Breast cancer remains the most frequently diagnosed cancer among women worldwide, and identifying modifiable environmental risk factors is a public health priority. Ambient air pollution has been classified as carcinogenic, yet its contribution to breast cancer risk is still not fully understood. Air pollution is a complex mixture of particles and chemically diverse constituents with potentially distinct biological effects (e.g., oxidative stress, inflammation, endocrine disruption). Recent classification approaches can identify multi-pollutant exposure profiles that may better capture exposure patterns and shared emission sources than single-pollutant analyses. We therefore investigated associations between fine particulate matter (PM) constituents and breast cancer risk using a multi-exposure clustering approach.
 

To identify multi-pollutant PM constituent profiles and assess their association with incident breast cancer risk.
 

 

We conducted a nested case–control study within the French E3N-Generations cohort, including 5,222 incident breast cancer cases and 5,222 matched controls, with geocoded residential histories (1990-2010). Annual average exposures to multiple PM constituents (ammonium, nitrates, sulfates, black carbon, cadmium, benzo[a]pyrene (BaP), dioxins, polychlorinated biphenyls (PCBs), and dust) were estimated using a chemistry transport model (CHIMERE) and averaged over the study period. We applied Bayesian Profile Regression (BPR) to cluster participants into multi-pollutant exposure profiles while jointly estimating profile-specific breast cancer risk. Models accounted for the case–control design by adjusting for matching factors and additional potential confounders. Odds ratios (ORs) and 95% credible intervals (CrIs) were estimated, using the lowest-exposure profile as the reference.
 

Sixteen multi-pollutant exposure profiles were identified, displaying spatial patterns consistent with major emission sources (agriculture, industrial activities, dust, and urban traffic/incineration). Compared with the lowest-exposure reference profile, modest associations were observed for agricultural profiles characterized by high levels of ammonium and nitrates exposure (e.g., Profile 5: OR=1.11; 95% CrI: 1.01–1.24). Stronger associations were found for urban–industrial profiles characterized by elevated black carbon, cadmium, and dioxins exposure, including Profile 7 (North-East France: OR=1.25; 95% CrI: 1.09–1.46), Profile 12 (Lyon area: OR=1.33; 95% CrI: 1.18–1.50), and Profile 13 (Paris metropolitan outskirts: OR=1.27; 95% CrI: 1.16–1.39). The highest risks were observed for central Île-de-France profiles with extreme black carbon, cadmium, dioxins, and PCBs, and high benzo[a]pyrene, notably Profile 14 (Paris: OR=1.46; 95% CrI: 1.28–1.69) and Profile 16 (East/South-East Paris area: OR=1.42; 95% CrI: 1.25–1.66).
 

Urban–industrial multi-pollutant profiles, likely reflecting traffic and incineration-related sources, showed elevated estimated breast cancer risks compared with the lowest-exposure reference profile. These findings support the added value of multi-exposure, profile-based approaches in environmental epidemiology and provide insight into how complex mixtures and source-related patterns may shape breast cancer risk.