Abstract:e21055 Background: While methods for detecting SNV and indel in circulating tumor DNA (ctDNA) with hybrid-capture based next generation sequencing (NGS) has been available, algorithm for calling copy number variation(CNV) is more challenging. Here we present a method enabling CNV detection in very low amount of ctDNA with high sensitivity and specificity. Methods: First of all, we built a background model with a customized 150 gene panel cfDNA sequencing data from 10 healthy people. This model was then applied to normalize read depth to further reduce background noise. The probe level log2ratio was segmented using circular binary segmentation. Segments with log2ratio above 3 fold standard deviation was treated as amplification. Then we used NCI-H1573 and HCC1954 cell culture supernatant, which harbored three CNVs: ERBB2 (HCC1954), EGFR and MET (NCI-H1573), to mimic human ctDNA. Both in-silicon and in-vitro simulations were performed to define the detection limit. We validated the CNV concordance of FFPE samples and matched controls with WES profiling and plasma samples with 150 gene panel profiling from 28 lung cancer patients. Additional 1495 ctDNA samples were also profiled with 150 gene panel. Results: For in-silicon spike-in, the detection limit of ctDNA fraction is about 0.3% for ERBB2 (2.2 copy), 5% for MET (2.5 copy), 1% for EGFR (2.2 copy). For in-vitro spike-in, the detection limit of ctDNA fraction is about 1% for ERBB2 (2.6 copy), 5% for MET (2.5 copy), 3% for EGFR (2.6 copy). Compared with WES CNV results, the specificity of CNV calling from ctDNA panel is 95.5% and sensitivity is 50% when copy number > 6, and 71.4% when copy number > 12. We profiled 1495 lung cancer ctDNA samples with 150 gene panel and detected 15 HER2 amplification(1%), 75 EGFR amplification(5%) and 30 MET amplification samples(2%). An advanced lung adenocarcinoma patient with acquired resistance after EGFR TKI treatment, MET amplification was detected in plasma ctDNA. Patient accepted crizotinib and reached partial response. Conclusions: We developed a pipeline to detect CNV in very low amount of ctDNA with high sensitivity and specificity, which improves the accuracy of CNV calling in noninvasive way and avoids unnecessary biopsy on tumor patients.

BayVarC: an ultra-sensitive ctDNA variant caller using Bayesian approach

A novel virtual barcode strategy for accurate panel-wide variant calling in circulating tumor DNA

Benchmarking UMI-aware and standard variant callers for low frequency ctDNA variant detection

Comprehensive detection of ctDNA variants at 0.1% allelic frequency using a broad targeted NGS panel for liquid biopsy research.

Evaluating the performance of low-frequency variant calling tools for the detection of variants from short-read deep sequencing data

Detecting and Quantitating Low Fraction DNA Variants with Low-Depth Sequencing

Empirical Bayes single nucleotide variant-calling for next-generation sequencing data

[MinerVa: A high performance bioinformatic algorithm for the detection of minimal residual disease in solid tumors].

LFMD: a new likelihood-based method to detect low-frequency mutations without molecular tags

Biological background of the genomic variations of cf-DNA in healthy individuals

Identification of DNA variants at ultra-low variant allele frequencies via Taq polymerase cleavage of wild-specific blockers

Ultra-deep sequencing with unique molecular identifier(UMI) for detection of ctDNA by fragment profiling using machine learning.

Abstract P19: Comprehensive Benchmark of Somatic Mutation Callers on Cell-Free DNA Next-Generation Sequencing Data in Cancer Liquid Biopsy

NanoVar: accurate characterization of patients' genomic structural variants using low-depth nanopore sequencing

Technical And Clinical Validation Of Somatic Cnv Detection From Circulating Tumor Dna.

RabbitVar: Ultra-Fast and Accurate Somatic Small-Variant Calling on Multi-Core Architectures

Copy Number Variation Detection In Whole-Genome Sequencing Data Using The Bayesian Information Criterion

Evaluating Bioinformatics Processing of Somatic Variant Detection in cfDNA Using Targeted Sequencing with UMIs

Benchmarking of Germline Copy Number Variant Callers from Whole Genome Sequencing Data for Clinical Applications

Ultrasensitive Detection of Circulating Tumor DNA in Lymphoma via Targeted Hybridization Capture and Deep Sequencing of Barcoded Libraries

Optimizing Accuracy and Efficiency in Analyzing Non-UMI Liquid Biopsy Datasets Using the Sentieon ctDNA Pipeline