Molecular Diagnosis & Therapy

[Link]

REVIEW ARTICLE

Detection of Solid Tumor Molecular Residual Disease (MRD) Using

Circulating Tumor DNA (ctDNA)
Re‑I Chin1 · Kevin Chen1 · Abul Usmani1 · Chanelle Chua1 · Peter K. Harris1 · Michael S. Binkley2 · Tej D. Azad2 ·
Jonathan C. Dudley3 · Aadel A. Chaudhuri1,4,5

© Springer Nature Switzerland AG 2019

Abstract
Circulating tumor DNA (ctDNA) is a component of cell-free DNA that is shed by malignant tumors into the bloodstream and
other bodily fluids. Levels of ctDNA are typically low, particularly in patients with localized disease, requiring highly sophis-
ticated methods for detection and quantification. Multiple liquid biopsy methods have been developed for ctDNA analysis in
solid tumor malignancies and are now enabling detection and assessment of earlier stages of disease, post-treatment molecular
residual disease (MRD), resistance to targeted systemic therapy, and tumor mutational burden. Understanding ctDNA biology,
mechanisms of release, and clearance and size characteristics, in conjunction with the application of molecular barcoding
and targeted error correction, have increased the sensitivity and specificity of ctDNA detection techniques. Combinatorial
approaches including integration of ctDNA data with circulating protein biomarkers may further improve assay sensitivity and
broaden the scope of ctDNA applications. Circulating viral DNA may be utilized to monitor disease in some virally induced
malignancies. In spite of increasingly accurate methods of ctDNA detection, results need to be interpreted with caution given
that somatic mosaicisms such as clonal hematopoiesis of indeterminate potential (CHIP) may give rise to genetic variants in
the bloodstream unrelated to solid tumors, and the limited concordance observed between different commercial platforms.
Overall, highly precise ctDNA detection and quantification methods have the potential to transform clinical practice via
non-invasive monitoring of solid tumor malignancies, residual disease detection at earlier timepoints than standard clinical
and/or imaging surveillance, and treatment personalization based on real-time assessment of the tumor genomic landscape.

Key Points

Circulating tumor DNA (ctDNA) is shed into the blood-

stream and other bodily fluids and serves as a potential
biomarker for post-treatment molecular residual dis-
Re-I Chin and Kevin Chen contributed equally as co-first authors
ease (MRD) for solid tumor malignancies such as lung,
of this review.
breast, colon, pancreatic, and bladder cancer.
* Aadel A. Chaudhuri
Different techniques can detect ctDNA with high sen-
aadel@[Link]
sitivity and specificity, including digital PCR, multi-
1
Department of Radiation Oncology, Washington University plex PCR-based next-generation sequencing (NGS),
School of Medicine, St. Louis, MO, USA targeted hybrid capture-based NGS, and combinatorial
2
Department of Radiation Oncology, Stanford University approaches.
School of Medicine, Stanford, CA, USA
3
Interpretation of ctDNA needs to account for sources of
Department of Pathology, Johns Hopkins University School
of Medicine, Baltimore, MD, USA
somatic mosaicism such as clonal hematopoiesis of inde-
4
terminate potential (CHIP) that result in genetic variants
Department of Computer Science and Engineering,
Washington University, St. Louis, MO, USA
in cell-free DNA unrelated to a solid tumor malignancy.
5
Alvin J. Siteman Cancer Center, Barnes-Jewish Hospital
and Washington University School of Medicine, St. Louis,
MO, USA

Vol.:(0123456789)
 R.-I. Chin et al.

1 Cell‑Free DNA [26, 27]. As a whole, cfDNA has a short half-life ranging
between 16 min and 2.5 h [26, 28]. The half-life of cfDNA
Mandel and Métais [1] first described cell-free DNA may be longer when it is bound to protein complexes or
(cfDNA) in 1948, referring to extracellular DNA found in inside membrane vesicles since the cfDNA is less vulnerable
the blood plasma. Present in various forms, cfDNA can be to degradation by phagocytes [2]. Accumulation of cfDNA
encapsulated in lipid membrane microvesicles, trapped by can be attributed to an excessive release of DNA caused by
leukocytes, or bound to nucleosomes, serum and/or lipo- massive cell death, inefficient removal of dead cells, or a
proteins [2]. cfDNA circulates mostly in blood plasma, but combination of the two [29].
can also be found in various bodily fluids, including urine In healthy individuals, most of the cfDNA originates
[3–8], cerebrospinal fluid [9–11], pleural fluid [12–14], from hemopoietic cells such as erythrocytes, leukocytes, and
ascites [14], and saliva [15, 16]. Passive release via apop- endothelial cells, because hemopoietic cells are abundant, turn
tosis, necrosis, and phagocytosis account for the primary over rapidly, and have ready access to the vasculature [30,
mechanisms of cfDNA release [17–19] (Fig. 1). Active 31]. Normal tissues that undergo damage by ischemia, trauma,
secretion via extracellular vesicles or protein complexes is infection, or inflammation can also release cfDNA [32, 33].
also thought to contribute to cfDNA [20–23], although the Specifically, cfDNA has been shown to be elevated in patients
exact mechanisms have not been elucidated (Fig. 1). with myocardial infarction [34], cerebral infarction [35, 36],
The clearance of cfDNA is also not fully understood. lung transplant [37], trauma [38], parasitic infections [39],
In 1963, Tsumita and Iwanaga [24] studied the kinetics of urinary tract infections [5], inflammatory conditions such as
foreign DNA clearance by injecting radioactive DNA into systemic inflammatory response syndrome [40], rheumatoid
mice. They showed that 99% of the radioactivity was cleared arthritis [41], systemic lupus erythematosus [42], as well as
from the bloodstream in 30 min, and the highest increase of in patients who are pregnant [43] or undergo intense exercise
radioactivity was in the kidneys, followed by liver and spleen [44–46]. For each specific tissue type, factors such as the rates
[24]. However, the plasma levels of cfDNA did not seem of proliferation and apoptosis (cell turnover), total mass/vol-
to be dramatically different in patients with chronic kidney ume, and vascularity are primary determinants in the degree
disease or patients on peritoneal dialysis or hemodialysis, of contribution from that tissue to the cfDNA pool [2].
suggesting that renal elimination may not be the main mech- cfDNA is typically comprised of fragments of double-
anism of cfDNA clearance [25]. Other studies indicated that stranded DNA ranging between 150 and 200 base pairs in
cfDNA is cleared from the circulation via nuclease action length [47]. The commonly observed length of 166 bp corre-
sponds to the length of DNA wrapped around a nucleosome

Fig. 1  ctDNA release into the bloodstream from solid tumors. passenger mutations can be released in these ways, as can normal ger-
Mechanisms of release into the bloodstream include cellular apop- mline DNA, which typically comprises the majority of cell-free DNA
tosis, necrosis, phagocytosis, and active secretion. ctDNA driver and even in cancer patients. ctDNA circulating tumor DNA
ctDNA and MRD Detection

(~ 147 bp) plus linker DNA associated with histone H1 3 Challenges of ctDNA Detection
[17, 48–50]. Compared to cfDNA derived from non-cancer
cells, circulating tumor DNA (ctDNA) has been shown to To utilize ctDNA effectively as a cancer biomarker in clini-
be shorter [17, 51–53]. In fact, Mouliere et al. [54] demon- cal settings, its measurement methods must be reliable and
strated that selecting for fragments between 90 and 150 bp accurate. For most solid tumor malignancies, ctDNA levels
improved the detection sensitivity of ctDNA, with more than in the blood plasma are low, with mutant allele fractions
twofold median enrichment in > 95% of cases and more than typically less than 10% in advanced metastatic disease [73,
fourfold enrichment in > 10% of cases. Similarly, the size 75], and less than 1% in locally advanced non-metastatic
difference between circulating fetal and maternal DNA has disease [19, 63, 73]. ctDNA levels decrease further in early-
been used to improve the sensitivity of non-invasive prenatal stage cancers [63, 64, 73, 91] and after curative-intent treat-
testing [48, 55–57]. These studies suggest that the cell of ment [63, 64, 83], where mutant allele fractions are often
origin of the cfDNA can affect its specific fragmentation less than 0.1% [63, 64, 82, 84, 91]. Thus, highly sensitive
patterns, which may offer clues regarding cell type and biol- methods are necessary to measure ctDNA both pre-treatment
ogy and enhance the sensitivity and specificity of ctDNA and post-treatment. In cancer patients with comorbidities
detection [23, 31, 51, 54, 58]. such as anemia and poor performance status, clinical sam-
ples can be limited, thus decreasing the number of cfDNA
molecules available for assessment and further complicating
2 Circulating Tumor DNA (ctDNA) as a Tumor ctDNA detection [60, 91].
Biomarker ctDNA detection can also be confounded by clonal hemat-
opoiesis of indeterminate potential (CHIP) [92, 93]. CHIP
In 1987, Stroun et al. [59] first reported a potential correla- arises when age-dependent acquired mutations accumulate
tion between cfDNA and cancer in a report where ten of 37 in hematopoietic progenitor cells, leading to the formation
patients with advanced cancer had quantifiable cfDNA in of a genetically distinct subpopulation that contributes dis-
their plasma, compared with zero of 50 normal controls. proportionately to the population of mature blood cells [94,
More recently, tumor-derived cfDNA, commonly referred to 95]. These distinct subclones have driver mutations and have
as ctDNA, has shown incredible potential as a highly sensi- been implicated in hematologic diseases, such as myelod-
tive and specific cancer biomarker [47, 60, 61]. Quantita- ysplastic syndrome (MDS) [96], acute myeloid leukemia
tive characterization of ctDNA via liquid biopsy has been [97], chronic lymphocytic leukemia (CLL) [98], and hairy
associated with clinical and pathologic features of cancer, cell leukemia (HCL) [99]. However, CHIP typically occurs
including stage, tumor burden, vascularization, and response in elderly healthy individuals in the absence of clinically
to therapy [47, 60–62]. ctDNA is reflective of the muta- apparent hematologic disease [100–103]. Next-generation
tions present within the tumor from which it arose and sequencing (NGS) studies have shown that blood cells
thus includes both driver and passenger mutations [63–65] derived from CHIP are rarely found in individuals younger
(Fig. 1). The short half-life of ctDNA ensures that its detec- than 60 years old, but can be as high as 5–18% for people
tion captures tumor burden in real-time [66, 67]. older than 70 years old [95, 104]. Recent studies applying
As a whole, ctDNA has emerged as a dynamic plasma- ultra-sensitive targeted sequencing suggest that rates of
based biomarker with the translational potential to facili- CHIP mutations are much higher, as high as 92–95%, with
tate molecular profiling at diagnosis [6, 63–65, 68–74], the majority of CHIP mutations being present at low mutant
targeted therapy selection [6, 63, 68, 74–80], molecular allele fractions [92, 103]. This highlights the prevalence of
residual disease (MRD) detection [6, 63, 64, 81–84], and CHIP, and underscores the specificity challenge of detecting
post-treatment tumor surveillance [6, 63, 64, 66, 68, 69, 74, solid tumor-derived ctDNA.
76, 85–87] (Fig. 2). The molecular precision of longitudinal In the measurement of ctDNA, CHIP can result in false-
tumor surveillance via serial ctDNA measurement enables positive results due to detection of non-reference variants in
the identification of mutations that drive cancer progression the blood plasma, which is especially problematic when the
and treatment resistance [63, 64, 66, 68, 74–80] (Fig. 2). ctDNA mutant allele fraction is low in the setting of MRD
Crucial to its potential role in guiding therapeutic decision- detection [92, 93]. Focusing on clonal mutations may lead to
making, ctDNA may better capture intra-tumoral heteroge- more specific ctDNA detection by avoiding low allele frac-
neity than invasive biopsy, although highly sensitive meth- tion CHIP variants. Assays that involve sequencing of paired
ods are required to detect subclonal mutations [64, 65, 74, peripheral blood mononuclear cells (PBMCs) may also lead
75, 88–90]. to more specific ctDNA detection by enabling filtration of
Here we explore the key ctDNA detection methods, high- variants present in both cfDNA and PBMCs [63, 69, 84,
light ctDNA applications in various solid tumors, and dis- 92]. Reassuringly, the majority of CHIP mutations involve
cuss the potential clinical utility of ctDNA in oncology.
 R.-I. Chin et al.

