Precision medicine for metastatic colorectal cancer in clinical practice

Abstract

Globally, metastatic colorectal cancer is one of the leading causes for cancer-related death. Treatment limited to conventional chemotherapeutics extended life for only a few months. However, advances in surgical approaches and medical treatment regimens have greatly increased survival, even leading to long-term remission in selected patients. Advances in multiomics analysis of tumors have built a foundation for molecular-targeted therapies. Furthermore, immunotherapies are on the edge of revolutionizing oncological practice. This review summarizes recent advances in the growing toolbox of personalized treatment for patients with metastatic colorectal cancer. We provide an overview of current multimodal therapy and explain novel immunotherapy and targeted therapy approaches in detail. We emphasize clinically relevant therapies, such as inhibitors of MAPK signaling, and give recommendations for clinical practice. Finally, we describe the potential predictive impact of molecular subtypes and provide an outlook on novel concepts, such as functional precision medicine.

Introduction

Colorectal cancer is the third most common cancer worldwide and the second leading cause of cancer-related death. ¹ In metastatic disease, the prognosis remains poor and most patients cannot be cured. ² In these patients, chemotherapy is still the mainstay of treatment. However, 5-fluorouracil (5-FU) has been in use since the 1960s and still represents the core of chemotherapy regimens in combination with oxaliplatin^3,4 and irinotecan.^5–7 These chemotherapies are further combined with antibodies against vascular endothelial growth factor (VEGF) signaling (bevacizumab,^8–10 ramucirumab, ¹¹ and aflibercept) ¹² or epidermal growth factor (EGF) receptors (cetuximab ¹³ and panitumumab) ¹⁴ in RAS wild-type (WT) disease. ¹⁵ In addition, trifluridine/tipiracil (TAS-102) ¹⁶ and regorafenib ¹⁷ are available as third-line/salvage therapy options. These treatments have led to considerable increase in patients’ survival to more currently than 30 months. ¹⁸ Furthermore, surgical resection has become a standard-of-care option for treating metastases. In fact, metastatic cancer was long believed to necessitate palliative therapy, while today up to 25% of patients with liver metastases have curative potential with 5-year survival of up to 50%. ¹⁹ Overall, survival rates have vastly improved with these multimodal concepts and long-term survival is observed in a considerable fraction of patients. ²⁰

Precision medicine aims to identify the ideal treatment for individual patients by considering the molecular characteristics and specific vulnerabilities of their disease. Different levels of molecular characterization, including immunohistochemical staining, polymerase chain reaction (PCR) tests, next-generation sequencing (panel sequencing, whole exome or whole genome sequencing, RNA-sequencing), and clinical characteristics of the patient, are taken into consideration. To this end, large-scale sequencing studies have revealed the landscape of molecular alterations present in colorectal cancer within the last decades.^21–24 The most frequent alterations in colorectal cancer, including APC, TP53, or most KRAS mutations, unfortunately cannot be exploited therapeutically yet. Nevertheless, there are several novel medical therapies targeting less frequent molecular alterations, novel immunotherapy strategies, and emerging concepts, such as molecular subtypes and functional precision medicine. In this review, we summarize the current state-of-the-art medical treatment for metastasized colorectal cancer and comprehensively discuss recent advances in precision medicine relevant to this disease, including molecular-targeted therapies and immunotherapies. We provide information on molecular backgrounds of novel therapies and emphasize applications that are relevant for clinical practice.

Current therapy principles for metastatic colorectal cancer

Resection of metastases and chemotherapy

The general condition of the patient and the ability to tolerate combination chemotherapy, the molecular factors RAS (KRAS, NRAS), BRAF and mismatch-repair status (MMR), and the location of the primary tumor (right- vs left-sided) factor into therapy planning. ²⁰ In addition, every patient must be evaluated by a specialist surgeon whether complete resection of all metastases can be achieved.

Patients with resectable metastases of the liver or lung should undergo surgery. Perioperative therapy is usually not performed since data suggest limited or no benefit.^25,26 However, ESMO guidelines suggest preoperative chemotherapy with an oxaliplatin-based regimen (FOLFOX or CAPOX) in patients with unfavorable or unclear prognostic factors, such as synchronous onset of metastases, high number of metastases, suspicion of extrahepatic disease, or high FONG-score.^20,27 In patients with potentially resectable metastases, a conversion therapy is indicated. As there are currently no clear criteria for potentially resectable disease, any patient should principally be considered and regularly reassessed during the course of treatment. Up to 40% of patients with liver metastases become resectable after conversion therapy ²⁸ and survival rates are favorable compared with chemotherapy alone, despite recurrence rates of up to 75%.^20,29 Since response rate is correlated to resection rate, ²⁸ a potent as possible regimen should be used. The exact regimen for this setting is not clearly defined, but usually a chemotherapy doublet (FOLFOX/FOLFIRI) with EGF receptor (EGFR) antibodies is recommended in RAS WT disease^30,31 and a doublet or triplet (FOLFOXIRI) with bevacizumab ³² in RAS mutated cases. ²⁰

