Molecular Targets for the Treatment of Metastatic Colorectal Cancer

Over the past years, colorectal cancer (CRC) was subtyped according to its molecular and genetic characteristics, allowing the development of therapeutic strategies, based on predictive biomarkers. Biomarkers such as microsatellite instability (MSI), RAS and BRAF mutations, HER2 amplification or NTRK fusions represent major tools for personalized therapeutic strategies. Moreover, the routine implementation of molecular predictive tests provides new perspectives and challenges for the therapeutic management of CRC patients, such as liquid biopsies and the reintroduction of anti-EGFR monoclonal antibodies. In this review, we summarize the current landscape of targeted therapies for metastatic CRC patients, with a focus on new developments for EGFR blockade and emerging biomarkers (MSI, HER2, NTRK).


Introduction
Colorectal cancer (CRC) is the third most common cancer and the second leading cause of death worldwide [1]. Median overall survival (OS) in patients with previously untreated and unresectable advanced colorectal (CRC) cancer now ranges from 25 to 30 months, when combining molecular targeted therapies and chemotherapy [2].
Whereas treatment for localized CRC rely on surgery and chemotherapy, therapeutic strategies in metastatic setting were improved through the development of biomarkers for targeted therapies, such as EGFR (epithelial growth factor receptor) inhibitors, BRAF serine/threonine-protein kinase B-Raf) inhibitors, HER2 (human epidermal growth factor receptor 2) inhibitors, immune checkpoint inhibitors, or NTRK (neurotrophic tyrosine receptor kinase) inhibitors. Predictive biomarkers for mCRC patients now include RAS (Kirsten rat sarcoma virus) and BRAF mutations, microsatellite instability (MSI) and mismatch repair deficiency (dMMR), HER2 amplifications and NTRK fusions. The challenge is now to implement these innovations in daily practice.
In this review, we discuss current predictive biomarkers for mCRC patients and their associated targeted therapy, including emerging parameters for anti-EGFR agents, such as primary tumor sidedness and longitudinal follow-up using circulating tumor DNA (ctDNA), immune checkpoint inhibitors for MSI/dMMR tumors, new developments for BRAF V600E -mutated cancers, anti-HER2 treatments and NTRK inhibitors. Other mechanisms of resistance are described. These might involve the EGFR itself with mutations of the EGFR ectodomain [7]. It can also be related to the overactivation of a protein downstream from the EGFR in the MAPK (mitogen-activated protein kinase) pathway, such as BRAF [8], although the BRAF V600E hotspot mutation was not formally demonstrated as a factor of resistance to anti-EGFR agents (see below). The constitutional activation of the PI3K/Akt/mTOR pathway through PIK3CA exon 20 mutation or PTEN loss, could participate in resistance to monoclonal antibodies (mAbs) [9,10]. Furthermore, amplifications of HER2, HER3, or MET and HER2 activating mutations seem associated with resistance to anti-EGFR therapy [11]. Recent reports brought to light the predictive value of the micro RNA miR-31-3p. Mir-31 plays a significant role in activating the RAS signaling pathway, and high miR-31-3p expression could be the witness of the tumor's EGFR independency and subsequently to its resistance to anti-EGFR therapy. Several post hoc analysis of randomized trials showed the level of expression of miR-31-3p is a valid biomarker of anti-EGFR efficacy [12][13][14].

