Circulating tumor DNA sequencing to reveal the genomic complexity of advanced EGFR-mutant lung adenocarcinoma.

Tumor Biology
Session Type and Session Title: 
This abstract will not be presented at the 2016 ASCO Annual Meeting but has been published in conjunction with the meeting.
Abstract Number: 


J Clin Oncol 34, 2016 (suppl; abstr e23079)
Collin M. Blakely, Kimberly C. Banks, Richard Burnham Lanman, Jonathan Riess, Philip C. Mack, Trever Grant Bivona; University of California, San Francisco, San Mateo, CA; Guardant Health, Inc., Redwood City, CA; University of California, Davis Comprehensive Cancer Center, Sacramento, CA; UC Davis Comprehensive Cancer Center, Sacramento, CA; University of California, San Francisco, San Francisco, CA

Abstract Disclosures


Background: Lung adenocarcinoma (LUAD)patients (pts)whose tumors harbor activating EGFR mutations (mt) respond initially to EGFR TKIs but responses are typically incomplete and acquired resistance (AR) occurs. EGFR T790M resistance mt are thought to drive AR in 60% of cases, yet responses to 3rd generation EGFR TKIs that can target this mt are also incomplete and transient. Additional concurrent genomic alterations may underlie EGFR TKI resistance. Methods: We performed next-generation sequencing (NGS) on circulating cell-free DNA (cfDNA) from 546 consecutive advanced LUAD pts with detectable EGFR mt. The analysis included pts prior to and during EGFR TKI therapy, and at TKI resistance. Using a 68-gene panel of cancer-related genes, we determined the frequency of mt in common oncogenic pathways concurrent with EGFR mt. Results: The most common mt co-occurring with EGFR were in TP53 (54%) and EGFR T790M (39%). Genetic alterations in > 1 cancer-relevant pathway were detected in 23% of T790M mt and 15% of T790wt cases (p = 0.06). Mt or copy number amplification (CNA) within genes involved in cell cycle control (CDKN2A, RB1, CDK4/6, CCNE1) were more common in samples that harbored T790M mt (20% vs. 12%, p = 0.02), as were mt in Wnt pathway genes (CTNNB1 and APC) (9.4% vs. 4.8%, p = 0.05). Mt in genes involved in MAPK (KRAS, NRAS, BRAF, NF1, GNAS) (14% vs. 17%), or PI3K signaling (PIK3CA and PTEN) (11% vs. 11%) were equivalent in T790M vs. wt cases. CNAs in pathways previously implicated in EGFR TKI acquired resistance were identified in T790M mt samples including MET (6.9%), and ERBB2 (3.9%). 3 pts with T790M were identified with co-occurring oncogene fusions (2 ALK, 1 NTRK1) at lower mt allele fractions. Cases in which serial samples from individual pts were analyzed will be presented. Conclusions: Both T790M and T790wt EGFR-mt LUAD can harbor additional mt in oncogenic pathways detectable in circulating cfDNA. Cases with detectable T790M mt were more likely to harbor mts in multiple oncogenic pathways. This may in part explain why treatment with EGFR TKI monotherapy results in only partial and temporary disease control across different classes of EGFR TKIs.