Evolution and clinical impact of genomic alterations detectable in circulating tumor DNA of 1150 advanced EGFR-mutant (mt) lung cancer patients.

Abstract

9009 Background: Advanced EGFR-mt lung adenocarcinomas (LUAD) frequently harbor additional genetic alterations. The clinical significance of these concurrent alterations and how they evolve with treatment are unclear. Methods: We performed next-generation sequencing (NGS) of ~70 cancer related genes from the circulating tumor DNA (ctDNA) of 1150 consecutive advanced LUAD patients (pts) with detectable EGFR mts. The analysis included 113 samples from 81 pts for whom clinical outcome data were known. Clinical response to EGFR TKI treatment was correlated to mutational status in 12 cancer-related pathways. Results: EGFR-mt cases contained an average of 3.6 genetic alterations (range 1-18). There was enrichment for co-alterations in TP53, CDK6, CTNNB1 and SMAD4 in EGFR-mt cases compared to a control cohort of 1008 EGFR WT cases. Enrichment for the EGFR T790M mutation was found upon progression to first-line EGFR TKI treatment, as expected. Analysis of 450 T790M-mt cases showed enrichment for co-alterations in genes controlling the cell cycle (28% vs. 21%, p = 0.01), DNA repair (12% vs. 8%, p = 0.03), and WNT (16% vs. 11%, p = 0.03) signaling. The number of genetic alterations increased during progression on each line of therapy in a manner unlinked to age, gender, or tobacco exposure (mean 3.2 pre-TKI vs. 6.4 after 2nd line, p = 0.001). Upon progression to second-line treatment, analysis revealed further selection for co-alterations in TP53, CCNE1, MYC and PIK3CA and the associated pathway-level classifications. We found an increased frequency of co-alteration in cell cycle genes in EGFR TKI non-responders versus responders (33% vs. 0%, p = 0.0005). Conclusions: Within the landscape of advanced EGFR-mt LUAD, we uncover features of evolutionary selection for multiple concurrent oncogenic pathway alterations including TP53, WNT, PI3K, MYC, and cell cycle genes. This large clinical and genetic dataset prompts a re-evaluation of the prevailing paradigm of monogenic-based molecular stratification to monotherapy, and highlights an alternative model of genetic collectives as a previously underappreciated determinant of lung cancer progression and therapy resistance.