Background
The carcinogenic potential of agricultural pesticides remains contentious, largely due to limitations of conventional toxicological paradigms that rely on single-agent, single-endpoint assessments. Real-world exposures involve complex mixtures, heterogeneous environmental dispersion, and context-dependent biological responses, collectively complicating causal inference and constraining effective public health action. Addressing this challenge calls for integrative frameworks capable of jointly interrogating environmental contamination, exposure heterogeneity, and molecular processes implicated in carcinogenesis.

Objectives
We aimed to develop and apply an integrative exposomic framework to characterise spatial, epidemiological, and molecular relationships between environmental pesticide exposure and cancer risk at the national scale in Peru. Specifically, we sought to identify cancer clusters associated with elevated environmental pesticide exposure, decipher mechanistic pathways of environmentally driven carcinogenesis, and assess broader socio-ecological implications for sustainability and health equity within a planetary health framework.

Methods
We implemented a high-resolution spatial Bayesian framework integrating process-based environmental risk modelling of 31 pesticide active ingredients with comprehensive, nationwide geocoded cancer registry data from Peru. Cancer cases were stratified by developmental lineage to enhance biological interpretability beyond conventional organ-based classifications. Bayesian spatial inference, implemented using integrated nested Laplace approximations (INLA), was applied to identify fine-scale spatial convergence between cumulative exposure risk and cancer incidence. In high-risk regions, exposomic profiling and transcriptomic analyses of liver tissue—a primary target site of chemical carcinogens—were conducted to characterise molecular signatures associated with pesticide exposure and to interrogate underlying regulatory networks.

Results
Our integrative approach revealed robust spatial associations between cumulative pesticide exposure risk and cancer incidence, identifying high-risk clusters across diverse agro-ecological landscapes. Lineage-based stratification uncovered ontogenetically coherent clustering patterns, notably highlighting a previously uncharacterised molecular subtype of liver cancer among Andean–Amazonian Indigenous populations. Transcriptomic analyses revealed a distinctive signature of chronic pesticide exposure, characterised by perturbation of lineage-specific master transcription factors and destabilisation of core regulatory circuits governing cell identity, implicating a non-genotoxic, mixture-driven mode of carcinogenesis. The convergence of geospatial, aggregated epidemiological, and molecular evidence provides compelling support for a mechanistic link between environmental pesticide exposure and cancer risk. Moreover, geospatial modelling indicated that high-risk zones disproportionately coincide with regions undergoing agricultural land-use pressure, environmental degradation, and socio-economic marginalisation, highlighting the interplay between environmental pollution, social inequity, and planetary health.

Conclusions/Implications for practice or policy
This study establishes a scalable and transferable exposomic framework that integrates geospatial analysis, cancer epidemiology, and molecular profiling to quantify cancer risk under real-world pesticide exposure scenarios. By transcending reductionist toxicology, our approach provides mechanistic insight into how complex pesticide mixtures disrupt lineage-dependent regulatory networks to promote carcinogenesis. These findings challenge prevailing assumptions of human non-carcinogenicity and argue for a paradigm shift in environmental risk assessment. Embedding exposomics within planetary health and sustainability frameworks is essential to inform evidence-based policy, mitigate environmental injustice, and protect vulnerable populations from environmentally driven cancers.