Current Management of Anti-EGFR Therapy
First, anti-EGFR mAbs do not improve outcomes as adjuvant treatment for resected stage III colon cancer [15]. Concerning patients with resectable mCRC, a warning signal concerning the use of anti-EGFR mAbs in the peri-operative setting was raised by the NEW EPOC trial. This study showed a detrimental effect of cetuximab, both on disease free survival (DFS) and overall survival (OS), Other mechanisms of resistance are described. These might involve the EGFR itself with mutations of the EGFR ectodomain [7]. It can also be related to the overactivation of a protein downstream from the EGFR in the MAPK (mitogen-activated protein kinase) pathway, such as BRAF [8], although the BRAF V600E hotspot mutation was not formally demonstrated as a factor of resistance to anti-EGFR agents (see below). The constitutional activation of the PI3K/Akt/mTOR pathway through PIK3CA exon 20 mutation or PTEN loss, could participate in resistance to monoclonal antibodies (mAbs) [9,10]. Furthermore, amplifications of HER2, HER3, or MET and HER2 activating mutations seem associated with resistance to anti-EGFR therapy [11]. Recent reports brought to light the predictive value of the micro RNA miR-31-3p. Mir-31 plays a significant role in activating the RAS signaling pathway, and high miR-31-3p expression could be the witness of the tumor's EGFR independency and subsequently to its resistance to anti-EGFR therapy. Several post hoc analysis of randomized trials showed the level of expression of miR-31-3p is a valid biomarker of anti-EGFR efficacy [12][13][14].

Current Management of Anti-EGFR Therapy
First, anti-EGFR mAbs do not improve outcomes as adjuvant treatment for resected stage III colon cancer [15]. Concerning patients with resectable mCRC, a warning signal concerning the use of anti-EGFR mAbs in the peri-operative setting was raised by the NEW EPOC trial. This study showed a detrimental effect of cetuximab, both on disease free survival (DFS) and overall survival (OS), when added to chemotherapy [16]. Although anti-EGFR mAbs should not be used as perioperative treatment for patients with resectable mCRC, these treatments could be useful as a converting therapy to shrink potentially resectable metastatic diseases [17].
Anti-EGFR mAbs (cetuximab and panitumumab) are associated with improved response rates, progression free survival (PFS), and OS in first-line setting, when combined with an oxaliplatin-based or irinotecan-based regimen, as well as in second or latter lines alone, or combined with cytotoxic agents [18][19][20][21][22][23][24][25][26][27][28][29]. Although the NORDIC VII and COIN studies did not show a significant effect of cetuximab in combination with an oxaliplatin-based regimen [30,31], recent data provided by the TAILOR trial demonstrate that it can be safely prescribed with FOLFOX for RAS WT mCRC patients [32]. There is no direct comparison between panitumumab and cetuximab, except for chemo-resistant patients with the ASPECCT trial that showed the non-inferiority of panitumumab in patients with chemotherapy-refractory KRAS WT (exon 2) mCRC [33].

Primary Tumor Sidedness
Primary tumor sidedness (PTS) seems to play a decisive role in the sensitivity to anti-EGFR mAbs. Indeed, there is growing evidence showing that the PTS is a prognostic factor in the RAS WT population, as well as a predictive factor of response to anti-EGFR mAbs [34]. A retrospective analysis of six randomized trials (CRYSTAL, FIRE-3, CALGB 80405, PRIME, PEAK, and 20050181) showed poorer outcomes (response rates, PFS, OS) for right-sided tumors, compared to left-sided tumors. This analysis also showed a predictive effect of PTS, with a greater effect of chemotherapy plus anti-EGFR mAbs, compared to chemotherapy ± bevacizumab in patients with left-sided tumors [35]. A recent retrospective analysis of the ARCAD database showed that the prognostic value of PTS was restricted to the KRAS WT population and confirmed the predictive value of PTS for cetuximab efficacy, with improved outcomes for left-sided mCRC patients [36]. On the contrary, patients with RAS WT right-sided mCRC seem to derive limited benefit from anti-EGFR treatments.
Although these results should be interpreted with caution due to their retrospective nature, they suggest that one should use anti-EGFR mAbs plus chemotherapy as the first line treatment only in patients with KRAS/NRAS WT and left-sided tumors, and might prefer chemotherapy plus an antiangiogenic agent for patients with right-sided mCRC [37]. Nevertheless, it is worth noting the optimal sequence with anti-EGFR mAbs in patients with RAS WT mCRC in the context of PTS, remains unclear. Multi-line therapy trials such as the STRATEGIC-1 phase III study (NCT01910610) would provide valuable information.

