Purpose:

Plasma genotyping may identify mutations in potentially “actionable” cancer genes, such as BRCA1/2, but their clinical significance is not well-defined. We evaluated the characteristics of somatically acquired BRCA1/2 mutations in patients with metastatic breast cancer (MBC).

Experimental Design:

Patients with MBC undergoing routine cell-free DNA (cfDNA) next-generation sequencing (73-gene panel) before starting a new therapy were included. Somatic BRCA1/2 mutations were classified as known germline pathogenic mutations or novel variants, and linked to clinicopathologic characteristics. The effect of the PARP inhibitor, olaparib, was assessed in vitro, using cultured circulating tumor cells (CTCs) from a patient with a somatically acquired BRCA1 mutation and a second patient with an acquired BRCA2 mutation.

Results:

Among 215 patients with MBC, 29 (13.5%) had somatic cfDNA BRCA1/2 mutations [nine (4%) known germline pathogenic and rest (9%) novel variants]. Known germline pathogenic BRCA1/2 mutations were common in younger patients (P = 0.008), those with triple-negative disease (P = 0.022), and they were more likely to be protein-truncating alterations and be associated with TP53 mutations. Functional analysis of a CTC culture harboring a somatic BRCA1 mutation demonstrated high sensitivity to PARP inhibition, while another CTC culture harboring a somatic BRCA2 mutation showed no differential sensitivity. Across the entire cohort, APOBEC mutational signatures (COSMIC Signatures 2 and 13) and the “BRCA” mutational signature (COSMIC Signature 3) were present in BRCA1/2-mutant and wild-type cases, demonstrating the high mutational burden associated with advanced MBC.

Conclusions:

Somatic BRCA1/2 mutations are readily detectable in MBC by cfDNA analysis, and may be present as both known germline pathogenic and novel variants.

Translational Relevance

Identification of somatic mutations using plasma genotyping assays in patients with metastatic breast cancer (MBC) represents an opportunity for novel therapy selection, and a challenge in distinguishing clinically impactful genetic variants. For BRCA1/2 mutations, pathogenic germline mutations are well-annotated, whereas the therapeutic significance of somatically acquired variants is not well-defined. We describe a cohort of patients with MBC, in whom we identified BRCA1/2 mutations using cell-free DNA (cfDNA) genotyping, with clinical correlates, and in selected cases conducted functional assays in cultured circulating tumor cells (CTCs). As many as 13.5% of patients with MBC harbor somatic BRCA1/2 mutations in cfDNA; 4% are known germline pathogenic variants. In CTC-derived models, certain cell lines with somatically acquired driver variants demonstrate increased sensitivity to PARP inhibitors, while others with somatic BRCA1/2 variants resulting from increased APOBEC-mediated mutagenesis do not, behaving as passenger mutations. Detection of BRCA1/2 mutations using cfDNA requires caution before PARP inhibitor application.

Tumor genotyping is the central tenet of precision oncology and is increasingly becoming part of routine clinical care in breast cancer to identify actionable mutations for potential therapeutic intervention. However, tumor tissue genotyping of the primary tumor alone does not identify clonal evolution and mutations that are acquired during the course of treatment, some of which may be therapeutically relevant. Obtaining serial tumor biopsies from metastatic sites has been applied as proof of principle to guide successive therapeutic choices, but it is limited by the risks to the patient of an invasive procedure and accessibility of the metastatic site to biopsy, as well as biased sampling of a single tumor lesion in the midst of widespread sites of disease. Consequently, cell-free DNA (cfDNA) analyses or so-called liquid biopsies have emerged as an important strategy to monitor acquired cancer mutations and there has been a major rise in the clinical utilization of cfDNA assays (1–6).

While cfDNA assays are widely used in the clinic, the interpretation of multiple subclonal mutations and novel variants represents a major diagnostic challenge. This is particularly important for genes that have matched therapy approved in more traditional germline or tumor genotyping contexts, such as PARP inhibitors for patients harboring germline BRCA1/2 mutations (7, 8). Cancer predisposing heterozygous germline BRCA1/2 mutations, leading to somatic BRCA-null phenotypes, have been well-studied in breast cancer (7–12), but de novo somatic BRCA1/2 mutations are thought to be rare in breast cancer. An analysis of The Cancer Genome Atlas (TCGA) noted the prevalence of somatic BRCA1 mutations in primary breast cancer as 1.55%, and somatic BRCA2 mutations as 1.68% (13). However, progression to metastatic breast cancer (MBC) is associated with an increased frequency of mutations, particularly in metastatic triple-negative breast cancer (TNBC), where mutations in components of the homologous recombination pathway are more common, and the percentage of somatic BRCA1 mutations is around 6% (14). Even in such cases of definitive acquired BRCA1/2 mutations identified by traditional tumor genotyping, the functional implications for sensitivity to PARP inhibition are not established. In addition, unlike germline BRCA1/2 genotyping, where large datasets have been curated to help interpret pathogenic and silent genetic variants within various populations, there are no such guidelines for interpreting the even more diverse potential variations in BRCA1/2 that may be acquired somatically. These challenges are further magnified in cfDNA by the variable allele fractions and subclonal tumor cell populations that they represent. Given the general availability of cfDNA genotyping, its noninvasiveness as a diagnostic tool, and the potential for identifying impactful acquired mutations, its rapidly expanding applications require careful review before they are used to trigger therapeutic interventions.

The primary objective of this study was to understand the clinical and functional characteristics of somatic BRCA1 and BRCA2 mutations detectable by cfDNA in patients with MBC.

Study population

Patients with MBC who underwent cfDNA analysis as part of routine clinical care at the Massachusetts General Hospital (Boston, MA) before starting a new therapy from February 2015 to July 2017 were identified. The subset of patients with BRCA1 and/or BRCA2 mutations detectable by cfDNA analysis [next-generation sequencing (NGS)/Guardant360] was determined. All consecutive patients with MBC who had Guardant360 testing during the aforementioned time interval were included, and no cases were excluded. A retrospective review of medical and pathology records, based on an institutional review board (IRB)–approved institutional protocol, was conducted to identify tumor subtype, patient demographics, and germline BRCA1/2 testing results via standard commercial germline testing, and subsequent review of all cfDNA results with Guardant360 to verify the somatic nature of the mutations identified, as a post hoc analysis, tumor genotyping results (NGS, institutional platform), and treatment outcomes after cfDNA testing. This research was conducted in accordance with recognized ethical guidelines, including the Declaration of Helsinki, and the retrospective review was conducted on the basis of an IRB-approved institutional protocol.

cfDNA analysis

cfDNA analysis was performed using Guardant360 testing, an NGS-based clinical assay evaluating 73 genes. Guardant360 employs massively parallel and deep sequencing, with an analytic sensitivity of 0.1% mutant allele fraction (MAF), with quoted specificity above 99.9%, and clinical sensitivity of 85.0% (compared with 80.7% tissue sensitivity; ref. 15). The average molecule count was about 8,000 molecules, and the average single read depth was approximately 15,000 molecules. For BRCA1 (chromosome 17q21), exons 2–23 were included, and for BRCA2 (chromosome 13q13), exons 2–27 were included. A retrospective chart review of the Guardant360 cfDNA reports was performed to determine the presence of somatic BRCA1 or BRCA2 mutations, to identify coexisting cfDNA mutations, and to characterize clonality. On the basis of the MAF of coexisting alterations, we defined BRCA1 or BRCA2 mutations as clonal (MAF ratio of BRCA1/2 mutation/gene mutation with highest MAF ≥ 0.25) or subclonal (MAF ratio of BRCA1/2 mutation/gene mutation with highest MAF < 0.25; ref. 16).

Guardant360 can identify both germline and somatic BRCA1/2 mutations in a single test. The germline versus somatic origin of a BRCA mutation was determined using a decision tree algorithm, which relies on annotation from external databases (ExAC, COSMIC, ClinVar, etc) and the observed MAF of the BRCA variant relative to other known germline variants. All identified variants were first annotated with information from external databases to identify those variants that are known germline variants. Thereafter, variants that had an insufficient annotation to determine their origin were evaluated further on the basis of their observed MAF relative to that of nearby known germline variants, and then a beta-binomial significance test was applied, and the variant was scored as germline or somatic. Notably, somatic BRCA1/2 mutations were usually at a variant allele fraction two orders of magnitude lower than germline BRCA1/2 mutations (17).

