%0 Journal Article
%A Gellrich, Frank Friedrich
%A Hufnagel, Cosima
%A Funk, Alexander M
%A Jonas, Sophie
%A Altmann, Heidi
%A Hobelsberger, Sarah
%A Steininger, Julian
%A Feige, Christina
%A Tasdogan, Alpaslan
%A Chavakis, Triantafyllos
%A Beissert, Stefan
%A Meier, Friedegund
%A Mirtschink, Peter
%A Steiner, Gerald
%T ¹H-NMR serum metabolomic profiling from clinical routine identifies signatures of progressive melanoma metastasis.
%J Scientific reports
%V nn
%@ 2045-2322
%C [London]
%I Springer Nature
%M DKFZ-2026-00254
%P nn
%D 2026
%Z #DKTKZFB9# / #NCTZFB9# / epub
%X Early detection of active melanoma metastasis is crucial. Serum metabolomics may offer non-invasive biomarkers, but real-world applicability needs validation. This study aimed to identify ¹H-NMR-based serum metabolic signatures for active metastasis in a large clinical cohort. Serum from 963 melanoma patients (1698 samples) underwent ¹H-NMR spectroscopy. Patients were classified by active metastasis status. OPLS-DA and RFE followed by logistic regression models were developed on a patient-level training/test split. Subgroup analyses assessed signatures related to Immune Checkpoint Inhibitor (ICI) therapy, brain metastases, and BRAF status. Models for active metastasis showed moderate test set discrimination (Area Under the Curve [AUCs]: OPLS-DA 0.609, RFE 0.630). The RFE-model highlighted seven significant metabolites: increased pyruvate, phenylalanine, acetoacetate, glutamate, glucose, and decreased histidine and citrate were associated with active metastasis. OPLS-DA yielded concordant metabolites. Subgroup analyses revealed distinct metabolic associations, e.g., for ICI therapy (citrate, RFE AUC 0.721) and BRAF status (acetate, RFE AUC 0.655), but limited performance for brain metastases (RFE AUC 0.553). ¹H-NMR serum metabolomics detects systemic metabolic alterations of active melanoma metastasis with moderate accuracy in a real-world setting. Identified disruptions in energy and amino acid metabolism offer pathobiological insights and warrant investigation for multimodal biomarker panels.
%K Biomarkers (Other)
%K Melanoma (Other)
%K Metabolic profiling (Other)
%K Predictive model (Other)
%K Tumor metabolism (Other)
%F PUB:(DE-HGF)16
%9 Journal Article
%$ pmid:41617901
%R 10.1038/s41598-026-37118-5
%U https://inrepo02.dkfz.de/record/309604