Dissecting the heterogeneity of localized prostate cancer risk groups through integration of percent of positive cores.

Abstract

To develop a modified risk stratification scheme for localized prostate cancer incorporating percent of positive cores (PPC). SEER database was accessed for eligible patients. Assessment of the prognostic value of PPC was conducted in a multivariate Cox regression model. A modified risk stratification scheme was proposed. In a multivariate model, higher PPC was associated with worse cancer-specific survival (p < 0.0001). A modified risk-stratification scheme was proposed incorporating PPC. Concordance index was evaluated and the results were: D'Amico model: 0.782 (SE: 0.014; 95%CI: 0.755-0.810); modified model: 0.809 (SE: <0.001; 95%CI: 0.781-0.837). Integration of PPC into the risk stratification model for localized prostate cancer improves its performance.