Liquid biopsy for colorectal cancer could detect tumor DNA in the blood and urine

A new study from Washington University School of Medicine in St. Louis demonstrates that a liquid biopsy examining blood or urine can help gauge the effectiveness of therapy for colorectal cancer that has just begun to spread beyond the original tumor.

Such a biopsy can detect lingering disease and could serve as a guide for deciding whether a patient should undergo further treatments due to some tumor cells evading an initial attempt to eradicate the cancer.

The study appears online Feb. 12 in the Journal of Clinical Oncology Precision Oncology, a journal of the American Society of Clinical Oncology.

While a few liquid biopsies have been approved by the Food and Drug Administration, mostly for lung, breast, ovarian and prostate cancers, none has been approved for colorectal cancer.

Patients in this study had what is known as oligometastatic colorectal cancer, meaning each patient’s cancers had spread beyond his or her original tumor but only to a small number of sites. Such patients undergo chemotherapy to shrink the tumors before having surgery to remove whatever remains of the primary tumor.

There is debate in the field about whether, after initial therapy, oligometastatic cancer should be treated like metastatic cancer, with more chemotherapy – or like localized cancer, with more surgery plus radiation at those limited sites.

Contributing to the problem is that doctors have a limited ability to predict how patients will respond to early chemotherapy, especially since most patients don’t have access to cancer genome sequencing to identify the DNA mutations in their original tumors.

“Being able to measure response to early chemotherapy without prior knowledge of the tumor’s mutations is a novel idea and important for being able to determine whether the patient responded well to the therapy,” said senior author Aadel A. Chaudhuri, MD, Ph.D., an assistant professor of radiation oncology.

“This can provide guidance on how to treat oligometastatic disease. For example, if the liquid biopsy indicates that a patient responded well to the early chemotherapy, perhaps they should be offered the possibility of more surgery, which could potentially cure their disease.

But if they didn’t respond well, it’s likely the cancer is too widespread and can’t be eradicated with surgery, so those patients should receive more chemotherapy to control their disease.”

Liquid biopsies for colorectal cancer detect tumor DNA that has broken free of the cancer and is circulating in the blood and, to a lesser extent, has collected in the urine.

The biopsies described in this study are unique compared with other liquid biopsies being developed for colorectal cancer in three major ways.

First, most such biopsies have been developed to track metastatic cancers or to verify that local cancers have not started to spread. Second, most liquid biopsies for cancer rely on knowledge of the original tumor’s mutations, to see if those mutations are still present in the blood after therapy.

But many patients don’t get the opportunity to have their original tumors sequenced. Instead, the new biopsies rely on detecting DNA mutations in the blood or urine and comparing them with DNA mutations measured in the treated primary tumor, after it’s surgically removed.

And finally, the urine biopsy is unique for colorectal cancer as most urine biopsies have been limited to use in cancers of the genitourinary system, especially bladder cancer.

“The levels of circulating tumor DNA that we were able to measure in urine were lower than what we measured in blood, but this is still a proof of concept that it is possible to measure residual disease in a nonurinary cancer in this totally noninvasive way,” said Chaudhuri, who also treats patients at Siteman Cancer Center at Barnes-Jewish Hospital and Washington University School of Medicine. “We will need to develop more sensitive techniques to detect colorectal tumor DNA in urine to make this a useful clinical test. But this is a promising start.”

The study showed that lower circulating tumor DNA levels correlated with better responses to early chemotherapy. Indeed, most patients who had undetectable levels of tumor DNA in blood samples also had no measurable cancer in their surgical specimens.

There was also evidence that the residual disease detected in liquid biopsies was more predictive of outcomes than residual disease found in the surgical specimens. For example, the researchers described the experience of one man who, after early chemotherapy to shrink or eliminate the tumor, still had detectable cancer removed during surgery. But his blood sample taken that same day showed no circulating tumor DNA.

He experienced long-term survival with no cancer recurrence. On the other hand, a woman with no detectable cancer cells in her surgical specimen, removed after early chemotherapy, was found to have circulating tumor DNA in her same-day blood sample. Eight months later, the cancer returned in her liver.

