Towards non-invasive diagnosis of glioblastoma: identifying metabolic biomarkers in liquid biopsies using a ROC-based approach

For over seven decades, the World Health Organization (WHO) has classified central nervous system tumors primarily based on histological criteria. While recent updates have incorporated molecular genetics and histochemistry, this approach still overlooks key aspects of tumor biology. Metabolomics has emerged as a promising tool to uncover metabolic alterations in glioblastoma, offering insights beyond current classification systems. However, this metabolic dimension remains largely underrepresented in WHO guidelines. This study highlights key glioblastoma-associated metabolites reported in the literature, particularly those with high diagnostic accuracy as measured by area under the curve values. From these, a panel of 21 metabolites was curated to develop a diagnostic model that could enhance diagnostic precision and clinical decision-making. In addition to its diagnostic focus, the collected findings reignite a long-standing hypothesis regarding shared metabolic traits between cancer and schizophrenia, suggesting a potential biological link that merits further exploration involving multidisciplinary approaches.