Abstract B031: Spatial profiling reveals distinct tumor and myeloid ecosystems across lung and liver metastases in pancreatic ductal adenocarcinoma

Abstract

Pancreatic ductal adenocarcinoma (PDA) exhibits marked organotropism, with the liver and lung representing the most common sites of metastatic recurrence. While PDA most often metastasizes to the liver, patients with only lung metastasis have a better overall prognosis, highlighting a role for both tumor-intrinsic programs and tissue-specific microenvironments in disease progression. Among the immune components shaping these processes, myeloid cells are particularly influential, as they regulate immune suppression, tissue remodeling, and metastatic niche formation across tissues. While recent studies have identified transcriptional differences across metastatic lesions, the relationship between tumor cell state and local myeloid composition remains incompletely understood. To characterize site-specific metastatic ecosystems, we analyzed a cohort of human liver and lung PDA metastases using the Akoya Phenocyler spatial proteomics platform (n = 12 patients, 6 with lung metastasis and 6 liver metastasis). Tumor cells were classified using established molecular subtype markers, including GATA6 for classical tumors and CK17/S100A2 for basal-like tumors. Myeloid populations were identified using lineage and functional markers, enabling characterization of macrophage, dendritic cell, monocyte, and granulocyte phenotypes across metastatic sites. We observed marked differences in tumor cell composition between organs. Specifically, we showed lung metastases are enriched for classical tumor cells whereas liver metastases are enriched for basal tumor cells – consistent with published findings. In parallel, myeloid phenotype analysis showed distinct polarization states across metastatic sites. Lung metastases were enriched for immune-active myeloid populations, including inflammatory macrophage states, whereas liver metastases exhibited increased representation of suppressive myeloid populations, including tolerogenic dendritic cells. To further investigate mechanisms underlying metastatic organotropism, ongoing analyses integrate our proteomics analyses with matched spatial transcriptomic data. These studies will define transcriptional programs associated with site-specific metastatic progression and identify interactions between tumor and myeloid populations that may contribute to divergent metastatic evolution. Collectively, our findings reveal distinct metastatic microenvironments in liver and lung characterized by coordinated differences in tumor subtype composition and myeloid phenotype. These results highlight the importance of local immune landscapes in shaping metastatic disease biology and provide a foundation for identifying mechanisms that drive organ-specific progression in PDA.