Cancer Treatment Reviews 119 (2023) 102595

Contents lists available at ScienceDirect

Cancer Treatment Reviews

journal homepage: www.elsevier.com/locate/ctrv

Laboratory-Clinic Interface

Strategies for improving detection of circulating tumor DNA using next

generation sequencing
Tébar-Martínez Roberto a, b, Martín-Arana Jorge a, c, Gimeno-Valiente Francisco d,
Tarazona Noelia a, e, Rentero-Garrido Pilar b, Cervantes Andrés a, e, *
a
Department of Medical Oncology, INCLIVA Health Research Institute, University of Valencia, C. de Menéndez y Pelayo, 4, 46010 Valencia, Spain
b
Precision Medicine Unit, INCLIVA Health Research Institute, C. de Menéndez y Pelayo, 4, 46010 Valencia, Spain
c
Bioinformatics Unit, INCLIVA Health Research Institute, C. de Menéndez y Pelayo, 4, 46010 Valencia, Spain
d
Cancer Evolution and Genome Instability Laboratory, University College of London Cancer Institute, 72 Huntley St, WC1E 6DD London, United Kingdom
e
Health Institute Carlos III, CIBERONC, C/ Sinesio Delgado, 4, 28029 Madrid, Spain

A R T I C L E I N F O A B S T R A C T

Keywords: Cancer has become a global health issue and liquid biopsy has emerged as a non-invasive tool for various ap­
ctDNA plications. In cancer, circulating tumor DNA (ctDNA) can be detected from cell-free DNA (cfDNA) obtained from
Molecular Barcoding plasma and has potential for early diagnosis, treatment, resistance, minimal residual disease detection, and tu­
UMI
moral heterogeneity identification. However, the low frequency of ctDNA requires techniques for accurate
Liquid biopsy
analysis. Multitarget assay such as Next Generation Sequencing (NGS) need improvement to achieve limits of
detection that can identify the low frequency variants present in the cfDNA. In this review, we provide a general
overview of the use of cfDNA and ctDNA in cancer, and discuss techniques developed to optimize NGS as a tool
for ctDNA detection. We also summarize the results obtained using NGS strategies in both investigational and
clinical contexts.

Introduction feasible or may not meet the conditions required for proper analysis,
which prevents some diagnosed or relapsing patients from the benefiting
The World Health Organization (WHO) defines cancer as a group of from the potential advantages of molecular studies [7].
diseases characterized by abnormal cell growth that can invade tissues Liquid biopsy has recently emerged as a non-invasive tool for real-
and organs in other parts of the body. According to GLOBOCAN statis­ time monitoring of biological conditions, enabling liquid samples from
tics, in 2020 there were 19.3 million new cancer cases and 10.0 million the body such as blood, cerebrospinal fluid, urine, or even pleural fluid
cancer deaths worldwide [1]. This numbers highlight the global health to be used to study specific biomarkers associated with diverse biolog­
threat posed by cancer. ical events [8,9]. Examples of these events include transplant moni­
According to Hanahan et al., cancer is characterized by ten hall­ toring [8], non-invasive prenatal testing [10], infectious disease
marks, including genome instability and mutation incidence, which are detection [11] and cancer research. In the case of cancer, the plasma
among the main causes of cancer progression along with epigenetic al­ fraction from blood is used to extract markers such as cell-free DNA
terations [2]. Mutations in genes such as APC in Colorectal cancer [3], (cfDNA) for detecting somatic mutations from the tumors. This specific
EGFR in Non-small cell lung cancer (NSCLC) [4] and the Philadelphia and altered DNA is called circulating tumor DNA (ctDNA).
chromosome translocation in hematological malignancies [5] are This review explores the utility of liquid biopsy in cancer, with a
involved in cancer development and can be determined by studying the focus on Next-Generation Sequencing (NGS) strategies to study cfDNA,
tumor tissue aberrations. Depending on the mutation type, this infor­ and outline some of the advances in NGS for ctDNA study.
mation may have prognostic or predictive value for disease course and
treatment response [6]. However, obtaining tumor samples is not always

* Corresponding author at: Department of Medical Oncology, INCLIVA Health Research Institute, University of Valencia, C. de Menéndez y Pelayo, 4, 46010
Valencia, Spain.
E-mail addresses: robertotebarmartinez@gmail.com (T.-M. Roberto), jmartin@incliva.es (M.-A. Jorge), fgimenovaliente@gmail.com (G.-V. Francisco), noetalla@
incliva.es (T. Noelia), prentero@incliva.es (R.-G. Pilar), andres.cervantes@uv.es (C. Andrés).

https://doi.org/10.1016/j.ctrv.2023.102595
Received 20 June 2023; Accepted 23 June 2023
Available online 25 June 2023
0305-7372/© 2023 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
T.-M. Roberto et al. Cancer Treatment Reviews 119 (2023) 102595

Role of liquid biopsy in cancer study et al., with more than 50 types of cancer, showed a sensitivity of 18% for
stage I and 43% for stage II [22], reflecting the need for advancements in
Fragmented DNA found in plasma may be the result of the death of early detection technologies before they can be applied in the clinical
both normal and tumor cells. DNA is secreted into the blood as result of setting. Another proposed use is for the detection and removal of post-
death processes such as necrosis and apoptosis, as well as the inter­ treatment tumor traces that could pose a risk of relapse; this is known
vention of macrophage-mediated digestion of cell debris and apoptotic as Minimal Residual Disease (MRD), and biopsy for detection has been
bodies [12]. Because of the protection of the nucleosome, DNA has a proposed for the period after the curative intent and later follow-up
peak length of around 160 bp, with higher peaks in multiples (320 bp, stages. MRD detection times range from 2 to 10 weeks [23,24] after
480 bp) as this length equals the bases wrapped around the nucleosome surgical resection or up to 4 months after initial treatment [25]. Studies
plus the linker DNA connecting to the next one [13,14]. These short have shown that post-operative ctDNA detection can predict relapse
fragments represent cfDNA. The term “circulating tumor DNA” refers to [24,26–28] and that ctDNA levels increases before metastasis can be
the specific fraction of fragments liberated from the tumor compared to detected by imaging [26,27,29]. Monitoring response to treatment by
the rest of the cfDNA. Studies have described differences in length be­ ctDNA detection is also under consideration, as ctDNA dynamics before
tween cfDNA from healthy tissue and ctDNA, reporting a shorter length and after chemotherapy could define patient subgroups with higher risk
in the latter [15]. The concentration of cfDNA has also been found to be of poor outcomes [24,28,30,31]. Specifically tracking resistance bio­
higher in cancer patients than healthy controls [16]. Healthy individuals markers such as EGFR p.T790M in NSCLC patients receiving tyrosine
typically have around 1500 copies per mL of plasma [14]. However, not kinase inhibitors can detect increasing resistance [4]. Tumoral hetero­
all cancer cases have detectable ctDNA, as patients may have shedding geneity detection has also been proposed for diagnosis and relapse
or non-shedding tumors. The differences between the two have not yet stages to identifying new mutations that were not found in the original
been defined, but ctDNA positivity at diagnosis has been related to tumor, potentially identifying patients eligible for targeted treatments
higher stages and signs of invasion [17–19]. [32,33].
As cfDNA can be obtained at any time during cancer development,
liquid biopsy has been proposed for various uses at any stage of the Detection of circulating tumor DNA
disease (Fig. 1). One of these roles is ctDNA detection in early diagnosis,
as early detection of the tumor could lead to a completely different As previously stated, ctDNA is found among a background of cfDNA.
disease course [20]. A retrospective study in early-stage disease in Most circulating DNA in plasma comes from hematopoietic cells found
several cancer types found that more than 50% of analyzed patients with in blood [34], resulting in ctDNA fractions below one copy across 1000
stage I or II had detectable ctDNA in samples obtained prior to resection cfDNA copies (0.1%) in situations such as MRD detection [35]. How­
but with already established diagnosis [21]. The study presented by Liu ever, ctDNA fraction could be higher depending on the stage of the

Fig. 1. Applications of cfDNA from plasma in cancer patients. cfDNA has been proposed for different aspects. Plasma from patients at early stage or even pre-
disease would be used to identify early mutations in plasma if highly sensitive and validated technologies were developed. In the same way, biomarker mutations for
treatment could be detected, allowing targeted therapy for the patient. Monitoring could also identify possible resistance mutations arising at a later time. At any
point of the disease course, ctDNA from different clones that could not be detected in a FFPE tumor specimen could be characterized due to non-represented clones of
the tumor or non-identified lesions. Lastly, ctDNA could be tracked during patient follow-up to detect residual traces of the tumor, which could help predict relapse
before imaging techniques could detect it.