Fig. 2  Potential diagnostic and therapeutic approaches for solid and oncogenomic profiling, personalized systemic therapy could
tumors based on ctDNA detection and oncogenomic profiling. potentially be offered, and patients could potentially be monitored
ctDNA may be measured pre- and post-definitive treatment of local- by ctDNA surveillance with subsequent treatment based on real-time
ized cancer using highly sensitive techniques. Assessment of both ctDNA-based oncogenomic profiling. ctDNA circulating tumor DNA,
driver and passenger mutations can maximize the clinical sensitiv- MRD molecular residual disease, tx treatment
ity of detection. Based on post-treatment ctDNA MRD detection

DNMT3A, TET2, and ASXL1, genes implicated in hema- important to refine ctDNA detection techniques to identify
tologic cancer but not commonly involved in solid tumor somatic mosaicism, such as by sequencing matched PBMCs
malignancy [93–95, 101–106]. However, TP53 is among and healthy donor plasma and using these results to filter
the next most commonly mutated gene in CHIP [92, 93, 95, variants detected in the cell-free compartment [63, 69, 84,
101–103, 106], accounting for up to ~10% of all impacted 92, 93].
genes [92], which is challenging given its high prevalence Given the multitude of factors affecting efficiency of
as a driver mutation in solid tumors [107] that is commonly ctDNA detection, sensitive and specific methods are crucial
tracked in ctDNA [108]. Thus, CHIP must be properly [60, 84, 91, 111]. Currently available ctDNA detection tech-
accounted for in order to specifically measure ctDNA, such niques are summarized in Table 1.
as by sequencing matched PBMCs to similar depth [92, 93],
especially when using ultra-sensitive assays that are capable
of achieving detection of low mutant allele fraction variants. 4 ctDNA Detection Methods
In addition to CHIP, other sources of somatic mosaicism
may serve as potential confounders in ctDNA analysis. 4.1 PCR Assays
Somatic mosaicism can arise from limitations in the fidelity
of genetic material inheritance, such as the intrinsic error Initial efforts to characterize ctDNA relied on allele-specific
rate of DNA replication, which generates base substitutions PCR (AS-PCR) [112], which demonstrated clinical utility
occurring at an estimated frequency of 1­ 0−9 per replicative [113, 114] but had limited detection sensitivity [115, 116].
cycle in vivo [109, 110]. Environmental factors can also Digital PCR (dPCR) was subsequently developed [117],
contribute to somaticism, leading to chromosome losses, which enabled highly sensitive detection of specific ctDNA
deletions, duplications, and inversions [110]. Since these mutant alleles [83, 118, 119]. While PCR-based techniques
variants may not represent malignant transformation, it is have been implemented clinically, including the US Food
Table 1  Summary of circulating tumor DNA detection techniques
Technique Example technologies Scale of analysis Method Detection limit Cost Assay personalization Advantages Limitations
(% of cfDNA)

AS-PCR ARMS [112] Single-locus or multi- Preferentially ampli- ~ 0.1–1 $ Some required Ease of use; ideal for Can only query small
plexed assays fies rare mutant detecting recur- number of variants
DNA molecules rent ‘hotspot’ point concurrently; cannot
ctDNA and MRD Detection

mutations detect mutations not

known a priori
dPCR dPCR [117] Single-locus or multi- Partitions target DNA ~ 0.01 $$ Some required High sensitivity Can only query small
ddPCR [119, plexed assays into different reac- number of variants
125–127] tions for massively concurrently; cannot
BEAMing [128, 129] parallel qPCR detect mutations not
known a priori
WGS WGS [143, 144] Genome-wide NGS of whole ~ 10 $$–$$$ Not required Entire genome is Low sensitivity; mostly
Plasma-Seq [139, genome interrogated limited to SCNA
140] detection
PARE [145]
Retrotransposon- FAST-SeqS [141] Genome-wide retro- PCR amplification ~5 $$ Not required Rapid aneuploidy Limited to aneuploidy
based amplicon mFAST-SeqS [152] transposon sites of retrotransposon assessment with detection
NGS WALDO [142] insertion sites prior lower cost than
to NGS analysis WGS
WES WES [79] Exome-wide NGS of whole exome ~ 5 $$$ Not required Entire exome is inter- Low sensitivity
rogated
Multiplex PCR-based TAm-Seq [155] Targeted sequencing PCR amplification ~ 0.01–2.0 $$ Some required High sensitivity Less comprehensive
NGS Enhanced TAm-Seq enriches targets (modern methods) than other NGS meth-
[156] prior to NGS ods; unable to detect
Safe-SeqS [111] analysis SCNAs; unable to
Natera® [64] detect fusions without
assay personalization
Hybrid Capture-based CAPP-Seq [70, 84] Targeted sequencing Subset of exome is ~ 0.001-0.5 $$ Not required High sensitivity; Less comprehensive
NGS TEC-Seq [73] hybridized to bioti- detects multiple than WGS or WES
Guardant360® [161, nylated probes and mutation types;
162] captured for NGS broadly applicable
FoundationOne® analysis without personali-
Liquid [163] zation
Combination CAPP-Seq + GRP Single-locus to Combines different Variable $$–$$$ Variable Improved detection Potentially more
approaches [149] genome-wide ctDNA detection compared to stand- time and resource
CancerSEEK [169] methods, sometimes ard ctDNA analysis intensive
UroSEEK [7] including protein alone in certain
biomarkers settings

ARMS Amplification Refractory Mutation System, AS-PCR allele-specific PCR, BEAMing beads, emulsion, amplification and magnetics, CAPP-Seq Cancer Personalized Profiling by Deep
Sequencing, cfDNA cell-free DNA, ctDNA circulating tumor DNA, ddPCR digital droplet PCR, dPCR digital PCR, FAST-SeqS Fast Aneuploidy Screening Test-Sequencing System, GRP
Genome Representation Profiling, mFAST-SeqS modified FAST-SeqS, NGS next-generation sequencing, PARE Personalized Analysis of Rearranged Ends, Plasma-Seq plasma sequencing, Safe-
SeqS Safe Sequencing System, SCNA somatic copy number alternation, TAm-Seq Tagged-Amplicon Sequencing, TEC-Seq Targeted Error Correction Sequencing, WALDO Within Sample Ane-
uploidy Detection, WES whole exome sequencing, WGS whole genome sequencing
 R.-I. Chin et al.

and Drug Administration (FDA)-approved ­Cobas® test to 134], melanoma [135], and bladder cancer [136], with a high
detect EGFR (epidermal growth factor receptor) mutations sensitivity detection threshold of ~ 0.01% [83]. While ddPCR
in lung cancer patients [120], they are limited by their abil- is highly sensitive, it requires assay customization, has very
ity to assess only a small number of pre-defined mutations limited multiplexing capacity, and cannot detect mutations
at a time [19]. not known a priori, thus limiting its ability to detect tumor
heterogeneity and emergent mutations [19, 137, 138].
4.1.1 Allele‑Specific PCR (AS‑PCR): Amplification Refractory
Mutation System (ARMS) or Amplification of Specific 4.2 Genome‑Wide Next‑Generation Sequencing
Alleles (PASA)‑PCR (NGS) Approaches

AS-PCR, also known as amplification refractory mutation NGS approaches have become prevalent for tumor sequenc-
system (ARMS) or amplification of specific alleles (PASA), ing and have also been applied to cfDNA for ctDNA detec-
is a modified form of PCR designed to detect recurrent hot- tion [19, 60]. Whole genome sequencing (WGS), retro-
spot point mutations or single nucleotide polymorphisms transposon-based amplicon sequencing, and whole exome
(SNPs) in plasma and serum [112]. Invented by Newton sequencing (WES) enable mutation assessment across broad
et al. [112] in the late 1980s, AS-PCR [113] gained popu- areas of the genome, and are not limited to querying variants
larity in the mid-2000s. Instead of selecting primers from known a priori [19, 60]. This is a clear advantage over PCR-
an invariant region of the genome to amplify a polymorphic based methods. However, WGS and WES are costly, which
area between them as in traditional PCR, AS-PCR selects limits sequencing depth and assay sensitivity [19, 60]. Typi-
primers from at least one polymorphic region, with the cally, WGS assays applied to cfDNA achieve a sequencing
mutations of interest located at (or near) its 3′-end [121, depth of ~ 0.1x  [60, 139, 140] and WES achieves a sequenc-
122]. DNA polymerase only amplifies DNA when the nucle- ing depth of ~ 100x [74, 79]. Additionally, retrotransposon-
otide at the 3′-end of the primer perfectly complements the based amplicon sequencing is limited to aneuploidy assess-
base at the variant, thereby preferentially amplifying mutant ment [141, 142]. These disadvantages make these methods
DNA molecules [123]. While AS-PCR is relatively afford- less practical for routine clinical use than more targeted
able, has a limit of detection of ~0.1-1% [115, 116], and has methods (Fig. 3).
successfully detected key mutations in non-small cell lung
cancer (NSCLC) [114], it can only query one genetic locus 4.2.1 Whole Genome Sequencing (WGS): Plasma‑Seq
per assay [112]. This drawback has limited the widescale and Personalized Analysis of Rearranged Ends (PARE)
application of AS-PCR for ctDNA detection.
WGS has demonstrated the ability to detect chromosomal
4.1.2 Digital PCR (dPCR) and Beads, Emulsions, alterations in the plasma of patients with various cancer
Amplification, and Magnetics (BEAM)‑ing PCR types [143, 144]. Although such studies suggest feasible
clinical application to some patients, the cost and time
In an effort to improve ctDNA detection sensitivity using required to perform WGS and the associated analysis
PCR, Vogelstein and Kinzler [117] introduced a novel tech- have proven prohibitive for routine clinical implementa-
nique in 1999 called digital PCR (dPCR) [117]. Compared tion [143, 144]. To mitigate these limitations, Heitzer et al.
to conventional real-time quantitative PCR (qPCR), dPCR [139] devised a shallow whole genome-wide NGS approach
offers improved identification, absolute quantification, and called Plasma-Seq [140]. This method uses Illumina MiSeq,
rare allele detection by partitioning samples into multi- a benchtop high-throughput sequencing instrument to expe-
ple parallel quantitative PCR reactions [117, 124]. These dite runtime, with the goal of detecting somatic copy num-
parallel reactions can take place in a single reaction tube, ber alterations (SCNAs) in plasma-derived cfDNA with a
with PCR performed within aqueous microdroplets in an sequencing depth of ~ 0.1x  [139, 140]. Using Plasma-Seq,
oil emulsion, thus earning the name digital droplet PCR investigators detected multiple SCNAs in chromosome arms
(ddPCR) [119, 125–127]. ddPCR maximizes sensitiv- 8q, 8p, and the androgen receptor locus in patients with met-
ity by having a single molecule per compartment, where astatic prostate cancer, reporting a sensitivity and specificity
some portions of these reactions contain the target molecule of > 80% when plasma ctDNA concentrations were ≥ 10%
(positive) while others do not (negative) [118, 119]. Varia- [139]. However, Plasma-Seq’s shallow sequencing depth
tions of ddPCR, such as Beads, Emulsions, Amplification, (~ 0.1x ) made it difficult to detect structural chromosomal
and Magnetics (BEAM)-ing PCR, have also been devised rearrangements with high confidence, and the sequencing
[128, 129]. Indeed, dPCR and ddPCR have been applied resolution was too low to identify single nucleotide variants
to detect ctDNA in various cancers, such as NSCLC [80, (SNVs) [139, 140]. Leary et al. [145] thus implemented a
130–132], colorectal cancer [129, 133], breast cancer [83, technique called Personalized Analysis of Rearranged Ends
ctDNA and MRD Detection

Fig. 3  Trade‑offs between

sequencing breadth, depth,
cost, and limit of detection.
In general, there is an inverse
correlation between sequencing
breadth versus assay cost, and
sequencing depth versus assay
limit of detection. Since ctDNA
mutant allele fractions are typi-
cally low, especially in the set-
tings of early cancer detection
and MRD detection, targeted
approaches including hybrid
capture-based NGS, PCR-based
NGS, and digital PCR are
favored over broader sequencing
methods such as whole exome
or whole genome sequencing.
ctDNA circulating tumor DNA,
LOD limit of detection, MRD
molecular residual disease, NGS
next-generation sequencing

(PARE), utilizing ligation-based WGS [146] of sheared In samples with high aneuploidy, amplification based
tumor DNA to identify patient-specific genomic rearrange- methods such as Fast-SeqS [141] or Within Sample Aneu-
ments, followed by the design of ddPCR assays target- ploidy Detection (WALDO) [142] can provide genome-wide
ing rearranged loci to quantify their levels within cfDNA mutation information cost effectively by limiting the hypoth-
for tumor monitoring. Other investigators have also used esis space to long interspersed nuclear elements (LINEs).
WGS-based techniques to successfully detect cancer-specific While FAST-SeqS relies on differences between read depths
genomic rearrangements and SCNAs in blood plasma to at specific genomic loci and the average read depth in those
quantify solid tumor disease burden [147–150]. loci in other samples to identify aneuploidy [141], WALDO
utilizes a matched normal or curated selection of control
4.2.2 Retrotransposon‑Based Amplicon Sequencing: Fast samples, eliminating batch effects that could otherwise skew
Aneuploidy Screening Test‑Sequencing System data [142]. WALDO also controls for amplification bias
(FAST‑SeqS), modified FAST‑SeqS (mFAST‑SeqS), derived from amplicon size and detects chromosomal aber-
and Within Sample Aneuploidy Detection (WALDO) rations with > 90% sensitivity and > 99% specificity when
neoplastic cell fraction is > 5% [142]. The additional use of
Plasma-Seq performs best in tumors with high ctDNA a Support Vector Machine allows for aneuploidy detection
mutant allele fractions (≥ 10%) because its sequencing with 78% sensitivity at 99% specificity in samples with as
depths are too low to reliably detect lower levels of ctDNA little as 1% tumor-derived DNA [142]. Due to their mecha-
[139, 151]. Since plasma ctDNA concentration can vary nisms of detection, WALDO and Fast-SeqS have lower per-
widely [69, 150, 151], Belic et al. [152] devised a method formance when applied to cancers with less aneuploidy.
to identify plasma samples with ≥ 10% ctDNA that are suit-
able for Plasma-Seq analysis. To estimate ctDNA percent- 4.2.3 Whole Exome Sequencing
ages in plasma, Belic et al. [152] modified a method called
Fast Aneuploidy Screening Test-Sequencing System (FAST- To improve detection sensitivity and reduce cost while main-
SeqS) described by Kinde et al. [141] that used specific taining comprehensive coverage of likely mutated genomic
primers to amplify L1 retrotransposon regions dispersed regions, researchers turned to WES, which restricts the
throughout the genome. Even with modified FAST-SeqS sequencing space to protein-coding exons. Although the
(mFast-SeqS) as a pre-screening technique, WGS remains exome only represents ~ 1.5% of the whole genome, it has
limited by high costs and low detection sensitivity (Fig. 3), been reported to be enriched for disease-causing somatic
thus limiting its potential clinical utilization [19, 60, 153]. mutations [154]. Murtaza et al. [79] published the first
application of this approach to cell-free DNA, using WES
 R.-I. Chin et al.

