Detection of genomic alterations in breast cancer with circulating tumour DNA sequencing

Dimitrios Kleftogiannis,Danliang Ho,Ann Siew-Gek Lee,Jun Xian Liew,Benita Kiat-Tee Tan,Kiang Hiong Tay,Gek San Tan,Anna Gan,Tony Kiat-Hon Lim,Sarah B Ng,Raymond Chee-Hui Ng,Chih Chuan Shih,Iain Beehuat Tan,Swee H Lim,Yoon-Sim Yap,Polly S Y Poon

doi:10.1038/s41598-020-72818-6

Abstract

Analysis of circulating cell-free DNA (cfDNA) has opened new opportunities for characterizing tumour mutational landscapes with many applications in genomic-driven oncology. We developed a customized targeted cfDNA sequencing approach for breast cancer (BC) using unique molecular identifiers (UMIs) for error correction. Our assay, spanning a 284.5 kb target region, is combined with a novel freely-licensed bioinformatics pipeline that provides detection of low-frequency variants, and reliable identification of copy number variations (CNVs) directly from plasma DNA. We first evaluated our pipeline on reference samples. Then in a cohort of 35 BC patients our approach detected actionable driver and clonal variants at low variant frequency levels in cfDNA that were concordant (77%) with sequencing of primary and/or metastatic solid tumour sites. We also detected ERRB2 gene CNVs used for HER2 subtype classification with 80% precision compared to immunohistochemistry. Further, we evaluated fragmentation profiles of cfDNA in BC and observed distinct differences compared to data from healthy individuals. Our results show that the developed assay addresses the majority of tumour associated aberrations directly from plasma DNA, and thus may be used to elucidate genomic alterations in liquid biopsy studies.

Highlights

Analysis of circulating cell-free DNA has opened new opportunities for characterizing tumour mutational landscapes with many applications in genomic-driven oncology
We developed a new error-corrected cell-free DNA (cfDNA) sequencing approach with a panel customized for breast cancer (BC), in combination with publicly available bioinformatics strategies for the identification of tumour-associated genomic alterations
Usage of unique molecular identifiers (UMIs) for error correction, combined with variant calling based on variants with double-strand support, allowed us to detect low variant allele frequency (VAF) variants in circulating tumour DNA (ctDNA) samples with reduced number of false positives

Summary

Introduction

Analysis of circulating cell-free DNA (cfDNA) has opened new opportunities for characterizing tumour mutational landscapes with many applications in genomic-driven oncology. Recent studies have demonstrated the translational potential of circulating cell-free DNA (cfDNA), or circulating tumour DNA (ctDNA) in cancer patients, for improving cancer m anagement[3,4]. Such liquid biopsy data measured directly from body fluids (e.g. plasma) can be used to detect tumour somatic alterations, with the ability to provide early prognostication and/or better molecular profiling of patients with cancer without the risk and discomfort of invasive b iopsies[5]. Analyses of tumour NGS data from cfDNA is challenging due to several biological reasons (e.g. low cfDNA abundance in the blood stream) and other technical artifacts (e.g. error rates of NGS) that restrict the analytical sensitivity of tumour detection in plasma DNA

Methods

Results

Conclusion