# Multi-Modal Metabolomics Deciphers Pan-Cancer Metabolic Landscapes and Spatial-Niche-Specific Alternations

## TL;DR

This study uses multi-modal metabolomics to uncover shared and cancer-specific metabolic patterns across tumors and their microenvironments.

## Contribution

The study introduces a pan-cancer, multi-scale integrative approach combining bulk and spatial metabolomics to reveal metabolic heterogeneity.

## Key findings

- Metabolic features vary both across cancer types and within spatial niches of individual tumors.
- Nineteen metabolites, including carnitine species, show consistent alterations across bulk, spatial, and blood samples.
- Spatial metabolomics reveals unique metabolic modules associated with tumor microenvironment architecture.

## Abstract

Background: Metabolic reprogramming is a hallmark of cancer and supports tumor growth and adaptation within the tumor microenvironment (TME). The complexity of this reprogramming manifests as both distinct variations across cancer types and spatial heterogeneity within individual tumors. The specificity of these metabolic alterations, whether to cancer type, spatial niche, or as shared features, remains unclear, highlighting a critical gap in our systematic, pan-cancer understanding of metabolic reprogramming. Methods: We integrated bulk metabolomics and spatial metabolomics to investigate pan-cancer metabolic features and used blood-based metabolomics and spatial transcriptomics data to validate key findings. Metabolic differences were compared between tumor and normal tissues across multiple cancer types at the bulk level to identify metabolic modules shared across cancers or specific to individual cancer types. A two-step clustering framework was applied to identify both local and global TME-associated spatial metabolic modules of spatial metabolomics data from various tumor tissue slices. Results: We have identified a spectrum of metabolic features, including those specific to individual cancer types or spatial architectures and others shared across cancers, with some features emerging only at bulk-level and others uniquely discernible through spatial metabolomics. Integrative analyses also identified 19 metabolites consistently altered in both bulk and spatial data, especially carnitine species, which also showed concordant changes in blood samples and spatial associations with genes involved in fatty acid metabolism. Conclusions: This pan-cancer, multi-scale integrative analysis highlights substantial metabolic heterogeneity within the TME and across cancer types and identifies metabolites with consistent alterations across analytical layers, providing candidate features for future studies of tumor metabolism and potential metabolic biomarkers.