to analyze serial plasma samples containing high percent- variants. The UMIs were used to group sequencing results
ages of ctDNA (between 5% and 55%) to track genomic into PCR families, which robustly distinguished biologic
evolution and response to therapy in patients with metastatic variants from PCR errors, enabling reduction of the error
cancer receiving systemic therapy. The average depth of rate to less than an average of 2.0 × 10−4 errors/bp [111].
sequencing coverage ranged from 31- to 160-fold across 19 Safe-SeqS was successfully applied in clinical cohorts where
plasma samples [79]. This study identified increased repre- it demonstrated the ability to detect > 75% patients with mul-
sentation of mutant alleles associated with the emergence of tiple advanced cancer types [69] and highly specific detec-
therapeutic resistance, including an activating mutation in tion of MRD after surgery for colorectal cancer [81, 82].
PIK3CA (phosphatidylinositol-4,5-bisphosphate 3-kinase, Abbosh et al. [64] applied a similar UMI-based multi-
catalytic subunit alpha) after treatment with paclitaxel, a plex PCR NGS approach that was patient-customized with
truncating mutation in RB1 (retinoblastoma 1) after treat- primers designed to detect SNVs identified by tumor exome
ment with cisplatin, and a truncating mutation in MED1 sequencing in conjunction with the company N ­ atera® [64].
(mediator complex subunit 1) after treatment with tamoxifen The authors reported results from TRACERx [Tracking
and trastuzumab [79]. The authors recommended utilizing Non-Small-Cell Lung Cancer (NSCLC) Evolution Through
WES for advanced or metastatic cancers, where the median Therapy (Rx)], a multicenter prospective study that followed
pre-treatment ctDNA mutant allele fraction in plasma is at patients with stage IA through to IIIA NSCLC treated with
least 5–10% [79], to identify acquired drug resistance and curative-intent surgical resection [90, 157]. High-depth,
evaluate clonal evolution non-invasively. multi-region tumor whole-exome sequencing (M-Seq) was
performed to account for intra-tumoral heterogeneity and to
4.3 Targeted NGS Approaches construct phylogenetic trees [90, 157]. Personalized, patient-
specific multiplex PCR assays targeting clonal and subclonal
To improve sequencing depth and enable low-frequency SNVs and incorporating UMIs were applied to cfDNA, then
mutant allele detection, NGS-based techniques have been followed by NGS [64]. These assays prioritized coverage of
tailored for targeted or ‘focused’ sequencing of specific gene SNVs in driver genes, but also included candidate passenger
panels [19, 60, 91], instead of untargeted or ‘broad’ WGS variants [64]. This technique yielded an analytical sensitiv-
or WES (Fig. 3). These focused NGS-based approaches ity of above 99% for detecting SNVs present at greater than
can be subdivided into amplicon- or hybrid capture-based 0.1% and a specificity of detecting a single SNV of 99.6%
sequencing. [64].
As a whole, amplicon-based targeted NGS methods are
4.3.1 Multiplex PCR‑based NGS: Tagged‑Amplicon robust and highly sensitive and specific. However PCR
Sequencing (TAm‑Seq), Enhanced TAm‑Seq, Safe amplification can potentially bias the observed mutant allele
Sequencing System (Safe‑SeqS), and Natera® fraction [158], and these techniques are limited to mutation
detection within the queried amplicon space [19, 60, 91].
In these approaches, PCR amplification of selected genes
enriches for the target regions of interest prior to NGS 4.3.2 Hybrid Capture‑Based NGS: Cancer Personalized
analysis. Forshew et al. [155] pioneered Tagged-Amplicon Profiling by Deep Sequencing (CAPP‑Seq),
Sequencing (TAm-Seq), demonstrating detection of cancer Targeted Error Correction Sequencing (TEC‑Seq),
mutations with allele frequencies as low as 2%, with sen- ­Guardant360®, and ­FoundationOne® Liquid
sitivity and specificity over 97%, and successfully applied
this technique to monitor cancer mutations in patients with Hybrid capture-based NGS was recently developed to
advanced ovarian cancer [155]. This method was subse- improve the detection of multiple cancer mutations with
quently applied to monitor ctDNA in patients with meta- high sensitivity and without significant prior knowledge of
static breast cancer [85]. Enhanced TAm-Seq, which incor- the alterations [19, 60, 91]. In this ‘hybrid capture’ approach,
porates more efficient library preparation and statistical relevant DNA sequences are hybridized to biotinylated
analysis, later demonstrated identification of > 97.66% point probes [70]. Biotin is bound to streptavidin beads and then
mutations at > 0.5% allele frequency [156]. unbound DNA is washed away to ‘pull down’ the bait DNA
To improve ctDNA detection, Kinde et al. [111] reported for NGS analysis [70].
a method of unique molecular identifier (UMI)-based multi- Using this principle, Newman et al. [70] implemented an
plex PCR followed by NGS called Safe-SeqS (Safe Sequenc- ultrasensitive ctDNA detection technique called Cancer Per-
ing System), where they tagged each template molecule with sonalized Profiling by Deep Sequencing (CAPP-Seq) [70].
a UMI, amplified each tagged molecule using PCR to create Recurrently mutated genomic regions in the population of a
UMI families, and performed NGS. This redundant sequenc- given cancer type are identified through bioinformatic analy-
ing approach enabled highly sensitive detection of rare sis of cancer WES and/or WGS data from databases such as
ctDNA and MRD Detection

The Cancer Genome Atlas (TCGA) and Catalog of Somatic specificity reported to be 97% [162]. Another commer-
Mutations in Cancer (COSMIC) [70]. Biotinylated probes cial assay that utilizes hybrid capture-based NGS technol-
are designed against these recurrently mutated regions, and ogy, Foundation Medicine Inc.’s ­FoundationOne® Liquid,
are referred to collectively as the ‘Selector’ [70]. The ‘Selec- includes 70 genes implicated in cancer growth and genomic
tor’ is applied to cfDNA of cancer patients, which is fol- biomarkers for microsatellite instability in its capture panel
lowed by NGS in order to quantitate ctDNA with a detection [163]. It is reported to have > 99% analytical sensitivity for
limit of ~ 0.02% [19, 70]. The detection limit was signifi- SNVs, indels, and gene rearrangements at mutant allele frac-
cantly improved to ~ 0.001% when a NGS error-correction tions > 0.5%, and 99% positive predictive value for all altera-
strategy called integrated digital error suppression (IDES) tions [163]. While these commercially available platforms
was implemented. This strategy utilized UMIs to reduce the have high self-reported accuracy, specificity, and sensitiv-
effect of PCR errors and a bioinformatic error correction ity to detect and quantify ctDNA, a comparative study of
step called ‘polishing’ to reduce the effect of stereotypical two hybrid capture-based NGS platforms demonstrated that
background artifacts [84]. An analogous hybrid capture 40% of patients had ctDNA alterations that were completely
NGS method with error correction, targeted error correction incongruent between the two assays [164]. Furthermore, by
sequencing (TEC-Seq), was developed by Phallen et al. [73] not incorporating sequencing information from matched
and demonstrated ability to detect early-stage cancers. A PBMCs in these commercial assays, their reported results
key difference, however, is that CAPP-Seq leverages duplex may be confounded by CHIP [92, 93]. Future comparative
sequencing in its detection approach while TEC-Seq does studies are needed to determine the reliability and poten-
not [73, 84], thus CAPP-Seq likely has a lower error rate per tial utility of these platforms in the clinical management of
base and may achieve a lower theoretical limit of detection patients with cancer.
than TEC-Seq.
CAPP-Seq has some unique advantages compared to 4.4 Combinatorial Approaches
other methods. It has the capability to detect SNVs, inser-
tions/deletions (indels), SCNAs, and genomic rearrange- While some tumor types such as lung cancer and melanoma
ments without assay personalization [70]. It incorporates the have high tumor mutational burden (TMB) with hotspot
Fusion And Chromosomal Translocation Enumeration and regions containing SNVs and indels that are amenable to
Recovery Algorithm (FACTERA), a highly sensitive and ctDNA analysis using the aforementioned techniques [69,
specific method for identifying gene rearrangements [159]. 160, 165], other cancer types may contain mutations at low
Additionally, CAPP-Seq is generalizable to nearly all tumor prevalence and/or harbor a more heterogeneous spectrum of
types and does not require patient-specific optimization [19, mutation types including genome-wide SCNAs [149, 165].
70, 72, 86, 149, 160]. Compared to amplicon-based NGS, To optimize ctDNA analysis in these scenarios, investigators
CAPP-Seq can more reliably detect copy number changes have incorporated approaches that combine multiple detec-
and allow for detection of fusion proteins [19, 70, 158, 159]. tion methods. For example, Przybyl et al. [149] combined
Finally, because the CAPP-Seq Selector is not customized CAPP-Seq and Genome Representation Profiling (GRP), a
per patient, but rather enables broad assessment within a shallow whole-genome sequence technique for the assess-
cancer type [19, 70], results from sequencing can reveal new ment of genome-wide SCNAs [166], to detect tumor-derived
and complex mechanisms of carcinogenesis and treatment alterations within the cell-free compartment of leiomyosar-
resistance [75, 76]. CAPP-Seq has been applied to the detec- coma patients. Springer et al. [7] developed a technique for
tion of MRD in localized lung cancer after treatment [63], detecting bladder cancer from urine sediment, called Uro-
early response assessment of diffuse large B cell lymphoma SEEK, that applied three separate tests to urine sediment:
(DLBCL) after the first cycle of chemotherapy [86], as well a ten-gene panel for Safe-SeqS (that included KRAS, TP53,
as to bladder cancer post-treatment surveillance using a ver- and PIK3CA), a TERT promoter-specific Safe-SeqS test, and
sion of the assay adapted to urinary cfDNA [6]. FAST-SeqS, which was used to evaluate genome-wide ane-
In addition to CAPP-Seq and Tec-Seq, several other uploidy using an unpaired PCR primer targeting ~ 38,000
ctDNA sequencing techniques utilize hybrid capture-based members of long interspersed nucleotide element-1 retro-
NGS. For instance, G ­ uardant360® (Guardant Health, Inc.) transposons (LINEs) [7].
utilizes a 150 kb panel encompassing 73 cancer-related In another recent study by O’Leary et al. [68], ctDNA
genes for hybrid capture, followed by NGS with an average was assessed at baseline and end of treatment from 195
sequencing depth of ~ 15,000x, noise filtering and molecular breast cancer patients in the PALOMA-3 (Palbociclib [PD-
tracking, and variant calling for SNVs, indels, CNAs, and 0332991] Combined with Fulvestrant in Hormone Recep-
fusions [161, 162]. Validation of the ­Guardant360® assay tor+ HER2-Negative Metastatic Breast Cancer After Endo-
in adult patients with advanced-stage solid tumors revealed crine Failure) randomized phase III trial [68, 167]. The
high clinical sensitivity (85.9%), and SNV detection authors developed a method to assess tumor DNA purity in
 R.-I. Chin et al.

plasma via a targeted amplicon panel against ~ 1000 SNPs in 5 Application of ctDNA Molecular Residual
regions commonly lost in breast cancer, and also performed Disease Detection to Solid Tumors
dPCR for PIK3CA and ESR1 mutations [68]. Using this
method, they identified plasma samples with tumor purity in Analysis of ctDNA has shown promising clinical potential
plasma > 10% at day 1 and end of treatment, and performed as a way to detect MRD for solid tumors following curative-
WES on these samples (from 14 patients) and PyClone [168] intent treatment and in advance of clinical or radiographic
analysis to detail clonal evolution non-invasively, including disease relapse [63]. Furthermore, in small studies MRD
treatment-resistant subclonal RB1 and FGFR2 mutations detection by ctDNA profiling correlated with worse progno-
that were confirmed by ddPCR [68]. The authors used their sis in patients with different types of solid tumor malignancy
exome sequencing data to design a targeted panel for error- [6, 63, 64, 81–83, 87, 91].
corrected amplicon-based NGS to assess the remainder of In Sects. 5.1–5.5, recent developments in ctDNA MRD
their cohort, thus more broadly defining the genomic land- detection for lung, breast, colon, pancreatic, and bladder
scape and clonal evolution in response to breast cancer treat- cancer are discussed with a focus on how early post-treat-
ment with palbociclib and fulvestrant, and validated several ment ctDNA detection may inform and potentially improve
of their findings with ddPCR [68]. Overall, the combina- adjuvant treatment strategies (Figs. 2 and 4).
torial approach employed by O’Leary et al. [68] capably
utilized the complementary strengths of WES, targeted 5.1 Lung Cancer
sequencing, and ddPCR, to enable a thorough tumor evolu-
tion analysis on an clinically meaningful cohort. In the treatment of lung cancer, routine clinical surveillance
Another combinatorial approach is to combine ctDNA involves radiographic imaging [170]. However, standard
detection with circulating protein biomarkers. Cohen et al. imaging methods can only detect macroscopic disease recur-
[169] utilized a multiplex PCR-based NGS assay to measure rence [171], can be equivocal because inflammation or fibro-
ctDNA using a 61-amplicon panel, and combined it with sis are difficult to distinguish from tumor tissue especially
protein biomarker assessment in a combined assay called after radiotherapy [172, 173], and outcomes are typically
CancerSEEK. The immunoassay platform for protein bio- poor after clinical disease progression [174].
markers was optimized to increase sensitivity while main- Chaudhuri et al. [63] hypothesized that ctDNA measured
taining high specificity in the detection of eight different shortly after curative-intent treatment could identify patients
cancer types (ovarian, liver, gastric, pancreatic, esophageal, with localized lung cancer who are at high risk for relapse,
colorectal, lung, and breast). When applied to 1005 patients and post-treatment ctDNA could outperform standard-of-
with non-metastatic cancer, CancerSEEK had a median care imaging surveillance. The authors used CAPP-Seq to
sensitivity and specificity of 70% and > 99%, respectively show that ctDNA detected within 4 months of treatment
[169]. Then, using machine learning to predict the cancer’s completion (defined as the MRD landmark) robustly identi-
tissue of origin, the assay localized the primary tumor to fied patients who eventually relapsed with 94% sensitivity
two anatomic sites in a median of 83% of patients and to and 100% specificity, due in part to the detection of multiple
a single site in a median of 63% of patients [169]. Using a somatic mutations (both driver and passenger) per patient
similar approach, Cohen et al. [74] combined Safe-SeqS for
KRAS mutations with carbohydrate antigen 19-9 (CA-19-9)
as well as other protein biomarkers including carcinoem-
bryonic antigen (CEA), hepatocyte growth factor (HGF),
and osteopontin (OPN) to develop a non-invasive blood test
for the detection of resectable pancreatic ductal adenocarci-
noma. The five-component combinatorial panel led to a sen-
sitivity of 64%, compared with 30% when KRAS was utilized
alone [74]. Together, these studies illustrate the potential
for combinatorial noninvasive assays to offer more sensitive
cancer detection.

