Predictive Modeling of Early-Life Agrochemical Exposure and Pediatric Cancer Risk among Children of U.S. Farmers: A Narrative Review
Henry Okorie Ugorji *
Data Modernization Initiative, Public Health Division, Oregon Health Authority, USA.
Alex Nnanyelugo Egbuchiem
Department of Environmental, Agriculture and Occupational Health, College of Public Health, University of Nebraska Medical Center, United States of America.
Gifty Dudzilah
Department of Physical Science, Eastern New Mexico University, USA.
Rachael Boluwatife Oke
Department of Biological Sciences, Alabama Agricultural and Mechanical University, Huntsville, Alabama, USA.
Tosin Abiodun Aderanti
Department of Chemical and Environmental Sciences, Babcock University, Ilishan-Remo, Nigeria.
*Author to whom correspondence should be addressed.
Abstract
Children residing in agricultural settings are uniquely vulnerable to early-life exposure to agrochemicals, which may contribute to the development of pediatric cancers. This narrative review, incorporating systematic search and appraisal elements, synthesizes current evidence on the relationship between prenatal and early childhood agrochemical exposures and pediatric cancer risk, with a focus on predictive and spatial modeling approaches. A structured literature search was conducted across PubMed, Web of Science, Scopus, and Embase to identify epidemiologic, exposure modeling, and mechanistic studies published between 2000 and 2025. Epidemiologic evidence consistently implicates leukemia, particularly acute lymphoblastic leukemia, as the malignancy most strongly associated with early-life agrochemical exposure, while findings for central nervous system tumors and rarer pediatric cancers remain less consistent. Mechanistic studies support biological plausibility through pathways including genotoxicity, endocrine disruption, epigenetic reprogramming, and immune dysregulation during critical developmental windows. Predictive modeling approaches—ranging from traditional regression and Bayesian hierarchical models to machine learning and GIS-based methods—enhance exposure estimation, risk stratification, and identification of high-risk subgroups, though challenges persist related to exposure misclassification, chemical mixture assessment, and integration of mechanistic evidence. Overall, this review highlights key methodological gaps and research priorities, emphasizing the need for longitudinal, biomarker-validated studies and hybrid causal–predictive modeling frameworks. Advancing interdisciplinary, prevention-oriented research in this area is essential for informing evidence-based interventions, reducing pediatric cancer burden in agricultural populations, and promoting equitable environmental health outcomes.
Keywords: Pediatric cancer, early-life exposure, agrochemicals, predictive modeling, GIS, environmental epidemiology, agricultural health