Background

Breast cancer diagnosed at a young age (e.g., <50 years) is more likely to present at an advanced tumour stage and has a higher mortality. However, the factors driving early-onset cancer are largely unknown, resulting in significant challenges in identifying who is at risk and developing preventive strategies. 

Methods

We conducted a research program on early-onset breast cancer, integrating epidemiological, family, omics and linked medical data for up to 7 million women from large-scale cohorts, research consortia, and biobanks. We applied contemporary and novel analytical approaches to investigate the causes and risk factors for early-onset breast cancer. 

Results

In the US, the incidence of early-onset breast cancer increased by 3% annually since 1992, while the incidence of cancer diagnosed >50 years remained stable. Compared with women born in 1950-1959, those born in 1990-1999 had 31% (95% CI: 17%, 47%) increased breast cancer risk. The 313-variant polygenic risk score explained 5% familial risk of early-onset breast cancer (c.f., 20% for late-onset cancer), without evidence of a causal effect at a young age. We identified 40 novel genomic regions implicated in breast cancer risk, half of which had a greater association with early-onset disease. Regarding non-genetic factors, exposure to CT scan ionising radiation before age 20 (excessive relative risk per 10mGy: 0.24; 95% CI: 0.05, 0.47) and greater lifetime endogenous oestrogen exposure (OR per standard deviation: 1.25; 95% CI: 1.14, 1.38) were associated with increased risk. Mendelian randomisation analysis found 83 exposures with a causal effect on breast cancer risk, with schizophrenia, triglyceride levels in high-density lipoprotein, body mass index having a stronger effect for early-onset disease.

Conclusion

Our program provides novel insights into the epidemiology, genetic susceptibility, causes and risk factors for early-onset breast cancer, with important implications for age-specific breast cancer risk assessment and prevention.