**6.4 Detecting microsatellite phenotype**

The use of immune-checkpoint inhibitors has drastically changed the therapeutic landscape for several solid tumours, including a mCRC subset that harbours mutations in DNA mismatch repair (MMR) genes (for example, mutations in *MLH1*, *MSH2*, *MSH6*, *PMS2*, or *MLH1* promoter hypermethylation) [120]. On a molecular level, impaired DNA MMR can lead to genomic hypermutability, including uncontrolled expansion or contraction in DNA microsatellite repeats, termed microsatellite instability (MSI); and the consequent development of malignant neoplasms which have an MSI-high (MSI-H) phenotype. The MSI-H phenotype is present in approximately 15% of all primary CRC and may occur as a result of either inherited (hereditary non-polyposis colon cancer or lynch syndrome) or sporadic abnormalities. It is now common for institutions to screen for this in tumour tissue, either by immunohistochemistry for deficient MMR (dMMR) or PCR for MSI [121]. The identification of the MSI-H phenotype in CRC patients has important prognostic and therapeutic implications, both in the adjuvant and advanced settings when considering conventional chemotherapeutic and targeted agents.

More recently, a small phase II clinical trial using pembrolizumab, an antiprogrammed cell death protein 1 (PD1) monoclonal antibody in dMMR mCRC

**67**

**7. Conclusion**

adopted into routine clinical practice.

*Current Utility and Future Applications of ctDNA in Colorectal Cancer*

although the data for such techniques are preliminary at this stage.

**6.5 Detecting ctDNA in other body fluids**

patients, demonstrated high rates of objective response (40%) and progression-free survival, while no responses were seen in proficient MMR patients [122]. Similarly, a phase II study of anti-PD1 antibody, nivolumab in dMMR/MSI-H mCRC showed 31% objective response rate and 69% disease control rate (12 weeks or longer) [123]. In addition to these encouraging results, multiple trials using anti-PD1 agents, with or without other targeted therapies are ongoing (for example, NCT02460198, NCT02563002, NCT02060188), and it is expected that immunotherapy will rapidly

In this setting, liquid biopsy might be useful in providing a potentially faster, costefficient, and safer approach compared to tumour biopsy sampling in patients with suspected MSI-H tumours. Therefore, such assays need to be optimised for routine use in the future. The novel ctDNA techniques described earlier in this chapter could be adapted to identify dMMR CRC in several ways, such as change in microsatellite length, loss of heterozygosity, mutations, or hypermethylation of MMR-related genes [124]. However, similar drawbacks regarding their sensitivity and specificity apply in this setting [125]. To overcome these limitations, several groups have developed enrichment techniques which are able to enhance the presence of altered microsatellites with enrichment probes and detect alterations at very low allele frequencies [126]. Moreover, ctDNA in the setting of immunotherapy can also be used (i) as a predictive marker to identify tumour mutational burden or specific response mutations (for example, PTEN loss or activating beta-catenin mutations), (ii) to monitor treatment response or resistance in conjunction with radiological imaging, and (iii) to identify neoepitopes and epigenetic or transcriptomic markers [124];

This chapter has predominantly focused on the utility of ctDNA in the peripheral blood. Multiple studies have also demonstrated the presence of tumour-derived nucleic acids in other body compartments, such as the urine, stool, saliva, cerebrospinal fluid, pleural fluid, and bronchial washings [40, 127–129]. Of course, topography of the primary tumour, and of any disseminated lesions, will have a significant effect on the concentration of ctDNA in different body fluids.

In a small study, Fujii et al. demonstrated the utility of detecting *KRAS* mutations in the urine of mCRC patients who were undergoing systemic treatment. Both NGS and enrichment PCR were used to detect KRAS in the urine, plasma and archival tumour tissue [128]. The results not only suggested good concordance between ctDNA in the urine and mutant *KRAS* in the tumour, but also demonstrated that ctDNA trends in the urine reflected the tumour dynamics in the plasma. As such, this may also represent an alternative approach to monitoring for therapeutic response or resistance.

The data generated from basic research, retrospective clinical studies, and limited prospective studies all support the potential role of ctDNA as a biomarker for early disease, minimal residual disease, recurrence, response to therapy, and emerging drug resistance mechanisms in the management of CRC. Nevertheless, multiple challenges need to be overcome before this promising technology can be

Firstly, a crucial question is whether the genomic aberrations detected in ctDNA

actually drive tumour progression. It is also still unknown whether ctDNA will ever be able to mirror the heterogeneity or molecular subclones of CRC in a given

*DOI: http://dx.doi.org/10.5772/intechopen.82316*

become standard of care in dMMR/MSI-H mCRC.

#### *Current Utility and Future Applications of ctDNA in Colorectal Cancer DOI: http://dx.doi.org/10.5772/intechopen.82316*

*Advances in the Molecular Understanding of Colorectal Cancer*

both of these biomarkers in a liquid biopsy.

