Spatial Transcriptomic Profiling Reveals a Fibroblast-Associated Extracellular Matrix Niche in KRAS G12C-Mutant Lung Adenocarcinoma Primary Tumors From Patients With Metastasis.

KRAS G12C is the most common KRAS alteration in lung adenocarcinoma and has emerging therapeutic relevance. However, the spatial programs within primary tumors that accompany metastasis remain poorly understood. We profiled primary tumor tissues from four patients with KRAS G12C-mutant lung adenocarcinoma, including two without metastasis and two with metastasis, using 10× Visium spatial transcriptomics. Spatial clustering, differential expression, functional enrichment, and Robust Cell Type Decomposition (RCTD)-based deconvolution using a published non-small cell lung cancer (NSCLC) single-cell atlas were performed. Candidate pathway relevance was assessed in The Cancer Genome Atlas (TCGA) KRAS G12C-mutant lung adenocarcinoma cases, and functional validation was performed using fibroblast-tumor co-culture assays. Spatial analysis identified 18 transcriptionally distinct clusters across 14,351 spots. In this exploratory cohort, metastatic-primary tumors showed descriptive expansion of selected spatial domains and reduction of epithelial/alveolar-associated regions. Deconvolution suggested reduced AT2-like signals and increased fibroblast-associated features in metastatic-primary tumors. Stepwise filtering of group-level, compartment-level and cluster-supported genes identified a fibroblast-associated extracellular matrix remodeling module consisting of COL1A1, COL1A2, COL3A1, SPARC, POSTN, and VCAN. In TCGA KRAS G12C-mutant lung cancer, VCAN expression correlated with collagen formation, extracellular matrix degradation, and EMT marker signatures. In vitro, VCAN-overexpressing MRC-5 fibroblasts enhanced H358 invasion without markedly altering short-term proliferation. These exploratory analyses identified a fibroblast-associated extracellular matrix remodeling niche associated with metastatic status in KRAS G12C-mutant lung adenocarcinoma. These preliminary findings support further investigation of primary-tumor stromal reprogramming in larger spatially profiled cohorts and mechanistic models.