A comprehensive spatial proteome landscape of ectopic lymphoid-like structures in metastatic colorectal cancer reveals specific signatures for predicting clinical outcomes.

We previously generated a global bulk-level proteogenomic atlas of metastatic colorectal cancer (CRC) in a Chinese cohort (CCRC). Using formalin-fixed paraffin-embedded tissues from this cohort, we identified 2095 ectopic lymphoid-like structures (ELSs) across 132 primary tumors. Both lymphoid aggregates (Aggs) and tertiary lymphoid structures (TLSs) were significantly associated with metastasis and clinical outcomes in CRC, in which the spatial and compositional heterogeneity of ELSs and their clinical relevance remain unknown. Here, we upgraded our LCM-MTA platform to version 2.0 for deep spatial proteomics and profiled spatial proteomic landscapes from matched mCRC and non-mCRC cases with comparable baseline clinical features. This analysis revealed pronounced spatial heterogeneity in protein expression, distinguishing ELSs from parenchymal regions in paired tumor and para-carcinoma tissues. Notably, TLSs displayed distinct functional states: mCRC-associated TLSs showed enrichment of anti-tumor immune activation pathways, whereas non-mCRC TLSs were enriched in acute inflammatory response pathways. Based on these findings, we developed a 13-protein TLS signature and scoring system to predict metastasis and prognosis in the CCRC cohort, validated in two independent CPTAC cohorts. We present a comprehensive spatial proteomic map of ELSs in metastatic CRC and identify a TLS-based signature with strong potential for predicting metastasis and clinical outcomes.