Metabolomic investigation of urinary extracellular vesicles for early detection and screening of lung cancer

• Published in: Journal of nanobiotechnology Vol. 21; no. 1; pp. 153 - 10
• Main Authors: Yang, Qinsi, Luo, Jiaxin, Xu, Hao, Huang, Liu, Zhu, Xinxi, Li, Hengrui, Yang, Rui, Peng, Bo, Sun, Da, Zhu, Qingfu, Liu, Fei
• Format: Journal Article
• Language: English
• Published: England BioMed Central Ltd 16.05.2023; BioMed Central; BMC
• Subjects: Analysis; Antibodies; Benzenoids; Biomarkers; Biomarkers, Tumor - metabolism; Carcinogenesis; Carcinogens; Care and treatment; Carnitine; Cell proliferation; Chromatography; Diagnosis; Early Detection of Cancer; Early diagnosis; Extracellular vesicles; Extracellular Vesicles - metabolism; Gene expression; Genetic aspects; Health aspects; Humans; Immune response; Invasiveness; Lipids; Lung cancer; Lung cancer, Non-small cell; Lung Neoplasms - diagnosis; Lung Neoplasms - metabolism; Machine learning; Marker panels; Metabolism; Metabolites; Metabolomics; Metastases; Organic acids; Proteins; Reproducibility of Results; Tumors; Urine; Vesicles; China; United Kingdom > UK; United States > US; Extracellular vesicles; Metabolomics; Early diagnosis; Lung cancer
• DOI: 10.1186/s12951-023-01908-0
• ISSN: 1477-3155

Lung cancer is a prevalent cancer type worldwide that often remains asymptomatic in its early stages and is frequently diagnosed at an advanced stage with a poor prognosis due to the lack of effective diagnostic techniques and molecular biomarkers. However, emerging evidence suggests that extracellular vesicles (EVs) may promote lung cancer cell proliferation and metastasis, and modulate the anti-tumor immune response in lung cancer carcinogenesis, making them potential biomarkers for early cancer detection. To investigate the potential of urinary EVs for non-invasive detection and screening of patients at early stages, we studied metabolomic signatures of lung cancer. Specifically, we conducted metabolomic analysis of 102 EV samples and identified metabolome profiles of urinary EVs, including organic acids and derivatives, lipids and lipid-like molecules, organheterocyclic compounds, and benzenoids. Using machine learning with a random forest model, we screened for potential markers of lung cancer and identified a marker panel consisting of Kanzonol Z, Xanthosine, Nervonyl carnitine, and 3,4-Dihydroxybenzaldehyde, which exhibited a diagnostic potency of 96% for the testing cohort (AUC value). Importantly, this marker panel also demonstrated effective prediction for the validation set, with an AUC value of 84%, indicating the reliability of the marker screening process. Our findings suggest that the metabolomic analysis of urinary EVs provides a promising source of non-invasive markers for lung cancer diagnostics. We believe that the EV metabolic signatures could be used to develop clinical applications for the early detection and screening of lung cancer, potentially improving patient outcomes.