来自Xenium RNA和蛋白质的无量纲、零校准空间指数用于乳腺IDC和肺腺癌

Spatial features of the tumor-immune microenvironment, including immune exclusion at tumor borders, tertiary lymphoid structures, and checkpoint engagement, shape local control and treatment response, yet routine single-marker IHC or bulk assays miss single-cell context and cell-cell interactions. Same-section spatial multi-omics now measure transcriptome (RNA) and proteome (protein) in interacting cells. Compact metrics are designed from these data to compare regions and cases. Two Xenium slides from FFPE tissue were profiled, each including paired primary tumor (PT) and adjacent normal (PN) regions from breast invasive ductal carcinoma and lung adenocarcinoma using a Human Immuno-Oncology panel plus six protein subpanels. Regions of interest (ROIs) were the analysis unit. Cell types were annotated with SingleR against public single-cell references, and spatial morphologies were inspected. Segmentations showed expected differences: PT breast was dominated by luminal epithelium with stromal and immune heterogeneity, whereas PN breast was fibroblast rich. Lung samples contained epithelial, fibroblast, endothelial, and immune populations with shifts between PT and PN, providing diverse microenvironments on which spatial indices were computed. Three dimensionless indices quantified complementary features of the tumor-immune microenvironment. The Exclusion Index (EI) integrates reduced CD8 T-cell penetration into tumor nests with peritumoral macrophage enrichment and diminished antigen-presenting-cell ingress. The TLS Score identifies B/T aggregates consistent with tertiary lymphoid structures, supported by local chemokine signal and simple maturity surrogates. The Checkpoint Contact Index (CCI) quantifies adjacency between PD-1-positive T cells and PD-L1-positive tumor or myeloid neighbors within a small radius compared with ROI label-shuffle spatial nulls. For all indices, values were centered on PN regions within each case, converted to z-scores within tumor type, and scaled 0-1. The pipeline produced coherent cell-typing and index summaries that aligned with visible histology and tumor-normal heterogeneity, providing a quantitative complement to descriptive reads. Grounding the indices in same-section RNA and protein and explicit spatial nulls yields a practical framework for reporting tumor-immune microenvironment on small-N slides and a common language transferable across studies and pathology teams.