Detection, frequency and actionability of recurrent copy number gains detected by non-invasive liquid biopsy of 3,942 lung and breast cancer samples.

Tumor Biology
Session Type and Session Title: 
Poster Session, Tumor Biology
Abstract Number: 


Poster Board Number: 
Board #238
J Clin Oncol 34, 2016 (suppl; abstr 11541)
Diana Abdueva, Darya Chudova, Richard Burnham Lanman, Kimberly C. Banks, Rebecca J Nagy, Stefanie Mortimer, AmirAli Talasaz, Helmy Eltoukhy, Razelle Kurzrock; Guardant Health Inc, Redwood City, CA; Guardant Health, Inc., Redwood City, CA; University of California, San Diego, La Jolla, CA

Abstract Disclosures


Background: Circulating tumor DNA (ctDNA) testing must include detection of single nucleotide variants (SNVs), indels, fusions, and copy number amplifications (CNAs) to identify all NCCN-recommended genomic targets. We examined CNA landscape in a large clinical cohort of breast and lung cancer ctDNA samples analyzed with Guardant360 (G360) test, classifying recurrent gene-level copy number changes across distinct oncogenic pathways and disease subtypes. Methods: We developed an improved CNA detection pipeline and evaluated its analytical and clinical performance on a set of cell-line spike-ins with known amplifications and 72 clinical plasma samples from patients with stage IV breast cancer with known ERBB2 amplification/overexpression. We then examined 1,935 lung adenocarcinoma and 2,007 breast carcinoma clinical cases: CNA patterns in samples with and without known driver mutations were compared using Fisher’s exact test to identify candidate driver amplification events. Results: Analytical sensitivity, specificity, and precision of the new CNA pipeline improved significantly - average coefficient of variation (CV) of copy number estimates across 18 genes on the panel was 1.2% for spike-in replicates, all spike-ins and copy number neutral samples were detected correctly, and signal-to-noise ratio increased 4.3-fold. We observed high concordance between tissue-verified amplification events in breast carcinoma subset when the mutant allele fraction of the driver mutation exceeded 0.5%. In the lung adenocarcinoma cohort, CNA-driver mutations were identified in ~7% of 1,935 samples, excluding cases known functional driver events. Focal amplification events in MET, CCND1, and ERBB2 were enriched in samples lacking other driver oncogene mutations (p-value < 0.001). Conclusions: Copy number variation is an important mechanism of oncogene activation in lung and breast cancers. Highly accurate quantification of copy number state in ctDNA allows for clinically meaningful identification of amplified therapeutic targets with otherwise wildtype DNA sequence.