Fig. 4  ctDNA MRD as a potential tool for adjuvant therapy clini‑

cal decision‑making. Postoperative ctDNA MRD analysis may be
useful for identifying patients who are candidates for adjuvant sys-
temic treatment versus those who can be safely observed, a paradigm
that is being tested in the c-TRAK-TN trial in patients with resectable
triple negative breast cancer (NCT03145961 [230]). ctDNA circulat-
ing tumor DNA, MRD molecular residual disease, tx treatment
ctDNA and MRD Detection

[63]. ctDNA levels at the MRD landmark were confidently 5.2 Breast Cancer
detected to mutant allele fractions as low as 0.003% [63].
Compared with ctDNA MRD-negative patients, ctDNA By using mutation tracking in postsurgical blood samples,
MRD-positive patients had significantly worse overall sur- Nicholas Turner’s group used a ddPCR approach to detect
vival (p < 0.001, hazard ratio [HR] 14.3, 95% confidence ctDNA and predict relapse in non-metastatic breast cancer
interval [CI] 3.2–64.1), while computed tomography (CT) patients treated with chemotherapy, surgery, and in some
imaging at the same MRD landmark was not prognostic [63]. cases radiotherapy [83]. Among patients who eventually
ctDNA MRD detection had a positive predictive value of relapsed, 50% had detectable ctDNA MRD in a single post-
100% and a negative predictive value of 93%. Furthermore, operative sample drawn at 2–4 weeks after surgery [83].
post-treatment ctDNA detection preceded radiographic pro- In 37 analyzed patients, ctDNA MRD-positive patients
gression by a median of 5.2 months in 72% of patients who exhibited significantly worse disease-free survival than
developed recurrence [63]. ctDNA MRD-negative patients (p < 0.0001, HR 25.1; 95%
In addition to using ctDNA levels to prognosticate, CI 4.08–130.5) [83]. Serial mutation tracking in 43 patients
molecular genomic information from ctDNA analysis could beyond the postoperative timepoint increased the sensitivity
potentially guide patient-directed therapy. As reported by of relapse prediction to 80% (p < 0.0001, HR 12.0, 95% CI
Chaudhuri et al. [63], some patients in the study cohort 3.36–43.07), with ctDNA detected at a median of 7.9 months
had activating EGFR mutations detectable in ctDNA at earlier than clinical relapse [83], thus opening a window
the MRD landmark. In these patients, ctDNA MRD detec- to potentially intervene sooner with second-line treatment.
tion might have enabled earlier targeted intervention with While the ddPCR approach to ctDNA MRD detection is
a first-generation EGFR tyrosine kinase inhibitor, such as highly sensitive, it requires patient-specific assay customiza-
erlotinib [175, 176]. The authors also showed that quan- tion via a priori knowledge of tumor-specific somatic muta-
tification of non-synonymous mutations in the CAPP-Seq tions [19, 60, 83]. The authors thus also applied a hybrid
Selector space allowed for estimation of exome-wide TMB capture-based NGS approach to assess ctDNA pre- and post-
from ctDNA [63], a finding also demonstrated by Gandara treatment in five patients, and showed examples of emergent
­ oundationOne® CDx NGS assay [178]
et al. [177] with the F mutations detected post-treatment that were not identified
using cfDNA derived from the POPLAR and OAK advanced pre-treatment [83]. The Turner group further optimized
NSCLC immunotherapy trials [179, 180]. This is clinically this technology and applied it to advanced-stage breast
relevant because elevated TMB has been shown to be a cancer samples from the PALOMA-3 randomized phase III
biomarker for response to immune checkpoint blockade in trial [68, 167], and used it to analyze clonal evolution non-
NSCLC [181–183]. invasively, including the identification of treatment-resistant
Abbosh et al. [64, 91] also demonstrated the ability to subclonal RB1 and FGFR2 mutations that were confirmed
detect MRD shortly after definitive-intent NSCLC treatment by ddPCR [68].
completion using a personalized UMI-based multiplex PCR
NGS approach in the TRACERx study. The authors showed 5.3 Colon Cancer
that 13 of the 14 early-stage NSCLC patients with post-sur-
gical relapse diagnosed radiographically were also positive In colon cancer, the Roche ­AVENIO ® targeted hybrid
for the presence of ctDNA in plasma either before or at the capture-based NGS platform has been applied to patients
time of diagnosed recurrence [64]. Five of the patients with with advanced disease to identify resistance mechanisms
ctDNA detection that preceded radiographic confirmation to anti-EGFR monoclonal antibodies and to predict time to
of tumor recurrence by over 100 days had a median clonal treatment failure [76]. The use of ctDNA to predict recur-
mutant allele fraction of < 0.1%, suggesting highly sensi- rence risk could not only benefit high-risk patients at risk
tive methods are necessary to robustly detect MRD [63, 64, for relapse, but also may spare low-risk patients from poten-
91]. The authors also used a modified version of PyClone to tially serious morbidities associated with adjuvant systemic
perform phylogenetic ctDNA analyses to track the subclonal therapy [184]. While up to 40% of patients with stage II
nature of lung cancer relapse and progression, including a colorectal cancer undergo adjuvant therapy after resection
case where the subclonal mutation implicated in driving [185], the absolute risk reduction is estimated to be only
cancer relapse was a HER2 amplification event, which may 3–5% [186]. Using the Safe-SeqS platform to detect ctDNA
have been amenable to targeted therapy [64]. Overall, the MRD after surgery in stage II colon cancer patients, Tie
highly sensitive methods employed by Abbosh et al. [64, 91] et al. [82] showed that detectable ctDNA MRD measured
and Chaudhuri et al. [63] suggest that ctDNA MRD can be 4–10 weeks postoperatively in a population of 230 patients
robustly detected in patients with localized lung cancer after portended eventual radiographic recurrence with a hazard
treatment, potentially enabling earlier clinical intervention ratio of 18 (95% CI 7.9–40) and significantly outperformed
and more informed treatment selection for relapse. standard-of-care clinicopathologic characteristics. This
 R.-I. Chin et al.

approach demonstrated a very high specificity, with 97% of analyzing the genetic material in pelleted urine sediment.
patients with negative ctDNA postoperatively exhibiting no Applied to a cohort of 570 patients at risk for develop-
eventual relapse, while sensitivity was more modest at ~ 40% ing bladder cancer, UroSEEK demonstrated a significant
[82]. The lower sensitivity may have been because only a increase in the sensitivity of cancer detection (83%) com-
single mutant allele (identified by SafeSeqS [111] applied pared to conventional cytology (43%) [7]. UroSEEK positiv-
to formalin-fixed paraffin-embedded tumor) was tracked ity preceded the diagnosis of bladder cancer by an average
in the plasma of each patient [82]. This notwithstanding, of 2.3 months, and by more than 1 year in eight cases [7].
results remained highly significant when ctDNA was meas- Applied to a separate surveillance cohort of 322 patients
ured shortly after adjuvant chemotherapy in patients who with bladder cancer where urine was collected following
received it [82]. On serial follow-up, ctDNA detection pre- surgery, UroSEEK identified recurrent cancer on average
ceded imaging recurrence by a median of 5.5 months, signif- 7 months earlier than standard-of-care, with a clinical sen-
icantly earlier than the median 2-month lead time observed sitivity of 68% and specificity of 80% [7]. In contrast, sur-
with the CEA protein biomarker [82]. In a subsequent pro- veillance cytology had a clinical sensitivity of only 25% [7].
spective study of locally advanced rectal cancer evaluating More recently, Dudley et al. [6] focused on the cell-free
159 patients, Tie et al. [81] showed that ctDNA measured component of urine (liquid component, excluding pellet)
after neoadjuvant chemoradiation was also highly prognostic for bladder cancer detection. The authors demonstrated that
(p < 0.001). These results support the use of ctDNA MRD a targeted hybrid capture-based NGS technique that they
analysis in colorectal cancer after treatment as a potential termed urine tumor DNA CAPP-Seq (uCAPP-Seq) enables
biomarker to guide risk-adapted adjuvant therapy. sensitive detection of early-stage bladder cancer, with up
to 93% of patients having detectable urinary tumor DNA
5.4 Pancreatic Cancer (utDNA) pre-treatment, and a specificity of 96–100% when
assessed in 67 healthy adults [6]. utDNA was also assessed
Pancreatic cancer carries a poor prognosis with the major- in the post-treatment surveillance setting, and was detected
ity of patients succumbing to disease even when they are in 91% of patients who ultimately recurred, with utDNA
candidates for curative-intent surgery [187]. Genomically, detection preceding clinical progression in 92% of patients
KRAS mutations are very common in pancreatic cancer and by a median of 2.7 months [6]. Post-treatment uCAPP-Seq
plasma KRAS mutations have been shown to improve pre- significantly outperformed standard-of-care urine cytology
treatment detection compared to CA-19-9 alone [74]. To surveillance, which was positive in only 37.8% of patients
address whether ctDNA can be utilized to detect relapse who developed recurrence [6]. Moreover, uCAPP-Seq
after surgical resection, Sausen et al. [87] used dPCR to detected 100% of recurrence cases identified by cytology
measure ctDNA after surgical resection of pancreatic adeno- as well as 82% of recurrence cases missed by cytology [6].
carcinoma and demonstrated a significant improvement in Overall, the Dudley et al. [6] data and the Springer et al.
progression-free survival in patients with no post-treatment [7] data suggest that DNA derived from urine can be used
ctDNA detected compared to those with post-treatment to identify bladder cancer earlier and more sensitively than
ctDNA detected. Post-treatment ctDNA was detected on standard-of-care urine cytology, and facilitate non-invasive
average 3.1 months after surgery, approximately 6.5 months detection, genotyping, and monitoring.
earlier than progression scored by CT imaging [87]. It is
unclear whether ctDNA MRD status also translated to an
overall survival difference, and which patients received 6 Cell‑Free Viral DNA as a Tumor Biomarker
adjuvant chemotherapy. Regardless, these data are promis-
ing and indicate potential prognostic utility for ctDNA MRD It is well-documented that oncogenic viruses can lead to
in patients with resectable pancreatic cancer. various human cancers, such as human papillomavirus
virus (HPV) causing cervical [188, 189], anal [190], and
5.5 Bladder Cancer oropharyngeal cancer (OPC) [191], Epstein-Barr virus
(EBV) causing nasopharyngeal cancer [192], and hepa-
As liquid biopsy is often focused on sampling ctDNA in titis B virus (HBV) and hepatitis C virus (HCV) causing
plasma for solid tumor MRD detection, the analysis of hepatocellular carcinoma [193–196]. In most cases, viral
ctDNA in other bodily fluids may be more robust for certain genome integration represents a key step in carcinogenesis
cancer types [8, 60]. For example, urine contains cfDNA [197–199]. Due to the crucial role of viral integration in
arising from circulation that passes into the urinary system carcinogenesis, some have proposed measuring circulating
by glomerular filtration, as well as DNA arising from cells viral DNA in plasma and serum as a prognostic biomarker
that make up the urinary tract [3, 8]. Springer et al. [7] devel- in cancer diagnosis [200–209], treatment monitoring [208,
oped UroSEEK to non-invasively detect bladder cancer by 210], and residual disease detection [207, 208, 211, 212].
ctDNA and MRD Detection

In nasopharyngeal carcinoma, the clinical utility of cancer screening [221]. Campitelli et al. demonstrated the
using circulating EBV DNA in diagnosis and progno- ability to detect tumor-specific HPVviral integration sites in
sis has been robustly demonstrated with multiple large the serum of 11 stage II to IV cervical cancer patients, and
cohorts of patients [204–207, 211–213]. For instance, in showed their correlation with tumor response to therapy in
Leung et al. [213], pre-treatment circulating EBV DNA two patients [208]. Additionally, Han et al. [222] reported
load was found to be an independent prognostic factor to that in a cohort of 19 women with HPV-positive cervical
Union for International Cancer Control (UICC) staging, cancer who received curative-intent chemoradiation, detect-
and combining EBV DNA data with UICC staging data able plasma HPV DNA at the end of therapy preceded the
led to improved risk discrimination, especially in patients clinical diagnosis of metastasis and was associated with infe-
with early-stage disease [213]. In Lin et al. [207], patients rior progression-free survival. Furthermore, 3-month plasma
with pre-treatment plasma EBV DNA concentrations of HPV DNA detection was more accurate than 3-month FDG-
at least 1500 copies/mL had worse overall survival and PET (fluorodeoxyglucose–positron emission tomography)
relapse-free survival, and patients with persistently detect- imaging for identifying MRD in these patients [222].
able plasma EBV DNA one week after completion of While multiple studies have detailed the role of circulat-
radiotherapy had significantly worse overall survival and ing viral DNA in the evaluation of various virus-induced
relapse-free survival [207]. Based on these studies, circu- cancers, limited progress has been made in using circulat-
lating EBV DNA has become one of the leading cfDNA ing HBV and HCV DNA in the setting of hepatocellular
blood tests for evaluating nasopharyngeal carcinoma in carcinoma. It may be difficult to use circulating viral DNA
China, Hong Kong, and Taiwan, where this cancer is in the diagnosis and monitoring of hepatocellular carcinoma
endemic [214]. because acute and/or chronic hepatitis may be difficult to
In OPC, HPV-positive disease is associated with distinguish from hepatocellular carcinoma [223]. In fact,
improved overall survival and progression-free survival this challenge limits the use of most viral ctDNA assays,
compared to HPV-negative disease [215, 216]. Dahlstrom especially for diagnostic screening [32]. Therefore, it is
et al. [217] showed that in HPV-positive OPC, there was important to establish clinically meaningful cut-off levels
a non-significant trend in pre-treatment circulating HPV and potentially incorporate viral ctDNA data with other
DNA detection portending worse progression-free survival. tumor biomarkers to improve the sensitivity and specificity
Hanna et al. [218] used ddPCR to assess five HPV strains at of these assays [32].
multiple timepoints throughout immunotherapy or chemo-
therapy treatment of 22 patients with recurrent, metastatic
HPV-positive OPC [218]. The authors demonstrated that 7 Discussion and Future Perspectives
tumor burden measured by imaging strongly correlated with
HPV levels in the cfDNA, that median plasma HPV cfDNA Growing evidence supports the use of ctDNA as a solid
levels negatively correlated with survival, and that all par- tumor MRD biomarker, with the potential to guide thera-
ticipants demonstrated a corresponding change in their peutic decision-making. These studies have predominantly
HPV cfDNA levels at a median of 16 days before restaging utilized blood plasma, although research suggests that analy-
scans confirmed treatment response or progression [218]. ses of other bodily fluids such as urine or saliva may hold
To better elucidate the heterogeneity of HPV-positive OPC, great potential [7, 8, 16]. Due to the low ctDNA concentra-
Gupta et al. [219] are leading an ongoing study that has tions associated with MRD, detection techniques in the clini-
revealed two broad patterns of HPV DNA in the blood: (1) cal workflow must consistently detect mutations in plasma
a favorable-risk cohort of patients characterized by very high at mutant allele fractions of < 0.1% and should incorporate
levels of pre-treatment HPV DNA in the blood, which drop strategies to control for sequencing artifacts [91]. Such strat-
to undetectable after treatment; and (2) an unfavorable-risk egies include multi-mutation detection and tracking, UMIs
cohort of patients with lower pre-treatment levels of HPV to reduce the effect of PCR errors, and approaches to reduce
DNA that rapidly increase after initiation of therapy. These stereotypic noise such as background polishing [64, 73, 84,
studies illustrate the molecular heterogeneity of HPV-posi- 111].
tive OPC, suggesting the amount of circulating HPV DNA While preliminary data on the clinical utility of ctDNA
in the blood as well as clearance kinetics could aid further in MRD detection is promising, the studies demonstrating
risk-stratification in a cohort of patients who already have a this are mostly small, restricted to limited applications, and
favorable prognosis based on molecular testing [215–219]. need to be validated using large independent cohorts [47,
In cervical cancer, HPV DNA testing from a cervical 60, 61]. Only with these further studies can we move for-
swab was shown to have greater sensitivity for the detec- ward to the next important question, which is whether act-
tion of cervical intraepithelial neoplasia than the Papanico- ing on a ctDNA MRD-positive result can improve clinical
laou screening test [220], the standard-of-care for cervical outcomes, or if ctDNA MRD can be used to more precisely
 R.-I. Chin et al.