2
T.-M. Roberto et al. Cancer Treatment Reviews 119 (2023) 102595

disease, with an approximate frequency of 0.3% in early stage diagnosed However, traditional NGS strategies have not demostrate the capability
cancer and near 10% in metastatic settings [36]. Therefore, optimal to detect variants below 1.0% [55–57]. One source of the errors found in
ctDNA study requires technologies with low limits of detection (LoD), NGS is related to a technical perspective, including potential sequencing
equivalent to 0.1% variant allele frequency (VAF) or below depending errors or frequent polymerase errors in specific nucleotide or sequence
on the application. context, or even overamplification of low DNA inputs that leads to
Even with the proper tools, ctDNA is not always found, depending on wrong quantifications of mutations. The other sources of errors are those
the sample of cfDNA obtained from plasma. One of the factors behind of stochastic nature such as wrong base callings under inappropriate
this situation is the “sampling” of cfDNA. It has been estimated a total of sequencing conditions (i.e., over-clustering on Illumina instruments) or
300 haploid copies of genome per ng of human DNA [37]. Under this early PCR errors introduced by polymerases [58]. Modifications have
assumption, at least 3.3 ng of cfDNA is needed to have at least 1 copy of a been developed to reduce background noise aiming to reach LoDs below
mutation at 0.1% in plasma. Therefore, higher inputs of cfDNA will be 0.1–0.5% VAF, as certain errors appear at this frequency [59]. Strategies
needed to enable stricter LoD, which could be accomplished with the are mainly based on molecular barcoding or background noise
concentration of plasma or an eluted sample or with extraction tech­ modeling.
niques that do not have a fixed plasma input, such as those based on
membranes. It is also noteworthy that not all techniques will have 100% NGS strategies for ctDNA detection
efficiency on cfDNA introduction into the protocol, so the input in which
the LoD is defined is relevant when considering a modified NGS Molecular barcoding
protocol. Standard NGS library protocols involve working with a mixture of
The initial methods employed for these purposes were PCR-based. nucleic acid from the samples, without discriminating the source of the
Subsequently, modifications led to the development of mutation- different PCR duplicates (molecules derived from the same original
specific assays that could detect limits lower than 1.0% VAF. Howev­ template). To address this, random-nucleotide molecular barcodes
er, most of these PCR-based protocols can only analyze one change or [known as Unique Molecular Identifiers (UMI) o Unique Identifiers
codon per assay, and the amplicon size must be adjusted to the length of (UID)] are added before any PCR step allowing all PCR duplicates to be
cfDNA. Due de abundance of 160 bp fragments, amplicons must be aggregated with the same identifier [55,56]. The primary goal is to
shorter, ideally around 80 bp and under 100 bp, since some techniques generate one consensus sequence from all chains with the same UMI or
requires the hybridization of a probe in the middle of the insert [38,39]. UMI family, removing the PCR error introduced during library genera­
The Amplification Refractory Mutation System-qPCR (ARMS-qPCR) tion (Fig. 2A). As a result, true variants should be present in all chains
leverages the 3′ end of the primer,which is necessary for amplification from the same UMI family. Increasing the UMI family size enhances the
[40,41]. ARMS-qPCR oligonucleotides (oligos) is useful for detecting confidence in the identified variants in the consensus sequence [60].
point mutations and short indels, and has the ability to detect fre­ Software such as UMI-Tools [61], Picard/MarkDuplicates [62], Caleb
quencies as low as 0.015% [42,43]. [63], UMI-VarCal [64], gencore [65], or UMIc [66] can be used to create
COLD-PCR is another technology that can adapt to the melting consensus. Molecular barcoding is also used in single-cell strategies for
temperature of the different alleles [44,45]. The techniqués modality is cell-of-origin identification [67,68] and in RNA-Seq for absolute quan­
determined by the type of change being studied, as a nucleotide change tification of transcripts [69] (Fig. 2B). As a limitation, early-PCR errors
can produce alterations of between 0.2 and 1.5oC [46]. COLD-PCR is a could be challenging to eliminate, as they could be found in high pro­
useful prior step for other detection techniques as it enables the portions of the family [60,70].
enrichment of one of the alleles [47]. It has been able to detect variants
at 0.01% from higher inputs above 1 ng [48–50]. Barcode introduction and region selection
Digital PCR is another interesting technique that involves creating UMI is typically introduced through ligation or low-cycle PCR
microreactors containing all the PCR reagents instead of a homogeneous (Fig. 3). Depending on the method used to incorporate UMIs and select
mix. The microreactors isolate one allele to create a specific signal for regions regions of interest, molecular barcoding methods can be classi­
each variant, which will only appear if one of the variants of interest is fied as Capture-Based, PCR-based, or Anchored PCR-based.
present [47]. To create the signal, two specific probes with different Capture-based strategies (Fig. 3A) involves a first stage of library
markers are needed, one for each allele. For instance, BEAMing PCR pre- preparation from the whole genome, with the correct size and
amplifies chains isolated on a microreactor generated through an sequencing adaptors already incorporated. The second step uses labeled
emulsion, and after breaking the emulsion, two probes with different probes against the region of interest, which are retrieved with magnetic
labels are used against each variant [51]. Similary, Droplet Digital PCR spheres that are bound to a specific ligand for the label. This permits
(ddPCR) employs an emulsion solution to create the microreactors. The recovery of very small regions (few kb) or complete exomes. UMI is
signal is created by one of the fluorophores derived from splitting a introduced by ligating an adaptor to the template. The random sequence
TaqMan probe that recognizes the specific allele [52]. In both cases, the can be read alongside the insert or as part of the index. It is also possible
label accumulates, enriching the microreactor in one of the signals that to distinguish between the two chains if the design allows for the
can be detected with appropriate equipment. Both ddPCR and BEAMing introduction of different UMIs. Safe-Sequencing System (Safe-SeqS)
PCR are useful for detecting point mutations and short indels, with proposed a design that did not incorporate UMIs but instead used the
ddPCR being able to detect frequencies as low as 0.003–0.005% [4,53]. nucleotides from the ends of the inserts as an “endogenous” UID to
Despite the sensitivity of the techniques described, there is a need to identify inserts from the same family, resulting in a 70-fold reduction in
study a large number of loci simultaneously, which has created a de­ errors. Other protocols, such as TEC-Seq, CAPP-Seq + iDES, Duplex
mand for alternative techniques. For instance, tumor dynamics have Sequencing, and Singleton Correction, make use of different UMIs at
demonstrated that metastasis can grow from cell clones that are not each end of the molecule to identify which chain of the template has
detectable in the primary tumor, so selecting a specific mutation needed been initially amplified, allowing for a consensus to be made for each
for these technologies may not always be the best approach for ctDNA chain and between both sense and antisense molecules. Commercial
detection[24,54], unless specific targetable or resistance mutations are assays like Guardant360 [71] and Avenio (based on CAPP-Seq) [72]
being tracked in cfDNA to evaluate treatment response. This issue can be make use of this protocol. Alcaide et al. proposed a strategy that uses
addressed by multitarget technologies such as NGS, which unlike PCR- semi-degenerate UMIs and three non-complementary nucleotides in the
based assays, can detect several types of alterations with only one ligated extreme of the adapter, allowing both strands to be identified,
locus analyzed. Depending on the requirements, NGS can be designed to and reaching LoDs of 0.02%. The TrUMIseq approach has reported
analyze any quantity, from a few loci to an entire exome or genome. frequencies below 1% VAF by selecting the sample index nearer to the

3
T.-M. Roberto et al. Cancer Treatment Reviews 119 (2023) 102595

Fig. 2. Function of molecular barcoding for NGS applications. A) DNA fragments (blue rectangles) with same UMI sequence are considered PCR duplicates.
During file processing, mutations that only appear in some sequences of the UMI family are considered PCR errors (red stars) and eliminated from the consensus
sequence, whereas true mutations (green stars) are seen in most sequences of the family. B) Two different RNA molecules were read 10 times. After consensus, the
count made using molecular barcoding shows that only two molecules of RNA A and four of RNA B were identified. (For interpretation of the references to colour in
this figure legend, the reader is referred to the web version of this article.)

