Immunophenotyping TCF1-expressing TILs: spatial profiling and prognostic value in operable non-small cell lung cancer

The spatial distribution and functional heterogeneity of tumor-infiltrating lymphocytes (TILs) significantly impact patient outcomes in non-small cell lung cancer (NSCLC). While T cell factor 1 (TCF1) expressing TILs have emerged as key players in sustaining anti-tumor immunity, their subset characterization, localization, and clinical significance within the tumor microenvironment remain poorly defined.

We performed multiplex immunohistochemistry and immunofluorescence to characterize TCF1⁺ immune cell subsets, in 102 NSCLC tumors, separately analyzing the tumor center (TC) and invasive front (IF). We integrated this data with publicly available single-cell RNA-sequencing datasets and clinical outcome analyses.

CD4⁺ T cells and CD79α⁺ B cells, dominate the TCF1⁺ landscape, while CD8⁺ T cells constitute a minority of TCF1⁺ immune cells, particularly in the TC. We demonstrated the presence of tumor-infiltrating IgG⁺/IgA⁺ plasma cells co-expressing TCF1. PD1⁺TCF1^- cells were more frequent than PD1⁺TCF1⁺ cells both in the TC and IF, reflecting that terminally differentiated exhausted TILs predominate within the tumor microenvironment. Survival analyses revealed significantly different prognostic impact of TILs including TCF1-expressing cells based on topography. Multivariate analysis showed that increased CD8⁺TCF1⁺ cells (HR: 2.5; p=0.039) and increased TCF1 expression by cancer cells (HR: 2,7; p=0.041) in the TC and CD4⁺TCF1⁺ cells (HR: 0.4; p=0.043) in the IF emerged as negative and positive independent prognostic markers for Disease-free survival (DFS), respectively. Integrating PD-L1 expression with TILs, PD-L1 immunopositivity was correlated with increased CD8⁺ and PD1⁺TCF1^- cell infiltration and was associated with favorable DFS especially in the TC.

Our findings support a more refined framework for TCF1⁺ TIL assessment and TCF1 expression across cellular populations in the tumor microenvironment, with implications for prognostication in operable NSCLC.