DNA damage repair proteins exist to facilitate the replication of normal cellular DNA [3]. The MSI phenotype results when this protective mechanism is lost through dMMR. Critical mismatch repair (MMR) proteins involved in proofreading and correction of insertion-deletion loops include MLH1, MSH2, MSH6 and PMS2. Deficiencies in MMR can occur through either a germline mutation in an MMR gene (MLH1, MSH2, MSH6 and PMS2), resulting in Hereditary Nonpolyposis Colorectal Cancer (HNPCC) also known as Lynch syndrome, however, this condition occurs more commonly through sporadic epigenetic inactivation of MLH1 [3]. The latter is generally associated with hypermethylation of promoter regions of cancer-specific genes known as the CIMP-H.

Currently, the routine use of adjuvant chemotherapy is not recommended in stage II CRC patients. Exceptions to this include those at higher risk of recurrence, for example tumors with adverse features such as poor differentiation, lymphovascular or perineural invasion and younger age patients. dMMR provides a molecular, tailored approach to stratifying patients based on their potential response to chemotherapy. Sargent and colleagues described dMMR as a predictive biomarker for poor response to FU-based adjuvant therapy in stage II and stage III colon cancer [4]. This is highlighted in their study of 457 colon cancer patients who were randomly assigned to either FU-based therapy or no post-surgical treatment. Patients with dMMR tumors receiving FU-based therapy had no improvements in DFS (hazard ratio [HR], 1.10; 95% CI, 0.42 to 2.91; p = 0.85), when compared to those assigned to surgery alone. In contrast, patients with microsatellite-stable or proficient MMR (pMMR) tumors, receiving adjuvant therapy, demonstrated significantly improved DFS (HR, 0.67; 95% CI, 0.48 to 0.93; p = 0.02) [4]. Similarly, in an Australian cohort study, patients with dMMR, despite not being given adjuvant chemotherapy, still had excellent outcomes [5]. These studies support the concept that adjuvant chemotherapy in stage II colon cancer patients with high dMMR results in minimal OS benefit (2–3%) and as such it is not routinely recommended ^[4][5][6][4,5,6].

MSI high tumors are associated with a better prognosis in curative settings, however, in mCRC, it appears to confer a negative prognosis. As a predictive biomarker, a large amount of evidence suggests possible resistance to 5-FU in MSI-H tumors [10]. This is due to the high mutational load eliciting an endogenous immune anti-tumor response, which is counterbalanced by the expression of immune inhibitory signals, such as PD-1 or PD-L1, resisting tumor elimination [11]. Based on these considerations, MSI-H CRCs are highly responsive to immunotherapy, such as anti-PD-1 [11]. Current guidelines recommend MSI testing in all CRC patients, to not only identify HNPCC but to guide adjuvant treatment decisions and to identify patients likely to benefit from immunotherapy in stage IV disease.

HNPCC is an autosomal dominant condition caused by genomic mutations in DNA MMR genes, MLH1, MSH2, MSH6 and PMS2 [7]. Inactivating mutations in the MMR result in a high level of MSI (MSI-H) and subsequently an increased risk of cancer, particularly colon and endometrial [8].

KRAS belongs to the RAS family of oncogenes and is mutated in 40–50% of CRCs [13], most commonly via point mutations ^[14][15][14,15]. Whilst some studies have suggested a prognostic role for RAS ^{[16][17][18][19]}[16,17,18,19], its main utility is as a predictive biomarker. Tumors with a mutation in codon 12 or 13 of exon 2 of the KRAS gene are essentially unresponsive to EGFR-antibody (EGFR-ab) therapy. Similarly, mutation in KRAS outside of exon 2 and mutation in NRAS are predictive for poor response to EGFR-ab therapy ^[20][21][20,21]. This is also highlighted in studies by Karapetis et al. and Amado et al., who demonstrated that KRAS predicts response to cetuximab and panitumumab in advanced CRC, respectively ^[68][69][68,69]. The study by Karapetis et al., correlated tumor mutation status of the KRAS gene with survival in advanced CRC patients receiving either cetuximab or supportive care. Their study found that for patients with wild-type KRAS tumors (KRAS^WT), treatment with cetuximab as compared with supportive care alone significantly improved OS (median, 9.5 vs. 4.8 months; HR for death, 0.55; 95% CI, 0.41 to 0.74; p < 0.001) and PFS (median, 3.7 months vs. 1.9 months; HR for progression or death, 0.40; 95% CI, 0.30 to 0.54; p < 0.001) [68]. Similarly, Amado and colleagues assessed the impact of KRAS mutations on PFS in mCRC patients on PFS following treatment with panitumumab. They found PFS was significantly greater in patients receiving panitumumab with KRAS^WT, (HR, 0.45; 95% CI: 0.34 to 0.59) than in the mutant group (HR, 0.99; 95% CI, 0.73 to 1.36) [69]. These practice-changing discoveries have defined restrictions for the use for EGFR-ab therapy to patients with mCRC with wild-type RAS (RAS^WT), sparing up to 60% of patients’ futile exposure to toxicity and saving needless cost [22].

