e16004 Background: Men receiving high dose external beam radiotherapy (EBRT) and androgen deprivation therapy (ADT) for high risk prostate cancer (HRPC) have other competing causes of mortality, however predictive schema do not account for patient-related co-morbidities. We aim to create nomograms estimating all-cause mortality (ACM) and prostate cancer-specific mortality (PCSM) in this population. Methods: 660 patients with HRPC defined by NCCN guidelines were treated with EBRT ≥74 Gy and ADT from 1996-2009. The probabilities of death from prostate cancer and other causes were estimated by cumulative incidence function. Multivariable Cox proportional hazards regression and competing risks regression analyses were used for modeling ACM and PCSM respectively. Deaths from other causes were treated as competing risks for PCSM. Missing values in the predictors were multiply imputed before conducting multivariable regression analysis. Variables investigated were age, clinical T stage, prostate specific antigen (PSA), Gleason score, race, family history, duration of ADT, body mass index (BMI), Charlson co-morbidity index score, coronary artery disease, and smoking pack-years. The stepdown method was used to make parsimonious models based on the rank of the predictive ability of each variable with respect to each endpoint. The final nomograms were internally validated by assessing the discrimination and calibration with bootstrap resamples. Results: At last follow up, there were 199 deaths. The 10-year cumulative incidence of death from prostate cancer was 14% and from other causes was 26%. The variables that predicted for 10-year ACM included age, PSA, BMI, Charlson score, and smoking pack-years. The ACM nomogram achieved a concordance index of 0.672. The variables that predicted for PCSM included Gleason score, PSA, race, and duration of ADT. The nomogram concordance index for PCSM was 0.673. The calibrations for both ACM and PCSM appear reasonable. Conclusions: We have developed nomograms that predict for ACM and PCSM in men with aggressive prostate cancer and competing risks of death. External validation may be useful.
Prostate cancer-specific mortality and the extent of therapy in healthy elderly men with high-risk prostate cancer