guide adjuvant therapy. Prospective trials testing the clinical therapy. By optimizing risk versus benefit analyses using
utility of ctDNA MRD-based therapy will be the ultimate ctDNA-based prognostication, treatment de-escalation
determinant of the significance of this research. could spare low-risk patients from the toxicity of unneces-
Accordingly, multiple trials are being set up in the ctDNA sary adjuvant therapy. On the other hand, treatment inten-
space, with a [Link] search revealing that 358 sification with aggressive adjuvant therapy in high-risk
studies involve ctDNA-based observation or intervention. patients could maximize efficacy while disease burden and
When querying only actively recruiting studies, there are clonal heterogeneity are minimal [19, 91]. By using this
204 such studies. These studies include, to name but a few, risk-stratification strategy, prospective studies aiming to
NCT02887612 [224], a study from Sun Yat-sen University demonstrate clinical benefit of treatment could thus maxi-
in China that is querying ctDNA for the prediction of post- mize their therapeutic index.
operative relapse in early and intermediate stage gastric Future studies should focus on whether intervention at
cancer; NCT03691012 [225], a study from Johns Hopkins the early ctDNA MRD timepoint improves clinical out-
University and the Walter and Eliza Hall Institute of Medical comes. Currently, c-TRAK-TN is a multicenter phase II
Research measuring ctDNA as a marker of residual disease study of patients with resectable triple negative breast can-
and response to adjuvant chemotherapy in ovarian cancer; cer (NCT03145961 [230]), which aims to assess whether
and NCT03774758 [226], a study from the University of ctDNA screening for one year after completing primary
California in San Francisco that will follow patients under- treatment enables the detection of MRD. Patients with
going lung cancer screening with ctDNA testing. detectable ctDNA MRD will be randomly assigned in a
Further research also is required to explore the pre- 2:1 ratio to receive either the immune checkpoint inhibitor
analytical considerations and analytical validity of ctDNA pembrolizumab or undergo continued observation (Fig. 4).
analyses. For instance, more studies are necessary to address The results of this trial will provide insight into the utility
the pre-analytical variables that may affect ctDNA testing, of ctDNA MRD for the selection of patients for adjuvant
including specimen type and quantity, sampling timepoints, systemic therapy. The clinical trial NRG-HN001 is applying
handling variables, storage conditions and duration, and a similar approach to locoregionally advanced nasopharyn-
patient-related biologic factors [227]. Additionally, the geal cancer treated with curative-intent chemoradiation, with
accuracy, sensitivity, specificity, and robustness of differ- adjuvant chemotherapy given based on post-treatment MRD
ent ctDNA detection methods remain to be optimized and detection; however, it restricts analysis to only plasma EBV
refined, with closer attention needed for variants near an viral DNA. Similar clinical trials are needed to demonstrate
assay’s reported lower limit of detection. Cross-validation the clinical utility of ctDNA MRD in other disease types
studies across different platforms using standardized sam- and settings.
ples are necessary to assess head-to-head performance [228, It will also be important to demonstrate that ctDNA can
229]. Since mutant allele fraction calculations may become be cleared following effective adjuvant treatment, and that
inflated or deflated in some approaches versus others, such ctDNA clearance corresponds with improved patient sur-
studies will also help normalize these variations across vival. Abbosh et al. [64] demonstrated an example of this
platforms. with patient CRUK0013, who had early-stage lung adeno-
Discordance between different ctDNA assays can also be carcinoma with detectable ctDNA after surgery that was
related to the different sets of mutations queried. Given the cleared following adjuvant chemoradiation. The patient
low sensitivity of WES or WGS for ctDNA detection [19, remained ctDNA undetectable on surveillance, and this
60], assays typically target specific loci and differences in correlated with long-term disease-free survival [64]. Simi-
the targeted sequencing space can account for some of the larly, Tie et al. [82] demonstrated three examples of colon
differences observed between commercial assays. Finally, cancer patients who were ctDNA MRD-positive after sur-
differences in library preparation techniques, UMIs, duplex gery that was cleared by adjuvant chemotherapy, and these
sequencing, variant calling, and targeted error correction patients remained recurrence-free long-term [82]. It will be
may account for some of the differences observed between important to systematically design future studies measur-
platforms. ing ctDNA post-operatively (or post-definitive radiotherapy)
Aside from factoring in the aforementioned pre-ana- and again during adjuvant systemic treatment to confirm
lytical variables and analytical validity, future ctDNA the capability to track clearance of postoperative ctDNA
analyses including MRD detection must focus on dem- MRD, and correlate this with clinical outcomes.
onstrating clinical validity and utility in oncologic care. Large studies that apply ctDNA based genomic pro-
Potential clinical applications include risk-stratification of filing to clinical cohorts are already underway. Zill et al.
patients based on molecular treatment response, tracking [108] compared genomic alteration patterns in ctDNA in
mutations associated with acquired therapeutic resistance a cohort of 21,807 advanced cancer patients with those in
and clonal evolution, and selecting appropriate targeted published sequencing studies compiled in TCGA [108].
ctDNA and MRD Detection

Their comparison revealed significant correlation, except for 2. Vietsch EE, Wellstein A. Circulating DNA in cancer diagnosis
treatment-induced secondary mutations that conferred resist- and prognosis. In: Dammacco F, Silvestris F, editors. Oncog-
enomics. London: Elsevier Inc.; 2019.
ance, such as EGFRT790M and EGFRC797S [231–233], which 3. Botezatu I, Serdyuk O, Potapova G, et al. Genetic analysis of
were only observed in ctDNA. This likely reflected cohort DNA excreted in urine: a new approach for detecting specific
differences, with TCGA patients being pre-treatment and the genomic DNA sequences from cells dying in an organism. Clin
ctDNA analysis patients typically being post-treatment with Chem. 2000;46:1078–84.
4. Chan KCA, Leung SF, Yeung SW, et al. Quantitative analysis of
the assay performed to query suspected therapeutic resist- the transrenal excretion of circulating EBV DNA in nasopharyn-
ance. This observation supports data discussed earlier [68, geal carcinoma patients. Clin Cancer Res. 2008;14:4809–13.
74–77], illustrating that ctDNA can identify resistant sub- 5. Burnham P, Dadhania D, Heyang M, et al. Urinary cell-free DNA
clones with the potential to guide clinical decision-making. is a versatile analyte for monitoring infections of the urinary
tract. Nat Commun. 2018;9:1–10.
6. Dudley JC, Schroers-Martin J, Lazzareschi DV, et al. Detec-
tion and surveillance of bladder cancer using urine tumor
DNA. Cancer Discov. 2018. https​://[Link]/10.1158/2159-8290.
8 Conclusion CD-18-0825.
7. Springer SU, Chen C-H, Rodriguez Pena MDC, et al. Non-inva-
sive detection of urothelial cancer through the analysis of driver
The emergence of highly sensitive ctDNA assays has the gene mutations and aneuploidy. Elife. 2018;7:1–27.
potential to revolutionize cancer management. Future studies 8. Lu T, Li J. Clinical applications of urinary cell-free DNA in
will be necessary to determine how to best integrate com- cancer: current insights and promising future. Am J Cancer Res.
2017;7:2318–32.
plementary information from standard solid tumor biopsies, 9. Wang Y, Springer S, Zhang M, et al. Detection of tumor-
clinical examination, and medical imaging with oncog- derived DNA in cerebrospinal fluid of patients with primary
enomic and MRD information from ctDNA-based liquid tumors of the brain and spinal cord. Proc Natl Acad Sci USA.
biopsies. Ultimately, the implementation of ctDNA MRD 2015;112:9704–9.
10. Pan W, Gu W, Nagpal S, et al. Brain tumor mutations detected
analysis into the clinical workflow has the potential to guide in cerebral spinal fluid. Clin Chem. 2015;61:514–22.
clinical decision-making and improve patient outcomes in 11. De Mattos-Arruda L, Mayor R, Ng CKY, et al. Cerebrospinal
this era of personalized precision medicine. fluid-derived circulating tumour DNA better represents the
genomic alterations of brain tumours than plasma. Nat Com-
mun. 2015;6:1–6.
Compliance with Ethical Standards 12. Sriram KB, Relan V, Clarke BE, et al. Pleural fluid cell-free
DNA integrity index to identify cytologically negative malig-
Funding This work was funded by the NCI under award number nant pleural effusions including mesotheliomas. BMC Cancer.
K08CA238711 (A. A. Chaudhuri), a Cancer Research Foundation 2012;12:1–12.
Young Investigator Award (A. A. Chaudhuri), the Washington Uni- 13. Soh J, Toyooka S, Aoe K, et al. Usefulness of EGFR muta-
versity SPORE in Pancreatic Cancer Career Enhancement Program tion screening in pleural fluid to predict the clinical outcome
(A. A. Chaudhuri), and the Conquer Cancer Foundation ASCO Young of gefitinib treated patients with lung cancer. Int J Cancer.
Investigator Award (A. A. Chaudhuri), supported by Takeda Pharma- 2006;119:2353–8.
ceuticals. Any opinions, findings and conclusions expressed in this 14. Husain H, Nykin D, Bui N, et al. Cell-free DNA from ascites and
material are those of the authors and do not necessarily reflect those pleural effusions: molecular insights into genomic aberrations
of the American Society of Clinical O ­ ncology®, Conquer C ­ ancer®, and disease biology. Mol Cancer Ther. 2017;16:948–55.
or ­Takeda®. 15. Mithani SK, Smith IM, Zhou S, et al. Mitochondrial rese-
quencing arrays detect tumor-specific mutations in salivary
Conflict of interest Aadel A. Chaudhuri has received speaker honorar- rinses of patients with head and neck cancer. Clin Cancer Res.
ia and travel support from Varian Medical Systems, Roche Sequencing 2007;13:7335–40.
Solutions, and Foundation Medicine, Inc., has research support from 16. Wei F, Lin C-C, Joon A, et al. Noninvasive saliva-based EGFR
Roche Sequencing Solutions, has served as a consultant for Tempus gene mutation detection in patients with lung cancer. Am J
Labs and for Oscar Health, and is a scientific advisor for Roche Se- Respir Crit Care Med. 2014;190:1117–26.
quencing Solutions and for Geneoscopy, LLC. Jonathan C. Dudley 17. Jahr S, Hentze H, Englisch S, et al. DNA fragments in the
has served as a consultant for Merck. Re-I Chin, Kevin Chen, Abul blood plasma of cancer patients: quantitations and evidence
Usmani, Chanelle Chua, Peter K. Harris, Michael S. Binkley, and Tej for their origin from apoptotic and necrotic cells. Cancer Res.
D. Azad declare that they have no conflicts of interest that might be 2001;61:1659–65.
relevant to the contents of this review. 18. Choi JJ, Reich CF, Pisetsky DS. The role of macrophages in
the in vitro generation of extracellular DNA from apoptotic and
necrotic cells. Immunology. 2005;115:55–62.
19. Chaudhuri AA, Binkley MS, Osmundson EC, et al. Predicting
References radiotherapy responses and treatment outcomes through analysis
of circulating tumor DNA. Semin Radiat Oncol. 2015;25:305–12.
1. Mandel P, Métais P. Les acides nucléiques du plasma sanguin 20. Thierry AR, El Messaoudi S, Gahan PB, et al. Origins, struc-
chez l’ homme. C R Seances Soc Biol Fil. 1948;142:241–3. tures, and functions of circulating DNA in oncology. Cancer
Metastasis Rev. 2016;35:347–76.
 R.-I. Chin et al.