While germline results were initially suppressed in Guardant360 testing reports, as a post hoc analysis, we worked with the Guardant360 team to verify the somatic nature of detected mutations.

Somatic BRCA1/2 mutations were further classified as either known germline pathogenic variants or as novel/unclassified variants by two independent genetic counselors, who were blinded to the Guardant360 reports. The genetic counselors evaluated the specific DNA variants seen in cfDNA, which were specifically requested from Guardant360 for this analysis. This classification was based on review of the ClinVar database (18) to identify variants that had high classification confidence (three and four star review status) as of October 2018. For variants that had moderate to low classification confidence (two stars or fewer), additional criteria such as classification reports from Clinical Laboratory Improvement Amendments–certified germline genetic testing laboratories or review by consortia were evaluated. Finally, for variants not currently in ClinVar, the likelihood of a loss-of-function variant (such as nonsense mutations, frameshifts, and mutations in the ±1 or 2 splice site locations) was assessed, as outlined in ACMG/AMP guidelines for DNA variant classification (19). All variants not categorized as known germline pathogenic by this analysis were then categorized as novel/unclassified, including the majority of missense mutations.

Tumor genotyping analysis

A chart review of tumor genotyping results from archival tumor tissue was performed to identify coexisting tumor mutations. An institutional NGS assay evaluating 98 genes for mutations and 91 genes for copy-number changes (SNaPshot) was utilized for tumor tissue genotyping (20). This anchored multiplex PCR assay detects gene rearrangements, insertions and deletions, single-nucleotide variants, and copy-number changes present at allelic frequencies at 5% or higher with 100% analytic sensitivity and 100% analytic specificity (20). BRCA1 exons 2–23 and BRCA2 exons 2–27 are included in the assay. The time interval between cfDNA collection and the tumor tissue biopsy used for tissue genotyping was determined.

Mutation signature analysis

Mutation signatures were analyzed using non-negative matrix factorization (NMF) as described previously (21–23). Mutations detected by Guardant360 in the 29 BRCA-mutant patients were combined into a single “virtual patient,” which was then analyzed together with 785 patients with breast cancer from TCGA. These mutation calls from whole-exome sequencing (WES) were obtained from TCGA Unified Ensemble “MC3” Call Set (24), the public, open-access dataset of somatic mutation calls produced by the MC3 calling effort [“Multi-Center Mutation Calling in Multiple Cancers”), downloaded from the following link: http://www.synapse.org/#!Synapse:syn7214402/wiki/405297 (the results here are in whole or part based upon data generated by the TCGA research network: http://cancergenome.nih.gov/ as outlined in the TCGA publications guidelines (http://cancergenome.nih.gov/publications/publicationguidelines)]. Joint analysis of this combined TCGA + Guardant360 dataset by NMF (k = 6) revealed mutation signatures corresponding to aging (COSMIC Signature 1), APOBEC (COSMIC Signatures 2+13), the “BRCA” signature (COSMIC Signature 3), and MSI, microsatellite instability (COSMIC Signatures 6+26). The number and fraction of mutations due to each signature were then estimated and reported.

Statistical analysis

The association between cfDNA BRCA1/2 mutation status and patient age was determined with the Wilcoxon rank-sum test, and associations with cfDNA BRCA1/2 mutation status and tumor subtypes, prior treatment, and first treatment after cfDNA testing were performed with the Pearson χ2 test. The impact of cfDNA BRCA1/2 mutation status on progression-free survival (PFS) on the first treatment after cfDNA testing and overall survival (OS) was determined with the log-rank test. Cox regression analysis was used to determine the hazard ratio of cfDNA BRCA1/2 mutation status on PFS and OS. In addition, a multivariate analysis correcting for age and number of prior therapies was performed to determine the impact of cfDNA BRCA1/2 mutation status on PFS and OS. For all analyses, P < 0.05 was considered statistically significant.

Establishing BRCA1-mutant ex vivo circulating tumor cell culture

A cell line (Brx401) was established from circulating tumor cells (CTC) enriched from a patient with a somatic BRCA1 mutation (in this study cohort No. 16). This patient had a known germline pathogenic BRCA1 mutation [splice site single nucleotide variant (SNV) ENST00000357654.3:c.5075-1G>C] detectable in the cfDNA, which was acquired after treatment with chemotherapy for metastatic TNBC. The patient had no known germline BRCA1/2 mutation. A second CTC cell line (Brx142) was established from a patient with hormone receptor–positive (HR+) MBC. In this case, an early CTC culture showed wild-type (WT) BRCA2, but a subsequent culture, acquired after treatment with an oral selective estrogen receptor degrader, identified a BRCA2 mutation (missense mutation E3071Q). Again, the germline testing showed no BRCA2 mutation. Additional CTC cultures were used as controls (WT BRCA1/2) as described previously (4). For all CTC collections, written and signed informed consent was obtained as per IRB-approved protocol. CTCs were isolated using the microfluidic CTC i-CHIP and ex vivo cultures were established as described previously (4). CTC cultures were routinely checked for Mycoplasma with the MycoAlert Lonza kit and authenticated against matched blood sample via short tandem repeat profiling by Genetica DNA Laboratories (a LabCorp brand) using the commercially available PowerPlex16HSamplification Lit (Promega Corporation; mouse marker included) and GeneMapper ID v3.2.1 Software (Applied Biosystems).

WES

For WES, the AllPrep DNA/RNA Mini Kit (Qiagen) was used for extraction of genomic DNA. DNA was quantified in triplicate using a standardized PicoGreen dsDNA Quantitation Reagent (Invitrogen) Assay. The quality control identification check was performed using fingerprint genotyping of 95 common SNPs by Fluidigm Genotyping (Fluidigm). Library construction was performed using the KAPA Library Prep kit, with palindromic forked adapters from Integrated DNA Technologies. All library construction, hybridization, and capture steps were automated on the Agilent Bravo liquid handling system. Flowcells were sequenced utilizing sequencing-by synthesis chemistry for HiSeq 4000 flowcells. Each pool of whole-exome libraries was sequenced on paired 76 cycle runs with two eight-cycle index reads across the number of lanes needed to meet coverage for all libraries in the pool (raw data available on request).

Somatic mutation calling from WES data

Exome sequencing data of CTC lines were used to identify somatic SNVs (sSNV) and somatic small insertions and deletions (sINDEL). Output from Illumina software was processed by the Picard and GATK Toolkits developed at the Broad Institute (Cambridge, MA). The BAM files were generated by aligning with bwa version 0.5.9 to the NCBI Human Reference Genome Build hg19. Prior to variant calling, the impact of oxidative damage (oxoG) to DNA during sequencing was quantified as described previously (25). The cross-sample contamination was measured with ContEst (26) based on the allele fraction of homozygous SNPs, and this measurement was used in MuTect. From the aligned BAM files, somatic alterations were identified using a set of tools developed at the Broad Institute (Cambridge, MA www.broadinstitute.org/cancer/cga). The details of the sequencing data processing have been described previously (27, 28). Following our standard procedure, sSNVs were detected using MuTect (version 1.1.6; ref. 28) and sINDELs were detected using Strelka (version 1.0.11; ref. 29). Then, an allele fraction–specific panel-of-normals (PoN) filter was applied to filter false positive germline variants and common artifacts from mutation calls, which compares the detected variants to a large panel of normal exomes or genomes and removes variants that were observed in the PoNs. All somatic mutations, insertions, and deletions were annotated using Oncotator (version 1.4.1; ref. 30). sSNVs and sINDELS in only cancer genes (Cancer Gene Census; ref. 31) were used for mutation status analysis.

Olaparib sensitivity studies

Three independent breast cancer CTC lines were tested for drug sensitivity: Brx401, harboring a somatically acquired known germline pathogenic BRCA1 mutation, Brx142, harboring a somatically acquired mutation in BRCA2, and Brx07, with WT BRCA1/2 alleles. CTC lines were seeded in 96-well ultralow attachment plates (Corning) at 1,000 cells per well. Increasing concentrations of olaparib (Selleckchem S1060) ranging from 0.01 to 50 μmol/L were added to quadruplicate wells. Cell viability was measured using CellTiter-Glo luminescent cell viability assay per the manufacturer's instructions at day 5.

