Sage Journals: Discover world-class research

Abstract

Tumor biopsy is currently the gold standard for diagnosis and in determining cell signaling pathways involved in the development of treatment resistance. However, there are major challenges with this technique, including the need for serial sampling to monitor treatment resistance, which is invasive and also has the potential for selection bias due to intra-tumoral and inter-tumoral heterogeneity. These challenges highlight the need for more effective methods for obtaining Tumor samples. Liquid biopsy analyzes genetic material or tumor cells shed into the blood from the primary tumor and metastatic sites and consequently provides a comprehensive, real-time picture of the tumor burden in an individual patient. Indeed, liquid biopsy has the potential to revolutionize cancer management. Here, we review recent studies on the potential clinical applications of liquid biopsy using circulating tumor DNA in colorectal cancer, including screening, diagnosis, detection of minimal residual disease after surgery, detection of recurrence, prognosis, predicting treatment response, monitoring tumor burden or response during treatment, and tracking resistance. We also discuss recent data demonstrating the utility of detecting KRAS-mutated circulating tumor DNA, both at diagnosis to determine an appropriate treatment strategy and during anti-epidermal growth factor receptor therapy to predict treatment resistance. The future integration of liquid biopsy into clinical practice is discussed, together with alternative approaches and key questions that need to be answered in future clinical studies before this technology can be implemented and used routinely.

Keywords

Liquid biopsy colorectal cancer personalized treatment biomarker circulating tumor DNA

Introduction

The clinical outcome of patients with cancer is mainly determined by the blood-borne dissemination of cancer cells from the primary site to distant organs and the eventual outgrowth of a largely unknown subset of these cells into metastases in their new microenvironment.¹ Almost all cancers treated with tumor-specific therapies acquire resistance due to tumor heterogeneity, clonal evolution, and selection.² Subclones may also arise during disease progression, leading to changes in the proportion and pattern of specific aberrations among the tumor cells composing the primary tumor as well as the tumor cells of the metastatic lesions.^3,4 Since therapy-related biomarkers may change throughout tumor progression, serial biomarker investigations may provide crucial information for therapy selection and, thus, personalized treatment.⁵

Although tumor tissue biopsy sampling is the gold standard approach for diagnosis and for determining molecular pathways involved in the development of resistance to targeted therapies, the technique poses major challenges: difficulties in obtaining sufficient tumor cells for analysis, the need for invasive serial sampling to monitor response, and the potential for selection bias due to intra-tumoral and inter-tumoral heterogeneity.^6–8 Such challenges highlight the need for more effective biopsy methods. “Liquid biopsy” analysis using circulating tumor cells (CTCs) or plasma-derived circulating tumor DNA (ctDNA) offers a non-invasive alternative to tissue biopsies. CTCs were identified in 1869, by Thomas Ashworth, and correspond to one of the stages of the metastatic dissemination process. They occur at very low levels in blood (1 in 10⁶–10⁷ nucleated blood cells), presenting a challenge for detection.⁹ Currently, the most commonly used detection system is CellSearch^TM (Veridex, Raritan, NJ, USA), which allows quantification of CTCs and also the isolation of viable CTCs, from which genetic material can potentially be extracted, amplified, and analyzed.

In this review, we focus on the potential clinical applications of liquid biopsy using ctDNA in colorectal cancer (CRC). Despite substantial survival gains over recent years, CRC remains a leading cause of death in developed countries, with over half of patients developing metastatic disease.^10,11 CRC may be considered as a genetic disease characterized by molecular heterogeneity and sequential genetic alterations.^11,12 As a consequence of better understanding of colorectal carcinogenesis, targeted agents have been developed which have improved outcomes in patients with metastatic CRC (mCRC). These include monoclonal antibodies (mAbs) targeting two major pathways critical to tumorigenesis, namely, the epidermal growth factor receptor (EGFR) and vascular endothelial growth factor (VEGF) signaling pathways. These mAbs led to a substantial increase in overall survival, which is over 30 months in recent studies. A strong emphasis on the clinical problem of primary and acquired resistance to anti-EGFR agents has heightened interest in the potential for liquid biopsy to predict and monitor resistance in CRC.¹¹ While recognizing the potential for CTCs to monitor tumor dynamics and provide insight into the characteristics of cells responsible for metastasis,¹³ at present there are limited data supporting the use of CTCs in clinical practice over the short term.¹¹ Accordingly, studies on the use of ctDNA in diagnosis and screening, detection of recurrence, its prognostic value, monitoring tumor burden or response during treatment, and tracking resistance are reviewed.

Liquid biopsy using ctDNA in blood

The presence of circulating cell-free DNA (cfDNAs) in human blood was first described in 1948.¹⁴ In 1977, cancer patients were found to have higher plasma cfDNA levels than healthy individuals.¹⁵ Current molecular biology techniques make it easy to detect DNA in the blood, paving the way for major research advances and new clinical applications.

The fraction of cfDNA that carries tumor-specific alterations is termed ctDNA.¹⁶ Tumor cells release ctDNA into the bloodstream by apoptosis and necrosis, and probably to a lesser extent, by active secretion.^16,17 Apoptotic tumor cells are thought to be the most important sources of ctDNA in the blood, as detectable ctDNA is typically highly fragmented (between 180 and 200 base pairs), which is indicative of apoptotic DNA.^16,18–20 It is not known if all clones in a multiclonal tumor with genetic instability will shed DNA into the circulation. Some clones may have more proliferative and less apoptotic activities than others. Indeed, the genetic profile detected using ctDNA probably mirrors clones with high apoptotic activity. In addition, detectable ctDNA levels in patients with cancer do not always correspond to the ability to detect CTCs, suggesting that these two biomarkers are distinct entities.^16,18–21 CTCs can potentially be assessed in all patients rather than a subset with a known gene mutation, but some patients may test positive for ctDNA and negative for CTCs due to the very low CTC counts or absence of CTC in blood.²² Furthermore, CTC counts are lower in CRC compared with other cancers, probably because many CTCs released from a primary tumor in the colon become trapped in the liver before reaching the systemic circulation.²³ Consistent with these observations and using current techniques, the diagnostic performance of ctDNA is superior to that of CTCs in CRC.

Isolation of ctDNA from both plasma and serum cfDNA is possible due to the technical ability to detect genetic alterations at low levels in cfDNA.¹⁹ These ctDNA fragments contain genetic defects identical to those seen in the original tumor, and virtually, all cancer-related molecular alterations can be detected, including somatic point mutations, translocations, and DNA methylation changes.^24–26 However, it is not currently possible to isolate ctDNA from other circulating DNAs or to confirm that cfDNA is tumor derived unless molecular alterations (genetic or epigenetic) are detected.

There are two main approaches for the detection of mutations in cfDNA. The first involves a targeted approach that includes the analysis of known genetic changes from a given tumor type in a small set of frequently occurring mutations (e.g. KRAS, NRAS, BRAF, and EGFR mutations in CRC). The second involves a non-targeted approach using total cfDNA quantification, which is less cost-effective. A non-targeted approach using genome-wide analysis of cfDNA or whole-exome sequencing is also possible.²⁷

Levels of ctDNA in blood can vary widely between patients due to differences in tumor DNA release mechanisms and the effectiveness of cfDNA clearance, with estimates of the proportion of ctDNA from total cfDNA varying from 0.01% to 90%.^20,28 Serum cfDNA levels are significantly higher than plasma levels^20,29–33 due to clotting of white blood cells in the collection tube, leading to their lysis.^20,32,33 Consequently, serum ctDNA is diluted by genomic DNA released from white blood cells, and plasma is preferred as a source of ctDNA.^19,34 In cancer patients, cfDNA levels increase further in metastatic disease, suggesting that total cfDNA is related to tumor burden.²⁰ However, cfDNA elevation also occurs in inflammation (which is associated with accelerated shedding of DNA into the bloodstream) and conditions involving organ necrosis (e.g. myocardial infarction), whereas ctDNA is specific to one or more cancer subtypes. Nevertheless, lung cancer patients were found to have significantly higher plasma cfDNA levels than patients with chronic respiratory inflammation, suggesting that cfDNA elevation in cancer may be mainly due to tumor development rather than inflammation.³⁵

