Dynamic Changes and Clinical Significance of Gut Microbiota and Serum Metabolites in Breast Cancer Onset, Progression and Chemotherapy Intervention

Objective Gut microbiota dysbiosis and host metabolic reprogramming are closely associated with breast cancer development and treatment response, yet integrative gut microbiome–metabolome studies spanning benign breast disease, malignant transformation, and the post-chemotherapy stage remain limited. This study aimed to systematically characterize dynamic changes along the gut microbiota–serum metabolome–breast tumor axis from benign breast disease (BBD) to breast cancer (BC) and post-chemotherapy breast cancer (PCBC), and to evaluate their diagnostic and disease-monitoring potential. Methods We enrolled 295 female participants, categorized into BBD (n = 83), BC (n = 100), and PCBC (n = 88) groups, including 24 patients who provided paired fecal samples before and after chemotherapy. Fecal samples underwent 16S rRNA sequencing, and serum samples were analyzed using LC–MS/MS-based untargeted metabolomics. Gut microbial diversity, taxonomic composition, and predicted functional profiles were compared across groups, and differential metabolites and enriched metabolic pathways were identified. In a subset with paired multi-omics data (BBD n = 19, BC n = 31, PCBC n = 34), Spearman correlation analysis, multi-omics PCA/PLS-DA, and random forest modeling were applied to integrate microbial and metabolic features. Results α-diversity did not differ among BBD, BC, and PCBC, but β-diversity showed clear community separation. BBD was enriched in SCFA-producing genera (Faecalibacterium, Roseburia), BC in inflammation/tumor-associated genera (Blautia, Fusobacterium, Sneathia, Prevotella), and PCBC in opportunistic pathogens (Phocaeicola, Sutterella, Enterococcus, Chlamydia), indicating a shift toward an inflammation- and pathogen-dominated dysbiosis. Untargeted metabolomics (> 3,000 metabolites) revealed distinct benign–malignant and pre–post chemotherapy profiles, mainly involving amino acid, lipid, purine, and drug metabolism. BC showed broad metabolic disruption with widespread depletion, while PCBC partially normalized these changes and activated estrogen, bile acid, and drug-metabolism pathways, yet retained persistent perturbations in purine/nucleotide, carbon, and amino acid metabolism. Key metabolites (torsemide, cortolone-3-glucuronide, trimethylselenonium) achieved AUC > 0.75, supporting diagnostic and treatment-monitoring potential. Correlation networks indicated a transition from probiotic–energy/amino acid/polyphenol coupling to a new state driven by pro-inflammatory/opportunistic taxa with lipid reprogramming, xenobiotic metabolism, and oxidative stress. Integrative multi-omics models improved discrimination across stages and nominated paired microbial–metabolite biomarker candidates, including anaerobic Gram-positive cocci and Lactobacillus-related taxa. Conclusion By integrating gut microbiome and serum metabolomics, we show that the gut microbiota–host metabolism–breast tumor axis is progressively reshaped from benign breast disease to breast cancer and after chemotherapy, marked by commensal depletion, opportunistic expansion, and energy/lipid metabolic reprogramming, with a persistent post-chemotherapy imprint featuring drug-metabolism and oxidative-stress activation. Multi-omics models improved stage discrimination, and the combined bacterial genera–metabolite signatures may support diagnostic stratification, treatment monitoring, and microbiome-targeted intervention in breast cancer.