Cancer Risk Linked to Ultra-Processed Foods: Epidemiological Findings, Potential Mechanisms, and Research Outlook

Epidemiological research on ultra-processed food (UPF) consumption and cancer has expanded rapidly; however, the evidence is uneven across cancer sites. Associations were relatively consistent for colorectal cancer, modest and partly attenuated after adjustment for breast cancer, and limited, single-cohort, or largely indirect for lung, pancreatic, renal, and prostate cancers. Recent umbrella reviews have rated the certainty of this evidence as low to very low, reflecting heterogeneous exposure classification, residual confounding, and inconsistent findings. This narrative review synthesizes the current epidemiological evidence on UPF consumption and overall and site-specific cancer risk, evaluates the biological plausibility of three mechanism clusters—obesity-driven metabolic and inflammatory pathways, processing-derived contaminants and food-contact chemicals, and food additives and their mixtures—while distinguishing human evidence from experimental models, and critically examines the NOVA classification as an exposure framework, including its application to non-Western diets. Compared with previous reviews, we appraised the evidence site by site, integrated the most recent umbrella and dose–response meta-analyses, and mapped each proposed mechanism to the tier of evidence supporting it. Standardized NOVA-based exposure assessment, repeated dietary measurements, integration of exposure biomarkers into prospective cohorts, and causal inference frameworks that distinguish confounding from mediation are prerequisites for evidence-based, UPF-specific cancer prevention recommendations.