Extracellular vesicles (including exosomes, microvesicles, and apoptotic bodies) are another source of tumor DNA and can be secreted by cancer cells into the blood. Exosomes are spherical, nano-sized vesicles released by exocytosis from multi-vesicular bodies of the late endosome.^36,37 They are exocytosed in a constitutive manner from both normal and tumor cells and are present in blood and other body fluids. Tumor-derived exosomes have pleiotropic roles affecting tumor growth and development. Exosomes contain proteins, lipids, and nucleic acids, including double-stranded DNA, single-stranded RNA, long non-coding RNA, and microRNA (miRNA). As a result, tumor mutations can be identified using exosome-derived nucleic acids. Tumor DNA from extracellular vesicles is unlikely to contribute to detectable plasma ctDNA levels since exosomes contain mostly miRNA, with low levels of double-stranded DNA. In addition, DNA release from exosomes in the blood is poorly understood but probably low.^38,39 The vesicles can be extracted from serum and may represent a promising future alternative to ctDNA for tumor DNA analysis, as tumor-derived genomic material from extracellular vesicles is more concentrated and better preserved.^36,37 Some studies have also evaluated circulating free miRNA (non-exosomal) as biomarkers, but results are preliminary; neither circulating free miRNA nor exosomal miRNA are ready to be used as a liquid biopsy in clinical practice.⁴⁰

Levels of circulating DNA are influenced by clearance, degradation, and other physiological filtering events in blood and lymphatic circulation.^41–43 Nucleic acids are cleared from the blood by the liver and kidneys and have a variable half-life in circulation, ranging from 15 min to 2 h.⁴¹ Defining the tumor genetic profile based on blood ctDNA levels of an individual patient has been made possible with the advent of increasingly sensitive techniques.^41–43 Early studies used Sanger sequencing to detect plasma ctDNA, but this technique has many shortcomings, including low sensitivity and laborious protocols. “Second-generation” sequencing techniques provide significantly higher sensitivity for ctDNA detection (Table 1). For example, the polymerase chain reaction (PCR)-based bidirectional pyrophosphorolysis-activated polymerization (Bi-PAP) technique can precisely detect a single-nucleotide change using a pyrophosphorolysis-activated oligonucleotide that is then extended by DNA polymerization.⁴⁴ As with most PCR-based techniques, it is very specific but has limited sensitivity for ctDNA detection. BEAMing (beads, emulsion, amplification, and magnetics) is a digital PCR technique in which the target DNA segment is amplified using primers containing known tag sequences and covalently bound to magnetic beads.²⁰ Following incubation with fluorescently labeled hybridization probes, magnetic flow cytometry is used to quantify beads containing the DNA mutation. Next-generation sequencing (NGS) is a broad term encompassing various high-throughput technologies with a reduced cost per assay. These include deep sequencing using a probe panel consisting of DNA oligonucleotides targeting frequently mutated regions in the gene of interest.^45,46 Systemic PCR and sequencing errors may compromise the ability of high-throughput techniques to detect rare mutations. To address this limitation, the Safe-Sequencing System (Safe-SeqS) uses a unique DNA sequence or “barcode” to tag each DNA fragment that is to be amplified, followed by redundant sequencing of the amplified fragments.^45,47 This approach allows confident identification of low-frequency mutations, but PCR sampling efficiency and sampling bias may be issues when using unique identifiers.⁴⁸ Development of “third-generation” whole-genome sequencing techniques is ongoing. These include personalized analysis of rearranged ends (PARE), a technique developed to detect unselected genetic events across the whole genome.⁴⁹ A more cost-effective approach for mutation detection would be whole-exome sequencing, which may have potential for identifying mutations associated with acquired drug resistance and studying clonal evolution.⁵⁰ A new technique known as DREAMing (Discrimination of Rare EpiAlleles by Melt) enables analysis of very rare epiallelic variants in liquid biopsy samples.⁵¹

Table 1.

Available techniques used for DNA analyses.

Principle of detection	Technique	Sensitivity	Application	Limitations	Advantages	Optimal clinical utility
PCR-based	Nested real-time PCR	0.1%–10%	Detection of SNV only	Low sensitivity	Easy for routine use, low cost	Tumor tissue
	ARMS PCR
	Bi-PAP amplification
Digital PCR	BEAMing	0.01%–0.1%	Detection of SNV and known genomic rearrangements only	Specific equipment, quite expensive	High sensitivity	ctDNA (detection of tumor recurrence, determining prognosis, monitoring tumor response, and tracking resistance)
	Droplet-based digital PCR
	Microfluidic digital PCR
Targeted deep sequencing	Safe-SeqS	0.01%–0.1%	Detection of SNV, CNV, and rearrangements across targeted regions only	PCR sampling bias and sequencing errors	High sensitivity, cost decreasing	ctDNA (cancer diagnosis, screening, and determining prognosis)
	TAmSeq
	Ion-AmpliSeq
	CAPP-Seq
Whole-genome sequencing	PAREWhole-exome sequencing	1%	Detection of genome-wide SNV, CNV, and rearrangements	Expensive, sensitivity improvement ongoing	Genome-wide applications	Currently inappropriate since improvements to the technique are ongoing Future: ctDNA (cancer diagnosis, screening, and tracking resistance)

ARMS: amplification-refractory mutation system; BEAMing: beads, emulsion, amplification, and magnetics; Bi-PAP: bidirectional pyrophosphorolysis-activated polymerization; CAPP-Seq: cancer personalized profiling by deep sequencing; CNV: copy number variation; PARE: personalized analysis of rearranged ends; PCR: polymerase chain reaction; Safe-SeqS: safe-sequencing system; SNV: single-nucleotide variation; TamSeq: tagged-amplicon deep sequencing; ctDNA: circulating tumor DNA.

Bettegowda et al.²² demonstrated the ability to obtain valuable information on a tumor’s genetic make-up from a blood sample. By directly comparing ctDNA in 187 tumors originating from 16 different tissue types, and evaluating which ctDNA measurements could prove clinically useful, the authors demonstrated that ctDNA levels varied among tumor types, with all patients with mCRC having detectable ctDNA, compared with under 10% of those with glioma. Investigating ctDNA could also help to account for molecular heterogeneity, because the peripheral blood is a pool of DNA derived from the primary tumor and different metastatic sites and may therefore provide a comprehensive real-time picture of the whole tumor burden in an individual patient. This approach has potential in several important clinical applications.

Potential clinical applications of liquid biopsy in colorectal cancer

Cancer diagnosis and screening

Common gene mutations in CRC include KRAS and BRAF V600E, with a frequency of approximately 40% and 5%–9%, respectively.^52–55 Highly sensitive techniques can detect nearly all such mutations, but a good screening or diagnostic test should identify at least 80% of cases, thus ctDNA detection of individual mutations has insufficient sensitivity for population screening (≈50% if using KRAS and BRAF ctDNA testing). In a recent study, the overall sensitivity for detection of RAS mutations was 63%, but was 100% for patients whose primary tumor remained in place versus 46% in those who had undergone resection of the primary tumor.⁵⁶ Regarding KRAS-mutated tumors, sensitivity to detect KRAS-mutated ctDNA ranges from 20% to 60% for non-metastatic disease (with higher sensitivity in stage III than in stages I–II) to 80%–100% for metastatic disease.^57,58 By comparison, the fecal immunochemical test (FIT) had a sensitivity of 79% (95% confidence interval (CI) = 69%–86%) and specificity of 94% (95% CI = 92%–95%) for CRC screening, based on data from 19 studies in asymptomatic, average-risk adults.⁵⁹

An alternative approach to identifying ctDNA is aberrant DNA methylation of specific promoter regions, which is more consistent between tumors than mutations.⁶⁰ DNA methylation is an early event in CRC carcinogenesis, and various studies have investigated the use of methylated ctDNA for CRC screening and diagnosis (Table 2).^61–69 Several studies evaluated the pretreatment methylation status of the Septin 9 (SEPT9) gene promoter. SEPT9 methylation demonstrated a sensitivity of about 70% (50%–90%) and a specificity of about 90% (70%–100%) to detect CRC, depending on disease stage (metastatic vs non-metastatic).^{43,63,64,68–70} However, plasma SEPT9 methylation demonstrated disappointing results in a large prospective study of 7941 patients undergoing screening colonoscopy, with a standardized sensitivity of only 48% (35% for stage I, 63% for stage II, and 46% for stage III).⁶²

Table 2.

Colorectal cancer screening or diagnosis by detecting methylated and/or mutated ctDNA prior to treatment.

