Are biochemical recurrence outcomes similar after radical prostatectomy and radiation therapy? Analysis of prostate cancer-specific mortality by nomogram-predicted risks of biochemical recurrence.

BACKGROUND Due to the protracted natural history of the clinical progression of prostate cancer, biochemical recurrence (BCR) is often used to compare treatment modalities. However, BCR definitions and posttreatment prostate-specific antigen kinetics vary considerably among treatments, calling into the question the validity of such comparisons. OBJECTIVE To analyze prostate cancer-specific mortality (PCSM) according to treatment-specific nomogram-predicted risk of BCR for men treated by radical prostatectomy (RP), external-beam radiation therapy (EBRT), and brachytherapy. DESIGN, SETTING, AND PARTICIPANTS A total of 13 803 men who underwent RP, EBRT, or brachytherapy at two US high-volume hospitals between 1995 and 2008. INTERVENTION RP, EBRT, and brachytherapy. OUTCOME MEASUREMENTS AND STATISTICAL ANALYSIS The 5-yr progression-free probability (5Y-PFP) was calculated for each patient based on the treatment received using a validated treatment-specific nomogram. Fine and Gray competing risk analysis was then used to estimate PCSM by a patient's predicted 5Y-PFP. Multivariable competing risk regression analysis was used to determine the association of treatment with PCSM after adjusting for nomogram-predicted 5Y-PFP. RESULTS AND LIMITATIONS Men receiving EBRT had higher 10-yr PCSM compared with those treated by RP across the range of nomogram-predicted risks of BCR: 5Y-PFP >75%, 3% versus 0.9%; 5Y-PFP 51-75%, 6.8% versus 5.9%; 5Y-PFP 26-50%, 12.2% versus 10.6%; and 5Y-PFP ≤25%, 26.6% versus 21.2%. After adjusting for nomogram-predicted 5Y-PFP, EBRT was associated with a significantly increased PCSM risk compared with RP (hazard ratio: 1.5; 95% confidence interval, 1.1-2.0; p=0.006). No statistically significant difference in PCSM was observed between patients treated by brachytherapy and RP, although patient selection factors and lack of statistical power limited this analysis. CONCLUSIONS EBRT patients with similar nomogram-predicted 5Y-PFP appear to have a significantly increased risk of PCSM compared with those treated by RP. Comparison of treatments using nomogram-predicted BCR end points may not be valid. PATIENT SUMMARY Biochemical recurrence (BCR) outcomes after external-beam radiation therapy and radical prostatectomy are associated with different risks of subsequent prostate cancer-specific mortality. Physicians and patients should cautiously interpret BCR end points when comparing treatments to make treatment decisions.