Rechallenge and Liquid Biopsies
Resistance to targeted therapies such as anti-EGFR mAbs emerges through the selection of tumor clones harboring an innate mutation of resistance. When the anti-EGFR mAb is interrupted, the selecting pressure on the sensitive clones is withdrawn and the tumor can regain sensitivity to the targeted therapy (see Figure 2). Several strategies can be used to overcome tumor resistance (rechallenge, reintroduction, sequential therapy, dose intensification), and rechallenge seems to be the most promising in the context of anti-EGFR mAbs [38]. Rechallenge of anti-EGFR therapy could be defined as retreatment after a progression while on treatment, for a tumor that initially displayed sensitivity to the therapy [39]. Longitudinal follow-up of mutant clones seems particularly seductive for rechallenge strategies. Studies with longitudinal monitoring of ctDNA showed that RAS mutant clones arise in blood during anti-EGFR therapy and decline exponentially upon withdrawal of treatment, with a half-life of 4 to 5 months [40]. The CRICKET phase II study (n = 38) was the first prospective Longitudinal follow-up of mutant clones seems particularly seductive for rechallenge strategies. Studies with longitudinal monitoring of ctDNA showed that RAS mutant clones arise in blood during anti-EGFR therapy and decline exponentially upon withdrawal of treatment, with a half-life of 4 to 5 months [40]. The CRICKET phase II study (n = 38) was the first prospective demonstration that a rechallenge strategy with cetuximab and irinotecan might be active in RAS/BRAF WT mCRC patients with acquired resistance to cetuximab. No RAS mutation was detected in blood samples from patients who experienced partial response [41,42]. Several clinical trials are currently evaluating the use of liquid biopsy in the context of anti-EGFR rechallenge (CHRONOS, NCT03227926; RASINTRO, NCT03259009) [41].

BRAF V600E Mutation and Targeted Therapies
BRAF V600E mutation, found in approximately 8-10% of mCRC (see Figure 1), induces am RAS-independent constitutional activation of the MAPK pathway, leading to cell proliferation and survival, and is associated with poor prognosis [43]. Although some BRAF mutations occur beyond the V600 hotspot, representing 22% of all BRAF mutations in CRC, they do not have the same biological, clinical, and therapeutic consequences as the V600E mutation [44]. These BRAF non-V600E mutated tumors are more likely to be left-sided, with a lower grade of differentiation, and associated with better prognosis and resistance to BRAF inhibitors, although some might be sensitive to EGFR [45,46]. To note, fusions of the BRAF gene were also reported. These genetic alterations are associated with right-sided tumors and seem to confer resistance to anti-EGFR therapy [11,47,48].
BRAF V600E CRC patients are more likely to be older, female, with right-sided tumors harboring mucinous component, associated with peritoneal and distant lymph node metastases but less pulmonary metastases [49]. Importantly about 20% of BRAF V600E CRC exhibit the MSI phenotype, which is predictive for the efficacy of immune checkpoint inhibitors, whatever the BRAF mutational status (see below) [43,50].
BRAF V600E mCRC patients have a lower probability to receive second-line treatments than the BRAF WT population. As a matter of fact, intensification strategies seem effective for these patients [51][52][53]. In the TRIBE study, first-line FOLFOXIRI (folinic acid, fluorouracil, oxaliplatin, and irinotecan) plus bevacizumab was associated with a non-significant improvement of overall survival for BRAF V600E mutants, compared to FOLFIRI (folinic acid, fluorouracil, and irinotecan) plus bevacizumab (n = 28) [54]. Despite the small population sample of this subgroup analysis, FOLFOXIRI-bevacizumab is considered to be a valid option for chemotherapy-naïve patients with BRAF V600E mCRC. Importantly, no survival benefit was reported for BRAF V600E patients in the subsequent TRIBE2 phase III trial, which tested FOLFOXIRI plus bevacizumab and reintroduction, upfront after progression, versus mFOLFOX6 plus bevacizumab, followed by FOLFIRI plus bevacizumab (33 patients in each arm) [55]. The triplet chemotherapy regimen, associated with either bevacizumab or cetuximab, is currently being tested in the AIO-KRK-0116 phase II trial (NCT04034459; see Table 1).

