ctDNA: detection, prognosis and progression monitoring in CRC
cfDNA is an emerging potential biomarker for guiding precision therapy in CRC [
59,
60]. In general, cfDNA analysis is used to discover point mutations or structural variants, copy-number aberrations, microsatellite alterations, differential cfDNA length, and methylation status [
61]. ctDNA arises from somatic tumor DNA fragments released into the blood circulation during cell death and has been found to contain tumor-specific molecular characteristics [
62]. In a recent prospective, multicenter cohort study, plasma ctDNA was analyzed in patients with stage I to III CRC. ctDNA was detectable in 108 of 122 patients (88.5%). Furthermore, they found that ctDNA-positive patients were 7 times more likely to relapse than ctDNA-negative patients at postoperative Day 30. Similarly, ctDNA-positive patients may relapse after adjuvant chemotherapy (ACT). After regular therapy, ctDNA-positive patients were more than 40 times more likely to suffer cancer relapse during surveillance than ctDNA-negative patients (HR, 43.5; 95% CI, 9.8–193.5
P < 0.001). It is believed that ctDNA may be useful for risk classification, ACT surveillance, and early recurrence detection in CRC [
63]. The results of this article are consistent with many previous studies [
64‐
67]. Distant metastasis is one of the reasons for the poor prognosis of CRC. Relatively high levels of ctDNA in plasma in CRLM patients are known [
68]. A single-center retrospective study was performed to investigate the impact of ctDNA on the OS of patients who underwent initial hepatectomy for resectable CRLM. In this study, they found that patients who underwent ctDNA detection before surgery had a high recurrence rate [
68]. Furthermore, in another study, it was found that detectable postoperative ctDNA in resected CRLM patients had a significantly lower RFS (HR 6.3; 95% CI 2.58 to 15.2;
P < 0.001) and OS (HR 4.2; 95% CI 1.5 to 11.8; P < 0.001) than patients with undetectable ctDNA [
69]. More recently, peritoneal fluid was used to detect ctDNA for CRC peritoneal metastases (CRC-PM), and the findings suggest that ctDNA detection in peritoneal fluid may be prior to ctDNA in plasma to monitor CRC-PM [
70]. Blood contains ctDNA, which can also be detected in other biological fluids, such as urine. Yu, H. et al. indicated that both plasma and urine cfDNA levels were higher in mCRC patients than in healthy individuals, which can be used to monitor disease progression in CRC patients [
71].
The cfDNA methylation profile enables early diagnosis, prognosis prediction, and screening for CRC [
72]. The cfDNA methylation profile enables early diagnosis, prognosis prediction, and screening for CRC [
73‐
78]. Xianrui and colleagues discovered a novel cfDNA methylation model based on 11 methylation biomarkers to improve the detection of early-stage CRC patients in their investigation [
79]. In another study, 159/267 (87%) mCRC patients showed positivity for methylated markers (e.g., EYA4, GRIA4, ITGA4, MAP3K14-AS1, MSCs) by ddPCR assays, suggesting that methylation can be used as a monitoring marker of tumor burden under different therapeutic regimens [
80]. Furthermore, the promoter hypermethylation of septin 9 (SEPT9) in cfDNA has been confirmed as a potent biomarker in CRC, and the Epi proColon 2.0 kit for cell-free circulating methylated SEPT9 detection approval by the FDA as the first blood-based CRC screening test [
81‐
83].
Almost all advanced CRC patients need further treatment after surgery, such as systemic chemotherapy, molecular targeted therapies or immunotherapies [
84,
85]. In a multicenter cohort study of 96 patients with stage III colon cancer, ctDNA was detected in 15 of 88 (17%) post-chemotherapy samples. When ctDNA was detectable after chemotherapy, the estimated 3-year recurrence-free interval was 30%, compared to 77% when ctDNA was undetectable (HR, 6.8; 95% CI, 11.0–157.0;
P < 0.001). This indicates that monitoring post-chemotherapy ctDNA can reveal information on minimal residual disease, therapeutic response, and recurrence in patients despite completion of standard adjuvant treatment [
64]. In line with the aforementioned, a trial within a cohort study (MEDOCC-CrEATE) [
86] and CIRCULATE-Japan clinical trials [
87], confirmed that ctDNA could be utilized as a predictor of tumor recurrence and to monitor the effectiveness of adjuvant chemotherapy. A comprehensive genome-wide analysis of somatic mutations in CRC was conducted by the Cancer Genome Atlas (TCGA) network. The most commonly altered genes were APC, TP53, KRAS, PIK3CA and BRAF in CRC. In recent years, with the rapid development of next-generation sequencing (NGS), the detection of somatic mutations has become feasible for clinical application. For instance, the somatic BRAF V600E mutation seems to have a short life expectancy and is a poor indicator of response to standard chemotherapy [
88]. Activated RAS mutation is the main reason for primary or secondary resistance to anti-EGFR therapy and predicts poor survival outcomes among CRC patients [
89‐
92]. Although KRAS mutation status has been found in studies to be a biomarker for CRC response to EGFR-targeted therapy, however, a global phase III ASPECCT study detected RAS ctDNA mutations in panitumumab-treated mCRC patients, and emergent ctDNA RAS mutations were not associated with poor prognosis in panitumumab-treated patients [
93]. A study carried out by E. I. Dumbrava et al. found that patients with high variant allele frequencies of PIK3CA-mutant ctDNA at baseline were associated with shorter OS [
94].