Immunoblot

Cell pellets were lysed in 100 mmol/L Tris pH 6.8 1% SDS, sonicated for 10 seconds using a 4710 Series Ultrasonic Homogenizer (Cole-Parmer), and incubated for 3 minutes at 95°C. Protein lysates were then quantified and normalized using Pierce BCA Protein Assay Kit (23227, Thermo Fisher Scientific). Lysates were then combined 1:1 with 2 × sample buffer (100 mmol/L Tris pH 6.8, 12% glycerol, 3.5% SDS, and 0.2 mol/L DTT) and 20 μg of protein was loaded onto 4%–12% Bolt Bis-Tris Plus Gels (NW04122BOX, Thermo Fisher Scientific) and transferred onto polyvinylidene difluoride membranes by liquid transfer with CBS Scientific Electrophoretic Blotting System (EBX-700, 100 V, 2 hours). Membranes were immunoblotted using BRCA1 (1:1,000, D-9, Santa Cruz Biotechnology), GAPDH (1:20,000, AB516, Millipore), and H3 (1:40,000, ab1791, Abcam) antibodies and horseradish peroxidase–conjugated secondary anti-mouse (1:5,000, 115-035-003, Jackson ImmunoResearch) and anti-rabbit (1:5,000, 111-035-003, Jackson ImmunoResearch) antibodies. Signals were detected using the Chemidoc Imaging System (Bio-Rad) with Image Lab v6.0.1 software.

Patient demographics

We identified 215 patients at the Massachusetts General Hospital (Boston, MA) with MBC who had undergone cfDNA analysis before the start of a new therapy (first-line or greater) from February 2015 to July 2017. Supplementary Fig. S1 provides a consort diagram delineating the study population. Of the total population with MBC, 29 (13.5%) had somatic BRCA1 or BRCA2 mutations detectable by cfDNA. In nine patients (4.2%), mutations previously described as known germline pathogenic were detected (as described in Supplementary Table S1), while in 20 (9.3%), novel variants (not previously reported in public databases) were identified (18).

Altogether, patients with cfDNA BRCA1 or BRCA2 mutations had similar age, cancer subtype distribution, and number of prior lines of chemotherapy, and went on to receive similar therapies after testing as those lacking cfDNA BRCA1/2 mutations (BRCA WT population; Supplementary Table S1). The majority of patients with somatic BRCA1/2 mutations had MBC which was recurrent (97%), rather than de novo. The characteristics of patients with either known germline pathogenic BRCA1/2 somatic mutations or novel variants are shown in Table 1. Interestingly, the patients with known germline pathogenic BRCA1/2 mutations were significantly younger (median age of 48 years vs. 55 years; P = 0.008) and more often had TNBC (44% vs. 5%; P = 0.022), compared with the novel variants, which were generally seen in HR+ and in some HER2+ MBC. The analyses of somatic BRCA1/2 status on patient outcomes is described in the Supplementary Data and Supplementary Fig. S2.

Table 1.

Clinical characteristics of patients with MBC and known germline pathogenic BRCA1/2 mutations or novel BRCA1/2 variants.

Clinical variablecfDNA BRCA1/2 mutation absent (BRCA WT; N = 186)cfDNA BRCA1/2 known germline pathogenic mutation present (BRCA known germline pathogenic mutant; N = 9)acfDNA BRCA1/2 novel variant mutation present (BRCA novel variant; N = 20)aP value for difference between BRCA WT and BRCA known germline pathogenic mutantbP value for difference between BRCA known germline pathogenic mutant and BRCA novel variantb
Median age at MBC diagnosis 57 (48–65) 48 (46–52) 55 (52–67) 0.059 0.008 
Tumor subtype    0.023 0.022 
 HER2+ 11 (5.9%) 0 (0%) 3 (15%)   
 HR+ 134 (72%) 4 (44.4%) 14 (70%)   
 TNBC 24 (12.9%) 4 (44.4%) 1 (5%)   
 Unknown 17 (9.1%) 1 (11.1%) 2 (10%)   
Number of prior lines of chemotherapy    0.98 0.26 
 0–1 124 (66.7%) 6 (66.7%) 17 (85%)   
 ≥2 61 (32.8%) 3 (33.3%) 3 (15%)   
 Unknown 1 (0.5%) 0 (0%) 0 (0%)   
First therapy after cfDNA testing    0.32 0.42 
 Endocrine 57 (30.6%) 3 (33.3%) 7 (35%)   
 HER2 therapy 13 (7.0%) 0 (0.0%) 2 (10%)   
 Immunotherapy 14 (7.5%) 2 (22.2%) 2 (10%)   
 Chemotherapy 49 (26.3%) 3 (33.3%) 3 (15%)   
 Other 36 (19.4%) 0 (0.0%) 3 (15%)   
 None 13 (7.0%) 0 (0.0%) 2 (10%)   
 Unknown 4 (2.2%) 1 (11.1%) 1 (5%)   
Clinical variablecfDNA BRCA1/2 mutation absent (BRCA WT; N = 186)cfDNA BRCA1/2 known germline pathogenic mutation present (BRCA known germline pathogenic mutant; N = 9)acfDNA BRCA1/2 novel variant mutation present (BRCA novel variant; N = 20)aP value for difference between BRCA WT and BRCA known germline pathogenic mutantbP value for difference between BRCA known germline pathogenic mutant and BRCA novel variantb
Median age at MBC diagnosis 57 (48–65) 48 (46–52) 55 (52–67) 0.059 0.008 
Tumor subtype    0.023 0.022 
 HER2+ 11 (5.9%) 0 (0%) 3 (15%)   
 HR+ 134 (72%) 4 (44.4%) 14 (70%)   
 TNBC 24 (12.9%) 4 (44.4%) 1 (5%)   
 Unknown 17 (9.1%) 1 (11.1%) 2 (10%)   
Number of prior lines of chemotherapy    0.98 0.26 
 0–1 124 (66.7%) 6 (66.7%) 17 (85%)   
 ≥2 61 (32.8%) 3 (33.3%) 3 (15%)   
 Unknown 1 (0.5%) 0 (0%) 0 (0%)   
First therapy after cfDNA testing    0.32 0.42 
 Endocrine 57 (30.6%) 3 (33.3%) 7 (35%)   
 HER2 therapy 13 (7.0%) 0 (0.0%) 2 (10%)   
 Immunotherapy 14 (7.5%) 2 (22.2%) 2 (10%)   
 Chemotherapy 49 (26.3%) 3 (33.3%) 3 (15%)   
 Other 36 (19.4%) 0 (0.0%) 3 (15%)   
 None 13 (7.0%) 0 (0.0%) 2 (10%)   
 Unknown 4 (2.2%) 1 (11.1%) 1 (5%)   

aPatients with both known germline pathogenic and novel variants present in cfDNA were included in the known germline pathogenic category for these analyses.

bFor the statistical analyses, the Wilcoxon rank-sum test (age variable) and Pearson χ2 test (all categorical variables) were used.

Characteristics of cfDNA BRCA1/2 mutations

There was significant heterogeneity in the type of mutation and clinico-genomic characteristics of somatic BRCA1/2 mutations, as depicted in Table 2. Four patients (13.8%) had polyclonal (≥2) BRCA1/2 mutations and three patients (10.3%) had both BRCA1 and BRCA2 cfDNA mutations.

Table 2.

Characteristics of cfDNA BRCA1/2 mutations.

