Metabolomics and liver cancer risk: A systematic review of prospective studies

Introduction Emerging metabolomic studies have identified circulating metabolites associated with liver cancer, yet evidence remains inconsistent across populations. To address these gaps, we conducted the most comprehensive systematic review of prospective metabolomic studies to date, integrating results across different populations to identify metabolic signatures of liver cancer. Methods We followed Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines to report our findings. We systematically searched PubMed through January 2025, for prospective cohort or nested case-control studies examining pre-diagnostic circulating metabolomic profiles with liver cancer. Two reviewers independently screened, extracted, and evaluated studies. Due to heterogeneity in study design and outcome reporting, findings were synthesized descriptively. Direction and effect size of metabolite associations was summarized. Enrichment analysis was conducted for those metabolites significantly associated with liver cancer. Results Twenty-four prospective studies were included, of which eight nested case-control or cohort studies with 1,626 cases of liver cancer, providing risk estimates for per standard deviation (SD) increase. Seven studies used liquid chromatography-mass spectrometry and one used nuclear magnetic resonance. Across these studies, 98 unique metabolites were evaluated, and consistent metabolic alterations preceding liver cancer were observed. Elevated aromatic amino acids (tyrosine, phenylalanine), glutamate, arginine, and conjugated bile acids were positively associated with liver cancer risk with risk ratios (RRs for per SD increase) ranging from 1.49 to 5.82, whereas branched-chain amino acids (leucine and isoleucine), lysine, glutamine, and lysophosphatidylcholines showed inverse associations with RRs ranging from 0.15 to 0.75. Saturated phosphatidylcholine, sphingosine, and select amino acid derivatives (cystathionine, kynurenine) were positively associated (RRs, 1.62-3.32), while androgen sulfates and creatine were inversely related (RRs, 0.20-0.56). The direction was generally consistent across the majority of the 24 prospective studies. Three most enriched pathways were taurine and hypotaurine metabolism, valine/leucine and isoleucine biosynthesis, and arginine biosynthesis. Conclusions Our systematic review indicates that liver cancer risk was associated with increased aromatic amino acids, bile acid conjugation, and lipid remodeling, and decreased branched-chain amino acids and androgenic steroids. The identified metabolic signatures may inform future efforts in risk prediction, early detection, and targeted prevention for liver cancer.