The study suggests that such liquid biopsies could help personalize treatment for oligometastatic colorectal cancer. Beyond identifying patients at high risk of recurrence and helping guide decisions about which traditional therapies should be given, the new study also identified patients who might benefit from immune therapies and other targeted treatments.

“Based on mutations in the blood biopsy, we could identify patients who might benefit from a type of immune therapy called immune checkpoint inhibitors after their initial therapy is complete,” Chaudhuri said. “We also found mutations that could be targeted with drugs approved for other cancers. Our current study is observational, but it paves the way for designing future clinical trials that could test some of these potential therapies.”

---

Colorectal carcinoma (CRC) is one of the most diagnosed cancers in the world and the second leading cause of cancer related deaths [1]. In high-income countries, or in countries with accessible health care, there are observable stabilizing trends in the incidence and mortality rates of CRC, but overall rates are still one of the highest [2].

Interestingly, adults below 50 years of age are the exception, where the incidence of CRC has increased. In many low-income and middle-income countries, there are distinguishing patterns indicating a rising incidence and mortality rate of CRC [3]. Interestingly, in a projection of global trends in CRC to the year 2035, colon cancer and rectal cancer mortality rates were predicted to decline.

However, ongoing demographic changes (population growth and ageing) may lead to a rise in the number of deaths in many countries, with a doubling of the number of predicted deaths by 2035 in some regions [4]. Furthermore, CRC causes a financial strain to a significant number of patients (~40%), which results in a lower quality of life [5]. Overall, CRC can be defined as one of the greatest challenges to public and global health in the present and most likely in the future.

Colorectal carcinoma (CRC) is often diagnosed in late stage due to nonspecific symptoms, such as a change in bowel movement, weight loss, abdominal pain, iron deficiency, anemia, or rectal bleeding [6]. The gold standard for detection of CRC is currently colonoscopy [7].

Furthermore, CRC is clinically categorized by anatomical location as right CRC (RCC) or left CRC (LCC). RCC is defined as the proximal two-thirds of the transverse colon, ascending colon, and caecum [8]. LCC includes the distal third of the transverse colon, splenic flexure, descending colon, sigmoid colon, and rectum.

In general, there is a higher incidence of RCC among older patients (Figure 1) [9,10]. Studies comparing screening with and without colonoscopy found a statistically insignificant difference between LCC and RCC [11]. The main concern with RCC is that the right colon has a wider lumen and more frequently flat tumor growths which lead to a longer period without clinical symptoms. Subsequently, this results in a greater time to disease detection and start of treatment [12].

The incidence of stage IV cancer with less differentiated cells is also higher in RCC than LCC [9]. With respect to molecular pathways, the same frequency of the oncogenes Kirsten rat sarcoma viral oncogene homolog (KRAS)and neuroblastoma rat sarcoma viral oncogene homolog (NRAS)are seen in LCC and RCC, but the rate of v-Raf murine sarcoma viral oncogene homolog B (BRAF) mutation has been shown to be significantly higher in RCC [9]. Taken together, RCC is associated with a higher risk of poor prognosis than LCC, despite being classified as the same primary cancer.

Colorectal carcinoma (CRC) is characteristic for wide intratumor heterogeneity and general genomic instability, which impacts the treatment and quality of life of the patient [13]. Accumulation of somatic mutations, which is associated with CRC tumor progression, can be explained with molecular changes that add to genomic instability.

Specifically, there are three major molecular pathways in CRC that produce these mutations: chromosomal instability (CIN) [14], microsatellite instability (MSI) [15], and CpG island methylator phenotype (CIMP) [16,17,18]. CIN, as a consequence of improper mitosis and spindle checkpoint activity, promotes tumor progression by increasing the rate of genetic aberrations [19] and it is observed in the majority of sporadic CRC (85%) [20].