Characteristics of cfDNA BRCA1/2 mutations (N = 29, overall cohort)
CharacteristicNumber of patients
BRCA1 or BRCA2 BRCA1: 15 (51.7%) 
 BRCA2: 11 (37.9%) 
 Both BRCA1 and BRCA2: 3 (10.3%) 
Previously known germline pathogenic vs. novel variants Known germline pathogenic: 9 (31%) 
 Novel variants: 20 (69%) 
Clonal vs. subclonal Clonal: 16 (45.7%) 
 Subclonal: 19 (54.3%) 
Prior platinum or anthracycline treatment before cfDNA testing Prior platinum: 4 (13.8%) 
 Prior anthracycline: 16 (55.2%) 
 None: 11 (37.9%) 
Coexisting germline BRCA1/2 mutation Germline BRCA1 mutation: 1 (3.4%) 
 Germline BRCA2 mutation: 0 (0%) 
 No known germline BRCA1 or BRCA2 mutation: 28 (96.6%) 
Coexisting BRCA1/2 mutation detectable by tumor tissue genotyping BRCA1: 3 (15.8%) 
 BRCA2: 0 (0%) 
 No BRCA1 or BRCA2 mutation detected by tumor tissue genotyping: 16 (84.2%) 
Characteristics of cfDNA previously known germline pathogenic BRCA1/2 mutations (N = 9)a 
BRCA1 or BRCA2 BRCA1: 5 (55.6%) 
 BRCA2: 3 (33.3%) 
 Both BRCA1 and BRCA2: 1 (11.1%) 
Clonal vs. subclonal Clonal: 4 (44.4%) 
 Subclonal: 5 (55.6%) 
Prior platinum or anthracycline treatment before cfDNA testing Prior platinum: 4 (44.4%) 
 Prior anthracycline: 5 (55.6%) 
 None: 2 (22.2%) 
Coexisting germline BRCA1/2 mutation Germline BRCA1 mutation: 1 (11.1%) 
 Germline BRCA2 mutation: 0 (0%) 
 No known germline BRCA1 or BRCA2 mutation: 8 (88.9%) 
Coexisting BRCA1/2 mutation detectable by tumor tissue genotyping BRCA1: 2 (40%) 
 BRCA2: 0 (0%) 
 No BRCA1 or BRCA2 mutation detected by tumor tissue genotyping: 3 (60%) 
Characteristics of cfDNA novel variant BRCA1/2 mutations (N = 20)a 
BRCA1 or BRCA2 BRCA1: 10 (50.0%) 
 BRCA2: 5 (25.0%) 
 Both BRCA1 and BRCA2: 5 (25.0%) 
Clonal vs. subclonal Clonal: 10 (50.0%) 
 Subclonal: 10 (50.0%) 
Prior platinum or anthracycline treatment before cfDNA testing Prior platinum: 0 (0%) 
 Prior anthracycline: 11 (55.0%) 
 None: 9 (45.0%) 
Coexisting germline BRCA1/2 mutation Germline BRCA1 mutation: 0 (0%) 
 Germline BRCA2 mutation: 0 (0%) 
 No known germline BRCA1 or BRCA2 mutation: 20 (100%) 
Coexisting BRCA1/2 mutation detectable by tumor tissue genotyping BRCA1: 1 (6.7%) 
 BRCA2: 0 (0%) 
 No BRCA1 or BRCA2 mutation detected by tumor tissue genotyping: 14 (93.3%) 
Characteristics of cfDNA BRCA1/2 mutations (N = 29, overall cohort)
CharacteristicNumber of patients
BRCA1 or BRCA2 BRCA1: 15 (51.7%) 
 BRCA2: 11 (37.9%) 
 Both BRCA1 and BRCA2: 3 (10.3%) 
Previously known germline pathogenic vs. novel variants Known germline pathogenic: 9 (31%) 
 Novel variants: 20 (69%) 
Clonal vs. subclonal Clonal: 16 (45.7%) 
 Subclonal: 19 (54.3%) 
Prior platinum or anthracycline treatment before cfDNA testing Prior platinum: 4 (13.8%) 
 Prior anthracycline: 16 (55.2%) 
 None: 11 (37.9%) 
Coexisting germline BRCA1/2 mutation Germline BRCA1 mutation: 1 (3.4%) 
 Germline BRCA2 mutation: 0 (0%) 
 No known germline BRCA1 or BRCA2 mutation: 28 (96.6%) 
Coexisting BRCA1/2 mutation detectable by tumor tissue genotyping BRCA1: 3 (15.8%) 
 BRCA2: 0 (0%) 
 No BRCA1 or BRCA2 mutation detected by tumor tissue genotyping: 16 (84.2%) 
Characteristics of cfDNA previously known germline pathogenic BRCA1/2 mutations (N = 9)a 
BRCA1 or BRCA2 BRCA1: 5 (55.6%) 
 BRCA2: 3 (33.3%) 
 Both BRCA1 and BRCA2: 1 (11.1%) 
Clonal vs. subclonal Clonal: 4 (44.4%) 
 Subclonal: 5 (55.6%) 
Prior platinum or anthracycline treatment before cfDNA testing Prior platinum: 4 (44.4%) 
 Prior anthracycline: 5 (55.6%) 
 None: 2 (22.2%) 
Coexisting germline BRCA1/2 mutation Germline BRCA1 mutation: 1 (11.1%) 
 Germline BRCA2 mutation: 0 (0%) 
 No known germline BRCA1 or BRCA2 mutation: 8 (88.9%) 
Coexisting BRCA1/2 mutation detectable by tumor tissue genotyping BRCA1: 2 (40%) 
 BRCA2: 0 (0%) 
 No BRCA1 or BRCA2 mutation detected by tumor tissue genotyping: 3 (60%) 
Characteristics of cfDNA novel variant BRCA1/2 mutations (N = 20)a 
BRCA1 or BRCA2 BRCA1: 10 (50.0%) 
 BRCA2: 5 (25.0%) 
 Both BRCA1 and BRCA2: 5 (25.0%) 
Clonal vs. subclonal Clonal: 10 (50.0%) 
 Subclonal: 10 (50.0%) 
Prior platinum or anthracycline treatment before cfDNA testing Prior platinum: 0 (0%) 
 Prior anthracycline: 11 (55.0%) 
 None: 9 (45.0%) 
Coexisting germline BRCA1/2 mutation Germline BRCA1 mutation: 0 (0%) 
 Germline BRCA2 mutation: 0 (0%) 
 No known germline BRCA1 or BRCA2 mutation: 20 (100%) 
Coexisting BRCA1/2 mutation detectable by tumor tissue genotyping BRCA1: 1 (6.7%) 
 BRCA2: 0 (0%) 
 No BRCA1 or BRCA2 mutation detected by tumor tissue genotyping: 14 (93.3%) 

aFor these analyses, patients with both known germline pathogenic and novel variants in cfDNA were included in the known germline pathogenic category.

Among the various BRCA1/2 mutations detected in the cohort, 11 (29%) were protein-truncating alterations (six frameshift insertions/deletions, two splice variants, and three nonsense mutations), all of which were predicted to be pathogenic. In contrast, 20 (53%) were missense point mutations, the majority of which were novel variants of unknown significance.

Altogether, 45.7% of the detected mutations were clonal (i.e., MAF ratio ≥ 25%) and 54.3% were subclonal (MAF ratio < 25%). In the entire BRCA1/2-mutant population, 62% of patients with a cfDNA BRCA1/2 mutation had received prior platinum and/or anthracycline therapy, and this treatment distribution was similar in patients with known germline pathogenic mutations. However, fewer patients with novel variants had received prior anthracycline or platinum therapy.

Of the 29 patients with somatic BRCA1 or BRCA2 mutations, 21 had archival tumor available for analysis (details in Supplementary Data, 52.4% on metastatic lesion at MBC diagnosis, 33.3% on a metastatic lesion after MBC diagnosis, and 14.3% on primary tumor specimen). Of 21, only three patients had somatic BRCA1 mutations detectable in the archival tumor tissue, all from metastatic specimens. The BRCA1 variants in these three cases were identical in the blood and metastatic tumor tissue (details in Supplementary Data). The detailed clinical history and timing of tissue versus blood genotyping is outlined in Supplementary Table S2.

One patient (patient ID No. 17) had a known coexisting germline BRCA1 mutation (c. 3875del4 mutation) as well as three additional somatic BRCA1 mutations in exon 10 which appeared to be reversion mutations, restoring the open reading frame in different ways. This patient had received platinum chemotherapy, which may have triggered the development of the BRCA1 reversion mutations that can restore BRCA1 function leading to acquired resistance to platinum and/or PARP inhibitors (32, 33).