Reference	Gene(s)	Technique used for ctDNA measurement	Number of CRC patients	Disease stage (%)	Detection rate in blood (% (n))
Cassinotti et al.⁶¹	CYCD2 HIC1 PAX5 RASSF1A RB1 SRBC All genes together	Microarray-mediated methylation assay	30	Stage M0 (100%)	97% (n = 29) 63% (n = 19) 87% (n = 26) 93% (n = 28) 90% (n = 27) 33% (n = 10) 84% (n = 25)
Church et al.⁶²	SEPT9	Bisulfite conversion and real time of methylated SEPT9	53	Stage M0 (91%) Stage M1 (9%)	51% (n = 27)
deVos et al.⁶³	SEPT9	Bisulfite conversion and real-time PCR of methylated SEPT9	187	Stage M0 (96%) Stage M1 (4%)	56% (n = 105)
Grutzmann et al.⁶⁴	SEPT9	Bisulfite conversion and real-time PCR of methylated SEPT9	378	Stage M0 (84%) Stage M1 (11%) Unknown (5%)	51% (n = 193)
Herbst et al.⁶⁵	NEUROG1	Bisulfite conversion and real-time PCR of methylated NEUROG1	95	Stage M0 (81%) Stage M1 (19%)	66% (n = 63)
Lee et al.⁶⁶	APC MGMT RASSF2A Wif-1 All genes together	Methylation-specific PCR	243	Stage M0 (100%)	27% (n = 66) 39% (n = 95) 58% (n = 141) 74% (n = 180) 86% (n = 210)
Lee et al.⁶⁷	SEPT	Bisulfite conversion and real-time PCR of methylated SEPT9	101	Stage M0 (83%) Stage M1 (17%)	37% (n = 37)
Lofton-Day et al.⁶⁸	TMEFF2 NGFR SEPT9	Bisulfite conversion and real-time PCR of methylated TMEFF2, NGFR, and SEPT9	133	Stage M0 (75%) Stage M1 (23%) Unknown (2%)	65% (n = 87) 51% (n = 68) 69% (n = 92)
Warren et al.⁶⁹	SEPT9	Bisulfite conversion and real-time PCR of methylated SEPT9	50	Stage M0 (90%) Stage M1 (10%)	90% (n = 45)

APC: adenomatous polyposis coli; CYCD2: cyclin-D2-1; HIC1: hypermethylated in cancer 1; MGMT: O-6-methylguanine-DNA methyltransferase; NEUROG1: neurogenin 1; NGFR: nerve growth factor receptor; PAX5: paired box 5; PCR: polymerase chain reaction; RASSF2A: Ras-association domain family 2A; RB1: retinoblastoma; SEPT9: septin 9; SRBC: serpentine receptor, class BC; TMEFF2: transmembrane protein with EGF-like and two follistatin-like domains 2; Wif-1: WNT inhibitory factor 1; ctDNA: circulating tumor DNA; CRC: colorectal cancer.

It is important to note that these findings are dependent on the sensitivity of the technique used. These techniques are complex and heterogeneous, with variation in both the segment of the SEPT9 gene analyzed and the cut-off level considered to be positive. These factors could account for the differing SEPT9 methylation sensitivity reported. A meta-analysis of three studies that directly compared methylated SEPT9 and fecal occult blood testing (FOBT) as a screening tool for CRC found that methylated SEPT9 had a pooled sensitivity of 76% (95% CI = 71%–80%), compared with 67% (95% CI = 60%–74%) for FOBT, leading to a better diagnostic efficiency for methylated SEPT9.⁷¹ A methylation SEPT9 assay was recently approved in the United States for screening of CRC (Epi proColon^®; Epigenomics, Inc., Germantown, MD, USA). In addition, in 2014, the US Food and Drug Administration approved Cologuard^®, the first stool-based colorectal screening test that detects the presence of red blood cells and DNA mutations that may indicate the presence of colon cancer or precursors to cancer.⁷² Others trials evaluating multitarget stool DNA testing are also ongoing. Nevertheless, we believe that further data and standardization are necessary before these approaches can be used in clinical practice.

Regarding future directions, recent studies have identified additional DNA methylation biomarkers that may be relevant for CRC diagnosis.^51,73,74 Promising results have also been reported using multigene methylation signatures for CRC screening in order to increase sensitivity and specificity.^61–69 For example, an approach using a BCAT1 and IKZF1 methylation assay identified approximately 70% of CRC.^75–77

Detection of tumor recurrence, minimal residual disease, and prognosis value

Surgery alone cures a large fraction of patients with localized CRC. However, there are currently no effective means to identify which patients have residual disease that will result in recurrence. Consequently, patients with high-risk clinicopathological features indiscriminately receive potentially toxic and unnecessary adjuvant therapy. As ctDNA is a potential marker of residual disease after resection, liquid biopsy may help determine which patients will experience recurrence and would benefit most from adjuvant treatment. To date, few studies have evaluated ctDNA as a marker of recurrence after curative surgery in CRC. However, the available studies have found that ctDNA detection after surgical resection is associated with minimal residual disease and a high recurrence rate.^20,78–82 Studies specifically evaluating the tumor recurrence rate after curative surgery are shown in Table 3.^20,79,81,82

Table 3.

Detection of colorectal cancer recurrence with ctDNA.

Reference	Gene(s)	Period of DNA extraction from plasma	Methods used for measurement of ctDNA	Treatment	Numbers of patients	Detection rate of recurrence in plasma (%, n-positive ctDNA/n recurrence)
Diehl et al.²⁰	APC, TP53, KRAS, and PI3K	Before and after surgery (mostly stage IV)	BEAMing	Surgery ± chemotherapy (61%)	18	100% (n = 15/15)
Ryan et al.⁷⁹	KRAS	After surgery at 1 week, 1 month, and then 3-monthly (mostly stages I–III)	Semi-nested mutant enrichment technique and direct sequencing	Surgery ± chemotherapy (53%)	94	91% (n = 10/11)
Frattini et al.⁸¹	Total DNA, KRAS, and p16INK4a	Before and after surgery at 4, 10, and 16 months	Mutant-enriched PCR and F-MSP	Surgery ± chemotherapy	70	100% (n = 18/18)
Tie et al.⁸²	Personalized assay	4-10 weeks after surgery (stage II), then every 3 months for up to 2 years in a subset of patients	Safe-SeqS	Surgery ± chemotherapy	230	41% (n = 11/27) in surgery-only group

APC: adenomatous polyposis coli; BEAMing: beads, emulsion, amplification, and magnetics; F-MSP: fluorescent methylation-specific PCR; PI3K: phosphoinositide 3-kinase; PCR: polymerase chain reaction; Safe-SeqS: safe-sequencing system; ctDNA: circulating tumor DNA.

Some studies have used KRAS-mutated ctDNA to detect CRC recurrence.⁷⁹ In one such study, 63% of patients with detectable KRAS-mutated ctDNA postoperatively developed recurrence, compared with only 2% of patients who remained serum mutant negative, giving a sensitivity to detect recurrence of 91% and specificity of 88%.⁷⁹ Detection of ctDNA during follow-up was an independent predictor of disease recurrence (hazard ratio (HR) = 6.37, 95% CI = 2.26–18.0; p < 0.001). Diehl et al. investigated the use of serial plasma ctDNA measurements to monitor recurrence and tumor burden dynamics over time in 18 patients with CRC and complete surgical resection. The specified ctDNA mutants (all tumors had mutations in one or more of the four genes assessed: APC, KRAS, TP53, and PIK3CA) were detectable at first follow-up (13–56 days after surgery) in 80% of cases, with recurrence occurring in 94% of this group. In marked contrast, no recurrence occurred in the 20% of patients with undetectable ctDNA.²⁰ A recent study using a personalized Safe-SeqS assay reported that detection of ctDNA after surgery for stage-II CRC identified a subset of patients at very high risk of recurrence if not treated with chemotherapy: radiologic recurrence was confirmed in 79% of patients who tested positive for ctDNA postoperatively, compared with 9.8% of those with negative ctDNA.⁸² The sensitivity and specificity of post-operative ctDNA in predicting recurrence at 3 years were 48% and 100%, respectively. These findings indicate that ctDNA may help to detect minimal residual disease after resection, to predict tumor recurrence with high sensitivity and to determine which patients will most benefit from adjuvant treatment. In this setting, ctDNA should be measured before surgery and again before the initiation of adjuvant therapy (typically 6 weeks after surgery) to facilitate therapeutic decision-making. More sensitive techniques have a higher probability of detecting ctDNA and could account for the differences in sensitivity and specificity observed across studies using the same biomarkers.

Several small studies investigating ctDNA as a prognostic marker in localized CRC have shown that ctDNA detection before surgery is associated with locoregionally advanced (stage III) disease, a high risk of recurrence, and poor prognosis (Table 4).^{24,67,80,81,84–89} In a cohort of 58 patients with primary CRC (stages I–IV), 39 patients with KRAS exon 2 mutations and/or p16 hypermethylation in the tumor underwent preoperative plasma screening. Among stage-I to stage-III patients, the 2-year recurrence-free survival rate differed according to whether ctDNA was detected in plasma (66%) or not (100%) (p = 0.044).⁹⁰ Similarly, the 2-year overall survival rate for patients with detectable ctDNA was 48%, compared to 100% for patients with no ctDNA (p < 0.03). Nevertheless, some studies have found no prognostic value for ctDNA detection before surgery.^{24,67,80,81,84–89,91} For example, in a recent study, KRAS, TP53, and p16INK4A mutations were used to detect ctDNA in 66 patients with CRC. Detection of mutant KRAS in the blood before surgery predicted recurrence (p < 0.01), but not overall survival.⁸⁰ Large prospective trials are needed to confirm the prognosis value of ctDNA.