Antiangiogenic Agents for the BRAF V600E Mutants
There is currently no predictive biomarker for the efficacy of antiangiogenics in the overall population of mCRC patients, and the activity of these agents are not specifically proven among BRAF V600E mCRC patients. The AVF2107g35 and AGITG MAX36 trials showed no survival improvement by adding bevacizumab to first-line IFL (bolus irinotecan, fluorouracil, and folinic acid) or capecitabine [56,57]. Yet, the RAISE study (FOLFIRI ± ramucirumab) and the VELOUR trial (FOLFIRI ± aflibercept) showed that the subset of BRAF V600E CCR tended to benefit from these antiangiogenics in second line, although the small size of the cohorts precluded statistical significance [58,59]. All in all, these retrospective analyses suggest that antiangiogenics could be beneficial to BRAF V600E mCRC patients, in the first-line or latter [60].

Anti-EGFR Agents for BRAF V600E Mutants
Data regarding the efficacy of anti-EGFR therapies for BRAF V600E patients are not clear, either as single agent or in combination with chemotherapy. Two meta-analyses were performed. A meta-analysis performed by Pietrantonio et al. suggested that anti-EGFR agents are not effective for BRAF V600E patients [61]. However, another meta-analysis by Rowland et al. failed to detect any significant difference in the effect of anti-EGFR agents between the BRAF V600E and BRAF WT population [62]. Additionally, a retrospective analysis of the BRAF V600E subgroup, the FIRE-3 study (first-line FOLFIRI plus cetuximab versus FOLFIRI plus bevacizumab in KRAS WT mCRC patients) identified a higher response rate in the cetuximab arm [63]. Moreover, a recent study (VOLFI AIO KRK0109) comparing the efficacy of first line FOLFOXIRI with or without panitumumab, found a significant increase in objective response in the subgroup of BRAF V600E patients (71% versus 22%, n = 14) [18]. the positive results of the BEACON study also confirm that anti-EGFR agents are valuable for the BRAF V600E population (see below) [64]. International guidelines did not contraindicate anti-EGFR agents for BRAF V600E mCRC patients [6,37].

BRAF Inhibitors
In contrast to melanoma, BRAF inhibitors as single agents were associated with disappointing results. One hypothesis is that BRAF inhibition induces feedback EGFR activation and might promote MAPK constitutive signaling. EGFR-mediated reactivation of downstream signaling pathways contributes to the inherent resistance of these tumors to BRAF inhibitor monotherapy [65]. With this issue in mind, several combinations of BRAF inhibitor, anti-EGFR agents, PI3K inhibitors, or MEK inhibitors were tested, with interesting results [66][67][68][69][70]. These studies supported the design of the phase III BEACON study, which evaluated encorafenib and cetuximab ± binimetinib versus chemotherapy (investigator choice regimen of cetuximab plus irinotecan or FOLFIRI). 665 BRAF V600E mCRC patients with disease progression after one or two previous lines were randomized. Median OS was 9.3 months in the triplet and the doublet experimental arms, compared to 5.9 months in the control arm (HR = 0.60, 95% confidence interval 0.47-0.75 and HR = 0.61, 95% confidence interval 0.48-0.77, respectively) [64,71]. The objective response rate was statistically improved (2% in the control group, compared to 20% and 26% in the doublet and triplet group, respectively). Toxicity was manageable, with grade 3 or higher toxicities being comparable between the three arms, with gastrointestinal and skin-related events occurring in the experimental groups. A complementary quality of life analysis found a reduced risk of quality of life deterioration by more than 40% in both the doublet and the triplet group [72]. Although the study was not designed to compare the two experimental arms, the benefit-risk balance might be in favor of the encorafenib plus cetuximab combination, without the MEK inhibitor.