Coexisting cfDNA mutations

As depicted in Fig. 1, a wide spectrum of coexisting mutations were detected with somatic BRCA1/2 mutations, highlighting genomic complexity and clonal heterogeneity. The most common mutations included PIK3CA (44.8%), TP53 (41.4%), NF1 (27.6%), ERBB2 (20.7%), MET (17.2%), ARID1A (17.2%), EGFR (13.8%), APC (13.8%), NOTCH1 (13.8%), RHOA (10.3%), ESR1 (10.3%), KIT (10.3%), and FGFR3 (6.9%).

Figure 1.

Coexisting cfDNA mutations in patients with MBC with cfDNA BRCA1/2 mutations. Blue depicts WT genes, yellow denotes known pathogenic mutations, and green signifies novel variants.

Figure 1.

Coexisting cfDNA mutations in patients with MBC with cfDNA BRCA1/2 mutations. Blue depicts WT genes, yellow denotes known pathogenic mutations, and green signifies novel variants.

Close modal

TP53 mutations were more common among patients with known germline pathogenic BRCA1/2 mutations (77%) as compared with patients with novel BRCA1/2 variants (30%). Supplementary Fig. S3 depicts the mutation spectrum by MAF for each patient in this cohort.

In terms of mutation signatures, comprehensive analysis revealed that both the “BRCA” signature (COSMIC Signature 3, associated with homologous recombinant deficiency) as well as APOBEC mutational signatures (COSMIC Signatures 2+13) were present in the BRCA1/2 cohort, highlighting the functional heterogeneity with somatic BRCA mutations (Fig. 2).

Figure 2.

APOBEC mutation signature in somatic BRCA-mutant patients. The “Lego” plot on the left is the study data (all somatic mutations in the cohort). The “Lego” plot on the right is a “reference” for comparison: the APOBEC mutation signature. It also has the plot axes labeled. The rows are not mutational signature, but rather the whole plot is a mutation signature. The APOBEC mutation signature was clearly observed in this cohort. Approximately 40% of mutations were assignable to the APOBEC mutation signature (back row of bars in the “Lego” plot, COSMIC signatures 2+13), summing mutations across the 29 patients that were found to carry somatic BRCA mutations. Other contributors to the mutations in these BRCA-mutant patients were the “aging” signature (COSMIC Signature 1) and the “BRCA” signature (COSMIC Signature 3).

Figure 2.

APOBEC mutation signature in somatic BRCA-mutant patients. The “Lego” plot on the left is the study data (all somatic mutations in the cohort). The “Lego” plot on the right is a “reference” for comparison: the APOBEC mutation signature. It also has the plot axes labeled. The rows are not mutational signature, but rather the whole plot is a mutation signature. The APOBEC mutation signature was clearly observed in this cohort. Approximately 40% of mutations were assignable to the APOBEC mutation signature (back row of bars in the “Lego” plot, COSMIC signatures 2+13), summing mutations across the 29 patients that were found to carry somatic BRCA mutations. Other contributors to the mutations in these BRCA-mutant patients were the “aging” signature (COSMIC Signature 1) and the “BRCA” signature (COSMIC Signature 3).

Close modal

BRCA protein expression and olaparib sensitivity in CTC culture lines

Finally, to evaluate the functional significance of somatic BRCA1 mutations, we analyzed gene expression and BRCA1 protein expression in the CTC lines Brx401 (harboring a known germline pathogenic somatic BRCA1 mutant derived from patient ID No. 16 in this cohort) and Brx07 (harboring BRCA1 WT) using Western blot analysis (Fig. 3A). Additional coexisting mutations in Brx401 included TSC2, TP53, and NOTCH2. No BRCA1 protein was seen in the cell line with a somatic BRCA1 mutation (Brx401), highlighting functional loss of BRCA1 protein, but full-length BRCA1 protein was seen in the cell line harboring BRCA1 WT (Brx07). In addition, we treated the Brx401 and Brx07 CTC lines with olaparib for 5 days and evaluated cell proliferation. The known germline pathogenic somatic BRCA1-mutant line (BRx401) demonstrated increased sensitivity to olaparib (IC50 6.48 μmol/L) compared with the BRCA WT line (BRx07; IC50 63.68 μmol/L; Fig. 3B). Indeed, the patient (patient ID No. 16 from whom the CTC culture line BRx401 was developed) derived therapeutic benefit with carboplatin (PFS ∼6 months), but not eribulin (PFS 3 months), further confirming that the somatic BRCA1 mutation was a likely driver mutation and consistent with the known platinum sensitivity of pathogenic BRCA1/2 mutations.

Figure 3.

BRCA protein expression and olaparib sensitivity in CTC culture lines. A, CTC lines Brx401 (acquired somatic BRCA1 mutant, MUT) and Brx07 (WT) were analyzed by Western blot for BRCA1 protein expression. No BRCA1 protein was detected in Brx401. The red arrow indicates full-length BRCA1 protein (220 kDa) detected in Brx07. B, Brx401, Brx07, and BRx142 (novel variant BRCA2 mutant acquired (Brx142 (draw 2) from baseline Brx142 (draw 1) after treatment with a serum estrogen receptor degrader)). CTC lines were treated with increasing concentrations of olaparib for 5 days and cell proliferation was evaluated. Brx401 (acquired somatic BRCA1 mutant, IC50: 6.48 μmol/L) was more sensitive to PARP inhibition compared with BRx07 (WT, IC50: 63.68 μmol/L) and the Brx142 lines.

Figure 3.

BRCA protein expression and olaparib sensitivity in CTC culture lines. A, CTC lines Brx401 (acquired somatic BRCA1 mutant, MUT) and Brx07 (WT) were analyzed by Western blot for BRCA1 protein expression. No BRCA1 protein was detected in Brx401. The red arrow indicates full-length BRCA1 protein (220 kDa) detected in Brx07. B, Brx401, Brx07, and BRx142 (novel variant BRCA2 mutant acquired (Brx142 (draw 2) from baseline Brx142 (draw 1) after treatment with a serum estrogen receptor degrader)). CTC lines were treated with increasing concentrations of olaparib for 5 days and cell proliferation was evaluated. Brx401 (acquired somatic BRCA1 mutant, IC50: 6.48 μmol/L) was more sensitive to PARP inhibition compared with BRx07 (WT, IC50: 63.68 μmol/L) and the Brx142 lines.

Close modal

Furthermore, in a third breast cancer CTC line harboring a novel somatic BRCA2 variant (Brx142) from a patient with HR+ MBC, we observed no increased sensitivity to olaparib compared with cells with BRCA WT. Interestingly, in this line, an APOBEC mutational signature was widely evident, and it encompassed the novel BRCA2 mutation itself. Additional coexisting mutations observed included SMARCA4 (p.1787M and p.E1606Q), CIC (p.E2258Q and p.K2423N), PI3KCA, BCLAF1, FAM135B, ALK, CSMD3, MYCN, FAT1, NF2, MUC16, MAFB, ZNF331, APC, and HIF1A, but a TP53 mutation was not present in this line. Thus, the somatic BRCA2 variant is likely a passenger mutation induced by increased APOBEC activity.

We report that a proportion of patients with MBC harbor somatically acquired BRCA1/2 mutations in cfDNA, but that there is significant diversity in their associated clinico-genomic characteristics. While some of these mutations may be pathogenic in nature, others may not have functional significance. We were able to showcase the differences among these in selected cases for which cultured CTCs could be generated, but in general, distinguishing between cases with pathogenic BRCA1/2 mutations likely to respond to PARP inhibition and those with passenger mutations will require careful interpretation of both mutational and clinical parameters, and ultimately confirmation in prospective clinical trials. We identified that many detected mutations are subclonal. The clinical utility of using PARP inhibition to treat subclones with acquired BRCA1/2 mutations within a heterogeneous cancer is not known. The advent of PARP inhibitors as an approved therapy for germline BRCA1/2-mutant advanced breast cancer and the efficacy of DNA-damaging agents in BRCA1/2 germline–mutant patients makes the identification of nonfamilial cases with tumors that have BRCA-like features important, because this may help extend the application of PARP inhibitors (7, 8, 34), as has been demonstrated in ovarian cancer where germline and somatic BRCA-mutant tumors have similar responses to PARP inhibition and platinum salts (35–43).