Table 4.

Prognostic value of preoperative ctDNA in patients undergoing colorectal cancer surgical resection.

Reference	Gene(s)	Period of DNA extraction from plasma	Methods used for measurement of ctDNA	Patients (N and stage)	Treatment	Detection rate in plasma % (n plasma positive/n tumors positive)	Prognostic value
Wang et al.²⁴	APC, KRAS, and TP53	At surgery	PCR-SSCP	104 mostly stages I–III	Surgery ± chemotherapy	46% (n = 36/78)	Predictive of recurrence after surgery
Lee et al.⁶⁷	SEPT9 methylation	Before surgery and after surgery (n = 27)	Bisulfite conversion and real-time PCR of methylated SEPT9	101 mostly stages I–III	Surgery ± chemotherapy	37% (n = 37/101)	Predictive of disease-free survival
Bazan et al.⁸⁰	KRAS, TP53, and p16INK4a	Before surgery	PCR-SSCP	66 mostly stages I–III	Surgery ± chemotherapy	36% (n = 18/50)	KRAS predictive of earlier recurrence after surgery No association with overall survival
Trevisiol et al.⁸³	KRAS	Before surgery (stages I–IV)	Mutant-enriched semi-nested PCR	86	Surgery ± chemotherapy (48%)	100% (n = 7/7)	Association with overall survival (better in the absence of circulating KRAS-mutated ctDNA)
Frattini, et al.⁸¹	Total DNA, KRAS, and p16INK4a	Before and after surgery at 4, 10, and 16 months	Mutant-enriched PCR and F-MSP	70 mostly stages II–III	Surgery ± chemotherapy	100% (n = 18/18)	No association with disease recurrence
Hsieh et al.⁸⁴	KRAS, TP53, and APC	Before surgery	PCR-SSCP	118 stages I–III	Surgery ± chemotherapy	45% (n = 41/91)	Predictive of recurrence after surgery
Leung et al.⁸⁵	APC, hMLH1, and HLTF	Before surgery	Quantitative methylation-specific PCR	49 stages I–III	Surgery ± chemotherapy	57% (n = 28/49)	Association with overall survival (trend better in the absence of circulating methylated ctDNA)
Wallner et al.⁸⁶	HPP1/TPEF, HLTF, and MLH1	Before surgery	Quantitative methylation-specific PCR	104 mostly stages I–III	Surgery ± chemotherapy	Not available	Association with overall survival (significant better in the absence of circulating methylated HPP1 and/or HLTF ctDNA)
Nakayama et al.⁸⁷	p16INK4a	Before surgery	Quantitative methylation-specific PCR	21 stages I–III	Surgery ± chemotherapy	61% (n = 8/13)	Predictive of recurrence after surgery
Nishio et al.⁸⁸	RUNX3	Before surgery	Quantitative methylation-specific PCR	119 stages I–III (344 sera)	Surgery ± chemotherapy	29% (n = 100/344)	Predictive of recurrence after surgery
Ito et al.⁸⁹	KRAS and p16INK4a	Before surgery	PCR, mismatch ligation assay	90 stages I–III	Surgery ± chemotherapy	35% (n = 22/63)	None
Lecomte et al.⁹⁰	KRAS and p16INK4a	Before surgery	KRAS codon 12 or 13 mutation analysis by MASA; p16INK4a methylation status by methylation-specific PCR	58 stages I–IV	Surgery ± chemotherapy	67% (n = 26/39)	Predictive of recurrence and overall survival

F-MSP: fluorescent methylation-specific PCR; MASA: mutant allele-specific amplification; SSCP: single-strand conformational polymorphism; PCR: polymerase chain reaction; APC: adenomatous polyposis coli; hMLH1: human MutL homolog 1; HLTF: helicase-like transcription factor; HPP1/TPEF: hyperplastic polyposis 1/transmembrane protein containing epidermal growth factor and follistatin domains; ctDNA: circulating tumor DNA.

Future studies evaluating post-operative ctDNA levels could generate personalized markers based on the unique mutational profile of resected tumors, offering exquisitely high specificity to predict recurrence and/or evaluate prognosis. Sensitivity will rely on the ability to detect low levels of ctDNA released from micrometastatic deposits not detectable by imaging or other diagnostic modalities. Prospective trials are needed to evaluate whether early detection of recurrence using ctDNA can improve survival compared with standard imaging and whether adjuvant chemotherapy can prevent recurrence in patients with high pre- and/or post-operative ctDNA levels. To conclude, detection of ctDNA after surgery is very useful and highly associated with disease recurrence, but requires tumors with a mutated gene that can be detected in cfDNA. In contrast, the prognostic value of ctDNA remains unclear.

Monitoring tumor burden or response during treatment

Measuring ctDNA may be useful in monitoring tumor burden, a central aspect of patient management that is currently assessed using imaging and serum protein biomarkers, such as carcino-embryonic antigen (CEA). As the ctDNA profile can be absolutely specific for each patient’s individual tumor,⁹² its use would avoid false-positive and false-negative results when assessing response. Another benefit would be the real-time picture of tumor burden, taking into account that ctDNA has a half-life of less than 2 h, compared with commonly used protein biomarkers, which can take weeks to clear from the blood.²²

In the aforementioned work by Diehl et al.,²⁰ ctDNA was detectable in all patients before surgery, and serial blood sampling revealed oscillations in the ctDNA level that correlated with the extent of surgical resection. In this cohort, ctDNA levels were a more reliable and sensitive indicator (100%) than CEA (56%). Serial ctDNA measurements were likewise used to monitor tumor burden in six consecutive patients undergoing adjuvant treatment for CRC (stages I–IIIB).⁹³ Tumor burden measured by variant allele frequencies and absolute ctDNA concentration decreased markedly after surgery and adjuvant chemotherapy, with a mean ctDNA concentration of 0.163 and 0.038 ng/mL before and after surgery, respectively. The only patient with disease recurrence had a 13-fold increase in ctDNA levels. Danish researchers used NGS and droplet digital PCR to identify patient-specific somatic structural variants, which were then used as personalized ctDNA assays to monitor the disease burden after CRC surgery.⁹⁴ Their technique enabled efficient temporal assessment of disease status, with detection of metastatic recurrence at a mean of 10 months earlier than conventional follow-up. The authors concluded that ctDNA analysis is an exquisitely specific and highly sensitive approach for monitoring disease load.

Liquid biopsy also has potential for monitoring the response to treatment (Table 5).^95–97 An analysis of 31 mCRC patients demonstrated that ctDNA detection (KRAS mutation and RASSF2A methylation) was highly predictive of clinical outcome under chemotherapy.⁹⁵ Median progression-free survival (PFS) was 5 months in patients positive for ctDNA versus 14 months in patients negative for ctDNA (p = 0.004). Separately, Spindler et al.^96,97 developed a sensitive PCR method to quantitatively assess cfDNA as well as KRAS- and BRAF-mutated ctDNA in plasma. Serial plasma samples were obtained from 108 heavily pretreated patients before and during treatment with cetuximab and irinotecan. Low baseline cfDNA levels (<25% quartile) were associated with a higher disease control rate.⁹⁷ Patients with KRAS-mutant ctDNA levels in the highest quartile (>75% quartile) had a disease control rate of 0%, compared with 42% in patients with levels below the 75% quartile (p = 0.048). Decreases in cfDNA of >50% at cycle 3 compared to baseline were associated with treatment response (40% vs 17% in those with a lesser decrease, p = 0.04).⁹⁶

Table 5.

Use of ctDNA to monitor tumor burden and response during treatment of mCRC.

Reference	Gene(s)	Period of DNA extraction from plasma	Methods used for measurement of ctDNA	Patients (n)	Treatment	Detection rate in plasma % (n plasma positive/n tumors positive)	Predictive value
Lefebure et al.⁹⁵	KRAS and RASSF2A methylation	Before chemotherapy	Real-time PCR and methyl-specific PCR after bisulfite treatment	31	Chemotherapy	52% (n = 12/23)	Shorter progression-free survival in patients with positive ctDNA
Spindler et al.⁹⁶	Total cfDNA and KRAS	Before and during chemotherapy	Site-directed mutated PCR	73	Cetuximab plus irinotecan	-	Decrease of total cfDNA and KRAS-mutated ctDNA predict treatment response
Spindler et al.⁹⁷	Total cfDNA and KRAS	Before chemotherapy	Site-directed mutated PCR	108	Cetuximab plus irinotecan	78% (n = 32/41)	Low cfDNA is associated with higher disease control rate

PCR: polymerase chain reaction; ctDNA: circulating tumor DNA; mCRC: metastatic colorectal cancer; cfDNA: circulating cell-free DNA.