Fig. 3. Methods for background noise reduc­

tion. Strategies for reaching low limits of detection
with molecular barcoding (A-C) or bioinformatic
analysis (D). A) UMIs are incorporated with a
ligation step, prior to any PCR. In posterior steps,
fragments of interest are retrieved with hybridiza­
tion probes. B) Primers with UMI are employed in a
PCR with a low number of cycles. Primers are
designed against the region of interest. C) After
adding the UMI and a universal adaptor, the region
is selected with a specific primer. The second
oligonucleotide (green primer) is complementary to
the universal adaptor. D) Non-mutated controls
could be used to create a model of the noise for the
tested panel for posterior identification of true
variants. The schematic diagram of DNA molecules
is represented as follows: blue: cfDNA; green:
adaptor; red: specific sites recognized by probes or
primers; violet: UMI. UMI: Unique Molecular
Identifier. (For interpretation of the references to
colour in this figure legend, the reader is referred to
the web version of this article.)

insert, as UMI is read at the index read on the sequencing instrument. cycles. Safe-SeqS makes use of this type of primers with a two-cycle PCR
Agilent chemistry has shown LoDs of 0.075% using a UID read at the to introduce an exogenous UID [55]. AmpliSeq has design primers with
index read [73]. the UMI at the 5′ end in both oligonucleotides for a three-cycle PCR,
Amplicon-based or PCR-based methods (Fig. 3B) use modified with a reporting mutations above 0.25% VAF [74–76]. Quantitative Amplicon
random sequence incorporated through PCR with a small number of Sequencing (QASeq) uses a two-cycle PCR with primers including UMIs,

4
T.-M. Roberto et al. Cancer Treatment Reviews 119 (2023) 102595

detecting CNVs in cfDNA with high sensitivity. Personalized Molecular Non-tumor-related positive mutation elimination
Barcode NGS (Personalized MB-NGS) uses a primer with the non-fully Plasma cfDNA may contain true-positive somatic mutations from
random sequence BDHVBDHVBDHVBDH, excluding one nucleotide tissues throughout the body, as well as false positives due to stochastic
from each position of the UID, reporting LoDs of 0.1%. [77–79]. Simple, errors [76,97,98]. The majority of plasma cfDNA originates from white
Multiplexed, PCR-based barcoding of DNA for Sensitive mutation blood cells and strategies have been developed to eliminate these from
detection using sequencing (SiMSen-Seq) reduces background noise to the cell population [97,99]. Clonal hematopoietic mutations have been
0.05% and detects variants above 0.1% VAF [39,80]. observed in the hematopoietic lineage during aging [100] and have been
Anchored PCR utilizes universal adaptors added to the ends of the associated with older patients [76,97,100]. These mutations are defined
insert to create an amplicon with a universal and specific primer against as clonal hematopoiesis of indeterminate potential (CHIP) and are
the region of interest (Fig. 3C). QIASeq-targeted DNA use this technol­ mainly found in genes such as DNMT3A, TET2, ASXL1, JAK2, PPM1D,
ogy with adaptor ligation to detect variants near 0.1% VAF [81–83]. The and TP53 [76,97,100,101].
Non-overlapping integrated reads sequencing system (NOIR-SS) uses a Various studies have attempted to describe these non-tumoral so­
similar protocol to analyze either one or both chains of the original matic mutations in plasma. One study detected 24.4% of mutations
molecule. NOIR-SS has reported a background signal of 0.01% and it has concordant with tumor tissue in metastatic patients, with half of the
also detected somatic variants at 0.02% [84–87]. Peng et al. proposed identified mutations associated with CHIP in patients with non-
using two non-complementary bases in the adaptor sequence for later hypermutated tumors [97]. Another study linked development of mu­
identification depending on the base pair read [88]. Adaptor Template tations in DNA damage response genes like TP53, PPM1D, and CHEK2
Oligo Mediated Sequencing (ATOM-Seq) uses a complex structure with cytotoxic agent or chemotherapy treatment [102]. Germline mu­
bound by a non-replicable linker, reporting a LoD of 0.1% VAF [89]. tations involved in tumorigenesis can also be found in plasma, and an
Finally, Quantitative Blocker Displacement Amplification (QBDA) existing strategy uses software analyses to distinguish true somatic
combinations PCR-based and Anchored PCR method for adding UMIs, mutations from germline and CHIP mutations without sequencing white
along with variant enrichment, reporting a 0.1% detection with 6–20 ng blood cells [103,104].
input and 0.001% LoD if a consensus coverage of 23000X was reached
[90]. Other strategies
Other studies have explored the increase in sensitivity of NGS
Methylation approaches through physical properties. The Tagged-amplicon Sequencing (Tam-
Methylation is an area of interest in liquid biopsy research, as bio­ Seq) and its modified version eTAm-Seq reportes frequencies over 0.1%
markers have been discovered in plasma [91,92]. One promising tech­ by combining preamplification with chain separation and amplification
nology is cfMethyl-Seq, which includes UMIs at both ends and targets in an array [105–107]. Proximity-Sequencing (Pro-Seq) generates PCR
methylated cytosines. This approach is not affected by bisulfite con­ replicates linked with PEG to reduce errors, reporting 0.01% VAF true
version and selectively enriches CpG islands through adaptor ligations to variants in controls [108].
MspI digested sites. Validation of the strategy using 5–40 ng of cfDNA Fragmentome analysis is other ctDNA detection tool that analyze
reached a sensitivity of 80.7% in all-stages cancer from colon, liver, aspects such as size distribution, variability in length, and jagged ends,
lung, and stomach [93]. Liu et al. developed a custom panel targeting all related to different fragmentation patterns. [109]. DELFI’s Whole
103,456 methylation regions that, after proper training in non- Genome Sequencing (WGS) approach showed a sensitivity ranging from
cancerous individuals, detected 43.9% of stage I-III cancers and 93% 57% to 99% on pre-operative plasma and multi-tumor samples, being
of stage IV cancers with inputs up to 70 ng. They were also able to 91% when combining mutation detection and fragmentome analysis
identify the tissue of origin in 93% with a detection signal [22]. Another [109]. DELFI’s tool combined with CNA and methylome was able to
technology, Methylated CpG tandem Amplifications and Sequencing anticipate relapse using follow-up samples in a small cohort of uveal
(MCTA-Seq), employs semi-random primers with UMIs in the first PCR melanoma patients [110].
after DNA conversion and a CGCGCGG primer for a selective amplifi­
cation of methylated regions. MCTA-Seq can detect hypermethylated WES and WGS in ctDNA samples
regions throughout the genome with inputs as low as 10 pg of methyl­ Studies have examined whole exome and genome sequencing of
ated DNA, but does not determine the degree of methylation [94]. cfDNA, but most strategies lack molecular barcoding for error reduction
Finally, cfMeDIP-Seq is another enrichment method that use immuno­ due to high genome coverage [111]. WES employs 50-200X median
precipitation with antibodies targeting 5-methylcytidine. This approach coverage [112–115], while WGS is around 0.3-35X [116–118]. UMIs
selectively enriches hypermethylated regions and may include UMIs to could be used for molecule quantification, and ctDNA quantification
potentially enhance mutation detection [95,96]. could be used to select suitable plasma samples for WES [115,119].
Targeted panels like CAPP-seq [120] and FundationONE [121] have
Background noise modeling been used for TMB determination in ctDNA studies, although TMB
Bioinformatic analysis can be used to differentiate true positives concordance between plasma and tissue is not always consistent [122].
from background noise in sequencing, with or without UMIs (Fig. 3D). WGS of cfDNA can detect CNA in different tumor types, and MRDetect is
Several methods use healthy plasma controls to study the background a tool that combines SNV and CNA detection for improved sensitivity
signal, such as the initial CAPP-Seq model and iDES-enhanced CAPP-Seq and specificity [118].
with UIDs. The latter detected 2.5 copies among 106, equivalent to
0.00025% VAF in 32 ng. Signatera and ERASE-Seq also use healthy ctDNA detection with NGS in the research setting
donor cfDNA to create a background model, without molecular barc­
odes. INVAR combines molecular barcoding, polishing analysis, and NGS has been used to detect ctDNA in patients with different cancer
position reading senses to determine or rule out the presence of ctDNA types, exploring its potential in various investigations. Low-cost strate­
with a reported LoD of 0.005%. The Tri-Nucleotide Error Reducer gies, such as amplicon sequencing, have been effective in identifying
(TNER) uses a hierarchical Bayesian method to generate a background treatment biomarkers in breast cancer [77,78]. TEC-seq has also been
model by studying trinucleotide context. These methods have limita­ successful in early detection, identifying concordant mutations at
tions as the background error suppression models are protocol and diagnosis in more than two thirds of patients with ovarian, breast, lung,
panel-specific, and modifications would require new modeling due to or CRC [21]. Leal et al. used TEC-Seq to identify somatic alterations in
potential false signals. plasma founding MRD at the post-operative time point in 20 patients
[26]. Tie et al. demonstrated the utility of ctDNA monitoring during