In addition to CTCs, cfDNA or ctDNA, as the most important source in liquid biopsies, they have been implemented in the field of immune-oncology [
95]. ctDNA detection is quantitative. Interestingly, the change in ctDNA levels during chemotherapy is related to tumor response or progression in several tumor types [
96‐
98]. In a prospective pilot study performed by L. Cabel et al., which included NSCLC, CRC and melanoma, the PFS and OS of patients with detectable ctDNA after anti-PD-1 ICI treatment were significantly shorter than those without ctDNA detected. It provides a theoretical basis for evaluating ctDNA prior to treatment initiation [
99]. In a nonrandomized, HLA-A status double-blinded study performed by Masahiro Kitahara et al., they assessed cfDNA levels in plasma by semiquantitative real-time PCR, which were collected from 93 mCRC patients (HLA-A2402 matched,
n = 49; and HLA-unmatched,
n = 44) prior to receiving immunochemotherapy. The PFS of patients with low cfDNA was significantly better than that of patients with high cfDNA (
P = 0.0027). Interestingly, in the HLA-A2402-matched group, patients with low plasma cfDNA had significantly better PFS, but there was no difference in the HLA-A2402-unmatched group. This suggests that cfDNA may be a useful predictive biomarker of the outcome of immunotherapy in metastatic colorectal cancer [
100]. MSI is the first cancer indication approved for ICB [
41]. A recent study showed that CRC with MSI-H detected by using cfDNA-based assays was correlated with a good response to immunotherapy [
101]. In addition, Tieng FYF et al. reviewed liquid biopsy-based tests to evaluate MSI in CRC [
102]. TMB from cfDNA is emerging as a novel biomarker for cancer immunotherapy in several tumors [
103,
104]. Le DT et al. demonstrated that CRC patients with high TMB commonly respond to PD-1/PD-L1 blockade [
41]. In conclusion, the detection of ctDNA can be used for the early diagnosis of cancer, monitoring response, evaluating potential drug resistance to the treatment and prognosis (Fig.
3B).
ctDNA: methodology and technical challenges
Normal and tumor cells both release cfDNA into the circulation, and ctDNA is the portion of cfDNA shed by cancer cells. The investigation of ctDNA can disclose details about a cancer’s biological profile and clinical progression. According to estimates, ctDNA usually accounts for 0.01–5% of total cfDNA in patients with cancer [
105]. While ctDNA has a two-hour half-life, it is cleared quickly after entering the circulation. As a corollary, ctDNA can act as a useful dynamic marker of tumor bulk and reflect therapy responses. Several technologies have emerged to detect ctDNA in recent years, including ultrasensitive targeted PCR-based approaches and next-generation sequencing (NGS) methods. The former includes digital PCR (dPCR) [
106], allele-specific amplification refractory mutation system PCR (ARMS) [
107], allele-specific PCR (AS-PCR) [
108], droplet digital PCR (ddPCR), bead emulsification amplification and magnetics (BEAMing) [
16] to detect mutations in prespecified cancer-associated mutations. And the latter such as tagged-amplicon deep sequencing (TAm-Seq) [
109], Safe-Sequencing System (Safe-SeqS) [
110] and personalized profiling by deep sequencing (CAPP-Seq) [
111], enables simultaneous detection of the genome and multiple rare mutations in ctDNA simultaneously without the requirement of primary tumor sequencing. And the untargeted techniques, such as WGS or WES, allow for the detection of novel, clinically significant genomic aberrations without need the information about primary tumor. In general, the advantage of PCR-based methods is cost-effective and rapid, and no specific bioinformatic skills are needed. However, the main disadvantage is that they can detect a limited number of prespecified mutations. Among the PCR-based approaches, ddPCR or BEAMing can detect extremely infrequent mutations with high sensitivity; nonetheless, the DNA region assessed must be restricted [
112]. Although dPCR is the most commonly used method for detecting ctDNA, the use of NGS for ctDNA detection is becoming increasingly prominent. Somatic single nucleotide variant allele frequencies (SNV VAFs), copy number aberrations (CNAs), or DNA methylation patterns are used in NGS-based methodologies to estimate ctDNA levels in plasma [
113,
114]. Hangyu Zhang et al. developed an NGS-based ctDNA assay and evaluated its sensitivity and specificity while using ddPCR as a control in cetuximab-treated CRC patients. In the study, NGS actually found more mutation information than ddPCR in disease progression patients [
115]. In general, NGS methods can detect a large number of mutations and analyze multiple genomic targets and alterations, but they are restricted by poorer sensitivity, higher input sample volume, and expensive and time-consuming procedures [
113,
116]. In a recent meta-analysis, they compared the diagnostic accuracy of digital PCR, ARMS and NGS for detecting KRAS mutations in the cfDNA of CRC patients, and next-generation sequencing had overall high accuracy [
117]. In summary, to be fit for clinical application, the ideal ctDNA assay should take into account the appropriate testing sensitivity, target scope, maximum sample throughput, and total annual expenditures.