Besides enabling the earlier detection of progression, another potential advantage of using ctDNA to assess treatment response is the ability to minimize cumulative exposure of patients to ionizing radiation associated with surveillance imaging during long-term follow-up.⁹⁸ Conversely, potential difficulties using ctDNA to monitor response may be envisaged in patients treated with drugs that induce apoptosis and tumor necrosis, thus increasing the shedding of ctDNA into circulation. As with other tumor markers, such as CEA, an initial increase in ctDNA is not necessarily associated with tumor progression; rather, it is the evolution over time that will be predictive of response (initial increase and then secondary decrease). Efficient ctDNA monitoring has the potential to become a practice-changing tool. Changes in ctDNA may predict for treatment response early in the course of therapy, thereby creating a critical window of opportunity for intervention (i.e. treatment modification). Nevertheless, no prospective trial has yet demonstrated that detection of earlier disease progression with ctDNA, compared with imaging, and more rapidly change treatment results in improved survival.

Tracking resistance and tailoring therapies

CRC patients with wild-type RAS disease who initially respond to anti-EGFR treatment will ultimately acquire secondary resistance after a few months. The molecular basis of acquired resistance is an active area of research. Mutations linked to acquisition of anti-EGFR resistance have been reported in the EGFR extracellular domain (S492R), KRAS, NRAS, BRAF, and PIK3CA genes, as well as two novel EGFR ectodomain mutations (R451C and K467T).⁹⁹ Amplification of KRAS and co-occurring amplifications, such as MET and ERBB2, are additional mechanisms of resistance to anti-EGFR therapies.^100–102 In this setting, serial analysis of ctDNA provides a valuable opportunity for early detection of molecular changes that potentially confer resistance,^47,103–105 thereby enabling early treatment modification, which may prevent disease progression.

A cornerstone of ctDNA analysis in CRC has been the identification of acquired RAS mutations associated with secondary resistance to anti-EGFR therapy. Misale et al.¹⁰⁶ analyzed metastatic tissue from patients who developed resistance to cetuximab or panitumumab, demonstrating the emergence of KRAS amplification in one tumor and acquisition of secondary KRAS mutations in six of 10 cases analyzed. The onset of secondary KRAS mutations was detected in serum ctDNA analysis as early as 10 months before radiographic documentation of disease progression. A cohort of patients treated with chemotherapy alone was analyzed as a control group; no acquired KRAS mutations were found at disease progression in their serum samples.

Among patients who received third-line treatment with cetuximab and irinotecan, Spindler et al.⁹⁶ identified a discrepancy in mutational status between the primary tumor, baseline plasma samples, and/or plasma status at the time of progression in 19% of patients. In all, 12 patients had primary KRAS-mutant tumors but the mutation was not detected in the baseline plasma analysis, 2 patients had primary mutant disease and different types of mutations detected later in the plasma, and interestingly, 5 patients acquired new KRAS or BRAF mutations at progression.⁹⁶ In another study, Diaz et al.^107,108 found that KRAS mutations in serum were present during panitumumab treatment in 38% of patients which have initially WT tumors, which were detected on average 5 months before radiological progression. Recently, blood samples from mCRC patients included in the CORRECT trial were reanalyzed using BEAMing analysis. KRAS-mutated ctDNA was identified in 48% of patients who had received anti-EGFR therapy and whose archival tumor tissue DNA was KRAS wild type.¹⁰⁹

Bettegowda et al.²² designed a multiplexed, sequencing-based assay to query known mutated hotspots of several genes in the EGFR pathway that were not present in the primary tumor of patients receiving panitumumab monotherapy. They evaluated 24 mCRC patients’ progressing after achieving an objective response to anti-EGFR treatment; of these, 23 (96%) had up to 70 novel plasma mutations in genes involved in the mitogen-activated protein kinase pathway. Moreover, this assay predicted treatment resistance before radiologic progression occurred. Bardelli et al.¹⁰⁰ also demonstrated amplification of the MET locus in ctDNA before radiographic progression.

We can deduce some peculiar features of mutations arising under anti-EGFR pressure which are detectable in blood by ctDNA analysis. First, CRC tumors that acquire resistance to anti-EGFR therapy do not show the emergence of a single mutated clone, but typically the concomitant presence of diverse genetic abnormalities.¹⁰² It is not yet known whether these molecular changes are exclusively due to a passive clonal selection mechanism or involve treatment-induced mutagenesis. Second, there is a marked overlap between molecular abnormalities associated with primary resistance (i.e. RAS and BRAF mutations) and those detected by liquid biopsy eventually leading to secondary resistance. One exception is the EGFR mutation, which has only been described in secondary resistance.¹¹⁰ Nevertheless, the prevalence of KRAS or NRAS codon 61 mutations is higher in acquired resistance compared to untreated mCRC.^22,106

In conclusion, early detection of resistance to EGFR-targeted agents by liquid biopsy may help clinicians in deciding when to stop drugs and identifying driving mutations underlying resistance that can be targeted.⁹⁹ However, no prospective trial has yet demonstrated improved survival through earlier detection of mutations that confer anti-EGFR resistance with ctDNA, as compared with standard imaging. In addition, therapeutic options remain limited for patients with tumoral RAS pathway mutations. Plasma DNA can also indicate possible rechallenge strategies, that is, disappearance of RAS-mutated ctDNA. Anti-EGFR rechallenge has not yet been studied in a prospective trial.

Discussion

Analysis of ctDNA is a promising technology with the potential to improve the representation of all tumor lesions and the assessment of cancer heterogeneity, without the invasiveness of current biopsy methods. Studies suggest that liquid biopsy may detect minimal residual disease and tumor recurrence after curative surgery with greater sensitivity and specificity than current imaging and serum tumor markers. Liquid biopsies are also expected to be useful in establishing a more accurate prognosis and to greatly facilitate ongoing monitoring. The most attractive benefit of liquid biopsies is the ability to obtain a dynamic, ongoing real-time picture of the tumor genomic landscape in a given patient, providing the opportunity to tailor therapies throughout the disease course, from diagnosis to the development of resistance.

Many questions remain. Despite the proliferation of ctDNA publications, most involve very small patient numbers and none provide definitive evidence to guide the clinical application of ctDNA analysis (e.g. we do not know the relevance of changing anti-EGFR treatment when KRAS mutation appears in the blood versus at the time of radiological progression). Large prospective trials are needed to address clinically relevant questions, with a focus on using ctDNA to detect minimal residual disease and to monitor the emergence of molecular resistance in order to allow early adaptation of further treatment lines, as well as on measuring the impact of ctDNA analysis on disease-free, progression-free, and overall survivals.

It is essential to standardize the techniques used to analyze ctDNA in future trials, as results are dependent on the sensitivity of the technique used, with more sensitive sequencing techniques having a higher probability of detecting minority subclones. As in the tumor, the clinical relevance of minority subclones detected in the blood is unknown (e.g. the potential impact on resistance to anti-EGFR therapy). Previous studies suggested that patients with mCRC tumors containing no more than 1%–2% of KRAS-mutated subclones may benefit from anti-EGFR therapies.^111,112 However, no cut-off level for KRAS-mutated ctDNA that predicts resistance to anti-EGFR therapies has yet been established, thus, it is not known, for example, whether detection of 1% of cfDNA harboring a RAS mutation in a clinically responding wild-type RAS tumor should prompt a change in treatment.

Further contributing to uncertainty over when ctDNA findings support a change in the treatment strategy, it is not known whether some genomic alterations detected using ctDNA drive tumor progression or are merely “passenger” mutations. Targeting these alterations with new drugs may be not relevant if there are not driver mutations. It remains unclear exactly what the mutation profile of ctDNA reflects; it could reflect the profile of cells that are already dead, rather than those that are going to cause tumor progression. The extent to which ctDNA accurately reflects the heterogeneity of all tumor lesions is also unknown. Additional research is needed to elucidate why ctDNA levels vary among cancer patients, such as determining whether this variability is due to tumor biology and/or other parameters such as DNA clearance from plasma. Greater knowledge of the dynamics of ctDNA release is necessary before designing prospective clinical trials—for example, to understand the best timing for ctDNA assessment relative to therapy.

