Prognostic Implications of Heterogeneity in Intra-tumoral Immune Composition for Recurrence in Early Stage Lung Cancer

Background: Studies in the past have identified selected immune cells that associate with different clinical outcomes in non-small cell lung cancer (NSCLC). Considering the fact that immune responses are heterogenous and that the clinical outcome could be influenced by the interplay of various immune cell types, it is imperative to evaluate multiple intra-tumoral immune cell types in the same set of patients. Objective: To evaluate the individual and combined effects of diverse intra-tumoral immune cell types on recurrence after complete surgical resection in early stage lung adenocarcinoma. Methods: We obtained NCBI GEO datasets for lung adenocarcinoma, the most prevalent histological subtype of NSCLC and re-analyzed the gene expression data of 292 patients with early stage cancer (IA/IB). CIBERSORT was used to resolve 22 immune cell types from the tumor transcriptomes. Survival analysis was carried out to assess the effect of immune cell types and genes associated with recurrence. Results: Out of the 22 cell types, a high proportion of Tregs and monocyte-macrophages in the tumors were associated with significantly increased probability of recurrence. Conversely, increased proportion of non-Treg CD4+ T cells and plasma cells were associated with a lower probability of recurrence. The higher expression of CCL20 (which can direct the migration of cells of B cell lineage), XCL1 (associated with prototypical Th1 responses) and the immunoglobulin chains IGHV4.34 and IGLV6.57 were associated with a significantly lower probability of recurrence. Importantly, the intra-tumoral immune phenotype comprising these four cell types varied among patients and differentially associated with recurrence depending on net levels of positive and negative prognostic factors. Despite a high level of intra-tumoral plasma cells, a concomitant high level of monocyte-macrophages reduced the freedom from recurrence from ~80 to ~50% at 80 months (p < 0.05). Furthermore, stratification of the patients on the basis of a score estimated from the levels of four cell types enabled the identification of patients with significantly increased probability of recurrence (~50%) after surgery. Significance: Our analysis suggests that concomitant levels of macrophages and plasma cells, in addition to the T regs and non-TregCD4+ T cells in tumors can identify patients with early stage lung cancer at greater risk of recurrence.