21. Stroun M, Lyautey J, Lederrey C, et al. About the possible in systemic lupus erythematosus patients. Cell Immunol.
origin and mechanism of circulating DNA apoptosis and active 2014;292:32–9.
DNA release. Ann N Y Acad Sci. 2006;906:161–8. 43. Lo D, Sargent IL, Redman CW, et al. Presence of fetal DNA in
22. Anker P, Stroun M, Maurice PA. Spontaneous extracellular maternal plasma and serum. Lancet. 1997;350:485–7.
synthesis of DNA released by human blood lymphocytes. Can- 44. Tug S, Helmig S, Deichmann ER, et al. Exercise-induced
cer Res. 1976;36:2832–9. increases in cell free DNA in human plasma originate predomi-
23. Bronkhorst AJ, Wentzel JF, Aucamp J, et al. Characterization nantly from cells of the haematopoietic lineage. Exerc Immunol
of the cell-free DNA released by cultured cancer cells. Bio- Rev. 2015;21:164–73.
chim Biophys Acta. 2016;1863:157–65. 45. Atamaniuk J, Vidotto C, Tschan H, et al. Increased concentra-
24. Tsumita T, Iwanaga M. Fate of injected deoxyribonucleic acid tions of cell-free plasma DNA after exhaustive exercise. Clin
in mice. Nature. 1963;198:1088–9. Chem. 2004;50:1664–8.
25. Korabecna M, Opatrna S, Wirth J, et al. Cell-free plasma DNA 46. Breitbach S, Tug S, Simon P. Circulating cell-free DNA: an up-
during peritoneal dialysis and hemodialysis and in patients with coming molecular marker in exercise physiology. Sport Med.
chronic kidney disease. Ann N Y Acad Sci. 2008;1137:296–301. 2012;42:565–86.
26. Lo YM, Zhang J, Leung TN, et al. Rapid clearance of fetal DNA 47. Corcoran RB, Chabner BA. Application of cell-free DNA analy-
from maternal plasma. Am J Hum Genet. 1999;64:218–24. sis to cancer treatment. N Engl J Med. 2018;379:1754–65.
27. Tamkovich SN, Cherepanova AV, Kolesnikova EV, et al. Circu- 48. Lo YMD, Chan KCA, Sun H, et al. Maternal plasma DNA
lating DNA and DNase activity in human blood. Ann N Y Acad sequencing reveals the genome-wide genetic and mutational
Sci. 2006;1075:191–6. profile of the fetus. Sci Transl Med. 2010;2:1–13.
28. Yao W, Mei C, Nan X, et al. Evaluation and comparison of 49. Chandrananda D, Thorne NP, Bahlo M. High-resolution charac-
in vitro degradation kinetics of DNA in serum, urine and saliva: terization of sequence signatures due to non-random cleavage of
a qualitative study. Gene. 2016;590:142–8. cell-free DNA. BMC Med Genom. 2015;8:1–19.
29. Elshimali YI, Khaddour H, Sarkissyan M, et al. The clinical uti- 50. Jiang P, Lo YMD. The long and short of circulating cell-free
lization of circulating cell free DNA (CCFDNA) in blood of DNA and the ins and outs of molecular diagnostics. Trends
cancer patients. Int J Mol Sci. 2013;14:18925–58. Genet. 2016;32:360–71.
30. Lui YYN, Chik KW, Chiu RWK, et al. Predominant hemat- 51. Underhill HR, Kitzman JO, Hellwig S, et al. Fragment length of
opoietic origin of cell-free DNA in plasma and serum after circulating tumor DNA. PLoS Genet. 2016;12:1–24.
sex-mismatched bone marrow transplantation. Clin Chem. 52. Mouliere F, El Messaoudi S, Gongora C, et al. Circulating cell-
2002;48:421–7. free DNA from colorectal cancer patients may reveal high KRAS
31. Snyder MW, Kircher M, Hill AJ, et al. Cell-free DNA comprises or BRAF mutation load. Transl Oncol. 2013;6:319–28.
an in vivo nucleosome footprint that informs its tissues-of-origin. 53. Mouliere F, Robert B, Peyrotte E, et al. High fragmentation
Cell. 2016;164:57–68. characterizes tumour-derived circulating DNA. PLoS One.
32. Schwarzenbach H, Hoon DSB, Pantel K. Cell-free nucleic 2011;6:e23418.
acids as biomarkers in cancer patients. Nat Rev Cancer. 54. Mouliere F, Chandrananda D, Piskorz AM, et al. Enhanced detec-
2011;11:426–37. tion of circulating tumor DNA by fragment size analysis. Sci
33. Lehmann-Werman R, Neiman D, Zemmour H, et al. Identifica- Transl Med. 2018;10:1–14.
tion of tissue-specific cell death using methylation patterns of 55. Yu SCY, Chan KCA, Zheng YWL, et al. Size-based molecular
circulating DNA. Proc Natl Acad Sci USA. 2016;113:E1826–34. diagnostics using plasma DNA for noninvasive prenatal testing.
34. Chang CPY, Chia RH, Wu TL, et al. Elevated cell-free serum Proc Natl Acad Sci USA. 2014;111:8583–8.
DNA detected in patients with myocardial infarction. Clin Chim 56. Lun FM, Tsui NB, Chan KA, et al. Noninvasive prenatal diagno-
Acta. 2003;327:95–101. sis of monogenic diseases by digital size selection and relative
35. Bustamante A, Mancha F, Macher HC, et al. Circulating cell-free mutation dosage on DNA in maternal plasma. Proc. Natl. Acad.
DNA is a predictor of short-term neurological outcome in stroke Sci. 2008;105:19920–19925.
patients treated with intravenous thrombolysis. J Circ Biomark- 57. Minarik G, Repiska G, Hyblova M, et al. Utilization of benchtop
ers. 2016;5:1–6. next generation sequencing platforms ion torrent PGM and miseq
36. O’Connell GC, Petrone AB, Tennant CS, et al. Circulating extra- in noninvasive prenatal testing for chromosome 21 trisomy and
cellular DNA levels are acutely elevated in ischaemic stroke and testing of impact of in silico and physical size selection on its
associated with innate immune system activation. Brain Inj. analytical performance. PLoS One. 2015;10:1–12.
2017;31:1369–75. 58. Ulz P, Thallinger GG, Auer M, et al. Inferring expressed genes
37. De Vlaminck I, Martin L, Kertesz M, et al. Noninvasive monitor- by whole-genome sequencing of plasma DNA. Nat Genet.
ing of infection and rejection after lung transplantation. Proc Natl 2016;48:1273–8.
Acad Sci USA. 2015;112:13336–41. 59. Stroun M, Anker P, Lyautey J, et al. Isolation and characterization
38. Lo YMD, Rainer TH, Chan LYS, et al. Plasma DNA as a prog- of DNA from the plasma of cancer patients. Eur J Cancer Clin
nostic marker in trauma patients. Clin Chem. 2000;46:319–23. Oncol. 1987;23:707–12.
39. Weerakoon KG, McManus DP. Cell-Free DNA as a diag- 60. Wan JCM, Massie C, Garcia-Corbacho J, et al. Liquid biopsies
nostic tool for human parasitic infections. Trends Parasitol. come of age: Towards implementation of circulating tumour
2016;32:378–91. DNA. Nat Rev Cancer. 2017;17:223–38.
40. Margraf S, Lögters T, Reipen J, et al. Neutrophil-derived circulat- 61. Heitzer E, Haque IS, Roberts CES, et al. Current and future per-
ing free DNA (CF-DNA/NETs): a potential prognostic marker spectives of liquid biopsies in genomics-driven oncology. Nat
for posttraumatic development of inflammatory second hit and Rev Genet. 2019;20(2):71–88.
sepsis. Shock. 2008;30:352–8. 62. Sirera R, Bremnes RM, Cabrera A, et al. Circulating DNA is a
41. Zhong XY, Von Mühlenen I, Li Y, et al. Increased concentra- useful prognostic factor in patients with advanced non-small cell
tions of antibody-bound circulatory cell-free DNA in rheumatoid lung cancer. J Thorac Oncol. 2011;6:286–90.
arthritis. Clin Chem. 2007;53:1609–14. 63. Chaudhuri AA, Chabon JJ, Lovejoy AF, et al. Early detection of
42. Tug S, Helmig S, Menke J, et al. Correlation between cell free molecular residual disease in localized lung cancer by circulating
DNA levels and medical evaluation of disease progression tumor DNA profiling. Cancer Discov. 2017;7:1394–403.
ctDNA and MRD Detection

64. Abbosh C, Birkbak NJ, Wilson GA, et al. Phylogenetic ctDNA 83. Garcia-Murillas I, Schiavon G, Weigelt B, et al. Mutation
analysis depicts early-stage lung cancer evolution. Nature. tracking in circulating tumor DNA predicts relapse in early
2017;545:446–51. breast cancer. Sci Transl Med. 2015;7:1–7.
65. Nong J, Gong Y, Guan Y, et al. Circulating tumor DNA analysis 84. Newman AM, Lovejoy AF, Klass DM, et al. Integrated digital
depicts subclonal architecture and genomic evolution of small error suppression for improved detection of circulating tumor
cell lung cancer. Nat Commun. 2018;9:1–8. DNA. Nat Biotechnol. 2016;34:547–55.
66. Diehl F, Schmidt K, Choti MA, et al. Circulating mutant DNA 85. Dawson S-J, Tsui DWY, Murtaza M, et al. Analysis of circulat-
to assess tumor dynamics. Nat Med. 2008;14:985–90. ing tumor DNA to monitor metastatic breast cancer. N Engl J
67. To EWH, Chan KCA, Leung S, et al. Rapid clearance of plasma Med. 2013;368:1199–209.
Epstein-Barr virus DNA after surgical treatment of nasopharyn- 86. Kurtz DM, Scherer F, Jin MC, et al. Circulating tumor DNA
geal carcinoma. Clin Cancer Res. 2003;9:3254–9. measurements as early outcome predictors in diffuse large
68. O’Leary B, Cutts RJ, Liu Y, et al. The genetic landscape B-cell lymphoma. J Clin Oncol. 2018;36(28):2845–53.
and clonal evolution of breast cancer resistance to palboci- 87. Sausen M, Phallen J, Adleff V, et al. Clinical implications of
clib plus fulvestrant in the PALOMA-3 trial. Cancer Discov. genomic alterations in the tumour and circulation of pancreatic
2018;8:1390–403. cancer patients. Nat Commun. 2015;6:1–6.
69. Bettegowda C, Sausen M, Leary RJ, et al. Detection of circulat- 88. Endesfelder D, Math D, Gronroos E, et al. Intratumor heteroge-
ing tumor DNA in early- and late-stage human malignancies. Sci neity and branched evolution revealed by multiregion sequenc-
Transl Med. 2014;6:224ra24. ing. N Engl J Med. 2012;366:883–92.
70. Newman AM, Bratman SV, To J, et al. An ultrasensitive method 89. De Mattos-Arruda L, Weigelt B, Cortes J, et al. Capturing
for quantitating circulating tumor DNA with broad patient cover- intra-tumor genetic heterogeneity by de novo mutation profil-
age. Nat Med. 2014;20:548–54. ing of circulating cell-free tumor DNA: a proof-of-principle.
71. Chung JH, Pavlick D, Hartmaier R, et al. Hybrid capture-based Ann Oncol. 2014;25:1729–35.
genomic profiling of circulating tumor DNA from patients with 90. Jamal-Hanjani M, Wilson GA, McGranahan N, et al. Tracking
estrogen receptor-positive metastatic breast cancer. Ann Oncol. the evolution of non–small-cell lung cancer. N Engl J Med.
2017;28:2866–73. 2017;376:2109–21.
72. Scherer F, Kurtz DM, Newman AM, et al. Distinct biological 91. Abbosh C, Birkbak NJ, Swanton C. Early stage NSCLC: chal-
subtypes and patterns of genome evolution in lymphoma revealed lenges to implementing ctDNA-based screening and MRD
by circulating tumor DNA. Sci Transl Med. 2016;8:1–11. detection. Nat Rev Clin Oncol. 2018;15(9):577–86.
73. Phallen J, Sausen M, Adleff V, et al. Direct detection of early- 92. Swanton C, Venn O, Aravanis A, et al. Prevalence of clonal
stage cancers using circulating tumor DNA. Sci Transl Med. hematopoiesis of indeterminate potential (CHIP) measured by
2017;9:1–12. an ultra-sensitive sequencing assay: exploratory analysis of the
74. Cohen JD, Javed AA, Thoburn C, et al. Combined circulating Circulating Cancer Genome Atlas (CCGA) study. J Clin Oncol.
tumor DNA and protein biomarker-based liquid biopsy for the 2018;36:12003.
earlier detection of pancreatic cancers. Proc Natl Acad Sci USA. 93. Hu Y, Ulrich BC, Supplee J, et al. False-positive plasma
2017;114:10203–7. genotyping due to clonal hematopoiesis. Clin Cancer Res.
75. Chabon JJ, Simmons AD, Lovejoy AF, et al. Circulating 2018;24:4437–43.
tumour DNA profiling reveals heterogeneity of EGFR inhibitor 94. Genovese G, Kähler AK, Handsaker RE, et al. Clonal hemat-
resistance mechanisms in lung cancer patients. Nat Commun. opoiesis and blood-cancer risk inferred from blood DNA
2016;7:1–14. sequence. N Engl J Med. 2014;371:2477–87.
76. Khan KH, Cunningham D, Werner B, et al. Longitudinal liquid 95. Jaiswal S, Fontanillas P, Flannick J, et al. Age-related clonal
biopsy and mathematical modeling of clonal evolution forecast hematopoiesis associated with adverse outcomes. N Engl J
time to treatment failure in the PROSPECT-C phase II colorectal Med. 2014;371:2488–98.
cancer clinical trial. Cancer Discov. 2018;8(10):1270–85. 96. Jaiswal S, Ebert BL. MDS is a stem cell disorder after all.
77. Oxnard GR, Thress KS, Alden RS, et al. Association between Cancer Cell. 2014;25:713–4.
plasma genotyping and outcomes of treatment with osimerti- 97. Potter NE, Greaves M. Cancer: persistence of leukaemic ances-
nib (AZD9291) in advanced non-small-cell lung cancer. J Clin tors. Nature. 2014;506:300–1.
Oncol. 2016;34:3375–82. 98. Damm F, Mylonas E, Cosson A, et al. Acquired initiating
78. Thress KS, Paweletz CP, Felip E, et al. Acquired EGFR C797S mutations in early hematopoietic cells of CLL patients. Cancer
mutation mediates resistance to AZD9291 in non-small cell lung Discov. 2014;4:1088–101.
cancer harboring EGFR T790M. Nat Med. 2015;21:560–2. 99. Chung SS, Kim E, Park JH, et al. BRAFV600E mutations in
79. Murtaza M, Dawson SJ, Tsui DWY, et al. Non-invasive analysis hematopoietic stem cells of hairy cell leukemia patients sug-
of acquired resistance to cancer therapy by sequencing of plasma gests a stem cell origin. Sci Transl Med. 2014;99:404–5.
DNA. Nature. 2013;497:108–12. 100. Steensma DP, Bejar R, Jaiswal S, et al. Perspective: clonal
80. Taniguchi K, Uchida J, Nishino K, et al. Quantitative detection hematopoiesis of indeterminate potential and its distinction
of EGFR mutations in circulating tumor DNA derived from lung from myelodysplastic syndromes. Blood. 2018;126:9–17.
adenocarcinomas. Clin Cancer Res. 2011;17:7808–15. 101. Acuna-Hidalgo R, Sengul H, Steehouwer M, et al. Ultra-sen-
81. Tie J, Cohen JD, Wang Y, et al. Serial circulating tumour DNA sitive sequencing identifies high prevalence of clonal hemat-
analysis during multimodality treatment of locally advanced rec- opoiesis-associated mutations throughout adult life. Am J Hum
tal cancer: a prospective biomarker study. Gut. 2019;68(4):663– Genet. 2017;101:50–64.
71. https​://[Link]/10.1136/gutjn​l-2017-31585​2. 102. Zink F, Stacey SN, Norddahl GL, et al. Clonal hematopoiesis,
82. Tie J, Wang Y, Tomasetti C, et al. Circulating tumor DNA with and without candidate driver mutations, is common in the
analysis detects minimal residual disease and predicts recur- elderly. Blood. 2017;130:742–52.
rence in patients with stage II colon cancer. Sci Transl Med. 103. Young AL, Challen GA, Birmann BM, et al. Clonal haemat-
2016;8:1–10. opoiesis harbouring AML-associated mutations is ubiquitous in
healthy adults. Nat Commun. 2016;22(7):12484.
 R.-I. Chin et al.