Exosomes are of great interest as an alternative source of circulating nucleic acids for liquid biopsy. Several studies have demonstrated that miRNA from circulating exosomes are similar to those of the originating cancer cells, suggesting a role for exosomal miRNA as potential biomarkers.¹¹³ Detection of exosomal miRNA may therefore offer a new diagnostic approach in CRC, as plasma miRNAs are promising markers for early detection of CRC tumors¹¹⁴ and can also predict CRC recurrence.¹¹⁵ Studies comparing ctDNA, free circulating miRNA, and exosomal miRNA are necessary to evaluate the performance of each marker.

When a promising new technology emerges, we tend to forget about existing technologies, including those that are currently under development. CTC analysis has a sensitivity of about 80% to detect CRC, while CTC levels correlate with both tumor stage^116–118 and recurrence after curative surgery.¹¹⁹ CTCs are also associated with treatment response and prognosis in mCRC.¹²⁰ However, in this indication, CTC data are less robust than those for ctDNA. To our knowledge, there have been no studies to date comparing CTCs and ctDNA in CRC. Of note, a study in multiple tumor types found no cases in which CTCs were detected and ctDNA was absent, whereas in 81% of cases in which ctDNA was detected, no CTCs were detectable.²² Demonstrating reproducibility of results comparing ctDNA and CTCs is a key challenge, and both techniques require standardized protocols and validation in large trials. Comparative studies are needed to establish the relative sensitivity of each technique and to determine whether combining both techniques is advantageous. It is likely that ctDNA will be superior to CTCs for diagnosis and detecting recurrence. Conversely, CTCs have advantages for analyzing treatment response and choosing targeted therapy since they are viable cells and can undergo additional testing, such as characterization of the current tumor phenotype using immunocytochemistry or fluorescence in situ hybridization (FISH). CTCs can also provide insight into tumor heterogeneity, and in breast cancer, CTCs were recently found to have dynamic interconverting phenotypes influenced by chemotherapy.¹²¹ Orthotopic CTC-derived tumor models allow the elucidation of signaling pathways involved in tumor progression and drug resistance.

Extensive data now support the potential of liquid biopsy as a novel technology with potential for optimizing CRC assessment, monitoring evolution, tracking resistance to targeted therapies, and designing tailored treatments. There is great interest in both ctDNA and CTCs, which may provide complementary information. Large, well-designed trials using standardized methodology are needed to compare the performance of ctDNA with CTCs and, in future, extracellular vesicles, in order to optimize the clinical application of each technique and complete the transition of blood-based assays from bench to bedside.

Footnotes

Acknowledgements

This review article was written on behalf of GI CONNECT; for more information, please visit . Editorial support was provided by Mark English, PhD, of COR2Ed. We also thank Gaëlle Tachon for her help to revise the manuscript.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

References

Pantel

Brakenhoff

. Dissecting the metastatic cascade. Nat Rev Cancer 2004; 4: 448–456.

Greaves

Maley

. Clonal evolution in cancer. Nature 2012; 481: 306–313.

Loeb

. Human cancers express mutator phenotypes: origin, consequences and targeting. Nat Rev Cancer 2011; 11: 450–457.

Klein

. Selection and adaptation during metastatic cancer progression. Nature 2013; 501: 365–372.

Tachiki

Kabeer

. Cancer genomic research at the crossroads: realizing the changing genetic landscape as intratumoral spatial and temporal heterogeneity becomes a confounding factor. Cancer Cell Int 2014; 14: 115.

Bedard

Hansen

Ratain

. Tumour heterogeneity in the clinic. Nature 2013; 501: 355–364.

Burrell

McGranahan

Bartek

. The causes and consequences of genetic heterogeneity in cancer evolution. Nature 2013; 501: 338–345.

Meacham

Morrison

. Tumour heterogeneity and cancer cell plasticity. Nature 2013; 501: 328–337.

Yang

Wang

Tian

. Application of circulating tumor cells scope technique on circulating tumor cell research. Mol Cell Ther 2014; 2: 8.

10.

National Comprehensive Cancer Network (NCCN). Clinical practice guidelines in oncology (NCCN Guidelines^®; Colon Cancer Version 2). Fort Washington, PA: NCCN, 2016

11.

Van Cutsem

Cervantes

Adam

. ESMO consensus guidelines for the management of patients with metastatic colorectal cancer. Ann Oncol 2016; 27: 1386–1422.

12.

Misale

Di Nicolantonio

Sartore-Bianchi

. Resistance to anti-EGFR therapy in colorectal cancer: from heterogeneity to convergent evolution. Cancer Discov 2014; 4: 1269–1280.

13.

Gazzaniga

Raimondi

Nicolazzo

. The rationale for liquid biopsy in colorectal cancer: a focus on circulating tumor cells. Expert Rev Mol Diagn 2015; 15: 925–932.

14.

Mandel

Metais

. Les acides nucléiques du plasma sanguin chez l’homme. C R Seances Soc Biol Fil 1948; 142: 241–243.

15.

Leon

Shapiro

Sklaroff

. Free DNA in the serum of cancer patients and the effect of therapy. Cancer Res 1977; 37: 646–650.

16.

Jahr

Hentze

Englisch

. DNA fragments in the blood plasma of cancer patients: quantitations and evidence for their origin from apoptotic and necrotic cells. Cancer Res 2001; 61: 1659–1665.

17.

Stroun

Lyautey

Lederrey

. About the possible origin and mechanism of circulating DNA apoptosis and active DNA release. Clin Chim Acta 2001; 313: 139–142.

18.

Fan

Blumenfeld

Chitkara

. Noninvasive diagnosis of fetal aneuploidy by shotgun sequencing DNA from maternal blood. Proc Natl Acad Sci U S A 2008; 105: 16266–16271.

19.

Diehl

Dressman

. Detection and quantification of mutations in the plasma of patients with colorectal tumors. Proc Natl Acad Sci U S A 2005; 102: 16368–16373.

20.

Diehl

Schmidt

Choti

. Circulating mutant DNA to assess tumor dynamics. Nat Med 2008; 14: 985–990.

21.

Kin

Kidess

Poultsides

. Colorectal cancer diagnostics: biomarkers, cell-free DNA, circulating tumor cells and defining heterogeneous populations by single-cell analysis. Expert Rev Mol Diagn 2013; 13: 581–599.

22.

Bettegowda

Sausen

Leary

. Detection of circulating tumor DNA in early- and late-stage human malignancies. Sci Transl Med 2014; 6: 224ra24.

23.

Gkountela

Szczerba

Donato

. Recent advances in the biology of human circulating tumour cells and metastasis. ESMO Open 2016; 1: e000078.

24.

Wang

Hsieh

Chang

. Molecular detection of APC, K-ras, and p53 mutations in the serum of colorectal cancer patients as circulating biomarkers. World J Surg 2004; 28: 721–726.

25.

Shaw

Smith

Walsh

. Microsatellite alterations plasma DNA of primary breast cancer patients. Clin Cancer Res 2000; 6: 1119–1124.

26.

Fujiwara

Fujimoto

Tabata

. Identification of epigenetic aberrant promoter methylation in serum DNA is useful for early detection of lung cancer. Clin Cancer Res 2005; 11: 1219–1225.

27.

Chan

Jiang

Zheng

. Cancer genome scanning in plasma: detection of tumor-associated copy number aberrations, single-nucleotide variants, and tumoral heterogeneity by massively parallel sequencing. Clin Chem 2013; 59: 211–224.

28.

Dawson

Tsui

Murtaza

. Analysis of circulating tumor DNA to monitor metastatic breast cancer. N Engl J Med 2013; 368: 1199–1209.

29.

Jing

Wang

Cui

. A sensitive method to quantify human cell-free circulating DNA in blood: relevance to myocardial infarction screening. Clin Biochem 2011; 44: 1074–1079.

30.

Shimony

Zahger

Gilutz

. Cell free DNA detected by a novel method in acute ST-elevation myocardial infarction patients. Acute Card Care 2010; 12: 109–111.

31.

Macher

Egea-Guerrero

Revuelto-Rey

. Role of early cell-free DNA levels decrease as a predictive marker of fatal outcome after severe traumatic brain injury. Clin Chim Acta 2012; 414: 12–17.

32.

Ohayon

Boyko

Saad

. Cell-free DNA as a marker for prediction of brain damage in traumatic brain injury in rats. J Neurotrauma 2012; 29: 261–267.

33.

Tsai

Lin

Chen

. The value of serial plasma nuclear and mitochondrial DNA levels in patients with acute ischemic stroke. Clin Chim Acta 2011; 412: 476–479.

34.

Holdhoff

Schmidt

Donehower

. Analysis of circulating tumor DNA to confirm somatic KRAS mutations. J Natl Cancer Inst 2009; 101: 1284–1285.

35.

Szpechcinski

Chorostowska-Wynimko

Struniawski

. Cell-free DNA levels in plasma of patients with non-small-cell lung cancer and inflammatory lung disease. Br J Cancer 2015; 113: 476–483.

