Serum 1H NMR-Based Untargeted Metabolomics for Noninvasive Early Detection of Breast Cancer

Breast cancer (BC) is the most common malignancy among women, and late-stage presentation remains common in Asia, highlighting the need for affordable and minimally invasive adjuncts to imaging. This study evaluated whether fasting serum 1H NMR metabolomics could distinguish Malaysian women with newly diagnosed treatment-naïve BC from healthy controls and identify perturbed metabolic pathways. After spectral quality control exclusion, 101 participants were analyzed (BC, n = 44; control, n = 57). PCA and oOPLS-DA showed group separation, and 14 metabolites were deferred significantly after false discovery rate correction. Endogenous metabolites evaluated in BC included 3-hydroxybutyrate, citrate, glutamine, malonate, phenylalanine, dimethyl sulfone, myo-inositol, and L-carnitine, and the levels of formate, tyrosine, histamine, and p-hydroxyphenylacetate were reduced. Acetylsalicylate and salicylate were classified as exogenous aspirin-related markers and were excluded from the endogenous biomarker panel. Pathway analysis indicated disruptions in the tricarboxylic acid (TCA) cycle, glyoxylate/dicarboxylate metabolism, butanoate metabolism, and amino acid pathways. Metabolite-based classifiers achieved 0.901 accuracy on an internal held-out test partition (AUC 0.942) and 0.902 ± 0.061 accuracy in repeated stratified 5-fold cross-validation. These internally validated findings define a BC-associated serum metabolomic signature in a Southeast Asian cohort and warrant independent external validation.