Preventing premature deaths through polygenic risk scores

Polygenic risk scores (PRS) have demonstrated predictive validity across a range of cohorts and diseases, but quantifying their clinical utility remains a challenge. As PRS can be derived from a single biological sample and remains stable throughout life, we explore the potential of PRS to optimize existing screening programs. Via an integrated modelling approach, we quantify the potential clinical benefits arising from a knowledge of PRS across seven diseases with existing screening programs (abdominal aortic aneurysm, breast cancer, colorectal cancer, coronary artery disease, hypertension, prostate cancer, and type 2 diabetes). We identify individuals at high genetic risk (PRS OR>2) and very high genetic risk (PRS OR>3) and estimate the optimal screening ages for these genetically high-risk individuals, based on the equivalent risk to population-level risk at recommended screening ages. We then leverage published data on differential mortality and other outcomes, with and without screening-based interventions, to assess the potential benefits of tailoring screening age based on genetic risk. Very high risk individuals reach the risk level associated with usual starting screening age on average 10.8 years earlier, high risk individuals 8.9 years earlier and reduced risk individuals (OR<0.5) 16.8 years later. During this time, case enrichment (the ratio of the percentage of cases in the high PRS risk group and in the total population) in the high risk group is between 1.7 and 3.0, depending on the disease. Across all seven diseases, appropriate interventions following PRS-guided screening could reduce premature deaths in high-risk individuals by 23.3%. Knowledge of genetic risk, measured using PRS, has the potential to deliver substantial public health benefits when aggregated across conditions, and could reduce premature mortality by tailoring existing screening programs.