Polygenic Risk Scores Across Genomic Platforms for Reliable Breast Cancer Risk Stratification

PurposeWe evaluated differences in a 313-variant breast cancer polygenic risk score (PRS313) across genomic platforms and their impact on risk stratification. MethodsWe compared PRS313 derived from genotyping arrays (Global Screening Array [GSA], OncoArray-500K [OncoArray], Global Diversity Array [GDA], custom Axiom_PrecipV1 array [ThermoFisher]) and low-coverage genome sequencing (lc-WGS) in 2 cell lines and 92 individuals. Probes were designed for all variants on ThermoFisher (success rate: 259/313). Sanger sequencing was performed to profile indels. Concordance of high-risk classification (PRSscore>0.6) across platforms was assessed using Kappa statistics. ResultsPRS313-lc-WGS was identical in the 4 cell line repeats. In saliva samples, indel concordance with Sanger sequencing varied widely (Kappa: 0.007-1.000). PRS313-ThermoFisher was predictable from other platforms using linear models, despite systematic differences. Greater agreement was observed between arrays with high imputation overlap (e.g., GDA[~]GSA slope=0.986). Pre-calibration agreement in high-risk classification was moderate (Fleiss Kappa=0.552) and improved post-calibration (Kappa=0.650). Arrays with similar designs showed higher pre-calibration agreement (Kappa=0.745). Calibration narrowed high-risk proportions from 4-45% to 15-21% -28% were high-risk by any platform, while 8% were high-risk across all five. ConclusionPlatform-specific biases affect PRS interpretation. Calibration enhances consistency in identifying high-risk individuals. STATEMENT OF SIGNIFICANCEThis study compares the performance of a validated 313-variant breast cancer polygenic risk score across platforms, revealing systematic biases in risk stratification and raising concerns about including inconsistent indels in the model.