An additional consideration in RAS^WT patients is the impact of tumor sidedness on targeted therapy. A recent study by Holch and colleagues investigated the prognostic and predictive relevance of primary tumor location. Their meta-analysis of first line clinical trials concluded that patients with left-sided RAS^WT mCRC had significantly greater survival benefit from anti-EGFR treatment compared with anti-VEGF treatment when added to standard chemotherapy (HR, 0.71; 95% CI, 0.58–0.85; p = 0.0003). In contrast, patients with right-sided RAS^WT mCRC demonstrated significantly improved PFS when treated with chemotherapy plus VEGF-ab therapy (HR, 1.53; 95% CI, 1.16–2.01; p = 0.003) [12]. Nonetheless, due to the molecular heterogeneity within left- and right-sided tumors, caution regarding treatment decisions needs to be exercised when basing therapy on the location of tumor in the colon ^[123][128][123,128].

The prognostic impact of the most common BRAF mutation in mCRC, BRAF^V600E, is well characterized [129]. The survival rate of patients carrying the mutant form is decreased by approximately 50% compared to patients with wild-type BRAF (BRAF^WT) ^{[130][131][132][133]}[130,131,132,133]. This is highlighted in a pooled analysis of the CAIRO, CAIRO2, COIN and Focus studies, where patients with BRAF mutations demonstrated worse median PFS and OS compared with patients that had BRAF^WT tumors (PFS: 6.2 vs. 7.7 months, respectively; HR, 1.34; 95% CI, 1.17–1.54; p < 0.001; OS: 11.4 vs. 17.2 months, respectively; HR, 1.91; 95% CI, 1.66–2.19; p < 0.001) [133]. However, it is important to note that not all BRAF mutations exhibit the same clinical behavior. One previous study has suggested that BRAF^non-V600E mutations have more favorable outcomes compared to BRAF^V600E mutation or BRAF^WT tumors in mCRC (60.7 vs 11.4 vs. 43.0 months, respectively; p < 0.001) [23]. In addition to prognostic implications, BRAF mutations may serve as a predictive biomarker for triplet combination therapy of mitogen-activated protein kinase (MEK), BRAF inhibition plus EGFR-targeted therapies. This is highlighted in a recent phase II study by Corcoran and colleagues who found patients with BRAF^V600E mCRC receiving triplet therapy had a 21% response rate (95% CI, 12.5–43.3%) compared to 10% response for patients in the dabrafenib plus panitumumab arm (95% CI, 1.2–31.7%) [134]. In addition, the ongoing phase III BEACON CRC trial, where patients are receiving triplet combination of binimetinib, encorafenib and cetuximab, demonstrated an overall response rate of 48% (95% CI, 29.4% to 67.5%), median PFS of 8 months (95% CI, 5.6 to 9.3 months) and median OS of 15.3 months (95% CI, 9.6 months to not reached) [135]. Based on these results, the US Food and Drug Administration (FDA) granted a Breakthrough Designation to this triplet therapy for BRAF^V600E CRC patients whom failed one or two prior lines of therapy for metastatic disease [136].

Somatic BRAF mutations, most frequently V600E, have been described in a significant proportion of sporadic MSI-H CRC but not in HNPCC. Thus, clinical BRAF mutation testing has been proposed as a means to identify MSI-H CRC cases that do not require germline MMR gene testing [24].

CEA is one of the most extensively used tumor markers worldwide [25]. Despite its poor sensitivity and specificity ^[26][27][26,27], a rising CEA post curative surgery often correlates with relapse. Thus, CEA is useful in providing early detection of recurrence and allows clinicians a means for early detection and surgical resection of metastases ^[25][28][25,28]. The benefits of CEA as a reliable predictor or recurrence and survival after curative surgery in patients with CRC is highlighted in a recent retrospective study by Baqar and colleagues. In their study of 623 CRC patients, elevated CEA (≥5 ng/mL) was a predictor of recurrence (HR, 1.8; 95% CI, 1.09–3.00; p = 0.002) and of OS (HR, 7.79; 95% CI, 1.00–3.19; p = 0.046) [28].

