Predictive Metabolic Biomarkers in Breast Cancer: A Systemic Review of Metabolites and Lipids Identified Through Mass Spectrometry

Abstract

Background Metabolic reprogramming is a hallmark of breast cancer (BC) and varies across molecular subtypes, creating opportunities for biomarker discovery and therapeutic targeting. Methods We conducted a PRISMA-guided systematic review of mass spectrometry–based metabolomics and lipidomics studies on human BC published from 2005 to 2025 in PubMed and Scopus. Inclusion required LC–MS/GC–MS analyses reporting specific metabolite/lipid alterations with diagnostic, prognostic, predictive, or mechanistic relevance. Thirty-nine studies met criteria. Extracted variables included specimen type, analytical platform, key metabolites/lipids, changes, and pathways. Metabolite lists (n=173 overall; subtype analyses n=104; TNBC n=73) were processed using MetaboAnalyst 6.0 for pathway enrichment/topology. Results Most studies analysed blood samples (plasma 54%, serum 23%); 13% used tissue. Frequently altered features included decreased lysophosphatidylcholines, increased ceramides, glutamine depletion with elevated glutamate, perturbed branched-chain amino acids, and variable choline/phosphocholine; lactate was commonly elevated. Six pathways were most significantly dysregulated: one-carbon pool by folate, arginine biosynthesis, glutathione metabolism, alanine/aspartate/glutamate metabolism, glycine/serine/threonine metabolism, and the TCA cycle. Subtype analyses indicated greater disruption of amino-acid and redox metabolism in HER2-positive and triple-negative BC (TNBC). TNBC showed enrichment of one-carbon and pyruvate metabolism, consistent with glycolytic reliance and redox stress. Integrated evidence supports concurrent aerobic glycolysis and active mitochondrial/TCA anaplerosis, notably via glutamine/aspartate fuelling. Conclusions MS-based metabolomics and lipidomics consistently reveal BC-relevant metabolic signatures with diagnostic and therapeutic promise, including ceramides, LPCs, glutamine/glutamate, and one-carbon/redox pathway nodes. Heterogeneity in methodology and limited external validation particularly insufficient reporting of retention times— remain barriers to translation. Standardized reporting, multi-ethnic validation cohorts, and multi-omics integration are recommended to accelerate precision metabolic biomarkers and subtype-tailored interventions.