Background: Immunotherapy (IO) has improved response rates for patients with advanced clear cell renal cell carcinoma (ccRCC), but most will develop resistance. We sought to utilize cellular-level spatial transcriptomics in the IO naïve and IO exposed settings to better understand IO resistance in ccRCC tumor immune microenvironment (TIME). Methods: Tissue was obtained from primary ccRCC kidney tumors. Matched tumor and stromal fields of view (FOV) were included for analysis. Spatial molecular imaging (SMI) was obtained for three tissue microarrays using Nanostring’s CosMx platform. Cells were phenotyped using Insitutype and the Kidney Cell Atlas as a reference. T cells and macrophages were further subtyped using subclustering and differential gene expression. Tumor cells were phenotyped using differential gene expression of proximal tubule cells with high VEGF expression and a LASSO regression model. Cell abundance and clustering by phenotype were then analyzed by treatment status. Clustering of all cell types was quantified using univariate Ripley’s K. Radii between 9 and 90um were visually compared to identify an appropriate search distance; a final radius of 27um was selected. Spatial gene set enrichment (GSE) analysis followed by a post hoc spatial analysis of associated transcripts from select enriched gene sets were performed. Global Moran’s I test was used to quantify spatial autocorrelation of ligand-receptor (LR) pairs. Multiplex immunofluorescence (mIF) validation testing was performed using antibody markers against proteins from significant LR pairs in the autocorrelation analysis. Analysis was performed in R using the spatialTIME and sfdep packages. Results: 15 IO naïve and 6 IO treated patients were evaluated. Compared to IO naive tumors, IO exposed tumors harbored more CD8+ T cells and neutrophils in the stromal FOVs (p < 0.001 for both), and more non-classical monocytes in the tumor FOV (p = 0.002). No univariate clustering changes were seen following IO. On spatial GSE, the endothelial to mesenchymal (EMT) pathway was enriched and two associated LR transcript pairs were significantly autocorrelated; COL4A1 (gene for collagen IV) and ITGAV (gene for integrin αv-subunit) in the stroma (p=0.024). Expression of these genes were highest amongst fibroblasts and tumor cells. On mIF validation testing, integrin αv positive cells were more abundant in the IO exposed samples compared to IO naïve samples (p=0.004). Potential therapeutics that target this pathway have not yet been tested in ccRCC. Conclusions: We found a shift in the abundance of immune cells in the ccRCC TIME following IO treatment. Additionally, we saw significant autocorrelation of two transcripts associated with the EMT pathway, ITGAV and COL4A1, amongst fibroblasts and tumor cells. Increased abundance of integrin αv positive cells was confirmed on mIF validation testing.