Toward Application of Polygenic Risk Scores to Both Enhance and Deintensify Breast Cancer Screening

Abstract

Breast cancer is the most common cancer in individuals assigned female sex at birth (here, referred to as women), and factors such as family history, reproductive history, self-identified race and genetic ancestry, and lifestyle factors can increase and decrease an individual woman's risk. 1 The National Comprehensive Cancer Network which provides guidelines for breast cancer screening programs in the United States defines average risk as approximately 12% lifetime risk and increased risk as a lifetime risk of ≥20%. 1 Average risk women begin yearly mammography at age 40 years (of note, the US Preventative Services Task Force recommendations differ); high-risk women undergo enhanced screening with yearly mammography and breast MRI, in addition to consideration of risk reducing medications and surgery depending on the risk. Although high-risk radiographic screening decreases mortality in women with BRCA1 variants, 2 the survival benefit of this strategy in other high-risk women is unknown. Furthermore, no recommendations exist to deintensify screening for women at lower-than-average risk to avoid overscreening.While <1% of unaffected women unselected for family history would be identified for high-risk screening programs because of a pathogenic variant (such as in BRCA1, BRCA2, or PALB2), a polygenic risk score (PRS) can be calculated in all and, therefore, holds promise to better stratify risk for both enhancing and deintensifying screening. A PRS is constructed using the genotypes of multiple common risk loci which individually confer small (typically <1.5-fold) risks of disease. There are multiple methods to construct and apply a PRS, which differ in selection of genetic loci, method of determination of effect sizes for each locus, and the type of statistical prediction model and adjusting variables used to predict risk.This study adds to a large body of observational studies 5,6 and evidence from cost-effectiveness modeling 7 that PRS can stratify women into risk categories. Although breast cancer screening already benefits from well-validated risk prediction models, such as the BOADICEA algorithm, 8