The standard of practice for selection of patients with metastatic CRC who are candidates for targeted therapies with anti-EGFR antibodies has been primarily based on the mutation status of KRAS [26]. However, as discussed above, the mutation status of BRAF, NRAS, PI3KCA, and other genes downstream of EGFR may affect the tumor response to anti-EGFR therapy [33]. Therefore, testing for mutations in other genes may be indicated in candidate patients, particularly in the setting of clinical trials, at the present time. EGFR mutation testing is not indicated for CRC since activating EGFR mutations in CRC are rare and were not demonstrated to confer sensitivity to tyrosine kinase inhibitors or to cetuximab therapy [34].

In addition to KRAS mutations in codons 12 and 13 of exon 2, KRAS mutations in codons 59 and 61 of exon 3 and codons 117 and 146 of exon 4 and similar mutation hot spots in NRAS need to be considered when selecting true RAS wild-type patients who will benefit from cetuximab therapy for metastatic CRC (Tables 31.2 and 31.3) [20, 21, 35]. Further, the addition of cetuximab therapy for patients with any RAS mutation is of no benefit [20, 21, 35]. Therefore, patients with metastatic CRC should undergo expanded testing for all known activating RAS mutations, to avoid use of cetuximab in patients whose CRCs will not respond [20, 21, 35]. Interestingly, in contrast to other activating mutations in KRAS, use of cetuximab for patients with chemotherapy-refractory CRC with the KRAS G13D mutation may be associated with longer overall and progression-free survival than with other KRAS activating mutations [18], although this has not been supported by some studies.

Regarding the choice of tissue for DNA mutation analysis, since KRAS mutations occur early in colorectal carcinogenesis, most clinical trials have tested the primary tumor. Published studies show a good correlation between KRAS mutation status in primary and metastatic CRC lesions with an average concordance of 93 % (76–100 %) [36, 37]. Therefore, testing tumor tissue from either the primary site or from a metastatic lesion is appropriate, and a metastatic lesion when available is preferred. Pathologists should select a block of formalin-fixed, paraffin-embedded (FFPE) tissue with the highest proportion of viable tumor and highest percentage of tumor cells as possible, understanding that all cells present in the tissue contribute DNA, so this percentage is of all cells, namely, those that normally compose the colonic wall and variable amounts of inflammatory cells. For small samples, adequate DNA amount can be obtained by pooling macro- or microdissected tumor tissue from multiple tissue levels. This procedure may also provide a larger sampling of the tumor, including areas of mutational heterogeneity, thereby improving mutation detection. Importantly, a pretreatment biopsy sample may be preferable for testing rather than a resection specimen removed after neoadjuvant therapy, as is the case in rectal cancers, where minimal numbers of residual viable tumor cells may persist, making these tissues inadequate for molecular testing. Cytology specimens, particularly from metastatic lesions, also provide an adequate source of tumor for DNA extraction and mutation analysis.

In summary, RAS mutation analysis of CRC tumor tissues is recommended as the standard of care in patients who are candidates for anti-EGFR antibody therapy [20, 21, 35]. Additional mutation testing of other EGFR signaling pathway genes may be helpful to better select patients for targeted therapies with improved outcomes, but a general consensus about which genes should be tested is not yet established. In large practice centers, the trend is to test all CRCs for RAS and BRAF V600E mutations and for MMR deficiency (with the use of immunohistochemical and/or MSI testing), thus allowing for selection of patients for conventional therapy as well as targeted therapy [38]. The use of gene panels for mutation analysis with platforms that permit detection of hundreds of mutations in a single sample, such as next-generation sequencing (NGS) gene panels, may provide further insight into the issue of which genes should be tested. However, determination of which mutations should be considered for CRC treatment selection or exclusion still needs to be better defined based on data from clinical trials. The practice of Genomic Tumor Boards to address the use of gene mutation results and other molecular test results is evolving and is being used in some centers to evaluate treatment options other than established treatment protocols by providing clinical trial options for individual patients.

Approximately 15 % of all CRCs have a deficient MMR system and are characterized by MSI. In 3–5 % of MMR-deficient/MSI-positive CRCs, patients have germline mutations in MMR genes diagnostic of HNPCC or Lynch syndrome, while the remaining are sporadic-type CRCs [39, 40]. In cells with deficient MMR, errors in DNA replication accumulate and are detectable in short tandem repeats of microsatellite regions but also in functional gene coding regions, which results in changes in the length of repeats at any one location (known as MSI) [41–43]. MSI, specifically when a tumor is identified to have a high level of MSI (MSI-H), is a genomic marker of deficient DNA MMR [43]. Six different genes (MSH2, MLH1, PMS1, PMS2, MSH6, and MLH3) encode the MMR system [44]. In HNPCC, as in other hereditary cancer syndromes, recessive mutation of one allele followed by somatic inactivation of the other is the main mechanism of gene silencing [45]. In contrast, in sporadic CRCs, the most frequent mechanism of DNA MMR gene downregulation is biallelic inactivation by CpG methylation and transcriptional silencing of the MLH1 promoter region [46, 47].