5
T.-M. Roberto et al. Cancer Treatment Reviews 119 (2023) 102595

treatment to track response in CRC using NGS with Safe-seq, as shown by Consent to participate: Not applicable.
reduced ctDNA levels during adjuvant chemotherapy [23, [30]. Reinert
et al. used the Signatera assay for MRD assessment, finding high possi­ References
bility of recurrence if ctDNA was detected post-operatively [28]. NGS in
liquid biopsy is also useful for identifying heterogeneity [24,123–125] [1] Sung H, Ferlay J, Siegel RL, et al. Global Cancer Statistics 2020: GLOBOCAN
Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries.
and cfDNA sequencing has revealed mutations developed during disease CA Cancer J Clin 2021;71(3):209–49. https://doi.org/10.3322/caac.21660.
progression that tissue sequencing could not identify [54]. Future clin­ [2] Hanahan D, Weinberg RA. Hallmarks of Cancer: The Next Generation. Cell 2011;
ical trials will determine the utility of these results in patient manage­ 144(5):646–74. https://doi.org/10.1016/j.cell.2011.02.013.
[3] Aghabozorgi AS, Bahreyni A, Soleimani A, et al. Role of adenomatous polyposis
ment according to ctDNA testing. Numerous ongoing clinical trials have coli (APC) gene mutations in the pathogenesis of colorectal cancer; current status
been reviewed elsewhere [25,126,127]. and perspectives. Biochimie 2019;157:64–71. https://doi.org/10.1016/j.
biochi.2018.11.003.
[4] Lee JY, Qing X, Xiumin W, et al. Longitudinal monitoring of EGFR mutations in
Conclusions plasma predicts outcomes of NSCLC patients treated with EGFR TKIs: Korean
Lung Cancer Consortium (KLCC-12-02). Oncotarget 2016;7:6984–93. https://doi.
Given the increasing interest in liquid biopsy for cancer research, it is org/10.18632/oncotarget.6874.
[5] Soverini S, Bassan R, Lion T. Treatment and monitoring of Philadelphia
essential to have adequate technologies and strategies for ctDNA anal­
chromosome-positive leukemia patients: recent advances and remaining
ysis. Overconfidence in the limit of detection and the number of targets challenges. J Hematol Oncol 2019;12(39). https://doi.org/10.1186/s13045-019-
in assays could lead to false positives or negatives. Fortunately, with 0729-2.
various approaches for error reduction and the versatility of library [6] Henry NL, Hayes DF. Cancer biomarkers. Mol Oncol 2012;6(2):140–6. https://
doi.org/10.1016/j.molonc.2012.01.010.
generation chemistry, NGS is a tool that offers multitarget assays [7] Mack PC, Banks KC, Espenschied CR, et al. Spectrum of driver mutations and
capable of detecting ctDNA in various clinical settings. However, further clinical impact of circulating tumor DNA analysis in non–small cell lung cancer:
advancements are necessary to increase sensitivity in critical contexts Analysis of over 8000 cases. Cancer 2020;126(14):3219–28. https://doi.org/
10.1002/cncr.32876.
such as early-stage detection or MRD detection. Prospective clinical [8] Zhou B, Xu K, Zheng X, et al. Application of exosomes as liquid biopsy in clinical
assays should explore the true capacity of cfDNA testing for patient diagnosis. Signal Transduct Target Ther 2020;5(144). https://doi.org/10.1038/
management. s41392-020-00258-9.
[9] Poulet G, Massias J, Taly V. Liquid Biopsy: General Concepts. Acta Cytol 2019;63:
449–55. https://doi.org/10.1159/000499337.
CRediT authorship contribution statement [10] Rego de Sousa MJ, Albuquerque M, Ribeiro R, et al. Evaluation of Noninvasive
Prenatal Testing (NIPT) guidelines using the AGREE II instrument. J Matern Fetal
Neonatal Med 2020;33(3):455–63. https://doi.org/10.1080/
Tébar-Martínez Roberto: Conceptualization, Writing – original 14767058.2018.1494716.
draft, Writing – review & editing. Martín-Arana Jorge: Writing – [11] Han D, Li R, Shi J, et al. Liquid biopsy for infectious diseases: A focus on microbial
original draft. Gimeno-Valiente Francisco: Writing – review & editing. cell-free DNA sequencing. Theranostics 2020;10(12):5501–13. https://doi.org/
10.7150/thno.45554.
Tarazona Noelia: Writing – review & editing. Rentero-Garrido Pilar:
[12] Choi JJ, Reich CF, Pisetsky DS. The role of macrophages in the in vitro generation
Writing – review & editing. Cervantes Andrés: Writing – review & of extracellular DNA from apoptotic and necrotic cells. Immunology 2005;115(1):
editing. 55–62. https://doi.org/10.1111/j.1365-2567.2005.02130.x.
[13] Snyder MW, Kircher M, Hill AJ, et al. Cell-free DNA Comprises an in Vivo
Nucleosome Footprint that Informs Its Tissues-Of-Origin. Cell 2016;164(1–2):
Declaration of Competing Interest 57–68. https://doi.org/10.1016/j.cell.2015.11.050.
[14] Alborelli I, Generali D, Jermann P, et al. Cell-free DNA analysis in healthy
The authors declare that they have no known competing financial individuals by next-generation sequencing: a proof of concept and technical
validation study. Cell Death Dis 2019;10(534). https://doi.org/10.1038/s41419-
interests or personal relationships that could have appeared to influence 019-1770-3.
the work reported in this paper. [15] Mouliere F, Chandrananda D, Piskorz AM, et al. Enhanced detection of circulating
tumor DNA by fragment size analysis. Sci Transl Med 2018;10(466):eaat4921.
https://doi.org/10.1126/scitranslmed.aat4921.
Acknowledgements [16] Meddeb R, Dache ZAA, Thezenas S, et al. Quantifying circulating cell-free DNA in
humans. Sci Rep 2019;9(5220):1–16. https://doi.org/10.1038/s41598-019-
The authors acknowledge the provincial associate board of Valencia 41593-4.
[17] Zhang Y, Yao Y, Xu Y, et al. Pan-cancer circulating tumor DNA detection in over
from Spanish Association Against Cancer (AECC) for the funding of the 10,000 Chinese patients. Nat Commun 2021;12(11). https://doi.org/10.1038/
contract of RT-M. Also, the authors wish to thank both Precision Med­ s41467-020-20162-8.
icine Unit and Bioinformatic and Biostatistics Unit of INCLIVA for their [18] Chen M, Zhao H. Next-generation sequencing in liquid biopsy: cancer screening
and early detection. Hum Genomics 2019;13(34). https://doi.org/10.1186/
support. s40246-019-0220-8.
Declarations [19] Ørntoft MBW, Jensen SØ, Øgaard N, et al. Age-stratified reference intervals
Funding: RT-M is supported by a grant from the Spanish Association unlock the clinical potential of circulating cell-free DNA as a biomarker of poor
outcome for healthy individuals and patients with colorectal cancer. Int J Cancer
Against Cancer (AECC) [grant number: PRDVA172011TEBA]. JM-A was
2021;148(7):1665–75. https://doi.org/10.1002/ijc.33434.
granted with an ACIF 2020 contract form Generalitat Valenciana [grant [20] Schiffman JD, Fisher PG, Gibbs P. Early Detection of Cancer: Past, Present, and
number: ACIF/2020/381]. NT and AC are supported by a grant from the Future. Am Soc Clin Oncol Educ Book 2015;35:57–65. https://doi.org/
Instituto de Salud Carlos III [grant number PI21/00689]. NT is also 10.14694/EdBook_AM.2015.35.57.
[21] Phallen J, Sausen M, Adleff V, et al. Direct detection of early-stage cancers using
supported by a Joan Rodés contract [grant number JR20/00005] from circulating tumor DNA. Sci Transl Med 2017;9(403). https://doi.org/10.1126/
the Instituto de Salud Carlos III. NT’s projects are funded by TTD Group, scitranslmed.aan2415.
Mutua Madrileña Foundation and SEOM. AC’s projects are funded by [22] Liu MC, Oxnard GR, Klein EA, et al. Sensitive and specific multi-cancer detection
and localization using methylation signatures in cell-free DNA. Ann Oncol 2020;
the Spanish Association Against Cancer (AECC). 31(6):745–59. https://doi.org/10.1016/j.annonc.2020.02.011.
NT declares advisory board or speaker fees from Amgen, Merck [23] Tie J, Wang Y, Tomasetti C, et al. Circulating tumor DNA analysis detects minimal
Serono, Pfizer and Servier in the last 5 years. AC declares institutional residual disease and predicts recurrence in patients with stage II colon cancer. Sci
Transl Med 2016;8(346). https://doi.org/10.1126/scitranslmed.aaf6219.
research funding from Genentech, Merck Serono, Bristol Myers Squibb, [24] Garcia-Murillas I, Schiavon G, Weigelt B, et al. Mutation tracking in circulating
Merck Sharp & Dohme, Roche, BeiGene, Bayer, Servier, Lilly, Gardant tumor DNA predicts relapse in early breast cancer. Sci Transl Med 2015;7(302).
Health, Natera, Novartis, Takeda, Astellas and FibroGen; and advisory https://doi.org/10.1126/scitranslmed.aab0021.
[25] Chen K, Shields MD, Chauhan PS, et al. Commercial ctDNA Assays for Minimal
board or speaker fees from Abbvie, Amgen, Merck Serono, Roche, Bayer, Residual Disease Detection of Solid Tumors. Mol Diagn Ther 2021;25:757–74.
Servier and Pierre Fabre in the last 5 years. The rest of the authors have https://doi.org/10.1007/s40291-021-00559-x.
declared no conflict of interest.
Ethics and consent: Not applicable.

