Single-cell pancancer analysis reveals epithelial–immune crosstalk in shaping the premetastatic niche of liver metastasis

Liver metastasis is a frequent and clinically important event across multiple solid tumors, yet a cross-cancer single-cell description of the metastatic immune ecosystem remains incomplete. Here, we integrated single-cell RNA-sequencing data from primary tumors and liver metastases across 12 cancer types and combined cell-state annotation, pseudotime ordering, ligand–receptor inference, and multiplex immunofluorescence as orthogonal tissue-level support. We identified a hepatocyte-like malignant epithelial state enriched in liver metastases, characterized by coexpression of hepatic genes, epithelial markers, oncogenic transcripts such as MYC and KRAS, and inferred copy-number-variation signals. CellChat analysis suggested that this epithelial state may engage RGS1-high and LILRB5-high macrophage states through a CCL3–CCR1-associated signaling axis. Reprocessed tissue imaging supported spatial proximity between TAT+/EPCAM+ epithelial cells and these macrophage populations across several tumor types. In addition, macrophage and T-cell composition differed between primary tumors and liver metastases, and sample-level associations linked RGS1-high macrophages to selected CD4 + T-cell states and LILRB5-high resident-like macrophages to selected CD8 + T-cell states. Overall, this study provides a pancancer descriptive atlas of epithelial–immune programs associated with liver metastasis and nominates candidate cellular associations for future functional testing.