Questioning the validity of clinically available breast cancer polygenic risk scores: comparison of two labs reveals discrepancies

Excerpt

Cancer genomic medicine has until now focused largely on high risk, rare genetic variants. However, common variants in aggregate can contribute significantly to the risk for cancers and can have significant impact on public health. Individuals in the top 2–5% of polygenic risk score (PRS) have risk equivalent to many of the rare variants currently used clinically for cancers such as breast or colon [1]. PRSs are a potential tool to refine risk stratification and improve screening, early detection, early intervention, and overall public health. However, there are limitations to PRSs. Presently, there are no universally accepted standardized and validated algorithms for PRS calculation, and current PRS algorithms are usually derived from data from individuals of European ancestry. Validation for other ancestral groups can be challenging given the lack of diversity in genomic data sets [2]. Furthermore, there is recent evidence that even within ancestral groups there may be variability in predictive accuracy across age, sex, or socio-economic status [3]. Presently, the National Comprehensive Cancer Network (NCCN) recommends that PRS not be used in clinical care until formal clinical trials have been conducted [4]. …