Spatially-guided metabolomics profiling of heterogeneous human tumor tissues

However, intra-tumor heterogeneity (IntraTH) convolutes the cellular origins and tumor pathologies associated with the detected metabolites, thus making the elucidation of reproducible metabolic pathways and/or biomarkers very challenging. Here, we present “Spatially-guided MEtabolomics (SgME) profiling”, a multi-modal metabolomics data analysis approach to delineate IntraTH by integrating spatial and bulk metabolomics profiles from the same tumors.

We applied SgME profiling to 117 tumor and adjacent normal tissues from 26 surgically resected primary liver tumors, and constructed SgME maps of metabolic regions (MERs) associated with key histopathological features. We used these maps to survey IntraTH and train regression models that accurately predict the MER compositions of bulk tumor samples. We also discovered a group of putative metabolites that increase in low-grade tumor regions but abruptly decrease in necrotic regions.