Radical prostatectomy for clinically localized, high risk prostate cancer: critical analysis of risk assessment methods.

PURPOSE: Standardized criteria are lacking to define high risk, clinically localized prostate cancer before definitive treatment. Reliance on simple risk stratification schemes to define high risk cancers has led many physicians and patients toward therapeutic nihilism, inappropriately selecting androgen deprivation instead of definitive local therapy. Of patients undergoing radical prostatectomy we identified those at high risk based on 8 previously described definitions. We examined pathological characteristics and prostate specific antigen outcomes.
MATERIALS AND METHODS: The study population included 4,708 men treated with radical prostatectomy alone between 1985 and 2004. Estimates of prostate specific antigen relapse for patients at high risk were generated with the Kaplan-Meier method. Cox proportional hazards regression was used to estimate the HR for recurrence in high risk vs nonhigh risk cohorts.
RESULTS: Depending on the definition used patients at high risk composed 3% to 38% of the study population. The proportion of patients with extracapsular extension, seminal vesicle invasion and lymph node metastasis among men with high risk cancer was 35% to 71%, 10% to 33% and 7% to 23%, respectively. Of the high risk tumors 22% to 63% proved to be confined to the prostate pathologically. While patients at high risk had a 1.8 to 4.8-fold increased hazard of prostate specific antigen relapse, their 5-year relapse-free probability after radical prostatectomy alone was 49% (95% CI 39 to 58) to 80% (95% CI 77 to 83). Of patients at high risk who had relapse 25% across all definitions experienced relapse more than 2 years after surgery and in 26% to 39% prostate specific antigen doubling time at recurrence was 10 months or greater.
CONCLUSIONS: Patients diagnosed with high risk cancer by currently available definitions do not have a uniformly poor prognosis after radical prostatectomy. Many cancers classified clinically as high risk are actually confined to the prostate pathologically. The risk of extraprostatic disease and prostate specific antigen relapse varies greatly depending on the definition used.