**6.4 Detecting microsatellite phenotype**

*6.3.2 Extracellular vesicles*

*6.3.3 miRNA*

detectable ctDNA levels in CRC patients do not always correspond to the ability to detect CTC [102, 107]. As such, it would be worth exploring the concomitant use of

The clinical utility for these small, membrane-bound cell fragments, which are also thought to originate from apoptotic, necrotic or proliferating tumour cells, has also recently been considered [108]. Depending on their size and content, they fall under the categories of exosomes, microvesicles, and apoptotic bodies. In particular, tumour-derived exosomes are constitutively formed and released from tumour cells and can be found in the peripheral circulation, other body fluids and interstitial spaces. They can contain concentrated forms of RNA, miRNA, long non-coding RNA, nucleic acids, protein and lipids, but only very small amounts of doublestranded DNA [109]. As such, exosome-derived nucleic acids from the serum of CRC patients may be used to identify genetic aberrations from the tumour that are not detectable by ctDNA, and therefore can be used in a complementary fashion with other biomarkers. To date, there is no published data that has alluded to their clinical application in CRC. Intriguingly, exosomes have recently been shown to have roles in

cell-cell signalling which may affect tumour growth and development [110].

However, owing to extensive heterogeneity between several studies, it has been difficult to gauge their clinical worth in terms of sensitivity and specificity, which has often been described as 'modest'. Encouragingly, recent studies have suggested that by using miRNA panels or signatures, the predictive accuracy of these assays can be significantly enhanced [117, 118]. In mCRC, only a few studies have addressed the role of circulating miRNAs as predictive biomarkers to systemic therapy [119]. Conceptually, miRNA assays could be used in conjunction with ctDNA, or with the aforementioned biomarkers, to facilitate accurate read-outs for improved sensitivity and specificity.

The use of immune-checkpoint inhibitors has drastically changed the therapeutic landscape for several solid tumours, including a mCRC subset that harbours mutations in DNA mismatch repair (MMR) genes (for example, mutations in *MLH1*, *MSH2*, *MSH6*, *PMS2*, or *MLH1* promoter hypermethylation) [120]. On a molecular level, impaired DNA MMR can lead to genomic hypermutability, including uncontrolled expansion or contraction in DNA microsatellite repeats, termed microsatellite instability (MSI); and the consequent development of malignant neoplasms which have an MSI-high (MSI-H) phenotype. The MSI-H phenotype is present in approximately 15% of all primary CRC and may occur as a result of either inherited (hereditary non-polyposis colon cancer or lynch syndrome) or sporadic abnormalities. It is now common for institutions to screen for this in tumour tissue, either by immunohistochemistry for deficient MMR (dMMR) or PCR for MSI [121]. The identification of the MSI-H phenotype in CRC patients has important prognostic and therapeutic implications, both in the adjuvant and advanced settings when

considering conventional chemotherapeutic and targeted agents.

More recently, a small phase II clinical trial using pembrolizumab, an antiprogrammed cell death protein 1 (PD1) monoclonal antibody in dMMR mCRC

The role of circulating, exosome-free miRNAs as potential diagnostic and prognostic biomarkers in CRC has been extensively investigated over the past 5 years [111–116].

**66**

patients, demonstrated high rates of objective response (40%) and progression-free survival, while no responses were seen in proficient MMR patients [122]. Similarly, a phase II study of anti-PD1 antibody, nivolumab in dMMR/MSI-H mCRC showed 31% objective response rate and 69% disease control rate (12 weeks or longer) [123]. In addition to these encouraging results, multiple trials using anti-PD1 agents, with or without other targeted therapies are ongoing (for example, NCT02460198, NCT02563002, NCT02060188), and it is expected that immunotherapy will rapidly become standard of care in dMMR/MSI-H mCRC.

In this setting, liquid biopsy might be useful in providing a potentially faster, costefficient, and safer approach compared to tumour biopsy sampling in patients with suspected MSI-H tumours. Therefore, such assays need to be optimised for routine use in the future. The novel ctDNA techniques described earlier in this chapter could be adapted to identify dMMR CRC in several ways, such as change in microsatellite length, loss of heterozygosity, mutations, or hypermethylation of MMR-related genes [124]. However, similar drawbacks regarding their sensitivity and specificity apply in this setting [125]. To overcome these limitations, several groups have developed enrichment techniques which are able to enhance the presence of altered microsatellites with enrichment probes and detect alterations at very low allele frequencies [126].

Moreover, ctDNA in the setting of immunotherapy can also be used (i) as a predictive marker to identify tumour mutational burden or specific response mutations (for example, PTEN loss or activating beta-catenin mutations), (ii) to monitor treatment response or resistance in conjunction with radiological imaging, and (iii) to identify neoepitopes and epigenetic or transcriptomic markers [124]; although the data for such techniques are preliminary at this stage.

## **6.5 Detecting ctDNA in other body fluids**

This chapter has predominantly focused on the utility of ctDNA in the peripheral blood. Multiple studies have also demonstrated the presence of tumour-derived nucleic acids in other body compartments, such as the urine, stool, saliva, cerebrospinal fluid, pleural fluid, and bronchial washings [40, 127–129]. Of course, topography of the primary tumour, and of any disseminated lesions, will have a significant effect on the concentration of ctDNA in different body fluids.

In a small study, Fujii et al. demonstrated the utility of detecting *KRAS* mutations in the urine of mCRC patients who were undergoing systemic treatment. Both NGS and enrichment PCR were used to detect KRAS in the urine, plasma and archival tumour tissue [128]. The results not only suggested good concordance between ctDNA in the urine and mutant *KRAS* in the tumour, but also demonstrated that ctDNA trends in the urine reflected the tumour dynamics in the plasma. As such, this may also represent an alternative approach to monitoring for therapeutic response or resistance.