36.

Cai

Jose

. Functional transferred DNA within extracellular vesicles. Exp Cell Res 2016; 349: 179–183.

37.

Gold

Cankovic

Furtado

. Do circulating tumor cells, exosomes, and circulating tumor nucleic acids have clinical utility? A report of the association for molecular pathology. J Mol Diagn 2015; 17: 209–224.

38.

Reclusa

Sirera

Araujo

. Exosomes genetic cargo in lung cancer: a truly Pandora’s box. Transl Lung Cancer Res 2016; 5: 483–491.

39.

Schwarzenbach

Nishida

Calin

. Clinical relevance of circulating cell-free microRNAs in cancer. Nat Rev Clin Oncol 2014; 11: 145–156.

40.

Uratani

Toiyama

Kitajima

. Diagnostic potential of cell-free and exosomal microRNAs in the identification of patients with high-risk colorectal adenomas. PLoS ONE 2016; 11: e0160722.

41.

Vogelstein

Kinzler

. Digital PCR. Proc Natl Acad Sci U S A 1999; 96: 9236–9241.

42.

Dressman

Yan

Traverso

. Transforming single DNA molecules into fluorescent magnetic particles for detection and enumeration of genetic variations. Proc Natl Acad Sci U S A 2003; 100: 8817–8822.

43.

Liu

Sommer

. Pyrophosphorolysis-activated polymerization (PAP): application to allele-specific amplification. Biotechniques 2000; 29: 1072–1076, 1078, 1080 passim.

44.

Madic

Piperno-Neumann

Servois

. Pyrophosphorolysis-activated polymerization detects circulating tumor DNA in metastatic uveal melanoma. Clin Cancer Res 2012; 18: 3934–3941.

45.

Forshew

Murtaza

Parkinson

. Noninvasive identification and monitoring of cancer mutations by targeted deep sequencing of plasma DNA. Sci Transl Med 2012; 4: 136ra68.

46.

Newman

Bratman

. An ultrasensitive method for quantitating circulating tumor DNA with broad patient coverage. Nat Med 2014; 20: 548–554.

47.

Kinde

Papadopoulos

. Detection and quantification of rare mutations with massively parallel sequencing. Proc Natl Acad Sci U S A 2011; 108: 9530–9535.

48.

Kou

Lam

Duan

. Benefits and challenges with applying unique molecular identifiers in next generation sequencing to detect low frequency mutations. PLoS ONE 2016; 11: e0146638.

49.

Leary

Sausen

Kinde

. Detection of chromosomal alterations in the circulation of cancer patients with whole-genome sequencing. Sci Transl Med 2012; 4: 162ra54.

50.

Murtaza

Dawson

Tsui

. Non-invasive analysis of acquired resistance to cancer therapy by sequencing of plasma DNA. Nature 2013; 497: 108–112.

51.

Pisanic

2nd Athamanolap

Poh

. DREAMing: a simple and ultrasensitive method for assessing intratumor epigenetic heterogeneity directly from liquid biopsies. Nucleic Acids Res 2015; 43: e154.

52.

Cancer Genome Atlas Network. Comprehensive molecular characterization of human colon and rectal cancer. Nature 2012; 487: 330–337.

53.

Douillard

Oliner

Siena

. Panitumumab-FOLFOX4 treatment and RAS mutations in colorectal cancer. N Engl J Med 2013; 369: 1023–1034.

54.

De Roock

Claes

Bernasconi

. Effects of KRAS, BRAF, NRAS, and PIK3CA mutations on the efficacy of cetuximab plus chemotherapy in chemotherapy-refractory metastatic colorectal cancer: a retrospective consortium analysis. Lancet Oncol 2010; 11: 753–762.

55.

Tol

Nagtegaal

Punt

. BRAF mutation in metastatic colorectal cancer. N Engl J Med 2009; 361: 98–99.

56.

Rachiglio

Abate

Sacco

. Limits and potential of targeted sequencing analysis of liquid biopsy in patients with lung and colon carcinoma. Oncotarget 2016; 7: 66595–66605.

57.

Taly

Pekin

Benhaim

. Multiplex picodroplet digital PCR to detect KRAS mutations in circulating DNA from the plasma of colorectal cancer patients. Clin Chem 2013; 59: 1722–1731.

58.

Lin

. Clinical relevance of alterations in quantity and quality of plasma DNA in colorectal cancer patients: based on the mutation spectra detected in primary tumors. Ann Surg Oncol 2014; 21(Suppl. 4): S680–S686.

59.

Song

. Current noninvasive tests for colorectal cancer screening: an overview of colorectal cancer screening tests. World J Gastrointest Oncol 2016; 8: 793–800.

60.

Warton

Mahon

Samimi

. Methylated circulating tumor DNA in blood: power in cancer prognosis and response. Endocr Relat Cancer 2016; 23: R157–R171.

61.

Cassinotti

Melson

Liggett

. DNA methylation patterns in blood of patients with colorectal cancer and adenomatous colorectal polyps. Int J Cancer 2012; 131: 1153–1157.

62.

Church

Wandell

Lofton-Day

. Prospective evaluation of methylated SEPT9 in plasma for detection of asymptomatic colorectal cancer. Gut 2014; 63: 317–325.

63.

deVos

Tetzner

Model

. Circulating methylated SEPT9 DNA in plasma is a biomarker for colorectal cancer. Clin Chem 2009; 55: 1337–1346.

64.

Grutzmann

Molnar

Pilarsky

. Sensitive detection of colorectal cancer in peripheral blood by septin 9 DNA methylation assay. PLoS ONE 2008; 3: e3759.

65.

Herbst

Rahmig

Stieber

. Methylation of NEUROG1 in serum is a sensitive marker for the detection of early colorectal cancer. Am J Gastroenterol 2011; 106: 1110–1118.

66.

Lee

Jung

. Aberrant methylation of APC, MGMT, RASSF2A, and Wif-1 genes in plasma as a biomarker for early detection of colorectal cancer. Clin Cancer Res 2009; 15: 6185–6191.

67.

Lee

Hwang

Kim

. Circulating methylated septin 9 nucleic acid in the plasma of patients with gastrointestinal cancer in the stomach and colon. Transl Oncol 2013; 6: 290–296.

68.

Lofton-Day

Model

Devos

. DNA methylation biomarkers for blood-based colorectal cancer screening. Clin Chem 2008; 54: 414–423.

69.

Warren

Xiong

Bunker

. Septin 9 methylated DNA is a sensitive and specific blood test for colorectal cancer. BMC Med 2011; 9: 133.

70.

Weirich

Erzberger

Barral

. The septin family of GTPases: architecture and dynamics. Nat Rev Mol Cell Biol 2008; 9: 478–489.

71.

Yan

Liu

. Diagnostic value of methylated septin9 for colorectal cancer screening: a meta-analysis. Med Sci Monit 2016; 22: 3409–3418.

72.

Imperiale

Ransohoff

Itzkowitz

. Multitarget stool DNA testing for colorectal-cancer screening. N Engl J Med 2014; 370: 1287–1297.

73.

Lam

Pan

Linnekamp

. DNA methylation based biomarkers in colorectal cancer: a systematic review. Biochim Biophys Acta 2016; 1866: 106–120.

74.

Rasmussen

Krarup

Sunesen

. Hypermethylated DNA as a biomarker for colorectal cancer: a systematic review. Colorectal Dis 2016; 18: 549–561.

75.

Pedersen

Symonds

Baker

. Evaluation of an assay for methylated BCAT1 and IKZF1 in plasma for detection of colorectal neoplasia. BMC Cancer 2015; 15: 654.

76.

Symonds

Pedersen

Baker

. A blood test for methylated BCAT1 and IKZF1 vs. a fecal immunochemical test for detection of colorectal neoplasia. Clin Transl Gastroenterol 2016; 7: e137.

77.

Young

Pedersen

Mansfield

. A cross-sectional study comparing a blood test for methylated BCAT1 and IKZF1 tumor-derived DNA with CEA for detection of recurrent colorectal cancer. Cancer Med 2016; 5: 2763–2772.

78.

Mouliere

Robert

Arnau Peyrotte

. High fragmentation characterizes tumour-derived circulating DNA. PLoS ONE 2011; 6: e23418.

79.

Ryan

Lefort

McManus

. A prospective study of circulating mutant KRAS2 in the serum of patients with colorectal neoplasia: strong prognostic indicator in postoperative follow up. Gut 2003; 52: 101–108.

80.

Bazan

Bruno

Augello

. Molecular detection of TP53, Ki-Ras and p16INK4A promoter methylation in plasma of patients with colorectal cancer and its association with prognosis. Results of a 3-year GOIM (Gruppo Oncologico dell’Italia Meridionale) prospective study. Ann Oncol 2006; 17(Suppl. 7): vii84–vii90.

