Analysis of immune checkpoint blockade biomarkers in elderly patients using large-scale cancer genomics data.

Abstract

2543 Background: Immune checkpoint blockade (ICB) immunotherapy in some cases elicits striking patient responses, but its efficacy appears to be dependent on several incompletely understood factors. Most studies of ICB therapies in elderly patients have concluded that they received no reduced benefit or even increased benefit compared to the younger patients analyzed, despite the systemic age-related immune changes that might be expected to produce a less effective immune response, such as loss of the capacity to generate new naive T cells. To understand and apply these results, it is necessary to investigate the relationship of age and the immune tumor microenvironment. Methods: We apply bioinformatics methods to genomic, transcriptomic, and clinical data from 9,523 patients across 31 cancer types from TCGA, 15,557 patients with breast, colon, or head and neck cancers from Caris Life Sciences, and 37,961 patients across 8 cancer types collected by GENIE. From these data we apply multivariate linear models across and within individual tumor types to estimate age-related associations to tumor mutational burden (TMB), T cell receptor diversity (miTCR), differential gene expression (edgeR), pathway enrichment (mSigDB and fgsea), and immune cell type infiltration (Quantiseq and MIXTURE). Results: Our analysis of large-scale molecular and clinical databases associates patient age with changes in several major biomarkers of ICB response. Notably, a robust correlation between increased tumor mutational burden and age was found across three different large cohorts (TCGA, Caris Life Sciences, and GENIE) in most ICB-approved cancer types. In the TCGA data, TMB increased with age pan-cancer (p < 1x10-16) and in 7 of 9 ICB-approved cancer types. These associations were validated in the larger cohort of patient samples in GENIE, which demonstrated correlations between increased TMB levels and patient age in all eight ICB-approved cancer types assayed (Table), as well as in the Caris colorectal (q < 0.001) and breast (q < 0.001) cancer cohorts. Significant associations of age to other biomarkers of ICB response (checkpoint gene expression, immune infiltration, and immune related pathway signaling) will be presented. Conclusions: These results provide context for the efficacy of ICB in elderly patients, highlight potential biomarkers for the treatment of elderly patients with immunotherapies, and strongly suggest the value of large-scale prospective study of elderly cancer patients treated with ICB.[Table: see text]