Metastatic colorectal cancer patients with unresectable (‘never-resectable’) disease are treated in palliative intent with the goals of prolonging survival while keeping good quality of life. Exposure to all active therapeutical substances in combination and in a sequential manner is of importance according to the continuum-of-care concept, which leads to superior survival rates compared with best supportive care. ³³ The best possible, that is, most active and best tolerable therapy regimen should be given first. In addition, symptomatic patients may need a more intensive regimen to induce tumor shrinkage, while disease control with a well tolerable therapy is used in never-resectable patients with comorbidities or older patients. ²⁰ Examples for typical treatment courses of colorectal cancer patients with non-resectable metastases are shown in Table 1.

Table 1.

Examples for possible courses of treatment for colorectal cancer patients with non-resectable metastases (adapted from). ²⁰

	Example 1	Example 2	Example 3	Example 4	Example 5	Example 6
	RAS WT and left colon		RAS mutation or right colon			Older patient
First line	FOLFOX + EGFR-Ab	FOLFIRI + EGFR-Ab	CapOx + Bvz	FOLFIRI + Bvz	FOLFOXIRI + Bvz	5-FU + Bvz
Maintenance	5-FU + Bvz	Pause	Cap + Bvz	5-FU + Bvz	5-FU + Bvz	Pause
Second line	FOLFIRI + ramucirumab	CapOx + Bvz	FOLFIRI + aflibercept	FOLFOX + Bvz	FOLFIRI + ramucirumab	FOLFOX (reduced dose)
Third line	TAS-102	Irinotecan + EGFR-Ab	TAS-102	TAS-102	TAS-102
Fourth line	FOLFOX + EGFR-Ab	TAS-102	Regorafenib	Regorafenib

Choice of first-line treatment: chemotherapy, RAS status, and primary tumor location

A chemotherapy doublet with a fluoropyrimidine (5-FU/leucovorin or capecitabine) in combination with irinotecan (FOLFIRI) or oxaliplatin (FOLFOX) is standard of care for most patients. Both FOLFIRI and FOLFOX are equally effective^34,35 but have different side effect profiles. A higher frequency of gastrointestinal toxicity is observed with irinotecan, while peripheral neuropathy is a typical limiting side effect associated with oxaliplatin. 5-FU/leucovorin bolus regimens are more toxic than infusional regimens and have become obsolete. ³⁶ The oral 5-FU prodrug capecitabine can be used instead of i.v. 5-FU/leucovorin in combination with oxaliplatin, ³⁷ but is usually not combined with irinotecan due to higher toxicity than FOLFIRI. ³⁸ Furthermore, EGFR antibodies have not shown a survival benefit with CAPOX (vs CAPOX alone) in the COIN trial and may thus not be used with capecitabine-based regimens. ³⁹ Before starting treatment, dihydropyrimidine dehydrogenase (DPD) polymorphisms can be tested to potentially avoid severe fluoropyrimidine side effects by dose adjustment in affected individuals.^40–42 Testing has been implemented in several centers (including ours); however, it is to this point controversial according to NCCN ⁴³ or ESMO guidelines. ²⁰ The chemotherapy triplet (FOLFOXIRI) might be more effective than doublet combinations. This may also apply to the combination with bevacizumab. However, data are conflicting and increased toxicity has to be taken into consideration.^44–47 This regimen is therefore only used in selected patients without severe comorbidities and with high need for tumor shrinkage. The composition of chemotherapy (oxaliplatin vs irinotecan, infusional 5-FU vs capecitabine, doublet vs triplet) thus needs to be carefully tailored by the oncologist, taking treatment goals, patients’ disease, comorbidities, and preferences into account. In contrast, biological factors determine the choice of molecular antibodies added to chemotherapy. Both VEGF (bevacizumab) and EGFR (cetuximab, panitumumab) antibodies improve outcome when combined with chemotherapy.^8–10,13,14 While there is no predictive biomarker for bevacizumab, EGFR antibodies must only be administered in RAS WT disease. ¹⁵ According to data from the FIRE-3 trial, the PEAK trial, and a meta-analysis, treatment with EGFR antibodies and chemotherapy was favorable to bevacizumab with chemotherapy in RAS WT disease.^48–50 In addition, the location of the primary tumor (sidedness) plays a role for the choice of antibody treatment. Tumors located in the right hemicolon (coecum, ascending colon, and transverse colon) are often associated with specific histological and molecular characteristics [low differentiation or medullary morphology, high mucin production, more frequent BRAF mutations and microsatellite instability (MSI)] and a comparably poorer general prognosis. ⁵¹ In addition, EGFR antibodies seem to have no benefit in right-sided cancers in first-line therapy. According to retrospective analyses of clinical trials,^52,53 cetuximab with FOLFIRI was superior to FOLFIRI alone in left-sided disease (HR 0.65) but not in right-sided cases (HR 1.08). ⁵² FOLFIRI with cetuximab had longer overall survival (OS, 38.3 months) than FOLFIRI with bevacizumab (28 months) in left-sided RAS WT disease. In comparison, right-sided tumors had generally shorter OS with cetuximab in combination with FOLFIRI (18.3 months) and worse outcome than with bevacizumab (23 months, not significant). ⁵² Hence, bevacizumab can be used in first line in all patients independent of RAS and sidedness, but cetuximab or panitumumab is preferred in RAS WT cancers originating from the distal colon.

