Polygenic Risk Scores for Prediction of Breast Cancer in Korean women

ABSTRACT Background Polygenic risk scores (PRSs) developed using European and Asian GWAS have been shown to have good discrimination in Asian women. However, prospective calibration of absolute risk prediction models based on a PRS or PRS combined with lifestyle clinical and environmental factors in Asian women is limited. We evaluate the discrimination and calibration of several breast cancer PRSs among Korean women; these PRSs were developed using Asian and/or European training samples and include between 11 and 947,621 variants. Methods For each PRS, we compared discrimination (area under the curve [AUC]) and calibration (expected-to-observed ratio [E/O]) of three absolute risk models among 41,031 women from the Korean Cancer Prevention Study (KCPS)-II Biobank: (i) a model using incidence, mortality, and risk factor distributions (reference inputs) among U.S. women and European relative risks; (ii) a recalibrated model, using Korean reference but European relative risks; and (iii) a fully Korean-based model using Korean reference and relative risk estimates from KCPS. Results All Asian and European PRS improved discrimination over lifestyle, clinical and environmental (Qx) factors in Korean women; a PRS trained using both European and Asian GWAS results led to the greatest improvement (Qx: AUC=0.65, Qx+PRS: AUC=0.72). U.S.- based absolute risk models overestimated the risks for women age ≥50 years, and this overestimation was larger for models that only included PRS (E/O=1.2 for women <50, E/O=2.7 for women ≥50). Recalibrated and Korean-based risk models had better calibration in the large, although the risk in the highest decile was consistently overestimated. Absolute risk projections suggest that risk-reducing lifestyle changes would lead to larger absolute risk reductions among women at higher PRS. Conclusions Incorporation of Asian and European PRS can improve discrimination in Korean women and may be useful for the risk-stratified interventions. Key Messages ⍰ Prospective validation of absolute risk prediction models combining lifestyle and polygenic risk scores in Asian women is limited. ⍰ We evaluated the calibration and discrimination of five PRSs developed using Asian and/or European training samples; two PRS were restricted to genome-wide significant SNPs, two included sub-genome-wide significant SNPs, and a multi-ancestry PRS using both European and Asian GWAS results. ⍰ Incorporation of PRS previously developed in Asian and European-ancestry populations can improve discrimination in Korean women. ⍰ Calibration improved for risk models that incorporate age-specific incidence rates from the target population relative to models that use external incidence rates ⍰ Our finding suggests that PRS may be useful for prioritizing individuals for targeted intervention on their lifestyle such as alcohol intake and obesity. ⍰ Further studies are needed to evaluate the value of incorporating PRS into risk models in ancestrally diverse populations.