GenProb-PCSM: A Simplified Weighted Germline Score for Prostate Cancer-Specific Mortality.

Background

We previously developed a tier-based germline classification using the National Comprehensive Cancer Network (NCCN)-recommended DNA damage repair (DDR) genes and KLK3 I179T to predict prostate cancer (PCa)-specific mortality (PCSM). To provide an easier-to-use single inherited risk score while preserving gene-specific effects, we developed GenProb-PCSM.

Methods

We analyzed 14,644 men with incident PCa from the UK Biobank. The cohort was randomly divided into training (60%) and independent testing (40%) datasets. GenProb-PCSM was developed in the training cohort by integrating pathogenic variants in ten NCCN-recommended DDR genes and the KLK3 I179T variant using gene-specific weights derived from Fine-Gray competing-risk models. Performance was evaluated in the independent testing cohort by discrimination, calibration, and risk stratification. Secondary analyzes evaluated metastatic progression and the composite endpoint of metastatic progression and/or PCSM.

Results

Among 14,644 men with incident PCa, 1,581 died from PCa. GenProb-PCSM remained significantly associated with PCSM in the independent testing cohort (HR per SD, 1.18; 95% CI, 1.12-1.24; p < 0.001). Using predefined risk thresholds derived from the training cohort, patients in the intermediate- and high-risk groups had significantly increased risks of PCSM compared with the low-risk group (HR 1.49, 95% CI 1.22-1.84; and HR 4.04, 95% CI 2.58-6.33, respectively). GenProb-PCSM also predicted independent metastatic progression and the composite endpoint of metastatic progression and/or PCSM.

Conclusions

GenProb-PCSM transforms complex germline findings into a single inherited risk score for PCSM and metastatic progression. Its simplicity and preservation of gene-specific effects may facilitate clinical implementation of germline prognostic assessment in PCa.