MSI is caused by the inactivity of the DNA mismatch repair (MMR) [21] and can be detected in approximately 15% of CRCs [22] and gives the disease distinctive pathological features. Tumors that are positive for MSI tend to be focal, poorly differentiated [23], right-side located, and are associated with production of extracellular mucin [24,25].

CIN and MSI are not mutually exclusive and MSI tumors can show evidence of CIN [26]. MSI is also associated with hereditary non-polyposis colon cancer (HNPCC), the common form of hereditary CRC. The majority of HNPCC-associated mutations affect two crucial genes for MMR, which are MSH2 and MLH1 [27].

Lastly, CIMP positive tumors are defined by transcriptional inactivation by DNA methylation at promoter CpG islands of tumor suppressor genes. CpG islands are regions of the DNA sequence, typically associated with transcriptional promoters, containing a high density of the dinucleotide sequence cytosine followed by guanine in the 5′ to 3′ direction.

These tumors have a strong positive correlation to BRAF mutation [18,28]. However, regardless of MSI or BRAF status, CIMP tumors are associated with a significantly lower mortality rate. The relation between CIMP positivity and lower mortality is consistent across all stages of CRC [29].

Specific clones of CpG islands are methylated exclusively in CRC, which has not been observed in normal colon tissue [30]. Undoubtedly, CRC is a well-researched disease from a genomic point of view. Coupled with information about key driver genes, the information creates an ideal background for further research and subsequently utilization of the liquid biopsy in clinical settings.

There are six key driver genes in CRC [31], being APC, TP53, KRAS, BRAF, PIK3CA, and SMAD4, with the TP53 alteration being selectively enriched in the metastatic setting [32]. There are also several signaling cascades which can be affected. The most common is the WNT pathway occurring with most mutated APC genes [33,34].

Overall, WNT pathway alteration occurs in 93% of MSI positive and 85% of MSI negative CRC [32]. The combination of mutations in the WNT pathway and in the signaling pathway associated with KRAS mutation are crucial for tumor progression in CRC [35,36].

A KRAS mutation (and BRAF mutation) drives tumorigenesis through constitutive activation of the MAPK pathway [37] and can induce hyperproliferation in colonic epithelium, but only in combination with a mutation from the WNT pathway [38]. KRAS mutations are also associated with the PI3K pathway [39], which plays a fundamental role in the tumor–host interaction, enhancing tumor-induced angiogenesis and facilitating the establishment of metastatic colonies [40].

BRAF mutations are present in 5–13% [41,42,43] of sporadic CRC (geographical variation may account for differences in the occurrence [43]) and are significantly associated with a higher metastatic rate and worse overall survival (OS) [44]. Subsequently, BRAF mutation can be used as a prognostic factor and in clinical decision-making regarding targeted therapy [41,42].

Additionally, since BRAF is a downstream molecule of KRAS [43], concomitant KRAS and BRAF mutations are rare and could be considered mutually exclusive [45]. Correspondingly, combined mutation of PIK3CA and TP53 is correlated with a shorter OS of stage II/III CRC patients receiving 5-fluorouracil-based therapy [46].

Colorectal carcinoma (CRC) is a malignant disease with severe impact on the general population. Genomic information of CRC implies the challenge of intratumor heterogeneity and related poor outcomes during treatment. Extensive information presented in this review will demonstrate the promising opportunity for the liquid biopsy to improve prognosis and patients’ quality of life.

Liquid Biopsy for CRC

The liquid biopsy is generally referred to as the analysis of tumor-derived material from peripheral blood. However, in the general sense, this can also include sampling from other bodily fluids such as urine [47], stool [48], cerebrospinal fluid [49], saliva [50], pleural fluid [51], and ascites [52]. In this review, circulating tumor cells (CTCs) and cell-free DNA (cfDNA) derived from peripheral blood will be the focus of this discussion.

Circulating Tumor Cells

Circulating tumor cells (CTCs) may be detectable in the cellular fraction of patient peripheral blood. CTCs are defined as tumor cells in circulation coming from the primary tumor or from a metastatic site [53]. On average, CTCs measure between 15 and 25 µm, which is generally larger [54,55] than white blood cells (6–10 µm) [56], and have a distinct morphological shape of the nucleus [57].