6
T.-M. Roberto et al. Cancer Treatment Reviews 119 (2023) 102595

[26] Leal A, van Grieken NCT, Palsgrove DN, et al. White blood cell and cell-free DNA variations. Proc Natl Acad Sci USA 2003;100(15):8817–22. https://doi.org/
analyses for detection of residual disease in gastric cancer. Nat Commun 2020;11 10.1073/pnas.1133470100.
(525). https://doi.org/10.1038/s41467-020-14310-3. [52] Hindson CM, Chevillet JR, Briggs HA, et al. Absolute quantification by droplet
[27] Tarazona N, Gimeno-Valiente F, Gambardella V, et al. Targeted next-generation digital PCR versus analog real-time PCR. Nat Methods 2013;10:1003–5. https://
sequencing of circulating-tumor DNA for tracking minimal residual disease in doi.org/10.1038/nmeth.2633.
localized colon cancer. Ann Oncol 2019;30(11):1804–12. https://doi.org/ [53] Oxnard GR, Paweletz CP, Kuang Y, et al. Noninvasive Detection of Response and
10.1093/annonc/mdz390. Resistance in EGFR -Mutant Lung Cancer Using Quantitative Next-Generation
[28] Reinert T, Henriksen TV, Christensen E, et al. Analysis of Plasma Cell-Free DNA Genotyping of Cell-Free Plasma DNA. Clin Cancer Res 2014;20(6):1698–705.
by Ultradeep Sequencing in Patients with Stages I to III Colorectal Cancer. JAMA https://doi.org/10.1158/1078-0432.CCR-13-2482.
Oncol 2019;5(8):1124–31. https://doi.org/10.1001/jamaoncol.2019.0528. [54] Woolston A, Khan K, Spain G, et al. Genomic and Transcriptomic Determinants of
[29] Garcia-Murillas I, Chopra N, Comino-Méndez I, et al. Assessment of Molecular Therapy Resistance and Immune Landscape Evolution during Anti-EGFR
Relapse Detection in Early-Stage Breast Cancer. JAMA Oncol 2019;5(10):1473–8. Treatment in Colorectal Cancer. Cancer Cell 2019;36(1):35–50.e9. https://doi.
https://doi.org/10.1001/jamaoncol.2019.1838. org/10.1016/j.ccell.2019.05.013.
[30] Tie J, Cohen JD, Wang Y, et al. Circulating tumor DNA analyses as markers of [55] Kinde I, Wu J, Papadopoulos N, et al. Detection and quantification of rare
recurrence risk and benefit of adjuvant therapy for stage III colon cancer. JAMA mutations with massively parallel sequencing. Proc Natl Acad Sci USA 2011;108
Oncol 2019;5(12):1710–7. https://doi.org/10.1001/jamaoncol.2019.3616. (23):9530–5. https://doi.org/10.1073/pnas.1105422108.
[31] Powles T, Assaf ZJ, Davarpanah N, et al. ctDNA guiding adjuvant immunotherapy [56] Kou R, Lam H, Duan H, et al. Benefits and challenges with applying unique
in urothelial carcinoma. Nature 2021;595:432–7. https://doi.org/10.1038/ molecular identifiers in next generation sequencing to detect low frequency
s41586-021-03642-9. mutations. PLoS One 2016;11(1):1–15. https://doi.org/10.1371/journal.
[32] Butler TM, Johnson-Camacho K, Peto M, et al. Exome sequencing of cell-free DNA pone.0146638.
from metastatic cancer patients identifies clinically actionable mutations distinct [57] Fox EJ, Reid-Bayliss KS, Emond MJ, Loeb LA. Novel Next-Generation Sequencing
from primary disease. PLoS One 2015;10(8):1–14. https://doi.org/10.1371/ Applications. Next Gener Seq Appl 2014;1. https://doi.org/10.4172/
journal.pone.0136407. jngsa.1000106.
[33] Fernandes MGO, Cruz-Martins N, Souto Moura C, et al. Clinical Application of [58] Blomquist T, Crawford EL, Yeo J, et al. Control for stochastic sampling variation
Next-Generation Sequencing of Plasma Cell-Free DNA for Genotyping Untreated and qualitative sequencing error in next generation sequencing. Biomol Detect
Advanced Non-Small Cell Lung Cancer. Cancers (Basel) 2021;13(11):2707. Quantif 2015;5:30–7. https://doi.org/10.1016/j.bdq.2015.08.003.
https://doi.org/10.3390/cancers13112707. [59] Ma X, Shao Y, Tian L, et al. Analysis of error profiles in deep next-generation
[34] Lui YYNN, Chik K-W-W, Chiu RWKK, et al. Predominant hematopoietic origin of sequencing data. Genome Biol 2019;20(50):1–15. https://doi.org/10.1186/
cell-free dna in plasma and serum after sex-mismatched bone marrow s13059-019-1659-6.
transplantation. Clin Chem 2002;48(3):421–7. https://doi.org/10.1093/ [60] Deveson IW, Gong B, Lai K, et al. Evaluating the analytical validity of circulating
clinchem/48.3.421. tumor DNA sequencing assays for precision oncology. Nat Biotechnol 2021;39:
[35] Papadopoulos N. Pathophysiology of ctDNA Release into the Circulation and Its 1115–28. https://doi.org/10.1038/s41587-021-00857-z.
Characteristics: What Is Important for Clinical Applications. In: Schaffener F, [61] Smith T, Heger A, Sudbery I. UMI-tools: Modeling sequencing errors in Unique
Merlin JL, von Bubnoff N, editors. Tumor Liquid Biopsies. Switzerland: Springer Molecular Identifiers to improve quantification accuracy. Genome Res 2017;27:
Cham; 2020. p. 163–80. https://doi.org/10.1007/978-3-030-26439-0. 491–9. https://doi.org/10.1101/gr.209601.116.
[36] Abbosh C, Birkbak NJ, Swanton C. Early stage NSCLC — challenges to [62] Ebbert MTW, Wadsworth ME, Staley LA, et al. Evaluating the necessity of PCR
implementing ctDNA-based screening and MRD detection. Nat Rev Clin Oncol duplicate removal from next-generation sequencing data and a comparison of
2018;15:577–86. https://doi.org/10.1038/s41571-018-0058-3. approaches. BMC Bioinf 2016;17(239). https://doi.org/10.1186/s12859-016-
[37] Alcaide M, Cheung M, Hillman J, et al. Evaluating the quantity, quality and size 1097-3.
distribution of cell-free DNA by multiplex droplet digital PCR. Sci Rep 2020;10 [63] Orabi B, Erhan E, McConeghy B, et al. Alignment-free clustering of UMI tagged
(12564):1–10. https://doi.org/10.1038/s41598-020-69432-x. DNA molecules. Bioinformatics 2019;35(11):1829–36. https://doi.org/10.1093/
[38] Egyud M, Tejani M, Pennathur A, et al. Detection of Circulating Tumor DNA in bioinformatics/bty888.