104. Xie M, Lu C, Wang J, et al. Age-related mutations associated 125. Taly V, Pekin D, El Abed A, et al. Detecting biomarkers with
with clonal hematopoietic expansion and malignancies. Nat Med. microdroplet technology. Trends Mol Med. 2012;18:405–16.
2014;20:1472–8. 126. Dressman D, Yan H, Traverso G, et al. Transforming single DNA
105. McKerrell T, Park N, Moreno T, et al. Leukemia-associated molecules into fluorescent magnetic particles for detection and
somatic mutations drive distinct patterns of age-related clonal enumeration of genetic variations. Proc Natl Acad Sci USA.
hemopoiesis. Cell Rep. 2015;10:1239–45. 2003;100:8817–22.
106. Coombs CC, Zehir A, Devlin SM, et al. Therapy-related clonal 127. Nakano M, Komatsu J, Matsuura SI, et al. Single-molecule PCR
hematopoiesis in patients with non-hematologic cancers is com- using water-in-oil emulsion. J Biotechnol. 2003;102:117–24.
mon and associated with adverse clinical outcomes. Cell Stem 128. Diehl F, Li M, He Y, et al. BEAMing: single-molecule PCR
Cell. 2017;21(374–382):e4. on microparticles in water-in-oil emulsions. Nat Methods.
107. Bailey MH, Tokheim C, Porta-Pardo E, et al. Comprehensive 2006;3:551–9.
characterization of cancer driver genes and mutations. Cell. 129. Taly V, Pekin D, Benhaim L, et al. Multiplex picodroplet
2018;173(371–385):e18. digital PCR to detect KRAS mutations in circulating DNA
108. Zill OA, Banks KC, Fairclough SR, et al. The landscape of from the plasma of colorectal cancer patients. Clin Chem.
actionable genomic alterations in cell-free circulating tumor 2013;59:1722–31.
DNA from 21,807 advanced cancer patients. Clin Cancer Res. 130. Zhu G, Ye X, Dong Z, et al. Highly sensitive droplet digital PCR
2018;24:3528–38. method for detection of EGFR-activating mutations in plasma
109. McCulloch SD, Kunkel TA. The fidelity of DNA synthesis by cell-free DNA from patients with advanced non-small cell lung
eukaryotic replicative and translesion synthesis polymerases. cancer. J Mol Diagnostics. 2015;17:265–72.
Cell Res. 2008;18:148–61. 131. Yung TKF, Chan KCA, Mok TSK, et al. Single-molecule detec-
110. Lynch M. Rate, molecular spectrum, and consequences of human tion of epidermal growth factor receptor mutations in plasma by
mutation. Proc Natl Acad Sci USA. 2010;107:961–8. microfluidics digital PCR in non-small cell lung cancer patients.
111. Kinde I, Wu J, Papadopoulos N, et al. Detection and quantifica- Clin Cancer Res. 2009;15:2076–84.
tion of rare mutations with massively parallel sequencing. Proc 132. Sacher AG, Paweletz C, Dahlberg SE, et al. Prospective valida-
Natl Acad Sci USA. 2011;108:9530–5. tion of rapid plasma genotyping as a sensitive and specific tool
112. Newton CR, Graham A, Heptinstall LE, et al. Analysis of any for guiding lung cancer care. JAMA Oncol. 2016;2:1014–22.
point mutation in DNA. The amplification refractory mutation 133. Diehl F, Li M, Dressman D, et al. Detection and quantification of
system (ARMS). Nucleic Acids Res. 1989;17:2503–16. mutations in the plasma of patients with colorectal tumors. Proc
113. Wu DY, Ugozzoli L, Pal BK, et al. Allele-specific enzymatic Natl Acad Sci USA. 2005;102:16368–73.
amplification of beta-globin genomic DNA for diagnosis of 134. Schiavon G, Hrebien S, Garcia-Murillas I, et al. Analysis of
sickle cell anemia. Proc Natl Acad Sci USA. 1989;86:2757–60. ESR1 mutation in circulating tumor DNA demonstrates evolu-
114. Kimura H, Kasahara K, Kawaishi M, et al. Detection of epider- tion during therapy for metastatic breast cancer. Sci Transl Med.
mal growth factor receptor mutations in serum as a predictor 2015;7:148–57.
of the response to gefitinib in patients with non-small-cell lung 135. Sanmamed MF, Fernandez-Landazuri S, Rodriguez C, et al.
cancer. Clin Cancer Res. 2006;12:3915–21. Quantitative cell-free circulating BRAFV600E mutation analy-
115. Khakoo S, Georgiou A, Gerlinger M, et al. Circulating tumour sis by use of droplet digital PCR in the follow-up of patients
DNA, a promising biomarker for the management of colorectal with melanoma being treated with BRAF inhibitors. Clin Chem.
cancer. Crit Rev Oncol Hematol. 2018;122:72–82. 2015;61:297–304.
116. Milbury C, Li J, Makrigiorgos G. PCR-based methods for the 136. Birkenkamp-Demtröder K, Nordentoft I, Christensen E, et al.
enrichment of minority alleles and mutations. Clin Chem. Genomic alterations in liquid biopsies from patients with bladder
2009;55:632–40. cancer. Eur Urol. 2016;70:75–82.
117. Vogelstein B, Kinzler KW. Digital PCR. Genetics. 137. Christensen E, Nordentoft I, Vang S, et al. Optimized targeted
1999;96:9236–41. sequencing of cell-free plasma DNA from bladder cancer
118. Gerdes L, Iwobi A, Busch U, et al. Optimization of digital droplet patients. Sci Rep. 2018;8:1–11.
polymerase chain reaction for quantification of genetically modi- 138. Alizadeh AA, Aranda V, Bardelli A, et al. Toward understanding
fied organisms. Biomol Detect Quantif. 2016;7:9–20. and exploiting tumor heterogeneity. Nat Med. 2015;21:846–53.
119. Hindson BJ, Ness KD, Masquelier DA, et al. High-throughput 139. Heitzer E, Ulz P, Belic J, et al. Tumor-associated copy num-
droplet digital PCR system for absolute quantitation of DNA ber changes in the circulation of patients with prostate cancer
copy number. Anal Chem. 2011;83:8604–10. identified through whole-genome sequencing. Genome Med.
120. Karlovich C, Goldman JW, Sun JM, et al. Assessment of EGFR 2013;5:1–16.
mutation status in matched plasma and tumor tissue of NSCLC 140. Farris C, Trimarchi JM. Plasma-seq: a novel strategy for meta-
patients from a phase I study of rociletinib (CO-1686). Clin Can- static prostate cancer analysis. Genome Med. 2013;5:35.
cer Res. 2016;22:2386–95. 141. Kinde I, Papadopoulos N, Kinzler KW, et al. FAST-SeqS: a sim-
121. Liu J, Huang S, Sun M, et al. An improved allele-specific PCR ple and efficient method for the detection of aneuploidy by mas-
primer design method for SNP marker analysis and its applica- sively parallel sequencing. PLoS One. 2012;7:1–8.
tion. Plant Methods. 2012;8:1–9. 142. Douville C, Springer S, Kinde I, et al. Detection of aneuploidy
122. Kwok S, Kellogg DE, Mckinney N, et al. Effects of primer- in patients with cancer through amplification of long inter-
template mismatches on the polymerase chain reaction: human spersed nucleotide elements (LINEs). Proc Natl Acad Sci USA.
immunodeficiency virus type 1 model studies. Nucleic Acids 2018;115:1871–6.
Res. 1990;18:999–1005. 143. Chan KCA, Jiang P, Zheng YWL, et al. Cancer genome scanning
123. Cha RS, Zarbl H, Keohavong P, et al. Mismatch amplifica- in plasma: detection of tumor-associated copy number aberra-
tion mutation assay (MAMA): application to the c-H-ras gene. tions, single-nucleotide variants, and tumoral heterogeneity by
Genome Res. 1992;2:14–20. massively parallel sequencing. Clin Chem. 2013;59:211–24.
124. Huggett JF, Whale A. Digital PCR as a novel technology and 144. Welch JS, Larson DE, Wallis J, et al. Use of whole-genome
its potential implications for molecular diagnostics. Clin Chem. sequencing to diagnose a cryptic fusion oncogene. JAMA.
2013;59:1691–3. 2011;305:1577–84.
ctDNA and MRD Detection

145. Leary RJ, Kinde I, Diehl F, et al. Development of personal- 166. Amant F, Verheecke M, Wlodarska I, et al. Presymptomatic
ized tumor biomarkers using massively parallel sequencing. Sci identification of cancers in pregnant women during noninvasive
Transl Med. 2010;2:1–15. prenatal testing. JAMA Oncol. 2015;1:814–9.
146. McKernan KJ, Peckham HE, Costa G, et al. Sequence and struc- 167. Cristofanilli M, Turner NC, Bondarenko I, Ro J, Im S-A, Masuda
tural variation in a human genome uncovered by short-read, mas- N, Colleoni M, DeMichele A, Loi S, Verma S, Iwata H, Harbeck
sively parallel ligation sequencing using two-base encoding. N, Zhang K, Theall KP, Jiang Y, Bartlett CH, Koehler M. Slamon
Genome Res. 2009;19:1527–41. D (2016) Fulvestrant plus palbociclib versus fulvestrant plus pla-
147. McBride DJ, Orpana AK, Sotiriou C, et al. Use of cancer-specific cebo for treatment of hormone-receptor-positive, HER2-negative
genomic rearrangements to quantify disease burden in plasma metastatic breast cancer that progressed on previous endocrine
from patients with solid tumors. Genes. Chromosomes Cancer. therapy (PALOMA-3): final analysis of the multicentre, double-
2010;49:1062–9. blind, phase 3 randomised controlled trial. The Lancet Oncology.
148. Kirkizlar E, Zimmermann B, Constantin T, et al. Detection of 2016;17:425–39.
clonal and subclonal copy-number variants in cell-free DNA 168. Roth A, Khattra J, Yap D, et al. PyClone: statistical infer-
from patients with breast cancer using a massively multiplexed ence of clonal population structure in cancer. Nat Methods.
PCR methodology. Transl Oncol. 2015;8:407–16. 2014;11:396–8.
149. Przybyl J, Chabon JJ, Spans L, et al. Combination approach for 169. Cohen JD, Li L, Wang Y, et al. Detection and localization of
detecting different types of alterations in circulating tumor DNA surgically resectable cancers with a multi-analyte blood test. Sci-
in leiomyosarcoma. Clin Cancer Res. 2018;24:2688–99. ence. 2018;359:926–30.
150. Leary RJ, Sausen M, Kinde I, et al. Detection of chromosomal 170. Ettinger DS, Wood DE, Aisner DL, et al. Non-small cell lung
alterations in the circulation of cancer patients with whole- cancer, version 5.2017, NCCN Clinical Practice Guidelines in
genome sequencing. Sci Transl Med. 2012;4:162ra154. Oncology. J Natl Compr Cancer Netw. 2017;15:504–35.
151. Heitzer E, Auer M, Hoffmann EM, et al. Establishment of tumor- 171. Billè A, Ahmad U, Woo KM, et al. Detection of recurrence pat-
specific copy number alterations from plasma DNA of patients terns after wedge resection for early stage lung cancer: rationale
with cancer. Int J Cancer. 2013;133:346–56. for radiologic follow-up. Ann Thorac Surg. 2016;102:1067–73.
152. Belic J, Koch M, Ulz P, et al. Rapid identification of plasma DNA 172. Kocak Z, Evans ES, Zhou SM, et al. Challenges in defining radia-
samples with increased ctDNA levels by a modified FAST-SeqS tion pneumonitis in patients with lung cancer. Int J Radiat Oncol
approach. Clin Chem. 2015;61:838–49. Biol Phys. 2005;62:635–8.
153. Han X, Wang J, Sun Y. Circulating tumor DNA as biomarkers for 173. Huang K, Dahele M, Senan S, et al. Radiographic changes after
cancer detection. Genom Proteom Bioinform. 2017;15:59–72. lung stereotactic ablative radiotherapy (SABR)—can we dis-
154. Choi M, Scholl UI, Ji W, et al. Genetic diagnosis by whole exome tinguish recurrence from fibrosis? A systematic review of the
capture and massively parallel DNA sequencing. Proc Natl Acad literature. Radiother Oncol. 2012;102:335–42.
Sci USA. 2009;106:19096–101. 174. Hung JJ, Hsu WH, Hsieh CC, et al. Post-recurrence survival in
155. Forshew T, Murtaza M, Parkinson C, et al. Noninvasive iden- completely resected stage I non-small cell lung cancer with local
tification and monitoring of cancer mutations by targeted deep recurrence. Thorax. 2009;64:192–6.
sequencing of plasma DNA. Sci Transl Med. 2012;4:136ra68. 175. Shepherd F, Pereira J, Ciuleanu T, et al. Erlotinib in previ-
156. Gale D, Plagnol V, Lawson A, et al. Analytical performance ously treated non-small-cell lung cancer. N Engl J Med.
and validation of an enhanced TAm-Seq circulating tumor DNA 2005;353:123–32.
sequencing assay [abstract]. In: AACR 107th annual meeting, 176. Pérez-Soler R, Chachoua A, Hammond LA, et al. Determinants
16–20 Apr 2016, New Orleans. of tumor response and survival with erlotinib in patients with
157. Jamal-Hanjani M, Hackshaw A, Ngai Y, et al. Tracking genomic non-small-cell lung cancer. J Clin Oncol. 2004;22:3238–47.
cancer evolution for precision medicine: the Lung TRACERx 177. Gandara DR, Paul SM, Kowanetz M, et al. Blood-based tumor
study. PLoS Biol. 2014;12:1–7. mutational burden as a predictor of clinical benefit in non-small-
158. Gagan J, Van Allen EM. Next-generation sequencing to guide cell lung cancer patients treated with atezolizumab. Nat Med.
cancer therapy. Genome Med. 2015;7:1–10. 2018;24(9):1441–8.
159. Newman AM, Bratman SV, Stehr H, et al. FACTERA: a practical 178. Frampton GM, Fichtenholtz A, Otto GA, et al. Develop-
method for the discovery of genomic rearrangements at break- ment and validation of a clinical cancer genomic profiling test
point resolution. Bioinformatics. 2014;30:3390–3. based on massively parallel DNA sequencing. Nat Biotechnol.
160. Bratman SV, Newman AM, Alizadeh AA, et al. Potential clinical 2013;31:1023–31.
utility of ultrasensitive circulating tumor DNA detection with 179. Fehrenbacher L, Spira A, Ballinger M, Kowanetz M, Vansteenk-
CAPP-Seq. Expert Rev Mol Diagn. 2015;15:715–9. iste J, Mazieres J, Park K, Smith D, Artal-Cortes A, Lewanski C,
161. Lanman RB, Mortimer SA, Zill OA, et al. Analytical and clinical Braiteh F, Waterkamp D, He P, Zou W, Chen DS, Yi J, Sandler
validation of a digital sequencing panel for quantitative, highly A, Rittmeyer A. Atezolizumab versus docetaxel for patients with
accurate evaluation of cell-free circulating tumor DNA. PLoS previously treated non-small-cell lung cancer (POPLAR): a mul-
One. 2015;10:1–27. ticentre, open-label, phase 2 randomised controlled trial. The
162. Odegaard JI, Vincent JJ, Mortimer S, et al. Validation of a Lancet. 2016;387:1837–46.
plasma-based comprehensive cancer genotyping assay utilizing 180. Rittmeyer A, Barlesi F, Waterkamp D, Park K, Ciardiello F, von
orthogonal tissue- and plasma-based methodologies. Clin Cancer Pawel J, Gadgeel SM, Hida T, Kowalski DM, Dols MC, Corti-
Res. 2018;24:3539–49. novis DL, Leach J, Polikoff J, Barrios C, Kabbinavar F, Frontera
163. FoundationOne® Liquid. Foundation Medicing Inc.; 2018. https​ OA, De Marinis F, Turna H, Lee J-S, Ballinger M, Kowanetz
://[Link]​ation​medic​[Link]/genom​ic-testi​ng/found​ation​ M, He P, Chen DS, Sandler Gandara DR. Atezolizumab versus
-one-liqui​d. docetaxel in patients with previously treated non-small-cell lung
164. Torga G, Pienta KJ. Patient-paired sample congruence between cancer (OAK): a phase 3, open-label, multicentre randomised
2 commercial liquid biopsy tests. JAMA Oncol. 2018;4:868–70. controlled trial. The Lancet. 2017;389:255–65.
165. Alexandrov LB, Nik-zainal S, Wedge DC, et al. Signatures of 181. Goodman AM, Kato S, Bazhenova L, et al. Tumor mutational
mutational processes in human cancer. Nature. 2013;500:415–21. burden as an independent predictor of response to immunother-
apy in diverse cancers. Mol Cancer Ther. 2017;16(11):2598–608.
 R.-I. Chin et al.