Although CTCs were discovered 150 years ago [58], it was not until the development of novel detection methods that they became the focus of new prognostic and diagnostic approaches. Individual CTCs can be characterized by their morphology [59], phenotype, and genotype [60]. Genomic information from CTCs may be used to better understand cancer cell biology, provide confidence around diagnostic [61] or treatment decisions [62], and to track treatment response [63].

Genomic information also allows for the assessment of clonality and the characterization of different cell populations [64,65]. CTCs can be detected in the peripheral blood as single cells or as cell clusters [66]. Steeg hypothesized that tumor cells invade the tissue surrounding the primary tumor, enter the lymphatics or the bloodstream, circulate in the body, then extravasate into a tissue, and form a secondary tumor at the new location [67].

During this metastatic process, CTCs can interact with immune cells [68]. For example, CTCs that express programmed death-ligand 1 (PD-L1) [69] may interact with circulating immune cells and have been associated with an increased two-year mortality risk [70].

Cell-Free DNA

The discovery of free DNA in blood dates back to 1948 by Mandel and Metais [71]. Typically, cfDNA in healthy individuals derives from the normal turnover of cells through apoptosis [72] or necrosis [72], or active release from lymphocytes [73,74,75] and can be increased by such events as infection or inflammation [76].

In cancer patients, the level of cfDNA can be increased through the turnover of tumor cells. This component can be distinguished from normal cfDNA by the presence of tumor-associated mutations and is often categorized separately as circulating tumor DNA or ctDNA. Although cfDNA can be found as fragments from 150 to 10,000 bp, the vast majority is found at a peak of 166 bp, corresponding to the size a single nucleosome, as is typically released during apoptosis [77,78].

Peripheral blood of a healthy individual contains only limited amount of cfDNA (up to 100 ng/mL) [79]. In contrast, the patients with metastatic CRC have been shown to have significantly higher levels of cfDNA in blood (up to 209 ng/mL) [79]. Recently, the same relationship was shown again in which patients with primary CRC had significantly higher levels of cfDNA than patients with intestinal polyps and healthy controls [80].

Analysis of cfDNA in CRC can be used for testing KRAS and BRAF mutations. In the case of KRAS mutation, cfDNA analysis may identify patients with a worse prognosis who may benefit from a more aggressive chemotherapy regimen [81]. Additionally, analysis of the BRAF mutation can be used as a negative prognostic biomarker for OS [82]. Likewise, the detection of BRAF mutation in cfDNA may have a role in treatment selection for patients [83].

Liquid Biopsy Platforms
Detection of Circulating Tumor Cells

The platforms used for detection and collection of CTCs vary widely in their technology and methodology. They can be separated into four different categories based on their technique of enrichment: immunocapture, physical characteristics (i.e., size and density), non-enrichment based, or a combination.

To embrace all of the possibilities of how CTCs can be utilized in clinical routine, it is necessary to understand the advantages and potential biases of those different technologies. Immunocapturing is one of the most common used techniques of detection and is based on targeting specific antigen(s) on the surface of cells.

A positive enrichment immunocapturing approach, often based on the epithelial cell adhesion molecule (EpCAM) antigen, targets the cells of interest. On the other hand, negative enrichment, typically based on the CD45 antigen, targets the removal of other cell populations, typically lymphocytes, and thereby enriches the cells of interest. In either case, only cells displaying these antigens are detected and consequently either collected or discarded.

Techniques based on physical properties of the cells are utilizing the size and or density difference between CTCs and other non-rare cells. As with immunocapturing, only cells with certain physical properties are collected. For this reason, both techniques carry inseparable bias. Non-enrichment techniques do not exclude cells from a sample.

This approach presents the possibility of avoiding potential bias during the identification process linked to the physical or biological properties of cells. Consequently, non-enrichment techniques can be more labor-intensive than the other methods. Further discussion about possible limitations of the liquid biopsy can be found in Section 5, Challenges.

