Discov Oncol. 2026 May 18. doi: 10.1007/s12672-026-05150-6. Online ahead of print.
ABSTRACT
BACKGROUND: The non-invasive detection of cancer via exhaled breath condensate (EBC) represents a promising frontier in oncology. This study explores the metabolic profiles of EBC to identify biomarkers for the early detection of thyroid cancer (TC) and breast cancer (BC), and to investigate potential metabolic interrelationships between them.
METHODS: We conducted gas chromatography-mass spectrometry (GC-MS) analysis of EBC samples from 74 individuals, including 65 cancer patients (TC and BC) and 9 non-cancer controls. Comparative statistical analyses and machine learning were employed to identify discriminant metabolites.
RESULTS: 305 metabolites were identified in total. Comparative analysis revealed 16 differential metabolites in cancer patients versus controls, with 14 specific to TC and 7 to BC. Notably, five metabolites were common to both cancers: 1,2-Bis(trimethylsilyl)benzene, 1,4-Phthalazinedione,2,3-dihydro-6-nitro-, Eicosane, Methyltris (trimethylsiloxy) silane, and Octadecane, highlighting metabolites that were commonly altered in both cancer types. ROC analysis demonstrated strong diagnostic potential: 1,2-Bis(trimethylsilyl)benzene effectively discriminated cancer from controls (AUC = 0.822) and identified TC (AUC = 0.866), while 1,4-Phthalazinedione,2,3-dihydro-6-nitro- detected BC (AUC = 0.783). Combinations of metabolites yielded AUCs > 0.7 for both cancers. However, limited discriminatory power was observed between TC and BC (maximum AUC = 0.663), indicating significant metabolic similarity. Furthermore, specific metabolite abundances correlated with conventional serum biomarkers, thyroid hormone levels, and lymphatic metastasis.
CONCLUSION: Our findings establish EBC metabolomics as a powerful, non-invasive tool for early cancer detection and monitoring. The identification of shared metabolic alterations between TC and BC suggests common metabolic features that warrant further investigation and paves the way for developing breath-based diagnostic assays.
PMID:42151663 | DOI:10.1007/s12672-026-05150-6
