Polygenic and classical risk integration reshapes breast cancer screening recommendations in 9,761 Asian women
Related Experiment Videos
Jingmei Li1,2,3, Nicole Sze Ting Lim4, Chi Hui Chin5
1Genome Institute of Singapore (GIS), Agency for Science, Technology and Research (A*STAR), 60 Biopolis St, Genome, #02-01, Singapore, 138672, Singapore. lijm1@a-star.edu.sg.
Background:
Age-based mammography screening fails to account for individual breast cancer risk variation, particularly in Asian populations where prospective evidence is absent.
Methods:
The BREAst Screening Tailored for HEr (BREATHE) study is a multicentre prospective cohort of 4,513 women aged 35-59 years recruited across six Singapore clinical sites (2021-2023). Risk-stratified screening recommendations were assigned using a five-year absolute risk threshold of 0.9% (equivalent to the population-average risk of an average 50-year-old Singaporean woman) to initiate screening based on polygenic risk score (PRS) and Gail model, as well as mammographic density, family history, abnormal prior screening. Two independent breast cancer case cohorts provided comparative analyses: Singapore Breast Cancer Cohort (SGBCC, n = 2,782) and Malaysian Breast Cancer Genetic Study (MyBrCa, n = 2,466); combined n = 9,761.
Results:
Among 2,003 healthy women aged 40-49 years currently recommended annual mammography, risk stratification downgraded 776 (38.7%) to no screening and 960 (47.9%) to biennial as they were not in the top 2.5th percentile of PRS and Gail risk and did not have any additional risk indicators. Of 41 breast cancers diagnosed after BREATHE enrolment, 35 (85.4%) occurred in women assigned to screening-eligible categories at baseline. Risk stratification combining PRS and Gail model captured 68-69% of cases aged 40-49 in SGBCC (n = 1,184) and MyBrCa (n = 1,122), versus ~ 54% by PRS alone. Relative to cases meeting screening criteria for both PRS and Gail model, PRS-missed cases were enriched for progesterone receptor negativity (OR 1.24; 95% CI 1.01-1.52), tumour size ≥ 50 mm (OR 1.43; 95% CI 1.02-2.00), and germline protein-truncating variants (PTVs) in breast cancer predisposition genes (OR 1.53; 95% CI 1.03-2.29).
Conclusions:
Risk stratification combining PRS with classical risk factors reshapes mammography recommendations in Asian women. However, due to the limited number of events in BREATHE, this finding requires evaluation in larger prospective studies. PRS-missed cancers exhibit a biologically aggressive profile, defining the current limits of common-variant prediction and identifying the biology that future precision screening tools must address.