Irinotecan hydrochloride is an anti-neoplastic topoisomerase inhibitor that is widely used in combination with 5-FU and leucovorin chemotherapy for first line treatment of mCRC and as a single agent in second-line salvage therapy of 5-FU refractory mCRC. The principle dose-limiting toxicities associated with irinotecan are delayed diarrhea and severe neutropenia; these toxicities are reversible, not cumulative and related to irinotecan dose [137]. Irinotecan is metabolized into toxic 7-ethyl-10-hydroxycamptothecin (SN-38) via the hepatic enzyme uridine diphosphate-glucuronosyltransferase 1A (UGT1A) and the inactivated byproduct, SN-38, excreted in bile. The effect of genetic polymorphisms of the UGT1A1 gene in predicting irinotecan-associated toxicity has gained interest. Currently, over 100 genetic variants of UGT1A1 exist, the wild-type allele, UGT1A1*1, being associated with normal enzyme activity and the most common variant allele, UGT1A1*28, being investigated as a cause for increased irinotecan toxicity [138]. The findings from four pharmacogenetic trials, assessing the impact of several UGT1A1 variants, found that patients homozygous for UGT1A1*28 experienced significantly more serious hematological side effects ^{[139][140][141][142]}[139,140,141,142]. Based on this evidence, the United States (US) Food and Drug Administration (FDA) amended the irinotecan label in 2005 to include UGT1A1*28/*28 as a risk factor for severe neutropenia, stating that when administered as a single-agent, a reduction in the starting dose by at least one level or irinotecan hydrochloride injection should be considered for patients known to be homozygous for the UGT1A1*28 allele [138]. In addition, Hoskins and colleagues performed a meta-analysis assessing the association of irinotecan dose with the risk of irinotecan-related hematologic toxicities (grade III-IV) based on UGT1A1 variants. Their findings concluded that the risk of toxicity was higher among patients with UGT1A1*28/*28 genotype than among those with UGT1A1*1/*1 or UGT1A1*1/*28 genotype for both medium (OR = 3.22; 95% CI, 1.52 to 6.81; p = 0.008) and high (OR = 27.8; 95% CI, 4.0 to 195; p = 0.005) doses of irinotecan, only. Despite black box warnings in the US by the FDA, these warnings have not been replicated in other jurisdictions such as Australia most likely due to conflicting studies ^{[143][144][145]}[143,144,145]. In summary, despite the significance of UGT1A1*28 as a potential biomarker for irinotecan toxicity, genotyping for UGT1A1 is not current clinical practice for determining risk of hematologic toxic effects. Instead, the current clinical protocol suggests close clinical monitoring for patients receiving irinotecan, particularly during the first cycle of chemotherapy, with drug doses adjusted based on standard clinical tests such as white blood cell counts.

The DPYD gene encodes the enzyme dihydropyrimidine dehydrogenase (DPD), which functions as the rate-limiting step in the metabolism of fluoropyrimidine chemotherapies [146]. Greater than 80% of 5-FU is metabolized by DPD, with factors such as age, race, comorbidities and concomitant therapies influencing metabolism. Reduced activity of DPD impacts on the ability to metabolize 5-FU at normal rates and may result in life threatening toxicity [30]. DPYD variants that do not affect DPD activity in a clinically relevant manner include c.85T > C, *9A, rs1801265, p.C29R; c.1627A > G, *5, rs1801159, p.I543V; c.2194G > A, *6, rs1801160, p.V7321. Conversely, variants that have been shown to have deleterious effects on DPD activity, resulting in 5-FU toxicity, include DYPD*2A and DPYD*13. While variants c.2846A > T and c.1129–5923C > G have been shown to have moderately reduced DPD activity [147]. A multicenter study of 17 hospitals assessing DPYD genotype-guided dosing in patients receiving fluoropyrimidines (capecitabine or fluorouracil) was carried out by Henricks and colleagues [148]. Their study found DPYD genotype-based dose reductions improved patient safety and fluoropyrimidine treatment. Specifically, patients with either the DPYD*2A or c.1679T > G variant benefited from an initial 50% dose reduction of fluoropyrimidines. While patients that were c.1236G > A or c.2846A > T carriers, a 25% dose reduction was not enough to lower the risk of severe toxicity and a larger dose reduction of 50% was suggested in these patients. The authors highlight the need for additional prospective studies to validate and further refine these findings [148]. Currently prospective testing for DPYD mutations is not routinely performed in clinical practice due to associated costs (approximately $300 in Australia) and long test turnaround times (3–4 weeks) which can be unsatisfactory for developing a therapeutic strategy for patients who require immediate treatment. Thus, for DPYD testing to be routinely used in clinical practice, the problematic turnaround time and lack of funding for tests are barriers that would need to be overcome.

APC is a tumor suppressor gene that is mutated in more than 80% of CRCs and is a common germline mutation in the autosomal dominant FAP syndrome [31]. This disease is characterized by numerous colonic adenomas which, without recognition and intervention, results in the development of early-onset CRC [32]. Clinical diagnosis of FAP is based on genetic testing of the APC gene via an in vitro synthesized-protein assay [33]. A positive test justifies surveillance and familial screening with colorectal endoscopy and aids in surgical management and planning.

As previously discussed in Section 4.1.1, diagnosis for HNPCC involves confirmation of a pathogenic germline mutation in one of several DNA MMR genes, including MLH1, MSH2, MSH6 and PMS2 and/or loss of DNA MMR proteins via immunohistochemistry (IHC) [9]. Germline testing is usually performed on patients with MSI as identified by IHC and in whom acquired methylation has been excluded.

Juvenile polyposis syndrome (JPS) is an autosomal dominant disorder characterized by the occurrence of juvenile polyps predominantly in the gastrointestinal tract, resulting in an increased risk of CRC [149]. Genetic testing for germline mutations of SMAD4 or BMPR1A genes are found in approximately 40–60% of JPS cases ^[150].