Plasma: A Potential Biomarker for Esophageal Adenocarcinoma. Ann Thorac Surg [64] Sater V, Viailly PJ, Viailly PJ, et al. UMI-VarCal: A new UMI-based variant caller
2019;108(2):343–9. https://doi.org/10.1016/j.athoracsur.2019.04.004. that efficiently improves low-frequency variant detection in paired-end
[39] Ståhlberg A, Krzyzanowski PM, Egyud M, et al. Simple multiplexed PCR-based sequencing NGS libraries. Bioinformatics 2020;36(9):2718–24. https://doi.org/
barcoding of DNA for ultrasensitive mutation detection by next-generation 10.1093/bioinformatics/btaa053.
sequencing. Nat Protoc 2017;12:664–82. https://doi.org/10.1038/ [65] Chen S, Zhou Y, Chen Y, et al. Gencore: An efficient tool to generate consensus
nprot.2017.006. reads for error suppressing and duplicate removing of NGS data. BMC Bioinf
[40] Little S. Amplification-Refractory Mutation System (ARMS) Analysis of Point 2019;20(606):1–8. https://doi.org/10.1186/s12859-019-3280-9.
Mutations. Curr Protoc Hum Genet 1995;7(1):1–12. https://doi.org/10.1002/ [66] Tsagiopoulou M, Maniou MC, Pechlivanis N, et al. UMIc: A Preprocessing Method
0471142905.hg0908s07. for UMI Deduplication and Reads Correction. Front Genet 2021;12:1–10. https://
[41] Lo YMD. The Amplification Refractory Mutation System. Clinical Applications of doi.org/10.3389/fgene.2021.660366.
PCR 2003;I:61–70. https://doi.org/10.1385/0-89603-499-2:61. [67] Islam S, Zeisel A, Joost S, et al. Quantitative single-cell RNA-seq with unique
[42] Zhu F, Peng Y, Qi L, et al (2016) Nested ARMS-qPCR is a fast and cost-saving molecular identifiers. Nat Methods 2014;11:163–6. https://doi.org/10.1038/
method for single nucleotide polymorphism genotyping in clinical service. Int J nmeth.2772.
Clin Exp Med 9:16292–16300. ISSN:1940-5901/IJCEM0029480. [68] Lee J, Hyeon DY, Hwang D. Single-cell multiomics: technologies and data analysis
[43] Zhang X, Chang N, Yang G, et al. A comparison of ARMS-Plus and droplet digital methods. Exp Mol Med 2020;52:1428–42. https://doi.org/10.1038/s12276-020-
PCR for detecting EGFR activating mutations in plasma. Oncotarget 2017;8: 0420-2.
112014–23. https://doi.org/10.18632/oncotarget.22997. [69] Kivioja T, Vähärautio A, Karlsson K, et al. Counting absolute numbers of
[44] Li J, Wang L, Mamon H, et al. Replacing PCR with COLD-PCR enriches variant molecules using unique molecular identifiers. Nat Methods 2012;9:72–4. https://
DNA sequences and redefines the sensitivity of genetic testing. Nat Med 2008;14: doi.org/10.1038/nmeth.1778.
579–84. https://doi.org/10.1038/nm1708. [70] Potapov V, Ong JL. Examining sources of error in PCR by single-molecule
[45] Milbury CA, Li J, Makrigiorgos GM. Pcr-based methods for the enrichment of sequencing. PLoS One 2017;12(7):1–19. https://doi.org/10.1371/journal.
minority alleles and mutations. Clin Chem 2009;55(4):632–40. https://doi.org/ pone.0169774.
10.1373/clinchem.2008.113035. [71] Lanman RB, Mortimer SA, Zill OA, et al. Analytical and clinical validation of a
[46] Lipsky RH, Mazzanti CM, Rudolph JG, et al. DNA Melting Analysis for Detection digital sequencing panel for quantitative, highly accurate evaluation of cell-free
of Single Nucleotide Polymorphisms. Clin Chem 2001;47(4):635–44. https://doi. circulating tumor DNA. PLoS One 2015;10(10):1–27. https://doi.org/10.1371/
org/10.1093/clinchem/47.4.635. journal.pone.0140712.
[47] Gilson P. Enrichment and Analysis of ctDNA. In: Schaffener F, Merlin JL, von [72] Verma S, Moore MW, Ringler R, et al. Analytical performance evaluation of a
Bubnoff N, editors. Tumor Liquid Biopsies. Switzerland: Springer Cham; 2020. commercial next generation sequencing liquid biopsy platform using plasma
p. 181–211. https://doi.org/10.1007/978-3-030-26439-0. ctDNA, reference standards, and synthetic serial dilution samples derived from
[48] How-Kit A, Tost J. Pyrosequencing®-Based Identification of Low-Frequency normal plasma. BMC Cancer 2020;20(945). https://doi.org/10.1186/s12885-
Mutations Enriched Through Enhanced-ice-COLD-PCR. In: Lehmann U, Tost J, 020-07445-5.
editors. Pyrosequencing Methods in Molecular Biology, vol 1315. New York, NY, [73] Mansukhani S, Barber LJ, Kleftogiannis D, et al. Ultra-Sensitive mutation
USA: Humana Press; 2015. p. 83–101. https://doi.org/10.1007/978-1-4939- detection and genome-wide DNA copy number reconstruction by error- corrected
2715-9_7. circulating tumor DNA sequencing. Clin Chem 2018;64(11):1626–35. https://doi.
[49] How-Kit A, Lebbé C, Bousard A, et al. Ultrasensitive detection and identification org/10.1373/clinchem.2018.289629.
of BRAF V600 mutations in fresh frozen, FFPE, and plasma samples of melanoma [74] Hirotsu Y, Otake S, Ohyama H, et al. Dual-molecular barcode sequencing detects
patients by E-ice-COLD-PCR. Anal Bioanal Chem 2014;406:5513–20. https://doi. rare variants in tumor and cell free DNA in plasma. Sci Rep 2020;10(3391).
org/10.1007/s00216-014-7975-5. https://doi.org/10.1038/s41598-020-60361-3.
[50] How Kit A, Mazaleyrat N, Daunay A, et al. Sensitive detection of KRAS mutations [75] Yeh YM, Lin PC, Lee CT, et al. Treatment monitoring of colorectal cancer by
using enhanced-ice-COLD-PCR mutation enrichment and direct sequence integrated analysis of plasma concentration and sequencing of circulating tumor
identification. Hum Mutat 2013;34(11):1568–80. https://doi.org/10.1002/ DNA. Mol Cancer 2020;19(150):1–6. https://doi.org/10.1186/s12943-020-
humu.22427. 01273-8.
[51] Dressman D, Yan H, Traverso G, et al. Transforming single DNA molecules into
fluorescent magnetic particles for detection and enumeration of genetic

