Introduction

The current pre-treatment nomograms for prostate cancer were developed and validated based on a patient population that primarily consisted of men with low or intermediate risk prostate cancer. This study aims to develop a preoperative nomogram that predicts for metastasis from a contemporary cohort of men with NCCN high (HR) or very high risk (VHR) prostate cancer.

Methods

We identified 1,241 men from 2005 to 2015 with NCCN HR or VHR prostate cancer from two large academic medical centers primarily treated with radical prostatectomy. The cohort was divided into training (n=620) and validation (n=621) cohorts. The primary endpoint of analysis was mets. Multivariable Cox proportional hazards regression analysis was used to model characteristics and outcomes in the training cohort. Predictive accuracy was assessed using the time-dependent area under the receiver operating characteristic curve (AUC) in the validation cohort.

Results

123 men (64 training and 59 validation) developed metastasis. The overall metatasis-free probability was 86.5% (95% CI 83.7%-89.4%) at 5-years. Predictive nomograms including age, ethnicity, PSA, Gleason grade, clinical stage, and the number of positive cores with Gleason 8-10 disease were developed. The AUC for the model was 0.75. In comparison, the MSKCC preoperative nomogram and CAPRA nomogram had AUCs of 0.66 and 0.67 respectively.

Conclusions

Individualized risk assessment is imperative for optimal decision making for both disease management and appropriate clinical trial design. The nomogram described here, created from a population composed of HR/VHR men, has greater predictive accuracy of mets up to 5 years after radical prostatectomy than those previously established on cohorts of primarily low and intermediate risk men. Thus, this nomogram may be a more suitable approach for predicting mets in men with HR or VHR disease.

Funding

None