We identified that 13.5% of patients with MBC had somatic BRCA1 or BRCA2 mutations detectable by cfDNA. This mutation frequency is higher than expected on the basis of the rates of somatic BRCA1/2 mutations in primary breast cancer (13), and possibly reflects the acquisition of mutations under therapeutic pressure (3, 4, 44). The majority of the BRCA1/2 mutations detected by cfDNA are not currently known to be pathogenic, and rather were novel variants, some of which might be passenger mutations resulting from APOBEC activity and other mutagenic conditions (14, 23, 45). A limitation of the Guardant360 assay used in these analyses is that tumor mutation burden cannot be calculated, so we could not determine whether the presence of cfDNA BRCA1/2 mutations may be linked to an increased mutation rate, although more sophisticated cfDNA assays to evaluate this association could be considered in the future.

There are currently no guidelines to determine the pathogenicity of somatic BRCA1/2 mutations. On the basis of our work, we advocate the approach summarized in Fig. 4 to determine the pathogenicity of a somatic BRCA1/2 mutation detected in cfDNA prior to consideration of PARP inhibitor therapy. We recommend initial review of detected somatic BRCA1/2 mutations by expert genetic counselors, utilizing open databases such as ClinVar (18) to understand the functional impact of genetic variants including splice site changes, frameshifts, stop codons, and missense mutations at different locations within the BRCA1/2 coding sequence. We recognize that this approach is limited by the uncertainty of extrapolating pathogenic mutation status from germline to somatic sequence analysis, and the fact that many mutations detected are likely to be novel variants whose functional significance is not known. The prospective development of large somatic genomic databases will be helpful in future classification. Clinical characteristics may help determine the potential for a somatic BRCA1/2 mutation to be pathogenic, such as TNBC histology and young age at diagnosis, criteria that are also characteristic of pathogenic germline BRCA1/2 mutations (46, 47). In contrast, most of the HR+ cases with cfDNA BRCA1/2 mutations were novel variants of uncertain significance. The coexisting genomic environment may provide clues such as coexisting TP53 mutations, which we observed more commonly in patients with pathogenic BRCA1/2 mutations, similar to the association between TP53 mutations and germline BRCA1/2 mutations (48). While these criteria may provide guidance in interpreting cfDNA BRCA1/2 mutations, ultimately prospective clinical trials of PARP inhibitors in nonfamilial breast cancer are needed (49, 50).

Figure 4.

Approach to establish the pathogenic nature of a somatic BRCA1/2 mutation in cfDNA. On the basis of our work, we recommend determining the functional impact on the BRCA1/2 protein (step 1), using clinical characteristics and the coexisting genomic landscape (steps 2–3) to help corroborate the presence of a pathogenic mutation. Future goals to aid in this assessment include developing real-time CTC culture for individualized preclinical testing of individual BRCA1/2 variants, and obtaining data on the utility of PARP inhibition for various cfDNA BRCA1/2 mutations from large prospective clinical trials.

Figure 4.

Approach to establish the pathogenic nature of a somatic BRCA1/2 mutation in cfDNA. On the basis of our work, we recommend determining the functional impact on the BRCA1/2 protein (step 1), using clinical characteristics and the coexisting genomic landscape (steps 2–3) to help corroborate the presence of a pathogenic mutation. Future goals to aid in this assessment include developing real-time CTC culture for individualized preclinical testing of individual BRCA1/2 variants, and obtaining data on the utility of PARP inhibition for various cfDNA BRCA1/2 mutations from large prospective clinical trials.

Close modal

N. Vidula reports grants from Pfizer (research funding to institution MGH and travel funding) outside the submitted work. B. Nagy reports personal fees from Guardant Health, Inc. (employee and shareholder) during the conduct of the study. S.J. Isakoff reports personal fees from Immunomedics, Mylan, Myriad, Puma, OncoPep, and AbbVie, grants from AbbVie (institution), AstraZeneca (institution), Merck (institution), OncoPep (institution), PharmaMar (institution), and Genentech (institution) outside the submitted work. D. Juric reports grants and personal fees from Novartis, Genentech, Eisai, EMD Serono, and Syros, and Petra Pharma outside the submitted work, personal fees from Ipsen, and Relay Therapeutics, MapKure, and Vibliome outside the submitted work, grants from Takeda, Amgen, Celgene, and Placon Therapeutics, and InventisBio and Infinity Pharmaceuticals outside the submitted work. S. Wander reports personal fees from Foundation Medicine (consulting) and Puma Biotechnology (consulting) outside the submitted work. L. Spring reports consulting fees from Novartis and Puma, research funding to institution from Merck and Tesaro, and travel reimbursement from Merck and Tesaro. B. Moy reports grants from PUMA Biotechnology (to institution) outside the submitted work. R. Lanman reports other from Guardant Health, Inc. (employee and stockholder) during the conduct of the study. A.J. Iafrate reports from ArcherDx (equity) during the conduct of the study, personal fees from Repare (consulting), grants from Sanofi (for brain tumor research) outside the submitted work, and has a patent for Anchored Multiplex PCR issued, licensed, and with royalties paid from ArcherDx. G. Getz reports grants from IBM and Pharmacyclics outside the submitted work; has a pending patent for MuTect about calling somatic mutations in cancer and owned by the Broad Institute, and a pending patent for MSMuTect on detecting indels in microsatellites and detecting MSI cancers, co-owned by Massachusetts General Hospital and the Broad Institute; is a founder, consultant, and holds privately held equity in Scorpion Therapeutics. D.A. Haber reports grants from NIH, HHMI, BCRF, and NFCR, and personal fees from Tell Bio (founder equity) during the conduct of the study, and reports that Massachusetts General Hospital has filed for patent protection for the microfluidic CTC isolation technology. A. Bardia reports grants and personal fees from Pfizer (grant to institution; consultant/advisory board), Genentech (grant to institution; consultant/advisory board), Novartis (grant to institution; consultant/advisory board), AstraZeneca/Daiichi (grant to institution; consultant/advisory board), Immunomedics (grant to institution; consultant/advisory board), Spectrum (grant to institution; consultant/advisory board), Merck (grant to institution; consultant/advisory board), Sanofi (grant to institution; consultant/advisory board), Radius Health (grant to institution; consultant/advisory board), and PUMA (grant to institution; consultant/advisory board), personal fees from Foundation Medicine (consultant/advisory board) and Phillips (consultant/advisory board), and grants from BioTheranostics outside the submitted work. No potential conflicts of interest were disclosed by the other authors.

Massachusetts General Hospital has applied for patents regarding the CTC-iChip technology and CTC detection signatures. D.A. Haber, M. Toner, and S. Maheswaran are cofounders and have equity in Tell Bio, which aims to commercialize the CTC-iChip technology.

N. Vidula: Conceptualization, data curation, formal analysis, supervision, investigation, methodology, writing-original draft, project administration, writing-review and editing. T. Dubash: Conceptualization, data curation, formal analysis, investigation, methodology, writing-original draft, writing-review and editing. M.S. Lawrence: Data curation, software, formal analysis, investigation, methodology, writing-original draft, writing-review and editing. A. Simoneau: Formal analysis, investigation, writing-review and editing. A. Niemierko: Software, formal analysis, investigation, writing-review and editing. E. Blouch: Data curation, formal analysis, investigation, writing-review and editing. B. Nagy: Data curation, formal analysis, investigation, writing-review and editing. W. Roh: Data curation, investigation, methodology, writing-review and editing. B. Chirn: Formal analysis, investigation, writing-review and editing. B.A. Reeves: Formal analysis, investigation, writing-review and editing. G. Malvarosa: Data curation, writing-review and editing. J. Lennerz: Data curation, investigation, writing-review and editing. S.J. Isakoff: Writing-review and editing. D. Juric: Writing-review and editing. D. Micalizzi: Writing-review and editing. S. Wander: Writing-review and editing. L. Spring: Writing-review and editing. B. Moy: Writing-review and editing. K. Shannon: Data curation, formal analysis, writing-review and editing. J. Younger: Writing-review and editing. R. Lanman: Writing-review and editing. M. Toner: Funding acquisition, investigation, writing-review and editing. A.J. Iafrate: Writing-review and editing. G. Getz: Funding acquisition, writing-review and editing. L. Zou: Writing-review and editing. L.W. Ellisen: Writing-review and editing. S. Maheswaran: Conceptualization, supervision, funding acquisition, investigation, writing-review and editing. D.A. Haber: Conceptualization, supervision, funding acquisition, writing-review and editing. A. Bardia: Conceptualization, formal analysis, supervision, investigation, methodology, writing-review and editing.

