Macrophage state in neoadjuvant and perioperative immunotherapy resistance in resectable NSCLC: a critical appraisal of strategies from CSF1R and CD47–SIRPα to CAR-macrophages

Neoadjuvant or perioperative chemoimmunotherapy has become a standard treatment option for selected patients with resectable non-small cell lung cancer (NSCLC), with eligibility determined by tumor stage, nodal involvement, oncogenic-driver status, regimen, and jurisdiction. Even so, most patients do not achieve a pathological complete response, and the determinants of resistance in early-stage disease remain incompletely understood. Tumor-associated macrophages are a key, and targetable, component of that resistance. We argue that abundance alone is an insufficient predictor of whether the myeloid compartment supports or suppresses a checkpoint-dependent response, and that macrophage state, namely the dominant subset, its spatial niche, and its trajectory under therapy, adds explanatory and therapeutic information whose weight remains to be established. Single-cell and spatial studies have identified metabolically distinct SPP1 + , CCL18 + , and oncofetal STAB1 + states associated with immunosuppressive niches in NSCLC, although these labels vary between datasets and their functional assignments are provisional. Limited serial neoadjuvant data from a single cohort suggest that treatment expands tissue-resident macrophages and shifts tumor-associated macrophages toward a neutral rather than antitumor phenotype, most evidently in tumors that reached a major pathological response. Current agents, including CSF1R and CD47–SIRPα blockade, TREM2, LILRB2 and PI3Kγ inhibitors, and chimeric antigen receptor (CAR) macrophages, act on programs that cut across these subsets rather than on the subsets themselves, and almost all efficacy evidence derives from advanced or preclinical disease. The resectable, biomarker-rich perioperative window is nonetheless the most informative setting in which to test subset-informed myeloid reprogramming.