A list of some current platforms is provided in Table 1. However, the field of liquid biopsy is rapidly developing and therefore it is not our intention to provide an exhaustive list of all platforms. In addition, brief details of three different approaches to detect CTCs can be found below. The specific platforms which are discussed were selected as an example of each enrichment technique to illustrate their practical use.

Table 1

Platforms used for the detection of circulating tumor cells (CTCs) in the peripheral blood. Three discussed platforms in text are marked with * EpCAM, epithelial cellular adhesion molecule; CD45, cluster of differentiation.

Platform	Company/Institution Details	Description	References
AdnaTest	Qiagen GmbH, Hilden, Germany	Antibody targeting EpCAM conjugated to magnetic beads for labeling tumor cells in sample	[100,101]
CanPatrol™ CTC	SurExam, Guangzhou, China	Filtration (and CD45+ depletion) 1	[102,103]
CellCollector®	Gilupi GmbH, Postdam, Germany	Nano guidewire inserted into patient cubital vein collecting cell expressing EpCAM	[104,105]
CellMax CMx	CellMax Life Inc., Sunnyvale, CA, USA	Blood passing through antibody-coated microfluidic chip targeting EpCAM	[106,107]
CellSearch® *	Menarini Silicon Biosystems Spa, Castel Maggiore, Italy	Antibody targeting EpCAM conjugated to magnetic beads for labeling tumor cells in sample	[108,109]
ClearCell® FX	Genomax Technologies, Bangkok, Thailand	Blood passing through microfluidic biochip with larger cells along the inner wall	[110,111]
Cytelligen®	Cytelligen Inc., San Diego, CA, USA	Antibody targeting CD45 conjugated to magnetic beads for labeling tumor cells in the blood sample	[112,113]
DEPArray™	Menarini Silicon Biosystems Spa, Castel Maggiore, Italy	Cell suspension loaded into microchip-based sorter using dielectrophoresis to trap cells	[114,115]
Easysep™	Stemcell Technologies Inc., Vancouver, BC, Canada	Antibody targeting EpCAM or CD45 conjugated to magnetic beads for labeling tumor cells in the blood sample	[116,117]
Epic Sciences/HDSCA *	Epic Sciences Inc., San Diego, CA, USA	After processing, cells are plated on the slide and subsequently characterized based on surface markers	[57,92]
Herringbone Chip	Massachusetts General Hospital, Boston, MA, USA	Blood processed through antibody-coated microfluidic chip targeting EpCAM	[118,119]
ISET® *	Rarecells Diagnostics SAS, Paris, France	Filtration on pressure-controlled system	[120,121]
MagSweeper™	Stanford University, Stanford, CA, USA	Antibody targeting EpCAM or CD133 conjugated to magnetic beads for labeling tumor cells in blood	[122,123]
MetaCell®	MetaCell s.r.o., Ostrava, Czech Republic	Capillary-action driven size-based separation	[124,125]
Oncoquick®	Greiner Bio-One International GmbH, Kremsmünster, Austria	The denser blood compartment migrates through the porous barrier of the polypropylene centrifugation tube	[126,127]

The CellSearch® platform (https://www.cellsearchctc.com/) is based on immunomagnetic enrichment of cells expressing EpCAM from peripheral blood samples. The circulating epithelial cells are selected for using immunomagnetic beads targeting EpCAM.

The enriched sample is then stained with 4′,6-diamidino-2-phenylindole (DAPI) and immunofluorescently labeled with a monoclonal antibody for CD45 to identify leukocytes, and with antibodies against cytokeratins (CK) 8, 18, 19 to identify epithelial cells. CTCs are then defined as CD45 negative and CK positive [84,85,86].

CellSearch® was the first approved CTC detection and capture platform by the U.S. Food and Drug Administration (FDA) for prognostics in cancer treatment and has been allowed for clinical use in breast [87], colorectal [85,88], and prostate cancer [86,89]. A prospective study using CellSearch® analyzed the dynamic change of CTCs between pre- and post-surgery 7.5 mL blood samples collected in a cohort of 44 metastatic CRC patients (43 pre-surgery and 38 post-surgery samples).