7
T.-M. Roberto et al. Cancer Treatment Reviews 119 (2023) 102595

[76] Chin YM, Takahashi Y, Chan HT, et al. Ultradeep targeted sequencing of [101] Gibson CJ, Steensma DP. New insights from studies of clonal hematopoiesis. Clin
circulating tumor DNA in plasma of early and advanced breast cancer. Cancer Sci Cancer Res 2018;24:4633–42. https://doi.org/10.1158/1078-0432.CCR-17-
2021;112(1):454–64. https://doi.org/10.1111/cas.14697. 3044.
[77] Masunaga N, Kagara N, Motooka D, et al. Molecular Barcode Sequencing of the [102] Bolton KL, Ptashkin RN, Gao T, et al. Cancer therapy shapes the fitness landscape
Whole Ligand Binding Domain of the ESR1 Gene in Cell-Free DNA from Patients of clonal hematopoiesis. Nat Genet 2020;52:1219–26. https://doi.org/10.1038/
with Metastatic Breast Cancer. Transl Oncol 2020;13(3):100735. https://doi.org/ s41588-020-00710-0.
10.1016/j.tranon.2019.12.007. [103] Mirabello L, Zhu B, Koster R, et al. Frequency of Pathogenic Germline Variants in
[78] Masunaga N, Kagara N, Motooka D, et al. Highly sensitive detection of ESR1 Cancer-Susceptibility Genes in Patients with Osteosarcoma. JAMA Oncol 2020;6
mutations in cell-free DNA from patients with metastatic breast cancer using (5):724–34. https://doi.org/10.1001/jamaoncol.2020.0197.
molecular barcode sequencing. Breast Cancer Res Treat 2018;167:49–58. https:// [104] Parikh AR, van Seventer EE, Siravegna G, et al. Minimal Residual Disease
doi.org/10.1007/s10549-017-4487-y. Detection using a Plasma-only Circulating Tumor DNA Assay in Patients with
[79] Yoshinami T, Kagara N, Motooka D, et al. Detection of ctDNA with Personalized Colorectal Cancer. Clin Cancer Res 2021;27(20):5586–94. https://doi.org/
Molecular Barcode NGS and Its Clinical Significance in Patients with Early Breast 10.1158/1078-0432.ccr-21-0410.
Cancer. Transl Oncol 2020;13(8):100787. https://doi.org/10.1016/j. [105] Forshew T, Murtaza M, Parkinson C, et al. Noninvasive identification and
tranon.2020.100787. monitoring of cancer mutations by targeted deep sequencing of plasma DNA. Sci
[80] Ståhlberg A, Krzyzanowski PM, Jackson JB, et al. Simple, multiplexed, PCR-based Transl Med 2012;4(136). https://doi.org/10.1126/scitranslmed.3003726.
barcoding of DNA enables sensitive mutation detection in liquid biopsies using [106] Plagnol V, Woodhouse S, Howarth K, et al. Analytical validation of a next
sequencing. Nucleic Acids Res 2016;44(11):e105–. https://doi.org/10.1093/nar/ generation sequencing liquid biopsy assay for high sensitivity broad molecular
gkw224. profiling. PLoS One 2018;13(3):1–18. https://doi.org/10.1371/journal.
[81] Calapre L, Giardina T, Robinson C, et al. Locus-specific concordance of genomic pone.0193802.
alterations between tissue and plasma circulating tumor DNA in metastatic [107] Gale D, Lawson ARJ, Howarth K, et al. Development of a highly sensitive liquid
melanoma. Mol Oncol 2019;13(2):171–84. https://doi.org/10.1002/1878- biopsy platform to detect clinically-relevant cancer mutations at low allele
0261.12391. fractions in cell free DNA. PLoS One 2018;13(3):1–18. https://doi.org/10.1371/
[82] Keup C, Benyaa K, Hauch S, et al. Targeted deep sequencing revealed variants in journal.pone.0194630.
cell-free DNA of hormone receptor-positive metastatic breast cancer patients. Cell [108] Pel J, Choi W, Leung A, et al. Duplex Proximity Sequencing (Pro-Seq): A method
Mol Life Sci 2020;77:497–509. https://doi.org/10.1007/s00018-019-03189-z. to improve DNA sequencing accuracy without the cost of molecular barcoding
[83] Crowgey EL, Mahajan N, Wong WH, et al. Error-corrected sequencing strategies redundancy. PLoS One 2017;13(10). https://doi.org/10.1371/journal.
enable comprehensive detection of leukemic mutations relevant for diagnosis and pone.0204265.
minimal residual disease monitoring. BMC Med Genomics 2020;13(32):1–11. [109] Cristiano S, Leal A, Phallen J, et al. Genome-wide cell-free DNA fragmentation in
https://doi.org/10.1186/s12920-020-0671-8. patients with cancer. Nature 2019;570:385–9. https://doi.org/10.1038/s41586-
[84] Nakanishi K, Kukita Y, Segawa H, et al. Characterization of the T-cell receptor 019-1272-6.
beta chain repertoire in tumor-infiltrating lymphocytes. Cancer Med 2016;5(9): [110] Wong D, Luo P, Znassi N, et al. Integrated, Longitudinal Analysis of Cell-free DNA
2513–21. https://doi.org/10.1002/cam4.828. in Uveal Melanoma. Cancer Research Communications 2023;3(2):267–80.
[85] Kukita Y, Ohkawa K, Takada R, et al. Selective identification of somatic mutations https://doi.org/10.1158/2767-9764.CRC-22-0456.
in pancreatic cancer cells through a combination of next-generation sequencing of [111] Bos MK, Angus L, Nasserinejad K, et al. Whole exome sequencing of cell-free DNA
plasma DNA using molecular barcodes and a bioinformatic variant filter. PLoS – A systematic review and Bayesian individual patient data meta-analysis. Cancer
One 2018;13(2):1–15. https://doi.org/10.1371/journal.pone.0192611. Treat Rev 2020;83:101951. https://doi.org/10.1016/j.ctrv.2019.101951.
[86] Kukita Y, Matoba R, Uchida J, et al. High-fidelity target sequencing of individual [112] Olmedillas-López S, García-Olmo DC, García-Arranz M, et al. Liquid biopsy by
molecules identified using barcode sequences: De novo detection and absolute NGS: differential presence of exons (DPE) in cell-free DNA reveals different
quantitation of mutations in plasma cell-free DNA from cancer patients. DNA Res patterns in metastatic and nonmetastatic colorectal cancer. Cancer Med 2018;7
2015;22(4):269–77. https://doi.org/10.1093/dnares/dsv010. (5):1706–16. https://doi.org/10.1002/cam4.1399.
[87] Hagi T, Kurokawa Y, Takahashi T, et al. Molecular Barcode Sequencing for Highly [113] Ahlborn LB, Rohrberg KS, Gabrielaite M, et al. Application of cell-free DNA for
Sensitive Detection of Circulating Tumor DNA in Patients with Esophageal genomic tumor profiling: A feasibility study. Oncotarget 2019;10:1388–98. htt
Squamous Cell Carcinoma. Oncology (Switzerland) 2020;98:222–9. https://doi. ps://doi.org/10.18632/oncotarget.26642.
