Mapping pesticide mixtures to cancer risk at the country scale with spatial exposomics.

Despite decades of concern over the carcinogenic potential of agricultural pesticides, toxicological studies relying on single endpoints have yet to establish a definitive link between environmental pesticide exposure and cancer in real-world contexts. Here we use an integrative spatial Bayesian framework that merges high-resolution environmental pesticide risk modelling with comprehensive cancer registry data to map pesticide-linked cancer clusters in Peru with unprecedented precision. Our process-based model, encompassing 31 key pesticide active ingredients, together with an innovative stratification of cancer cases by developmental lineage, reveals a robust spatial association between environmental pesticide exposure risk and cancer incidence. In pesticide-associated cancer hotspots, exposomic profiling of liver tissue-a primary target of chemical carcinogens-uncovers a distinct transcriptomic signature of pesticide exposure, implicating a non-genotoxic mode of action that disrupts core regulatory circuitries sustaining cell identity. Collectively, these findings strongly support a mechanistic link between pesticide exposure and cancer, challenging assumptions of human non-carcinogenicity derived from reductionist experimental models. This study redefines the exposome as a lineage-conditioned, mechanistically tractable framework and shows how complex pesticide mixtures can contribute to carcinogenic trajectories, with profound and far-reaching implications for global health policy and socio-ecological equity.