RAS Mutants and Targeted Therapies
Approximately 50% of mCRC are reported to have KRAS/NRAS pathogenic mutation. The treatment of these tumors of poor prognosis was proven to be highly challenging. They are Cancers 2020, 12, 2350 6 of 17 intrinsically resistant to anti-EGFR mAbs (see above). Antiangiogenics (bevacizumab, aflibercept, ramucirumab) seem to be effective in this population but predictive biomarkers for the efficacy of these compounds are lacking [59,73,74].
KRAS G12D (glycine 12 to aspartic acid) is the most common KRAS mutation in colorectal cancer. A new generation of KRAS inhibitor might be a game changer for this population [75]. Recently, promising results of a direct KRAS G12C inhibitor were reported. AMG 510 is a novel small molecule that specifically and irreversibly inhibits KRAS G12C , by locking it in an inactive guanosine diphosphate (GDP)-bound state. In the first in-human study evaluating AMG 510 in adult patients with locally advanced or metastatic KRAS G12C mutant solid tumors (CodeBreak-100; NCT03600883), ORR and DCR were 12.0% (3/25) and 80.0% (20/25), in patients with KRAS G12D -mutated mCRC [76]. Further confirmation is needed.

Colorectal Cancers, Mismatch Repair Deficiency, and Microsatellite Instability
While most CRCs develop via the chromosomal instability pathway (aneuploidy and loss of genetic material), 10-15% of them arise from the microsatellite instability (MSI) pathway. MSI is caused by the deficiency of the DNA mismatch repair (dMMR) system, resulting from a germline mutation in MMR genes (MLH1, PMS2, MSH2, MSH6) predisposing to Lynch syndrome or from an epigenetic inactivation of MLH1 (i.e., sporadic cancers). These sporadic cases are frequently associated with the BRAF V600E mutation [77]. MSI/dMMR is observed in approximately 10-15% of localized CRC and 4-5% of mCRC (see Figure 1) [43,78]. MSI/dMMR CRCs mainly arise from the proximal colon and display specific features such as poor differentiation, abundant tumor-infiltrating lymphocytes, and particular patterns of metastatic spread with frequent distant lymph node metastases and peritoneal involvement [49]. In localized CRC, MSI/dMMR is associated with favorable prognosis [79,80]. Data are more controversial in metastatic setting. Still, the available literature suggests MSI/dMMR mCRC are less responsive to conventional chemotherapy, compared to microsatellite stable/MMR-proficient (MSS/pMMR) tumors [43,[81][82][83].

Targeting the Immune System
MSI/dMMR has emerged as a major predictive biomarker for the efficacy of ICIs, especially for mCRC patients. While pMMR/MSS CRCs are mainly resistant to ICIs (i.e., cold tumors), MSI/dMMR tumors were associated with high sensitivity to immunotherapy (i.e., hot tumors). The activity of ICIs for patients with chemoresistant MSI/dMMR mCRC was demonstrated in several phase II trials, with objective response rates ranging from 33% to 58% and 12-month PFS rates between 31% and 71% [50,[89][90][91][92][93][94]. Importantly, results of the non-randomized CheckMate-142 trial suggest that combinations of anti-PD1 and anti-CTLA4 mAbs might be more effective than anti-PD1 or anti-PDL1 agents alone. Pembrolizumab and nivolumab +/− ipilimumab are approved by the Food and Drug Administration, for the treatment of patients with MSI/dMMR mCRC, who progressed after fluoropyrimidine, oxaliplatin, and irinotecan-based treatment.
The efficacy of ICIs was also demonstrated as front-line treatment for patients with chemotherapy-naive MSI/dMMR mCRC. In a third cohort of the CheckMate-142 trial, 45 patients were treated with nivolumab plus ipilimumab as the first-line treatment. ORR was 77%, the 1-year PFS estimate was 77% [95]. Of major importance, the phase III KEYNOTE-177 trial recently showed pembrolizumab monotherapy to be superior to standard of care first-line chemotherapy (investigator's choice of FOLFOX or FOLFIRI, with or without bevacizumab or cetuximab) in terms of PFS for MSI/dMMR mCRC patients. Medians of PFS were 16.5 months versus 8.2 months (HR = 0.60, 95%CI 0.45-0.80). The 12-and 24-month PFS rates were 55% and 48% with pembrolizumab versus 37% and 19% with chemotherapy. It is highly likely that pembrolizumab will become the standard of care for patients with newly diagnosed MSI/dMMR mCRC [96].
ICIs are currently evaluated for patients with localized MSI/dMMR colon cancer. The NICHE phase II study paved the way for their development in this setting and might refine therapeutic strategies for early-stage MSI/dMMR CC [97]. Indeed, in this trial testing nivolumab plus ipilimumab as neoadjuvant treatment, all 21 dMMR CC patients had a pathological response, with 95% of major responses including 12 complete pathological responses. These impressive results highlight neoadjuvant immunotherapy as a promising strategy that warrants further research. ICIs are also evaluated in combination with adjuvant chemotherapy for stage III MSI/dMMR colon cancer patients-the ATOMIC trial (NCT02912559; FOLFOX +/− atezolizumab), and the POLEM trial (NCT03827044; 24 weeks of single agent fluoropyrimidine chemotherapy or 12 weeks of oxaliplatin-based chemotherapy +/− avelumab) [28].

