Abstract Breast cancer is and remains among the most common types of cancer, estimated to be 16% of all new cancers diagnosed in 2025. While a variety of area-level social and environmental exposures have been implicated in breast cancer outcomes, comparatively less work has been done to study how multiple concurrent exposures to these factors impact outcomes. Importantly, a growing literature shows a high correlation of socioeconomic status (SES) and environmental exposures, highlighting the need for suitable exposure measurement methods which can account for this high correlation. Additionally, uncertainty in how many different SES and environmental factors may exist, an approach which does not require a priori specification of the number of clusters is also necessary. We identified Bayesian Profile Regression as a suitable clustering approach for our analysis which meets this set of needs. We utilized data from the United States Census American Community Survey 5-year estimates for 2020 and the United States Environmental Protection Agency Air Toxics Screening Assessment (AirToxScreen) estimates for 2020. Using area-level estimates of factors from all census tracts within Pennsylvania and New Jersey, we first identified clusters of SES risk factors—specifically, proportion of households that are single-headed, rented, crowded and without a vehicle, and proportion of people on public assistance, w/less than a high school diploma/equivalent (25 and older), unemployed (16 and older), and in poverty); identifying 21 unique neighborhood SES profiles, 11 of which had at least one component variable 1 standard deviation (SD) above the global mean. Using the same approach, we then identified clusters of environmental exposures, specifically, ambient exposure to: 8 volatile organic compounds, 4 heavy metals, and diesel particulate matter; identifying 29 unique neighborhood pollution profiles, 16 of which had at least one component variable that was 1 standard deviation above the global mean. When restricting to tracts within the Sidney Kimmel Comprehensive Cancer Center (SKCCC) catchment area, there was evidence of high global spatial autocorrelation for latent pollution clusters (Moran’s I of 0.82 (p.001) and SES clusters (Moran’s I of 0.39, p.001). Census tracts with poor SES profiles as well as tracts with greater than average pollution for one or more pollutants tended to cluster primarily in urban areas. By identifying area-level clusters of both environmental and SES risk factors for breast cancer, we can subsequently ask questions about how joint exposure to these factors may impact breast cancer incidence, mortality, survival, and stage at diagnosis. Importantly, by focusing on the smaller area level, these estimates and their impact on outcomes may also provide crucial evidentiary support for more targeted outreach efforts by the SKCCC Community Outreach and Engagement (COE) team and will inform research priorities for the region. Citation Format: Zachary H. Fusfeld, Terry Hyslop. Constructing latent clusters of joint exposure to census tract level socioeconomic and environmental breast cancer risk factors abstract. In: Proceedings of the 18th AACR Conference on the Science of Cancer Health Disparities; 2025 Sep 18-21; Baltimore, MD. Philadelphia (PA): AACR; Cancer Epidemiol Biomarkers Prev 2025;34(9 Suppl):Abstract nr A099.
Fusfeld et al. (Thu,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: