Combining 33 genetic variants with prostate-specific antigen for prediction of prostate cancer: longitudinal study.

The aim of this study was to investigate if a genetic risk score including 33 common genetic variants improves prediction of prostate cancer when added to measures of prostate-specific antigen (PSA). We conducted a case-control study nested within the Northern Sweden Health and Disease Cohort (NSHDC), a prospective cohort in northern Sweden. A total of 520 cases and 988 controls matched for age, and date of blood draw were identified by linkage between the regional cancer register and the NSHDC. Receiver operating characteristic curves with area under curve (AUC) estimates were used as measures of prostate cancer prediction. The AUC for the genetic risk score was 64.3% [95% confidence interval (CI) = 61.4-67.2], and the AUC for total PSA and the ratio of free to total PSA was 86.2% (95% CI = 84.4-88.1). A model including the genetic risk score, total PSA and the ratio of free to total PSA increased the AUC to 87.2% (95% CI = 85.4-89.0, p difference = 0.002). The addition of a genetic risk score to PSA resulted in a marginal improvement in prostate cancer prediction that would not seem useful for clinical risk assessment.