Polygenic hazard score predicts aggressive and fatal prostate cancer in multi-ethnic populations

2019 
Background: Prostate cancer causes substantial morbidity and mortality worldwide. Recently, a polygenic hazard score (PHS1)--the weighted sum of 54 single-nucleotide polymorphism (SNP) genotypes--was developed and validated to predict age of aggressive prostate cancer onset in Caucasians. We evaluated the performance of the PHS for prediction of aggressive prostate cancer and of death from prostate cancer across diverse ethnic populations. Methods: 80,491 men of various self-reported race/ethnicities were included (30,575 controls, 49,916 cases. Previously determined genetic ancestries for these men were: 71,856 European, 6,253 African, and 2,382 Asian. Where applicable, age of prostate cancer diagnosis, age at last follow-up, TNM stage, PSA, Gleason score, and cause of death were also determined. Median age at last follow-up was 70 years (IQR 63-76). 3,983 men died from prostate cancer, 5,806 died from non-prostate cancer causes, and 70,702 were still alive at the end of follow-up. Patient samples were previously genotyped on a cancer-specific array; PHS1 was adapted for compatibility with this array (PHS2) and tested in the multi-ethnic dataset via Cox proportional hazards models for age of aggressive prostate cancer onset and for age at prostate-cancer-specific death. Results: PHS2 had 46 SNPs: 24 directly genotyped and 22 acceptable proxies (r2>=0.94). PHS2 was predictive of age of aggressive prostate cancer onset in the independent, multi-ethnic dataset (z=48, p<10-16) and in each genetic ancestry: European (z=46, p<10-16), Asian (z=44, p<10-16), and African (z=24, p<10-16). PHS2 was also predictive of age of prostate cancer death in the multi-ethnic dataset (z=16, p<10-16). Comparing the 80th and 20th percentiles of genetic risk, men with high PHS had a hazard ratio of 5.9 for aggressive prostate cancer and 5.7 for prostate cancer death. Within each genetic ancestry group, analogous hazard ratios for aggressive prostate cancer were 5.6, 5.2, and 2.4 for men of European, Asian, and African ancestry, respectively. Conclusions: PHS2 is highly predictive of both age of onset of aggressive prostate cancer and age of prostate-cancer-specific death in a multi-ethnic dataset. Genetic risk stratification via PHS could guide individualized screening and treatment strategies for fatal and potentially fatal prostate cancer for men of European, Asian, and African genetic ancestry.
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    49
    References
    7
    Citations
    NaN
    KQI
    []