Predictive Biomarkers among MSI/dMMR CRC Patients Treated with Immunotherapy
There is currently no robust predictive biomarker for the efficacy of ICIs among MSI/MMR CRC patients. Despite the high clinical activity of ICIs in this population, some patients do not benefit from these treatments, and there is currently no clear explanation for these cases of primary resistance. It is noteworthy that misdiagnosis of MSI/dMMR status is responsible for a significant amount of cases with primary resistance to ICIs [98,99].
First, MSI/dMMR tumors harboring the BRAF V600E mutation seem to be quite sensitive to MSI/dMMR BRAF WT diseases [50]. Second, PD-1 expression, beta-2-microglobulin mutations, and major histocompatibility complex class I expression were not found to be associated with resistance to ICI [50,90,100]. Loss-of-function mutations in Janus kinases JAK1/2 might lead to primary resistance of MSI/dMMR mCRC to ICI [101]. Interestingly, the tumor mutational load was reported to predict ICI efficacy in 2 small cohort studies (<33 patients) [102,103]. Data are also emerging for the predictive impact of the immune infiltrate for the efficacy of ICIs. Indeed, in a recent work by Loupakis and colleagues, the level of T cell infiltration was linked to better response, PFS, and OS [98]. All these results deserve confirmation in larger prospective studies.
The diagnostic method for HER2 testing in CRC was standardized by the Heracles diagnostic criteria, with first-line immunohistochemistry (IHC) analysis followed, if appropriate, by FISH analysis (fluorescence in situ hybridization). Positivity is defined by an IHC 3+ score or an IHC 2+ score associated with FISH positivity [112].
Evidence is growing for the efficacy of anti-HER2 agents for patients with HER2-positive mCRC. Combinations of trastuzumab plus lapatinib, trastuzumab with pertuzumab, and trastuzumab plus tucatinib were tested in phase II studies (Heracles-A, MyPathway, and Mountaneer, respectively). They showed promising response rates of 30 and 32, and a 55% and median PFS of 4.7, 2.9, and 6.2 months, respectively [109,113,114]. Patients with HER2-positive and KRAS-mutated mCRC were excluded from the Mountaneer and Heracles-A trials, but it is worthy to note that one HER2-positive and KRAS-mutated mCRC patient experienced an objective response in the MyPathway study [109,113,114]. In contrast, the Heracles-B trial, with the association of pertuzumab and trastuzumab emtansine, failed to reach its primary endpoint (ORR), even if the median PFS was 4.7 months [115]. Recent report of the DESTINY-CRC01 phase II trial has brought to light trastuzumab-deruxtecan as a potential game changer. Fifty patients with chemoresistant HER2-positive mCRC were treated with this antibody-drug conjugate (composed of an anti-HER2 antibody and topoisomerase I inhibitor). The confirmed objective response rate was 45% (24/53). Even patients with prior anti-HER2 agents derived benefit from this treatment, with an objective response rate of 43.8% (7/16 pts). Two deaths from drug-related interstitial lung disease were reported.
Overall, anti-HER2 agents are highly seductive therapies for the HER2-positive population, but data from randomized trials are lacking in rigorous evaluation of their added value. The only ongoing randomized study is a phase II trial comparing trastuzumab and pertuzumab to irinotecan and cetuximab, in patients with HER2-positive RAS/RAF WT mCRC (NCT03365882).

