Incorporating Polygenic Risk Scores and Nongenetic Risk Factors for Breast Cancer Risk Prediction among Asian Women, Results from Asia Breast Cancer Consortium

Importance Polygenic risk scores (PRSs) have shown promises in breast cancer risk prediction; however, limited studies have been conducted among Asian women. Objective To develop breast cancer risk prediction models for Asian women incorporating PRSs and nongenetic risk factors. Design PRSs were developed using data from genome-wide association studies (GWAS) of breast cancer conducted among 123 041 Asian-ancestry women (including 18 650 cases) using three approaches (1) reported PRS for European-ancestry women; (2) breast cancer-associated single-nucleotide polymorphisms (SNPs) identified by fine-mapping of GWAS-identified risk loci; (3) genome-wide risk prediction algorithms. A nongenetic risk score (NgRS) was built including six well-established nongenetic risk factors using data from 1974 Asian women. Integrated risk scores (IRSs) were constructed using PRSs and the NgRS. PRSs were initially validated in an independent dataset including 1426 cases and 1323 controls and further evaluated, along with the NgRS and IRSs, in the second dataset including 368 cases and 736 controls nested withing a prospective cohort study. Setting Case-control and prospective cohort studies. Participants 20 444 breast cancer cases and 106 450 controls from the Asia Breast Cancer Consortium. Main Outcomes and Measures Logistic regression was used to examine associations of risk scores with breast cancer risk to estimate odds ratios (ORs) with 95% confidence intervals (CIs) and area under the receiver operating characteristic curve (AUC). Results In the prospective cohort, PRS111, a PRS with 111 SNPs, developed using the fine-mapping approach showed a prediction performance comparable to a genome-wide PRS including over 855,000 SNPs. The OR per standard deviation increase of PRS111 was 1.67 (95% CI=1.46-1.92) with an AUC of 0.639 (95% CI=0.604-0.674). The NgRS had a limited predictive ability (AUC=0.565; 95% CI=0.529-0.601); while IRS111, the combination of PRS111 and NgRS, achieved the highest prediction accuracy (AUC=0.650; 95% CI=0.616-0.685). Compared with the average risk group (40th-60th percentile), women in the top 5% of PRS111 and IRS111 were at a 3.84-folded (95% CI=2.30-6.46) and 4.25- folded (95% CI=2.57-7.11) elevated risk of breast cancer, respectively. Conclusions and Relevance PRSs derived using breast cancer-associated risk SNPs have similar prediction performance in Asian and European descendants. Including nongenetic risk factors in models further improved prediction accuracy. Our findings support the utility of these models in developing personalized screening and prevention strategies.