Background – Approximately 30% of patients (pts) with non-small cell lung cancer (NSCLC) have alterations in the KRAS oncogene. In NSCLC KRAS mutational variants are diverse and therapeutically relevant. However, it is unclear how each variant is associated with the tumor biology, including lipid metabolism and the immune microenvironment. Since perturbed tumor immune infiltration and lipid metabolism are linked to NSCLC outcomes, we evaluated and characterized KRAS variants and association with lipid metabolism and immune infiltration.

Methods – Using the Tempus Database, 5,925 de-identified records from pts with NSCLC and KRAS alterations (G12A (n=464), G12C (n=2,510), C12D (n=937), G12R (n=61), G12V (n=1,251), G13C (n=244), G13D (n=185), Q61H (n=273)) were retrospectively analyzed. Samples were sequenced with the Tempus xT and xR RNA assays. Tumor microenvironment cell proportions were estimated using QuanTIseq. Single-sample gene set enrichment analysis based on 775 lipid metabolic genes was used to compute enrichment scores (ES) for each pt. We analyzed neoantigen tumor burden (NTB) and tumor mutational burden (TMB) as mutations per megabase (mut/Mb). P-values were calculated using Pearson’s Chi-squared and Kruskal-Wallis rank sum test. Pairwise comparisons of median ES were performed using the Wilcoxon test and the FDR method was used to correct for multiple comparisons.

Results – Among the overall KRAS cohort, KRAS G12C and G12V represented over 60% of cases. Pts with KRAS G12D alterations had the highest proportion (23%) of never-smokers (p<0.001). We identified small but significant differences in the median lipid ES in pts with KRAS G12C variants compared to pts with G12V and G12D, with G12C having lower median lipid ES (0.89 versus 0.91 and 0.91 respectively, both p<0.001). Differences were also identified in pts with KRAS G13C compared to pts with G12D (0.9 versus 0.91, p=0.03). Median TMB and NTB differed amongst pts with KRAS variants with a higher TMB and NTB found in G12C, G13C and G13D cohorts while the lowest TMB and NTB was observed in the G12D cohort (p<0.001). KRAS G12C variant tumors had the highest median percentage of CD8 T cells while G12D and Q61H variants had the lowest median percentage of CD8 T cells (p=0.016). Interestingly, G12D variant tumors had the highest median percentage of infiltrating B cells (p=0.006).

Conclusions – Our data show KRAS variants are associated with changes in the lipid and tumor immune landscape in NSCLC. There were significant differences in the proportion of tumor immune cells amongst tumors with different KRAS variants. Notably, pts with KRAS G12D variants had a less immunogenic immune microenvironment indicated by low proportion of TMB, NTB and CD8 T cells compared to G12C variants, which could affect immunotherapy efficacy. Further analysis is needed to determine if immunotherapy efficacy is altered by the distinct molecular phenotypes in pts with KRAS variants.

VIEW THE PUBLICATION