We thank the patients who participated in this study. This work was supported by grants from NIH (2O1CA129933 to D.A. Haber, 2U1EB012493 to M. Toner, D.A. Haber, and S. Maheswaran, and 5P41EB002503 to M. Toner), the Howard Hughes Medical Institute (to D.A. Haber), National Foundation for Cancer Research (to D.A. Haber), ESSCP Breast Cancer Research Fund (to S. Maheswaran), MGH Cancer Center startup funds (to M.S. Lawrence), and IBM/Broad collaboration on cancer genetics (to G. Getz).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Bardia
A
,
Haber
DA
. 
Solidifying liquid biopsies: can circulating tumor cell monitoring guide treatment selection in breast cancer?
J Clin Oncol
2014
;
32
:
3470
1
.
2.
Somlo
G
,
Lau
SK
,
Frankel
P
,
Hsieh
HB
,
Liu
X
,
Yang
L
, et al
Multiple biomarker expression on circulating tumor cells in comparison to tumor tissues from primary and metastatic sites in patients with locally advanced/inflammatory, and stage IV breast cancer, using a novel detection technology
.
Breast Cancer Res Treat
2011
;
128
:
155
63
.
3.
Cristofanilli
M
,
Budd
GT
,
Ellis
MJ
,
Stopeck
A
,
Matera
J
,
Miller
MC
, et al
Circulating tumor cells, disease progression, and survival in metastatic breast cancer
.
N Engl J Med
2004
;
351
:
781
91
.
4.
Yu
M
,
Bardia
A
,
Aceto
N
,
Bersani
F
,
Madden
MW
,
Donaldson
MC
, et al
Cancer therapy. Ex vivo culture of circulating breast tumor cells for individualized testing of drug susceptibility
.
Science
2014
;
345
:
216
20
.
5.
Vidula
N
,
Juric
D
,
Niemierko
A
,
Spring
L
,
Moy
B
,
Malvarosa
G
, et al
Comparison of tissue genotyping (TG) vs circulating tumor DNA (ctDNA) for selection of matched therapy and impact on clinical outcomes among patients with metastatic breast cancer (MBC)
.
J Clin Oncol
36
:
15s
, 
2018
(
suppl; abstr 1020
).
6.
Aggarwal
C
,
Thompson
JC
,
Black
TA
,
Katz
SI
,
Fan
R
,
Yee
SS
, et al
Clinical implications of plasma-based genotyping with the delivery of personalized therapy in metastatic non-small cell lung cancer
.
JAMA Oncol
2019
;
5
:
173
80
.
7.
Robson
M
,
Im
SA
,
Senkus
E
,
Xu
B
,
Domchek
SM
,
Masuda
N
, et al
Olaparib for metastatic breast cancer in patients with a germline BRCA mutation
.
N Engl J Med
2017
;
377
:
523
33
.
8.
Litton
JK
,
Rugo
HS
,
Ettl
J
,
Hurvitz
SA
,
Goncalves
A
,
Lee
KH
, et al
Talazoparib in patients with advanced breast cancer and germline BRCA mutation
.
N Engl J Med
2018
;
379
:
753
63
.
9.
Campeau
PM
,
Foulkes
WD
,
Tischkowitz
MD
. 
Hereditary breast cancer: new genetic developments, new therapeutic avenues
.
Hum Genet
2008
;
124
:
31
42
.
10.
Tutt
A
,
Robson
M
,
Garber
JE
,
Domchek
SM
,
Audeh
MW
,
Weitzel
JN
, et al
Oral poly(ADP-ribose) polymerase inhibitor olaparib in patients with BRCA1 or BRCA2 mutations and advanced breast cancer: a proof-of-concept trial
.
Lancet
2010
;
376
:
235
44
.
11.
Kaufman
B
,
Shapira-Frommer
R
,
Schmutzler
RK
,
Audeh
MW
,
Friedlander
M
,
Balmana
J
, et al
Olaparib monotherapy in patients with advanced cancer and a germline BRCA1/2 mutation
.
J Clin Oncol
2015
;
33
:
244
50
.
12.
Turner
NC
,
Telli
ML
,
Rugo
HS
,
Mailliez
A
,
Ettl
J
,
Grischke
E
, et al
Final results of a phase II study of talazoparib (TALA) following platinum or multiple cytotoxic regimens in advanced breast cancer patients with germline BRCA 1/2 mutations (ABRAZO)
.
J Clin Oncol
2017
;
35
:
1007
.
13.
Wang
X
. 
An exploration of mutation status of cancer genes in breast cancers
. J
Next Generat Sequenc Applic
2014
;
1
:
103
.
14.
Bertucci
F
,
Ng
CKY
,
Patsouris
A
,
Droin
N
,
Piscuoglio
S
,
Carbuccia
N
, et al
Genomic characterization of metastatic breast cancers
.
Nature
2019
;
569
:
560
4
.
15.
Lanman
RB
,
Mortimer
SA
,
Zill
OA
,
Sebisanovic
D
,
Lopez
R
,
Blau
S
, et al
Analytical and clinical validation of a digital sequencing panel for quantitative, highly accurate evaluation of cell-free circulating tumor DNA
.
PLoS One
2015
;
10
:
e0140712
.
16.
Taavitsainen
S
,
Annala
M
,
Ledet
E
,
Beja
K
,
Miller
PJ
,
Moses
M
, et al
Evaluation of commercial circulating tumor DNA test in metastatic prostate cancer
.
JCO Precis Oncol
2019
;
3
:
1
9
.
17.
Nance
T
,
Helman
E
,
Artieri
C
,
Yen
J
,
Slavin
TP
,
Chudova
D
, et al
A novel approach to differentiation of somatic vs. germline variants in liquid biopsies using a betabinomial model
. In:
Proceedings of the American Association for Cancer Research Annual Meeting; 2018 Apr 14–18; Chicago, IL
.
Philadelphia (PA)
:
AACR
; 
2018
.
Abstract nr 4272
.
18.
National Center for Biotechnology Information
. 
U.S. National Library of Medicine
.
ClinVar
Available from
: https://www.ncbi.nlm.nih.gov/clinvar/.
19.
Richards
S
,
Aziz
N
,
Bale
S
,
Bick
D
,
Das
S
,
Gastier-Foster
J
, et al
Standards and guidelines for the interpretation of sequence variants: a joint consensus recommendation of the American college of medical genetics and genomics and the association for molecular pathology
.
Genet Med
2015
;
17
:
405
24
.
20.
Zheng
Z
,
Liebers
M
,
Zhelyazkova
B
,
Cao
Y
,
Panditi
D
,
Lynch
KD
, et al
Anchored multiplex PCR for targeted next-generation sequencing
.
Nat Med
2014
;
20
:
1479
84
.
21.
Lawrence
MS
,
Stojanov
P
,
Polak
P
,
Kryukov
GV
,
Cibulskis
K
,
Sivachenko
A
, et al
Mutational heterogeneity in cancer and the search for new cancer-associated genes
.
Nature
2013
;
499
:
214
8
.
22.
Alexandrov
LB
,
Nik-Zainal
S
,
Wedge
DC
,
Aparicio
SA
,
Behjati
S
,
Biankin
AV
, et al
Signatures of mutational processes in human cancer
.
Nature
2013
;
500
:
415
21
.
23.
Buisson
R
,
Langenbucher
A
,
Bowen
D
,
Kwan
EE
,
Benes
CH
,
Zou
L
, et al
Passenger hotspot mutations in cancer driven by APOBEC3A and mesoscale genomic features
.
Science
2019
;
364
:
eaaw2872
.
24.
Ellrott
K
,
Bailey
MH
,
Saksena
G
,
Covington
KR
,
Kandoth
C
,
Stewart
C
, et al
Scalable open science approach for mutation calling of tumor exomes using multiple genomic pipelines
.
Cell Syst
2018
;
6
:
271
81.e7
.
25.
Costello
M
,
