment technique used.
The current data suggest that interinstitutional and interspecialty comparisons of treatment outcome for patients with prostate carcinoma are possible but that results must be based on all major prognostic variables to be meaningful. Analyzed in this fashion, 5-year PSA results were similar for patients in low-risk and intermediate-risk groups, regardless of the form of therapy. Findings from prospective, randomized trials using survival (cause specific and overall) as the end point for judging treatment efficacy and longer follow-up will be needed to validate these findings and to identify the most appropriate management option for patients with all stages of disease. Cancer 2002;95:2126–35. © 2002 American Cancer Society.