Abstract
BACKGROUND: We derived estimates of overdiagnosis by polygenic risk groups and examined whether polygenic risk-stratified screening for prostate cancer reduces overdiagnosis. METHODS: We calculated the polygenic risk score based on genotypes of 66 known prostate cancer loci for 4967 men from the Finnish section of the European Randomised Study of Screening for Prostate Cancer. We stratified the 72 072 men in the trial into those with polygenic risk below and above the median. Using a maximum likelihood method based on interval cancers, we estimated the mean sojourn time (MST) and episode sensitivity. For each polygenic risk group, we estimated the proportion of screen-detected cancers that are likely to be overdiagnosed from the difference between the observed and expected number of screen-detected cancers. RESULTS: Of the prostate cancers, 74% occurred among men with polygenic risk above population median. The sensitivity was 0.55 (95% confidence interval (CI) 0.45-0.65) and MST 6.3 (95% CI 4.2-8.3) years. The overall overdiagnosis was 42% (95% CI 37-52) of the screen-detected cancers, with 58% (95% CI 54-65) in men with the lower and 37% (95% CI 31-47) in those with higher polygenic risk. CONCLUSION: Targeting screening to men at higher polygenic risk could reduce the proportion of cancers overdiagnosed.