The spatial immune geography of resectable non-small-cell lung cancer: ecosystem architecture of perioperative immunotherapy response and relapse

Perioperative immune-checkpoint blockade has improved pathological and event-free outcomes in resectable non-small-cell lung cancer (NSCLC), but benefit remains heterogeneous and current biomarkers do not reliably distinguish tumor eradication from transient control or future relapse. This limitation partly reflects how tumors are measured: programed death ligand 1 expression, tumor mutational burden and bulk immune-cell abundance compress a spatially organized ecosystem into summary values. Here, we define the spatial immune geography of resectable NSCLC across four coupled scales: intratumor territories, cellular neighborhoods, regional immune structures—including tertiary lymphoid structures and tumor-draining lymph nodes—and longitudinal transitions from pretreatment biopsy to resection, molecular residual disease and recurrence. We critically synthesize human evidence linking T-cell access, antigen-presentation circuits, B-cell and tertiary-lymphoid organization, and stromal, myeloid and hypoxic niches to perioperative immunotherapy response. We further distinguish pretreatment predictors from treatment-induced pharmacodynamic changes and explain why pathological response and biological cure are related but non-equivalent endpoints. On this basis, we propose a testable spatial-immune-resilience framework comprising eradication-permissive, persistence-prone and escape-permissive ecosystems. We also evaluate how multiplex imaging, spatial transcriptomics, digital spatial profiling, anatomically registered single-cell analysis and computational pathology differ in biological reach and translational readiness. Finally, we outline an assay-to-decision validation pathway that integrates standardized multiregion tissue sampling, tumor-draining lymph nodes and circulating tumor DNA. Spatial profiling is unlikely to transform perioperative care through maximal molecular resolution alone; its clinical value will depend on reproducible, parsimonious measurements that improve decisions beyond stage, pathology and molecular residual disease.