org/10.1159/000504808. [114] Song J, Yang Z. Case report: Whole exome sequencing of circulating cell-free
[88] Peng Q, Xu C, Kim D, et al. Targeted Single Primer Enrichment Sequencing with tumor DNA in a follicular thyroid carcinoma patient with lung and bone
Single End Duplex-UMI. Sci Rep 2019;9(4810):1–10. https://doi.org/10.1038/ metastases. J Circ Biomark 2018;7(1):1–6. https://doi.org/10.1177/
s41598-019-41215-z. 1849454418763725.
[89] Dunwell TL, Dailey SC, Ottestad AL, et al. Adaptor Template Oligo-Mediated [115] Adalsteinsson VA, Ha G, Freeman SS, et al. Scalable whole-exome sequencing of
Sequencing (ATOM-Seq) is a new ultra-sensitive UMI-based NGS library cell-free DNA reveals high concordance with metastatic tumors. Nat Commun
preparation technology for use with cfDNA and cfRNA. Sci Rep 2021;11(3138): 2017;8(1324). https://doi.org/10.1038/s41467-017-00965-y.
1–13. https://doi.org/10.1038/s41598-021-82737-9. [116] Aguilar-Mahecha A, Lafleur J, Brousse S, et al. Early, on-treatment levels and
[90] Dai P, Wu LR, Chen SX, et al. Calibration-free NGS quantitation of mutations dynamic changes of genomic instability in circulating tumor DNA predict
below 0.01% VAF. Nat Commun 2021;12(6123):1–9. https://doi.org/10.1038/ response to treatment and outcome in metastatic breast cancer patients. Cancers
s41467-021-26308-6. (Basel) 2021;13(6):1331. https://doi.org/10.3390/cancers13061331.
[91] Hu J, Hu B, Gui YC, et al. Diagnostic value and clinical significance of methylated [117] Chen X, Chang CW, Spoerke JM, et al. Low-pass whole-genome sequencing of
SEPT9 for colorectal cancer: A meta-analysis. Med Sci Monit 2019;25:5813–22. circulating cell-free DNA demonstrates dynamic changes in genomic copy number
https://doi.org/10.12659/MSM.915472. in a squamous lung cancer clinical cohort. Clin Cancer Res 2019;25(7):2254–63.
[92] Constâncio V, Nunes SP, Moreira-Barbosa C, et al. Early detection of the major https://doi.org/10.1158/1078-0432.CCR-18-1593.
male cancer types in blood-based liquid biopsies using a DNA methylation panel. [118] Zviran A, Schulman RC, Shah M, et al. Genome-wide cell-free DNA mutational
Clin Epigenetics 2019;11(175):1–18. https://doi.org/10.1186/s13148-019-0779- integration enables ultra-sensitive cancer monitoring. Nat Med 2020;26:1114–24.
x. https://doi.org/10.1038/s41591-020-0915-3.
[93] Stackpole ML, Zeng W, Li S, et al. Cost-effective methylome sequencing of cell- [119] Vandekerkhove G, Todenhöfer T, Annala M, et al. Circulating Tumor DNA Reveals
free DNA for accurately detecting and locating cancer. Nat Commun 2022;13 Clinically Actionable Somatic Genome of Metastatic Bladder Cancer. Clin Cancer
(5566). https://doi.org/10.1038/s41467-022-32995-6. Res 2017;23(21):6487–97. https://doi.org/10.1158/1078-0432.CCR-17-1140.
[94] Wen L, Li J, Guo H, et al. Genome-scale detection of hypermethylated CpG islands [120] Chaudhuri AA, Chabon JJ, Lovejoy AF, et al. Early Detection of Molecular
in circulating cell-free DNA of hepatocellular carcinoma patients. Cell Res 2015; Residual Disease in Localized Lung Cancer by Circulating Tumor DNA Profiling.
25:1250–64. https://doi.org/10.1038/cr.2015.126. Cancer Discov 2017;7(12):1394–403. https://doi.org/10.1158/2159-8290.CD-
[95] Shen SY, Singhania R, Fehringer G, et al. Sensitive tumour detection and 17-0716.
classification using plasma cell-free DNA methylomes. Nature 2018;563:579–83. [121] Gandara DR, Paul SM, Kowanetz M, et al. Blood-based tumor mutational burden
https://doi.org/10.1038/s41586-018-0703-0. as a predictor of clinical benefit in non-small-cell lung cancer patients treated
[96] Shen SY, Burgener JM, Bratman S v, de Carvalho DD (2019) Preparation of with atezolizumab. Nat Med 2018;24:1441–8. https://doi.org/10.1038/s41591-
cfMeDIP-seq libraries for methylome profiling of plasma cell-free DNA. Nat 018-0134-3.
Protoc 14:2749–2780. https://doi.org/10.1038/s41596-019-0202-2. [122] Fridland S, Choi J, Nam M, et al. Assessing tumor heterogeneity: integrating tissue
[97] Razavi P, Li BT, Brown DN, et al. High-intensity sequencing reveals the sources of and circulating tumor DNA (ctDNA) analysis in the era of immuno-oncology -
plasma circulating cell-free DNA variants. Nat Med 2019;25:1928–37. https:// blood TMB is not the same as tissue TMB. J Immunother Cancer 2021;9:e002551.
doi.org/10.1038/s41591-019-0652-7. [123] Marusyk A, Janiszewska M, Polyak K. Intratumor Heterogeneity: The Rosetta
[98] Jiang T, Ren S, Zhou C. Multi-cancer blood testing combined with PET-CT: road Stone of Therapy Resistance. Cancer Cell 2020;37(4):471–84. https://doi.org/
for hope to screen for cancer and guide intervention. Signal Transduct Target 10.1016/j.ccell.2020.03.007.
Ther 2020;5(95). https://doi.org/10.1038/s41392-020-0210-2. [124] Siravegna G, Mussolin B, Buscarino M, et al. Clonal evolution and resistance to
[99] Park SJ, Bejar R. Clonal hematopoiesis in cancer. Exp Hematol 2020;83:105–12. EGFR blockade in the blood of colorectal cancer patients. Nat Med 2015;21:
https://doi.org/10.1016/j.exphem.2020.02.001. 795–801. https://doi.org/10.1038/nm.3870.
[100] Zink F, Stacey SN, Norddahl GL, et al. Clonal hematopoiesis, with and without
candidate driver mutations, is common in the elderly. Blood 2017;130:742–52.
https://doi.org/10.1182/blood-2017-02-769869.

8
T.-M. Roberto et al. Cancer Treatment Reviews 119 (2023) 102595

[125] Dentro SC, Leshchiner I, Haase K, et al. Characterizing genetic intra-tumor [127] Puccini A, Martelli V, Pastorino A, et al. ctDNA to Guide Treatment of Colorectal
heterogeneity across 2,658 human cancer genomes. Cell 2021;184(8):2239–2254. Cancer: Ready for Standard of Care? Curr Treat Options Oncol 2023;24:76–92.
e39. https://doi.org/10.1016/j.cell.2021.03.009. https://doi.org/10.1007/s11864-022-01048-x.
[126] Kasi PM, Fehringer G, Taniguchi H, et al. Impact of Circulating Tumor
DNA–Based Detection of Molecular Residual Disease on the Conduct and Design
of Clinical Trials for Solid Tumors. JCO Precis Oncol 2022;6:e2100181.