Predictors of overall and cancer-free survival of patients with localized prostate cancer treated with primary androgen suppression therapy: Results from the prostate cancer outcomes study

Purpose: Primary androgen suppression therapy for clinically localized prostate cancer is increasingly common in the United States despite a lack of supportive evidence for its use. We determined which demographic and clinical factors predict overall and cancer specific survival with this treatment strategy in patients enrolled in the Prostate Cancer Outcomes Study. Materials and Methods: In 1994 to 1995 the Prostate Cancer Outcomes Study recruited 3,533 men diagnosed with prostate cancer. Clinical and treatment information was abstracted from medical records and demographic characteristics were obtained from patient surveys 6, 12, 24 and 60 months after diagnosis. Overall and cancer specific mortality was analyzed through December 2002 using the Kaplan-Meier method and Cox regression. Results: A total of 276 patients had organ confined (cTl-2) prostatic adenocarcinoma and received primary androgen suppression therapy within 1 year of diagnosis. Median followup for censored patients was 7.6 years (range 1.1 to 8.1). Five-year overall and cancer specific survival was 66% (95% CI 59-72) and 91% (95% CI 86-94), respectively. Independent predictors of shorter overall survival were patient age 75 years or older, prostate specific antigen 20 ng/ml or greater, Gleason score 7 or greater and abnormal digital rectal examination. Gleason score 7 or greater, prostate specific antigen 20 ng/ml or greater and a low comorbidity index were independent predictors of shorter cancer specific survival. Conclusions: The use of primary androgen suppression therapy in the Prostate Cancer Outcomes Study data set resulted in 91% 5-year cancer specific survival. Advanced age, and factors that reflect tumor burden and biology were predictive of overall survival, while cancer specific survival was predicted by tumor factors and the burden of comorbid conditions. A nomogram for predicting overall survival at 5 years was constructed.