Pan-cancer analysis of spatial transcriptomics reveals heterogeneous tumor spatial microenvironment

Tumors are complex systems comprising diverse cell types that form the tumor spatial microenvironment (TSME). We present a pan-cancer spatial transcriptomic analysis of 373 samples across 12 cancer types and identify 56 local cellular programs (LCPs) and 13 recurrent niches. Ligand-receptor analysis reveals niche-shared and niche-specific interactions that drive spatial organization. Notably, gene expression in tumor cells and macrophages depends heavily on their specific location. Furthermore, niches associate significantly with clinical outcomes: macrophages colocalized with tumor cells (Niche₄) correlate with poor prognosis and immunotherapy resistance, while those colocalized with immune cells (Niche₁1) predict better survival and treatment response. This systematic dissection of the TSME provides deeper insights into cellular communication and the structural influences governing complex tumor ecosystems.