Abstract LB-178: Performance of breast cancer polygenic risk score (PRS) in a Ghanaian population

Abstract

Abstract Background: Three breast cancer PRSs comprised of 313 single nucleotide polymorphisms (SNPs) have been developed and validated in women of European ancestry for predicting risk for overall, estrogen receptor-positive (ER+), and ER-negative (ER-) disease[1]. We evaluated the associations of these PRSs with breast cancer risk in Ghanaian women Methods: Analyses included 899 cases (296 ER+, 277 ER-, and 326 unknown ER status) diagnosed with invasive (n=888) or in situ (n=11) breast cancer and 1,630 population-based controls from the Ghana Breast Health Study (GBHS). The 313 SNPs were determined using the Illumina Global Screening Array (66 SNPs genotyped and 247 SNPs imputed to the 1000 genome phase 3 reference panel). Three PRSs were calculated based on 313 SNPs for overall, ER+ and ER- disease using log odds ratios (OR) published by the Breast Cancer Association Consortium (BCAC)[1]. Multivariable logistic regression was used to estimate ORs and 95% confidence intervals (95% CI) between percentiles of PRS and breast cancer risk. Results: The mean PRS was higher in GBHS controls than BCAC controls for overall PRS (GBHS: mean= 0.14, standard deviation (SD)=0.51; BCAC: mean=-0.42, SD=0.61), ER+ PRS (GBHS: mean= 0.10, SD=0.55; BCAC: mean=-0.41, SD=0.65), and ER- PRS (GBHS: mean= 0.30, SD=0.51, BCAC: mean=-0.31; SD=0.59). The shift in mean was explained by protective alleles tending to be more common in BCAC and risk alleles with the largest effect sizes tending to be more common in GBHS. The OR per SD in GBHS was 1.22 (95% CI=1.13-1.33) for overall, 1.34 (1.19-1.52) for ER+, and 1.14 (1.00-1.29) for ER- disease. These risk estimates are lower than previously reported in BCAC [1]. Based on the distributions of each of the three PRSs in the GBHS controls, women at the top 5% PRS percentile compared with women at average risk (40-60% PRS percentile) were at higher risk for overall (OR=2.17 (95% CI=1.51-3.11)), ER+(2.71 (1.63-4.50)), and ER- disease (1.47 (0.81-2.49)). On the other hand, women at the lowest 5% PRS percentile were not at a significantly decreased risk, as would have been expected, for overall (1.11 (0.73-1.69)), ER+(0.91 (0.46-1.83)), and ER- disease (0.99 (0.55-1.80)). Conclusion: Our results show that the breast cancer 313-SNP PRS developed in women of European ancestry predicts risk of breast cancer in Ghanaian women; however, the level of risk stratification is smaller, particularly for women with lower PRSs. These results are likely due to differences in linkage disequilibrium, allele frequencies, and/or genetic architecture between populations. Further work to develop population-specific PRSs is needed to improve risk stratification in women of West African ancestry. 1. Mavaddat, N., et al., Polygenic Risk Scores for Prediction of Breast Cancer and Breast Cancer Subtypes. Am J Hum Genet, 2018. Citation Format: Thomas U. Ahearn, Andriy Derkach, Joel Yarney, Beatrice Wiafe Addai, Baffour Awuah, Louise Brinton, Nilanjan Chatterjee, Jonine D. Figueroa, Montserrat Garcia-Closas. Performance of breast cancer polygenic risk score (PRS) in a Ghanaian population [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr LB-178.