MSI status of CRC can be tested by immunohistochemistry (IHC) of CRC tissue sections to assess expression of the MMR proteins or by PCR-based DNA testing for MSI to detect instability at microsatellite sequences [43, 48]. One advantage of IHC is that the protein identified as showing lost or reduced expression in tumor cell nuclei can be an indicator of which MMR gene has a mutation, which is particularly useful in the workup of HNPCC patients, helping to focus the germline mutation analysis on the most likely affected MMR gene. In addition, combining testing for the BRAF V600E activating mutation and CpG island methylation status of the promoter region of MLH1 can be done to help determine whether a MSI-positive CRC with loss of MLH1 expression is likely to be an inherited HNPCC CRC (BRAF mutation negative and MLH1 promoter methylation negative) or sporadic-type CRC (BRAF mutation positive in up to 70 % of cases and MLH1 promoter methylation positive) [49].

MSI positivity in CRCs is a prognostic marker for improved survival [42]. MSI was detected in 17 % of CRCs of patients < 50 years of age and was associated with (1) a lower likelihood of tumor metastasis to regional lymph nodes as well as distant organs, contributing to improved survival through tumor downstaging, and (2) an overall survival advantage independent of stage of disease [50]. In addition, no significant difference in survival was found between patients with MSI due to hereditary (HNPCC) CRC compared to sporadic CRC [50, 51]. A reported survival benefit for patients with HNPCC CRC is mainly determined by a diagnosis at younger age and less advanced tumor stage as compared with sporadic MSI-H CRC patients [51].

Deficient DNA MMR and MSI status are predictive of response to therapy in some subsets of deficient MMR CRCs [52]. Both the tumor stage and whether the CRC is sporadic or hereditary (HNPCC) are important factors for selection of adjuvant therapy for CRC patients. Patients with stage II CRCs with MMR-deficient/MSI-positive status receiving 5-FU have no improvement in disease-free survival, and treatment is associated with reduced overall survival [53]. The effect of deficient MMR/MSI in stage III CRC on response to therapy has been more difficult to establish, which may be related to heterogeneity of the studied populations in terms of the proportions of sporadic and HNPCC cases. Some studies showed improved outcomes with chemotherapy for advanced stage III CRCs that were MSI-H [54], but other studies did not [55]. In patients with stages II and III CRC, MMR-deficient CRC patients receiving 5-FU had no improvement in disease-free survival [53]. Sinicrope et al. reported that distant recurrences were reduced by 5-FU-based adjuvant treatment in MMR-deficient stage III CRCs, and a subset analysis suggested that any treatment benefit was restricted to suspected inherited CRCs as compared to sporadic tumors [56]. In summary, larger trials are needed to determine with certainty the utility of MMR and MSI status for treatment selection in routine patient care [56, 57]. Current data support the concept that MMR-deficient/MSI-H stage II and III CRC patients do not benefit from 5-FU-based adjuvant therapy, although one study of stage III MMR-deficient CRC patients showed a treatment benefit restricted to patients with suspected hereditary (HNPCC) CRC.

Many technical approaches and assays are used for mutation testing by clinical molecular laboratories. Laboratory-developed tests (LDTs) are frequently used and follow requirements for validation and interpretation as dictated by CLIA and state regulations, while some commercial in vitro diagnostic (IVD) test kits also are used. Assay methods for mutation detection include Sanger sequencing, pyrosequencing, multiplex PCR with primer extension, fluorescent bead detection assays, MassARRAY matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) mass spectrometry, allele-specific PCR such as amplification refractory mutation system/scorpions (ARMS/S), melt curve analysis (as performed in Taq-Man assays using real-time quantitative PCR (qPCR)), and, most recently, NGS [58–62]. Table 31.4 compares three technologies commonly used for mutation detection in cancer-related genes in CRC. Until a few years ago, Sanger sequencing of PCR products was a commonly used method for mutation detection. Recent studies comparing Sanger sequencing with other approaches have highlighted its limited analytical sensitivity for clinical testing of DNA from solid tumor samples, requiring 20–50 % mutant alleles in the amplified products [62, 63].

A number of commercial IVD and LDT assays are available to detect the most common mutations in KRAS and BRAF. Many of the early assays only tested for a small number of mutations, limiting the correlation between mutation status and drug response in early clinical trials. Of note, early assays for KRAS mutation analysis performed both in clinical laboratories and as part of clinical trials were limited to codons 12 and 13 in exon 2, excluding other mutation hot spots. Clinical trials have demonstrated the importance of mutations in previously excluded codons of KRAS, in exons 2, 3, and 4, as well as the corresponding codons in NRAS. Additionally, mutations in other genes downstream of EGFR in the EGFR signaling pathway also are clinically significant for patient management decisions, as discussed above.

For mutation testing, genomic DNA can be extracted from fresh, frozen, or FFPE tissue using standard molecular laboratory methods. For solid tumor testing, histologic slides should be reviewed by an anatomic pathologist to assess the percent of tumor cells relative to total cells on the slide and to prioritize areas for macro- or microdissection, if necessary, to increase the percentage of tumor cells in the sample used for testing.