182. Rizvi NA, Hellmann MD, Snyder A, et al. Mutational landscape 203. Reimers N, Kasper HU, Weissenborn SJ, et al. Combined analy-
determines sensitivity to PD-1 blockade in non-small cell lung sis of HPV-DNA, p16 and EGFR expression to predict prognosis
cancer. Science. 2015;348:124–8. in oropharyngeal cancer. Int J Cancer. 2007;120:1731–8.
183. Hellmann MD, Ciuleanu T-E, Pluzanski A, et al. Nivolumab plus 204. Feinmesser R, Miyazaki I, Cheung R, et al. Diagnosis of naso-
ipilimumab in lung cancer with a high tumor mutational burden. pharyngeal carcinoma by DNA amplification of tissue obtained
N Engl J Med. 2018;378:2093–104. by fine-needle aspiration. N Engl J Med. 2015;326:17–21.
184. Hoeben KWJ, van Steenbergen LN, van de Wouw AJ, et al. 205. Chan KCA, Hung ECW, Woo JKS, et al. Early detection of
Treatment and complications in elderly stage III colon cancer nasopharyngeal carcinoma by plasma Epstein–Barr virus DNA
patients in the Netherlands. Ann Oncol. 2013;24:974–9. analysis in a surveillance program. Cancer. 2013;119:1838–44.
185. Benson AB, Venook AP, Al-Hawary MM, et al. NCCN guide- 206. Chan KCA, Woo JKS, King A, et al. Analysis of plasma Epstein–
lines insights: colon cancer, version 2.2018. J Natl Compr Can- Barr virus DNA to screen for nasopharyngeal cancer. N Engl J
cer Netw. 2018;16:359–69. Med. 2017;377:513–22.
186. Wirtzfeld DA, Mikula L, Gryfe R, et al. Concordance with clin- 207. Lin J-C, Wang W-Y, Chen KY, et al. Quantification of plasma
ical practice guidelines for adjuvant chemotherapy in patients Epstein–Barr virus DNA in patients with advanced nasopharyn-
with stage I–III colon cancer: experience in 2 Canadian prov- geal carcinoma. N Engl J Med. 2004;350:2461–70.
inces. Can J Surg. 2009;52:92–7. 208. Campitelli M, Jeannot E, Peter M, et al. Human papillomavirus
187. Perini MV, Montagnini AL, Jukemura J, et al. Clinical and mutational insertion: specific marker of circulating tumor DNA
pathologic prognostic factors for curative resection for pan- in cervical cancer patients. PLoS One. 2012;7:e43393.
creatic cancer. HPB (Oxford). 2008;10:356–62. 209. Jeannot E, Becette V, Campitelli M, et al. Circulating human pap-
188. Walboomers JMM, Jacobs MV, Manos MM, et al. Human pap- illomavirus DNA detected using droplet digital PCR in the serum
illomavirus is a necessary cause of invasive cervical cancer of patients diagnosed with early stage human papillomavirus-
worldwide. J Pathol. 1999;189:12–9. associated invasive carcinoma. J Pathol Clin Res. 2016;2:201–9.
189. Muñoz N, Bosch FX, de Sanjosé S, et al. Epidemiologic clas- 210. Lo YMD, Leung S, Chan LYS, et al. Kinetics of plasma Epstein–
sification of human papillomavirus types associated with cervi- Barr virus DNA during radiation therapy for nasopharyngeal car-
cal cancer. N Engl J Med. 2003;348:518–27. cinoma. J Cancer Res. 2000;60:2351–5.
190. Frisch M, Glimelius B, van den Brule AJC, et al. Sexually 211. Wang W-Y, Twu C-W, Chen H-H, et al. Long-term survival
transmitted infection as a cause of anal cancer. N Engl J Med. analysis of nasopharyngeal carcinoma by plasma Epstein–Barr
1997;337:1350–8. virus DNA levels. Cancer. 2013;119:963–70.
191. Chaturvedi AK, Engels EA, Pfeiffer RM, et al. Human papil- 212. Le Q, Zhang Q, Cao H, et al. An international collaboration to
lomavirus and rising oropharyngeal cancer incidence in the harmonize the quantitative plasma Epstein–Barr Virus (EBV)
United States. J Clin Oncol. 2011;29:4294–301. DNA assay for future biomarker-guided trials in nasopharyngeal
192. zur Hausen H, Schulte-Holthausen H, Klein G, et al. EBV DNA carcinoma. Clin Cancer Res. 2013;19:2208–15.
in biopsies of Burkitt tumours and anaplastic carcinomas of the 213. Leung SF, Zee B, Ma BB, et al. Plasma Epstein–Barr viral deoxy-
nasopharynx. Nature. 1970;228:1056–8. ribonucleic acid quantitation complements tumor-node-metasta-
193. Kim C-M, Koike K, Saito I, et al. HBx gene of hepati- sis staging prognostication in nasopharyngeal carcinoma. J Clin
tis B virus induces liver cancer in transgenic mice. Nature. Oncol. 2006;24:5414–8.
1988;351:317–20. 214. Chua MLK, Wee JTS, Hui EP, et al. Nasopharyngeal carcinoma.
194. Beasley RP. Hepatitis B virus. The major etiology of hepatocel- Lancet. 2016;387:1012–24.
lular carcinoma. Cancer. 2016;61:1942–56. 215. Fakhry C, Westra WH, Li S, et al. Improved survival of patients
195. Saito I, Miyamura T, Ohbayashi A, et al. Hepatitis C virus with human papillomavirus-positive head and neck squamous
infection is associated with the development of hepatocellular cell carcinoma in a prospective clinical trial. J Natl Cancer Inst.
carcinoma. Proc Natl Acad Sci USA. 1990;87:6547–9. 2008;100:261–9.
196. Moriya K, Fujie H, Shintani Y, et al. The core protein of hepa- 216. Ang KK, Harris J, Wheeler R, et al. Human papillomavirus and
titis C virus induces hepatocellular carcinoma in transgenic survival of patients with oropharyngeal cancer. N Engl J Med.
mice. Nat Med. 1998;4:1065–7. 2010;363:24–35.
197. Peitsaro P, Johansson B, Syrjänen S. Integrated human papil- 217. Dahlstrom KR, Li G, Hussey CS, et al. Circulating human
lomavirus type 16 is frequently found in cervical cancer pre- papillomavirus DNA as a marker for disease extent and recur-
cursors as demonstrated by a novel quantitative real-time PCR rence among patients with oropharyngeal cancer. Cancer.
technique. J Clin Microbiol. 2002;40:886–91. 2015;121:3455–64.
198. Wiest T, Schwarz E, Enders C, Flechtenmacher C, Bosch FX. 218. Hanna GJ, Supplee JG, Kuang Y, et al. Plasma HPV cell-free
Involvement of intact HPV16 E6/E7 gene expression in head DNA monitoring in advanced HPV-associated oropharyngeal
and neck cancers with unaltered p53 status and perturbed pRb cancer. Ann Oncol. 2018;29:1980–6.
cell cycle control. Oncogene. 2002;21:1510–7. 219. Gupta GP, Kumar S, Marron D, et al. Circulating tumor HPV16
199. Raab-Traub N. Epstein–Barr virus in the pathogenesis of NPC. DNA as a biomarker of tumor genomics and disease control
Semin Cancer Biol. 2002;14:423–30. in HPV-associated oropharyngeal squamous cell carcinoma
200. Mayrand M-H, Duarte-Franco E, Rodrigues I, et al. Human [abstract]. Radiat Oncol. 2018;100(5):1310.
papillomavirus DNA versus Papanicolaou screening tests for 220. Koliopoulos G, Nyaga VN, Santesso N, et al. Cytology versus
cervical cancer. New Engl. J Med. 2010;357:1579–1588. HPV testing for cervical cancer screening in the general popula-
201. Manos MM, Kinney WK, Hurley LB, et al. Identifying tion. Cochrane Database Syst Rev. 2017;8:CD008587.
women with cervical neoplasia using human papillomavi- 221. Koss LG. The Papanicolaou test for cervical cancer detection: a
rus DNA testing for equivocal Papanicolaou results. JAMA. triumph and a tragedy. JAMA. 1989;261:737–43.
1999;281:1605–10. 222. Han K, Leung E, Barbera L, et al. Circulating human papilloma-
202. Ndiaye C, Mena M, Alemany L, et al. HPV DNA, E6/ virus DNA as a biomarker of response in patients with locally
E7 mRNA, and p16INK4a detection in head and neck can- advanced cervical cancer treated with definitive chemoradiation.
cers: a systematic review and meta-analysis. Lancet Oncol. JCO Precis Oncol. 2018;2:1–8.
2014;15:1319–31.
ctDNA and MRD Detection

223. El-Serag HB, Rudolph KL. Hepatocellular carcinoma: epi- tumor DNA from non-small cell lung carcinoma patient plasma.
demiology and molecular carcinogenesis. Gastroenterology. Theranostics. 2017;7:1437–46.
2007;132:2557–76. 229. Thress KS, Brant R, Carr TH, et al. EGFR mutation detection in
224. Sun Yat-sen University. ctDNA for prediction of relapse in gastric ctDNA from NSCLC patient plasma: a cross-platform compari-
cancer [[Link] identifier NCT02887612]. National son of leading technologies to support the clinical development
Institutes of Health, [Link]. https​://clini​caltr​[Link]. of AZD9291. Lung Cancer. 2015;90:509–15.
Accessed 11 Feb 2019. 230. Institute of Cancer Research, United Kingdom. A trial using
225. Johns Hopkins University and the Walter and Eliza Hall Insti- ctDNA blood tests to detect cancer cells after standard treat-
tute of Medical Research. Circulating tumour DNA as a marker ment to trigger additional treatment in early stage triple negative
of residual disease and response to adjuvant chemotherapy breast cancer patients (c-TRAK-TN) [[Link] identifier
in stage I–IV ovarian cancer [[Link] identifier NCT03145961]. National Institutes of Health, [Link].
NCT03691012]. National Institutes of Health, ClinicalTrials. https​://clini​caltr​[Link]. Accessed 11 Feb 2019.
gov. https​://clini​caltr​[Link]. Accessed 11 Feb 2019. 231. Pao W, Miller VA, Politi KA, et al. Acquired resistance of lung
226. University of California, San Francisco. Biomarkers for risk adenocarcinomas to gefitinib or erlotinib is associated with a
stratification in lung cancer [[Link] identifier second mutation in the EGFR kinase domain. PLoS Med.
NCT03774758]. National Institutes of Health, ClinicalTrials. 2005;2:0225–35.
gov. https​://clini​caltr​[Link]. Accessed 11 Feb 2019. 232. Kobayashi S, Boggon T, Dayaram T, et al. EGFR mutation and
227. Merker JD, Oxnard GR, Compton C, et al. Circulating tumor resistance of non-small-cell lung cancer to gefitinib. N Engl J
DNA analysis in patients with cancer: American Society of Clini- Med. 2005;352:786–92.
cal Oncology and College of American Pathologists joint review. 233. Soria J-S, Ohe Y, Vansteenkiste J, et al. Osimertinib in untreated
J Clin Oncol. 2018;36:1631–41. EGFR-mutated advanced non–small-cell lung cancer. NEJM.
228. Xu T, Kang X, You X, et al. Cross-platform comparison of four 2018;378:113–25.
leading technologies for detecting EGFR mutations in circulating