Abstract
INTRODUCTION: The role of specific immune cell types within the tumor immune microenvironment in non-small cell lung cancer survival is unclear. The potential of these immune cells to become predictive biomarkers of prognosis, and to define subpopulations who will benefit of additional treatment is urgently needed. METHODS: Stage I to IIIA non-small cell lung cancer patients who underwent surgical resection were queried from the Cancer Genome Atlas; RNAseq data as well as clinical information was extracted. Sample-specific scores for different immune cells were computed via xCell. The association between each cell type and survival was assessed with Cox regression, both unadjusted and adjusted for sex, stage, smoking status, and tumor purity. Models were stratified by lung adenocarcinoma and lung squamous cell carcinoma. RESULTS: There were 383 lung adenocarcinoma and 328 lung squamous cell carcinoma samples, and 161 (42%) and 124 (38%) deaths respectively. There was no association between any immune cell infiltrations and survival in the combined unadjusted Cox regression model. After adjustment, the presence of CD8+ cytotoxic T cells (adjusted hazard ratio [HRajd]: 0.84; 95% confidence interval [CI]: 0.71-0.99; P=0.03), CD4+ helper T cells (HRajd: 0.79; 95% CI: 0.66-0.95; P=0.01) and CD20+ B cells (HRajd: 0.80; 95% CI: 0.66-0.97; P=0.02) were significant predictors of decreased risk of death. CONCLUSIONS: This study shows that the adjustment for clinical characteristics is key when evaluating tumor immune infiltration and its association with cancer outcomes. Adjustment for confounding factors modified the prognostic significance of specific immune cell populations in early-stage surgically resected NSCLC cases; clinical attributes may have high relevance on immune infiltration composition.