81.

Frattini

Gallino

Signoroni

. Quantitative and qualitative characterization of plasma DNA identifies primary and recurrent colorectal cancer. Cancer Lett 2008; 263: 170–181.

82.

Tie

Wang

Tomasetti

. Circulating tumor DNA analysis detects minimal residual disease and predicts recurrence in patients with stage II colon cancer. Sci Transl Med 2016; 8: 346ra92.

83.

Trevisiol

Di Fabio

Nascimbeni

. Prognostic value of circulating KRAS2 gene mutations in colorectal cancer with distant metastases. Int J Biol Markers 2006; 21: 223–228.

84.

Hsieh

Lin

Chang

. APC, K-ras, and p53 gene mutations in colorectal cancer patients: correlation to clinicopathologic features and postoperative surveillance. Am Surg 2005; 71: 336–343.

85.

Leung

Man

. Quantitative detection of promoter hypermethylation in multiple genes in the serum of patients with colorectal cancer. Am J Gastroenterol 2005; 100: 2274–2279.

86.

Wallner

Herbst

Behrens

. Methylation of serum DNA is an independent prognostic marker in colorectal cancer. Clin Cancer Res 2006; 12: 7347–7352.

87.

Nakayama

Kodera

Ohashi

. p16INK4a methylation in serum as a follow-up marker for recurrence of colorectal cancer. Anticancer Res 2011; 31: 1643–1646.

88.

Nishio

Sakakura

Nagata

. RUNX3 promoter methylation in colorectal cancer: its relationship with microsatellite instability and its suitability as a novel serum tumor marker. Anticancer Res 2010; 30: 2673–2682.

89.

Ito

Hibi

Nakayama

. Detection of tumor DNA in serum of colorectal cancer patients. Jpn J Cancer Res 2002; 93: 1266–1299.

90.

Lecomte

Berger

Zinzindohoue

. Detection of free-circulating tumor-associated DNA in plasma of colorectal cancer patients and its association with prognosis. Int J Cancer 2002; 100: 542–548.

91.

Lin

. Clinical relevance of plasma DNA methylation in colorectal cancer patients identified by using a genome-wide high-resolution array. Ann Surg Oncol 2015; 22(Suppl. 3): S1419–S1427.

92.

Yong

. Cancer biomarkers: written in blood. Nature 2014; 511: 524–526.

93.

Zhou

Chang

Guan

. Application of circulating tumor DNA as a non-invasive tool for monitoring the progression of colorectal cancer. PLoS ONE 2016; 11: e0159708.

94.

Reinert

Scholer

Thomsen

. Analysis of circulating tumour DNA to monitor disease burden following colorectal cancer surgery. Gut 2016; 65: 625–634.

95.

Lefebure

Charbonnier

Di Fiore

. Prognostic value of circulating mutant DNA in unresectable metastatic colorectal cancer. Ann Surg 2010; 251: 275–280.

96.

Spindler

Pallisgaard

Andersen

. Changes in mutational status during third-line treatment for metastatic colorectal cancer—results of consecutive measurement of cell free DNA, KRAS and BRAF in the plasma. Int J Cancer 2014; 135: 2215–2222.

97.

Spindler

Pallisgaard

Vogelius

. Quantitative cell-free DNA, KRAS, and BRAF mutations in plasma from patients with metastatic colorectal cancer during treatment with cetuximab and irinotecan. Clin Cancer Res 2012; 18: 1177–1185.

98.

Chaudhuri

Binkley

Osmundson

. Predicting radiotherapy responses and treatment outcomes through analysis of circulating tumor DNA. Semin Radiat Oncol 2015; 25: 305–312.

99.

Arena

Bellosillo

Siravegna

. Emergence of multiple EGFR extracellular mutations during cetuximab treatment in colorectal cancer. Clin Cancer Res 2015; 21: 2157–2166.

100.

Bardelli

Corso

Bertotti

. Amplification of the MET receptor drives resistance to anti-EGFR therapies in colorectal cancer. Cancer Discov 2013; 3: 658–673.

101.

Mohan

Heitzer

Ulz

. Changes in colorectal carcinoma genomes under anti-EGFR therapy identified by whole-genome plasma DNA sequencing. PLoS Genet 2014; 10: e1004271.

102.

Misale

Arena

Lamba

. Blockade of EGFR and MEK intercepts heterogeneous mechanisms of acquired resistance to anti-EGFR therapies in colorectal cancer. Sci Transl Med 2014; 6: 224ra26.

103.

Destouni

Vrettou

Antonatos

. Cell-free DNA levels in acute myocardial infarction patients during hospitalization. Acta Cardiol 2009; 64: 51–57.

104.

Kim

Park

. Cell-free DNA level as a prognostic biomarker for epithelial ovarian cancer. Anticancer Res 2012; 32: 3467–3471.

105.

Salvianti

Pinzani

Verderio

. Multiparametric analysis of cell-free DNA in melanoma patients. PLoS ONE 2012; 7: e49843.

106.

Misale

Yaeger

Hobor

. Emergence of KRAS mutations and acquired resistance to anti-EGFR therapy in colorectal cancer. Nature 2012; 486: 532–536.

107.

Diaz

Jr Bardelli

. Liquid biopsies: genotyping circulating tumor DNA. J Clin Oncol 2014; 32: 579–586.

108.

Diaz

Jr Williams

. The molecular evolution of acquired resistance to targeted EGFR blockade in colorectal cancers. Nature 2012; 486: 537–540.

109.

Tabernero

Lenz

Siena

. Analysis of circulating DNA and protein biomarkers to predict the clinical activity of regorafenib and assess prognosis in patients with metastatic colorectal cancer: a retrospective, exploratory analysis of the CORRECT trial. Lancet Oncol 2015; 16: 937–948.

110.

Esposito

Rachiglio

La Porta

. The S492R EGFR ectodomain mutation is never detected in KRAS wild-type colorectal carcinoma before exposure to EGFR monoclonal antibodies. Cancer Biol Ther 2013; 14: 1143–1146.

111.

Laurent-Puig

Pekin

Normand

. Clinical relevance of KRAS-mutated subclones detected with picodroplet digital PCR in advanced colorectal cancer treated with anti-EGFR therapy. Clin Cancer Res 2015; 21: 1087–1097.

112.

Tougeron

Lecomte

Pages

. Effect of low-frequency KRAS mutations on the response to anti-EGFR therapy in metastatic colorectal cancer. Ann Oncol 2013; 24: 1267–1273.

113.

Nedaeinia

Manian

Jazayeri

. Circulating exosomes and exosomal microRNAs as biomarkers in gastrointestinal cancer. Cancer Gene Ther 2016; 24: 48–56.

114.

Huang

. Plasma microRNAs are promising novel biomarkers for early detection of colorectal cancer. Int J Cancer 2010; 127: 118–126.

115.

Matsumura

Sugimachi

Iinuma

. Exosomal microRNA in serum is a novel biomarker of recurrence in human colorectal cancer. Br J Cancer 2015; 113: 275–281.

116.

Wong

Chan

. Clinical significance of cytokeratin 20-positive circulating tumor cells detected by a refined immunomagnetic enrichment assay in colorectal cancer patients. Clin Cancer Res 2009; 15: 1005–1012.

117.

Sastre

Maestro

Puente

. Circulating tumor cells in colorectal cancer: correlation with clinical and pathological variables. Ann Oncol 2008; 19: 935–938.

118.

Cohen

Alpaugh

Gross

. Isolation and characterization of circulating tumor cells in patients with metastatic colorectal cancer. Clin Colorectal Cancer 2006; 6: 125–132.

119.

Cohen

Punt

Iannotti

. Relationship of circulating tumor cells to tumor response, progression-free survival, and overall survival in patients with metastatic colorectal cancer. J Clin Oncol 2008; 26: 3213–3221.

120.

Tol

Koopman

Miller

. Circulating tumour cells early predict progression-free and overall survival in advanced colorectal cancer patients treated with chemotherapy and targeted agents. Ann Oncol 2010; 21: 1006–1012.

121.

Jordan

Bardia

Wittner

. HER2 expression identifies dynamic functional states within circulating breast cancer cells. Nature 2016; 537: 102–106.

Future perspectives of circulating tumor DNA in colorectal cancer

Abstract

Keywords

Introduction

Liquid biopsy using ctDNA in blood

Potential clinical applications of liquid biopsy in colorectal cancer

Cancer diagnosis and screening

Detection of tumor recurrence, minimal residual disease, and prognosis value

Monitoring tumor burden or response during treatment

Tracking resistance and tailoring therapies

Discussion

Footnotes

Acknowledgements

Declaration of conflicting interests

Funding

References