Pugh
TJ
,
Fennell
TJ
,
Stewart
C
,
Lichtenstein
L
,
Meldrim
JC
, et al
Discovery and characterization of artifactual mutations in deep coverage targeted capture sequencing data due to oxidative DNA damage during sample preparation
.
Nucleic Acids Res
2013
;
41
:
e67
.
26.
Cibulskis
K
,
McKenna
A
,
Fennell
T
,
Banks
E
,
DePristo
M
,
Getz
G
. 
ContEst: estimating cross-contamination of human samples in next-generation sequencing data
.
Bioinformatics
2011
;
27
:
2601
2
.
27.
Berger
MF
,
Lawrence
MS
,
Demichelis
F
,
Drier
Y
,
Cibulskis
K
,
Sivachenko
AY
, et al
The genomic complexity of primary human prostate cancer
.
Nature
2011
;
470
:
214
20
.
28.
Chapman
MA
,
Lawrence
MS
,
Keats
JJ
,
Cibulskis
K
,
Sougnez
C
,
Schinzel
AC
, et al
Initial genome sequencing and analysis of multiple myeloma
.
Nature
2011
;
471
:
467
72
.
29.
Saunders
CT
,
Wong
WS
,
Swamy
S
,
Becq
J
,
Murray
LJ
,
Cheetham
RK
. 
Strelka: accurate somatic small-variant calling from sequenced tumor-normal sample pairs
.
Bioinformatics
2012
;
28
:
1811
7
.
30.
Ramos
AH
,
Lichtenstein
L
,
Gupta
M
,
Lawrence
MS
,
Pugh
TJ
,
Saksena
G
, et al
Oncotator: cancer variant annotation tool
.
Hum Mutat
2015
;
36
:
E2423
9
.
31.
Tate
JG
,
Bamford
S
,
Jubb
HC
,
Sondka
Z
,
Beare
DM
,
Bindal
N
, et al
COSMIC: the catalogue of somatic mutations in cancer
.
Nucleic Acids Res
2019
;
47
:
D941
D7
.
32.
Lin
KK
,
Harrell
MI
,
Oza
AM
,
Oaknin
A
,
Ray-Coquard
I
,
Tinker
AV
, et al
BRCA reversion mutations in circulating tumor DNA predict primary and acquired resistance to the PARP inhibitor rucaparib in high-grade ovarian carcinoma
.
Cancer Discov
2019
;
9
:
210
9
.
33.
Weigelt
B
,
Comino-Mendez
I
,
de Bruijn
I
,
Tian
L
,
Meisel
JL
,
Garcia-Murillas
I
, et al
Diverse BRCA1 and BRCA2 reversion mutations in circulating cell-free DNA of therapy-resistant breast or ovarian cancer
.
Clin Cancer Res
2017
;
23
:
6708
20
.
34.
Tutt
A
,
Tovey
H
,
Cheang
MCU
,
Kernaghan
S
,
Kilbum
L
,
Gazinska
P
, et al
Carboplatin in BRCA1/2-mutated and triple-negative breast cancer BRCAness subgroups: the TNT trial
.
Nat Med
2018
;
24
:
628
37
.
35.
Lord
CJ
,
Ashworth
A
. 
PARP inhibitors: synthetic lethality in the clinic
.
Science
2017
;
355
:
1152
8
.
36.
Dougherty
BA
,
Lai
Z
,
Hodgson
DR
,
Orr
MCM
,
Hawryluk
M
,
Sun
J
, et al
Biological and clinical evidence for somatic mutations in BRCA1 and BRCA2 as predictive markers for olaparib response in high-grade serous ovarian cancers in the maintenance setting
.
Oncotarget
2017
;
8
:
43653
61
.
37.
Zhao
Q
,
Yang
J
,
Li
L
,
Cao
D
,
Yu
M
,
Shen
K
, et al
Germline and somatic mutations in homologous recombination genes among Chinese ovarian cancer patients detected using next-generation sequencing
.
J Gynecol Oncol
2017
;
28
:
e39
.
38.
Biranese
RC
,
Nakamura
KDM
,
Almeida
FGDSR
,
Ramalho
RF
,
Barros
BDF
,
Ferreira
ENE
, et al
BRCA1 deficiency is a recurrent event in early-onset triple-negative breast cancer: a comprehensive analysis of germline mutations and somatic promoter methylation
.
Breast Cancer Res Treat
2018 Feb
;
167
:
803
14
.
39.
Oza
AM
,
Tinker
AV
,
Oaknin
A
,
Shapira-Frommer
R
,
McNeish
IA
,
Swisher
EM
, et al
Antitumor activity and safety of the PARP inhibitor rucaparib in patients with high-grade ovarian carcinoma and a germline or somatic BRCA1 or BRCA2 mutation: integrated analysis of data from study 10 and ARIEL2
.
Gynecol Oncol
2017
;
147
:
267
75
.
40.
McCabe
N
,
Turner
NC
,
Lord
CJ
,
Kluzek
K
,
Bialkowska
A
,
Swift
S
, et al
Deficiency in the repair of DNA damage by homologous recombination and sensitivity to poly(ADP-ribose) polymerase inhibition
.
Cancer Res
2006
;
66
:
8109
15
.
41.
George
A
,
Banerjee
S
,
Kaye
S
. 
Olaparib and somatic BRCA mutations
.
Oncotarget
2017
;
8
:
43598
9
.
42.
Hennessy
BT
,
Timms
KM
,
Carey
MS
,
Gutin
A
,
Meyer
LA
,
Flake
DD
 II
, et al
Somatic mutations in BRCA1 and BRCA2 could expand the number of patients that benefit from poly (ADP ribose) polymerase inhibitors in ovarian cancer
.
J Clin Oncol
2010
;
28
:
3570
6
.
43.
Lheureux
S
,
Bruce
JP
,
Burnier
JV
,
Karakasis
K
,
Shaw
PA
,
Clarke
BA
, et al
Somatic BRCA1/2 recovery as a resistance mechanism after exceptional response to poly (ADP-ribose) polymerase inhibition
.
J Clin Oncol
2017
;
35
:
1240
9
.
44.
Patsouris
A
,
Tredan
O
,
Campion
L
,
Goncalves
A
,
Arnedos
M
,
Sablin
MP
, et al
An open-label, phase Ii study of rucaparib, a PARP inhibitor, in HER2-metastatic breast cancer patients with high genomic loss of heterozygosity
.
J Clin Oncol
36
:
15s
, 
2017
(
suppl; abstr TPS1112
).
45.
Andre
F
,
Filleron
T
,
Ng
C
,
Bertucci
F
,
Letourneau
C
,
Jacquet
A
, et al
Genomic characterization of metastatic breast cancer
.
[abstract]
. In:
Proceedings of the 2018 San Antonio Breast Cancer Symposium; 2018 Dec 4–8; San Antonio, TX
.
Philadelphia (PA)
:
AACR
; 
2019
.
Abstract nr GS1-08
.
46.
Greenup
R
,
Buchanan
A
,
Lorizio
W
,
Rhoads
K
,
Chan
S
,
Leedom
T
, et al
Prevalence of BRCA mutations among women with triple-negative breast cancer (TNBC) in a genetic counseling cohort
.
Ann Surg Oncol
2013
;
20
:
3254
8
.
47.
Copson
ER
,
Maishman
TC
,
Tapper
WJ
,
Cutress
RI
,
Greville-Heygate
S
,
Altman
DG
, et al
Germline BRCA mutation and outcome in young-onset breast cancer (POSH): a prospective cohort study
.
Lancet Oncol
2018
;
19
:
169
80
.
48.
Greenblatt
MS
,
Chappuis
PO
,
Bond
JP
,
Hamel
N
,
Foulkes
WD
. 
TP53 mutations in breast cancer associated with BRCA1 or BRCA2 germ-line mutations: distinctive spectrum and structural distribution
.
Cancer Res
2001
;
61
:
4092
7
.
49.
U.S. National Library of Medicine
. 
Evaluation of talazoparib, a PARP inhibitor, in patients with somatic BRCA mutant metastatic breast cancer: genotyping based clinical trial
.
Available from
: https://clinicaltrials.gov/ct2/show/NCT03990896.
50.
U.S. National Library of Medicine
. 
Olaparib in metastatic breast cancer
.
Available from
: https://clinicaltrials.gov/ct2/show/NCT03344965.