Results showed that all patients with pre-surgery positive samples (two or more CTCs per sample) relapsed. In contrast, 65% of patients with pre-surgery negative samples relapsed. Similarly, 68% and 85% of patients with post-surgery negative and positive samples, respectively relapsed. The study also showed significant difference in median OS between patients with positive and negative pre-surgery samples with 17 and 69 months, respectively [90].

Epic Sciences/High-Definition Single-Cell Assay

The high-definition single-cell assay (HD-SCA) workflow is a non-enrichment method developed for high-resolution characterization of CTCs, which provides intact cells enabling downstream single cell molecular characterization [57,91,92,93]. Biospecimens are processed within 24 to 48 h after collection, which involves erythrocyte lysis and centrifugation.

The HD-SCA workflow is a “no cell left behind”™ approach in which all nucleated cells from the blood sample are plated on custom glass slides, approximately 3 million nucleated cells per slide. Slides are subsequently stained with DAPI to identify the nuclei and antibodies against pan CK and CD45.

There is the possibility of adding a 4th channel for further characterization of the cellular populations. This approach allows for the functional profiling of all nucleated cells from a patient’s blood without enrichment bias [92,94] based on their morphological features, biomarker expression, and nuclear integrity [95].

A study using this platform, with a primary goal of establishing concordance among morphology of liquid and solid biopsy cells, gave substantial evidence by analyzing more than 1000 single cells. Out of 43 metastatic CRC patients, 15 patients had more than 4 HD-CTC per mL (CK positive, CD45 negative, morphologically distinct cells with a nuclear size larger than surrounding white blood cells) (35% positivity) in their pre-surgery blood. Additionally, higher concentration of HD-CTC was observed by patients with necrotic hepatic tumors.

The level of HD-CTCs also decreases in post-surgery blood of patients. The results demonstrated that liquid biopsy cells are associated with solid tumor cells in metastatic CRC and the relationship can be potentially used for diagnostics or monitoring disease burden over time [92].

The Isolation by Size of Epithelial Tumor Cells®

The isolation by size of epithelial tumor cells (ISET®, https://www.isetbyrarecells.com/iset-story/) platform is based on size enrichment, which operates on the presumption that tumor cells derived from carcinomas have a significantly larger size compared to peripheral blood leukocytes [96]. Blood samples are filtered through a module with up to 12 wells, where each well contains a 0.6 cm diameter membrane [96] with 8 μm diameter cylindrical pores [97].

Each well can then be further analyzed with immunohistochemical characterization of cells isolated by ISET® [59], fluorescence in situ hybridization, or genomic sequencing [98]. A study was conducted to establish correlation between ISET® CTC count and both progression-free survival (PFS) and OS in patients with metastatic CRC. Three serial collections were analyzed via ISET® and the CTC number varied from 0 to 46 cells per sample. Based on the variation in number of CTCs found between the collections, it suggests possible risk stratification [99].

cfDNA Analysis

Measurement of cfDNA, and subsequently ctDNA (circulating tumor DNA) detection, can be challenging due to low concentration, especially in early stages of cancer [128]. Generally, ctDNA can be analyzed with a focus on tumor-specific mutations [129], genome wide analysis for copy number alterations (CNAs) [130,131], point mutations by whole-genome sequencing (WGS), or whole exome sequencing (WES) [132].

In addition, CNAs are theoretically easier to detect than point mutations or epigenetic changes in cfDNA [133], because CNA analysis requires only sparse sequence coverage [131]. cfDNA-based CNAs show clinical validity as promising biomarkers for cancer diagnosis and prognosis, especially for late-stage cancers [134].

An important aspect in cfDNA targeted mutational analysis is mutant allele fraction (MAF). MAF is defined as a ratio between mutant alleles and all targeted alleles in a sample. Thus, sensitivity of MAF is a significant characteristic for platforms detecting ctDNA. In a case of cfDNA analysis, targeted methods are more suitable and applicable in the clinical setting. Classical quantitative PCR (qPCR) methods have sensitivity of between 10–20%.

