Abstract 5768: Genomic characteristics of classic and uncommon EGFR activating mutations explains different responses to first-generation EGFR-TKIs

Abstract

Abstract Background: Classic sensitizing mutations of epidermal growth factor receptor (cEGFR) (i.e., exon 19 deletions and exon 21 L858R) have shown positive responses to first-generation TKIs in non-small cell lung cancer (NSCLC). For patients with uncommon EGFR mutations (uEGFR), such as G719X, S768I, L861Q and exon 20 insertions, first-generation TKIs might still be one of the mainstay choices in clinic. However, they exhibited heterogeneity in treatment responses. Here, we aimed to elucidate different genomic profiles of tumors carrying uncommon EGFR activating mutations to explain differential treatment responses. Method: We screened 6669 NSCLC patients that underwent targeted NGS of 416 cancer-related genes. Total 2280 EGFR-positive NSCLCs were selected for analysis. We compared somatic mutations, tumor mutation burden (TMB), mutational signature and canonical oncogenic pathways between cEGFR and uEGFR. A validation set of 159 NSCLCs from the Cancer Genome Atlas (TCGA) was used. Adjusted p value less than 0.1 were considered statistically significant. Results: Only patients carrying known EGFR activating mutations and without complex mutations were included. The final cohort comprised 1022 patients with L858R (44.8%), 973 with exon 19 deletion (42.7%), 139 with exon 20 insertion (6.1%), 59 with G719X (2.6%), 51 with L861Q (2.2%), and 36 with S768I (1.6%). In general, clinical features were equally distributed except for higher prevalence of G719X and S768I in male. Most frequently co-mutated genes were similar between cEGFR and uEGFR subgroups, including TP53 (56% vs 60%), PIK3CA (8% vs 11%), and RB1 (8% vs 9%). Despite these, we found enrichment of EPHA3 and MED12 with S768I(both 5/36, 14%)as compared to with cEGFR (2% vs 3%, padj<0.1), while GATA6 predominantly co-occurred with L861Q (4/51, 8%,). Overall, patients carrying uEGFR had significantly higher TMB than those with cEGFR (median TMB 5.07 vs 3.80, p=0.0076), indicating more complicated clonal backgrounds that potentially led to the poor TKI response. Among the ten oncogenic pathways, the hippo pathway markedly more altered more in those with uEGFR than with cEGFR (p=0.0025). Moreover, we found higher fraction of immunoglobulin-related signature (signature 9) and BRCA1/2-associated signature (signature 3) in tumors with cEGFR than with uEGFR (p = 0. 0003 and 0. 02, respectively). Importantly, smoking induced mutations (signature 4) were enriched in uEGFR subgroup (p = 0. 0015), which was also verified in TCGA samples. Further, 28 patients had confirmed treatment with first-generation TKIs, where uEGFR were associated with poor outcomes (HR = 1.43, 95% CI 0.81-2.52). Conclusion: This comprehensive analysis of EGFR-positive mutational landscape revealed specific genomic characteristics associated with uncommon EGFR mutations that might explain their poor response to first-generation TKIs. Citation Format: Xue Wu, Junli Zhang, Jingwen Liu, Yedan Chen, Yang Shao. Genomic characteristics of classic and uncommon EGFR activating mutations explains different responses to first-generation EGFR-TKIs [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 5768.