NTRK Gene Fusions and TRK Inhibitors
NTRK gene fusions recently emerged as a highly seductive therapeutic target for cancer patients. TRK inhibitors (larotrectinib, entrectinib) demonstrated an impressive clinical activity in this population, regardless of histological type. In single-arm studies, larotrectinib showed an objective response rate of 75%, with a time of response superior to 6 months in 73% of cases, and entrectinib, an ORR of 57%, with a time of response superior to 6 months in 68% of cases [116,117]. These results led to accelerated approvals by the FDA for the treatment of refractory solid tumors with the NTRK gene fusion, whatever the tumor type.
Strategies for NTRK fusion screening are based on immunohistochemistry fluorescence in situ hybridization, RT-PCR (reverse transcription polymerase chain reaction) and next-generation sequencing, depending on the probability of NTRK fusion [118,119]. NTRK fusions are rare in colorectal cancers, with an incidence of 0.23-0.97% (see Figure 1) [48,[120][121][122][123]. It is, therefore, necessary to select the population to be screened.
Characteristics of NTRK fusion-positive CRC patients are female, right-sided primary tumor location, RAS/RAF WT status, and MSI phenotype [121]. Intriguingly, the MSI phenotype was consistently reported in association with NTRK fusions. More precisely, these genetic alterations seemed to occur in the context of BRAF WT tumors with hypermethylation of the MLH1 gene promoter [124]. The incidence of NTRK fusions was estimated at approximately 42% in this molecularly selected population [48]. There is currently no data about the efficacy of NTRK inhibitors or immune checkpoint inhibitors in this specific biological entity.

Conclusions
Substantial advances were made in the individualization of therapeutic strategies for mCRC patients over the last 10 years. Responders to anti-EGFR therapies can be specifically selected with an enlarged panel of biomarker, and the therapeutic strategies can be optimized with the longitudinal follow-up of ctDNA. Patients with BRAF V600E mCRC, who were long left behind, now have effective therapeutic options. Beyond highly seductive but quite rare targets, such as HER2 amplification and NTRK fusions, the most striking revolution for targeted therapies in CRC patients comes from ICIs that were a breakthrough for patients with MSI/dMMR tumors. Patients' outcomes were dramatically improved, and this enforced clinicians and researchers to conceptualize CRC as at least 2 distinct diseases-the MSI/dMMR tumors, and the others (Table 2). Importantly, the development of ICIs is associated with methodological issues in relation with the pseudoprogression phenomenon and long-term survivals. This observation highlights the necessity to develop new study designs and to anticipate such issues in pre-planned statistical analyses. This evolution brings to light the international CMS classification, that clustered CRC into 4 distinct consensus molecular subtypes (CMS): CMS1 (MSI Immune), hypermutated, microsatellite unstable, BRAF V600E mutated, with strong immune activation; CMS2 (Canonical), epithelial, chromosomally unstable, with marked WNT and MYC signaling activation; CMS3 (Metabolic), epithelial, with evident metabolic dysregulation; and CMS4 (Mesenchymal), prominent TGF-β activation, stromal invasion, and angiogenesis. The CMS classification might help identify new pathways to target, such as the TGF-β pathway or others. Although it is currently used in research investigation, the role of CMS classification for therapeutic strategies remains to be defined, and its feasibility in routine practice to be demonstrated.
To conclude, translational research seems more necessary than ever to understand the biological specificities of these various types of CRC, the innate and acquired mechanisms of resistance to targeted therapies, to bring progress to all patients.