Advanced targeted methods for detecting MAF in cfDNA are digital PCR (dPCR), which include droplet digital PCR (ddPCR) [135,136] and BEAMing (beads, emulsions, amplification, and magnetics) [137]. Studies that used BEAMing were able to detect MAF of 0.1% [138,139]. Detection sensitivity of ddPCR assays is 0.04% [140].

Furthermore, ddPCR is not solely used for point mutation detection, but is able to reveal indels and frequently observed cancer-related CNA [141]. Through PCR enrichment, with suitably short amplicons, amplicon-based sequencing can achieve sensitivity comparable to ddPCR [142].

Tagged-amplicon deep sequencing (TAm-Seq) technology and its enhanced version could detect MAF of 0.14% and 0.02%, respectively [142,143]. Another targeted sequencing method able to detect cfDNA in small amounts is duplex sequencing with unique molecular identifiers, which can distinguish MAF of 0.1%.

The duplex method exploits the knowledge of complementarity of both DNA strands. Adapters tag duplex DNA, and afterwards, it is possible to compare the strand with its counterpart. Mutation originating from the tumor must be present in both strands [144].

Another targeted sequencing method is cancer personalized profiling by deep sequencing (CAPP-Seq) [145] or safe-sequencing system (Safe-SeqS) [146]. CAPP-Seq combines identifying recurrent mutated regions in given cancer type via bioinformatic methods and sequencing with the objective to improve sensitivity and approximate patient-orientated approach [147].

CAPP and its enhanced methods are able to detect MAF of 0.2% and 0.04%, respectively [148]. Safe-SeqS is similar to duplex sequencing by assigning an unique molecular identifier. In contrast to duplex sequencing, these identifiers tag each template molecule. Safe-SeqS is able to detect MAF of 0.01% [149].

On the other hand, non-targeted methods for cfDNA analysis may harbor some advantages. They do not depend upon specific knowledge about the primary tumor, neither genetic nor epigenetic. However, genome-wide sequencing methods, like whole-genome sequencing or whole-exome sequencing, require a relatively high fraction of ctDNA/cfDNA for detection (5–10% MAF) [150]. For this reason, these methods are better suited for research than clinical routine.

There are some commercial cfDNA assays which can be used for clinical applications in CRC. In 2016, the FDA granted premarket approval for the first cfDNA-based liquid biopsy for cancer screening, Epigenomics’ Epi proColon® assay, using qPCR. This assay is based on the presence of aberrantly methylated SEPT9 DNA in the plasma and has shown promising results [151].

Out of 50 untreated CRC patients prior to surgery, the assay was able to detect positive 45 patients (90% sensitivity) for presence of methylated SEPT9 DNA. More interestingly, the assay was able to detect 33 out of 38 (87% sensitivity) patients with early stage disease (stages I and II) [152]. Since then, another cfDNA assay, Signatera™, received breakthrough device designation by the FDA.

Another commercial platform focused on CRC is the Idylla™, also using qPCR assays for KRAS, [153], NRAS, and BRAF mutations [153] or characterization of MSI [154]. The platform OncoBeam™, using BEAMing (beads, emulsification, amplification, and magnetics), offers panels for the detection of various CRC specific mutations, including KRAS, NRAS, BRAF, and PIK3CA [155].

Another interesting approach was taken by Wan et al. from Freenome, which relied on a machine learning approach using cfDNA from 546 CRC patients, the majority of which were in stage I and II (81%), to develop a computational approach to identify relation between the cfDNA profile and CRC stage. Blood samples were collected at unspecified timepoints. A mean sensitivity of 85% was achieved, showing promising preliminary performance for detection of CRC in early stages. [156].

reference link: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7352156/

---

More information: Pellini et al. ctDNA MRD detection and personalized oncogenomic analysis in oligometastatic colorectal cancer from plasma and urine. JCO Precision Oncology. Feb. 12, 2021.