Maintenance therapy

After first-line induction therapy, patients who respond to treatment but cannot undergo resection are usually switched to maintenance therapy after 4–6 months because of the toxicity of combination regimens. This is of special importance in case of oxaliplatin-based therapies, due to peripheral neuropathy. Maintenance strategies after induction therapy with oxaliplatin-based regimens have consequently been investigated in several trials. It was shown that a less toxic maintenance treatment with a fluoropyrimidine together with bevacizumab led to favorable progression-free survival (PFS) compared with drug holidays and to similar PFS compared with continuous treatment. The data on OS were less clear, trending toward an advantage of maintenance therapy versus drug holidays.^54–58 However, at least according to one meta-analysis, no clinically significant reduction of OS was reported for intermittent versus continuous treatment strategies. ⁵⁸ Hence, maintenance therapy is generally recommended for most patients, but treatment discontinuation may be discussed with the patient in selected cases.

Later-line treatments and therapy sequence

A second-line treatment is generally recommended for most patients, as it has been associated with prolonged survival. ⁵⁹ Therapeutic strategies are based on patient- and disease-related factors, and the previous therapy (reviewed in detail by Giordano et al. ⁶⁰ ). Regarding the chemotherapy backbone, usually a switch is recommended: patients who received irinotecan (+5-FU) switch to oxaliplatin (and fluoropyrimidine) and patients who received oxaliplatin (and fluoropyrimidine) switch to irinotecan together with 5-FU. ³⁴ Irinotecan (without 5-FU) may also be given as chemotherapy in second- or later-line treatments if appropriate. With respect to antibody therapy, patients having received anti-EGFR antibodies in first line usually switch to anti-angiogenesis in second line. Bevacizumab has been shown to be effective in second line in combination with chemotherapy and can be given beyond progression in first line.^61–63 Other anti-angiogenic strategies in second line include aflibercept ¹² or ramucirumab ¹¹ in combination with FOLFIRI. Anti-EGFR antibodies have also shown activity in second line in RAS WT disease, both as single agents and together with chemotherapy,^64,65 but are not continued after treatment failure in first line. As mentioned above, the sequence EGFR antibody (first line) followed by VEGF(R) antibody (second line) seems to be favorable over VEGF➔EGFR in RAS WT disease^66,67 Sidedness may also be predictive for EGFR antibody response in second and later therapy lines,^68,69 but evidence is currently less definte, ⁷⁰ so that cetuximab or panitumumab may be given in right-sided cases in later treatment lines at this point. ⁴³

‘Last line’ therapy options include nucleoside-analog TAS-102¹⁶ and multi-tyrosine kinase inhibitor regorafenib. ¹⁷ Combination therapies of TAS-102 with anti-angiogenic antibodies have shown promising results in phase II clinical trials,^71,72 and combinations are evaluated also in earlier treatment lines. TAS-102, alone or in combination with bevacizumab, is therefore a recommended treatment option in patients who have progressed beyond standard therapies. ⁴³ Beyond or alternatively, rechallenge and reintroduction remain frequently used options in the later-line setting of patients with good performance status. Reintroduction is the administration of a former treatment regimen the patient benefited from, but that was terminated (mostly due to toxicities), while rechallenge is the readministration of a treatment regimen toward the patient has previously developed a resistance. ⁷³ Data supporting these concepts are, however, limited. Further options, including immunotherapies or targeted therapies based on molecular characterization and molecular tumor boards, have become available recently for selected patients. These will be described in the following paragraphs.

MSI and checkpoint inhibitors

Targeting the immune system has become the third mainstay of medical oncology next to chemotherapy and targeted therapies. Clinically approved concepts are based on antibodies directed against immune checkpoints, such as the programmed cell death 1 (PD-1), its ligand (PD-L1), or cytotoxic T-lymphocyte-associated protein 4 (CTLA4). These checkpoints transduce inhibitory signals to T cells leading to immune evasion of cancers, while antibodies directed against them conversely disinhibit T cell function leading to cancer cell killing. ⁷⁴ Response to immune checkpoint antibodies has been hypothesized to be associated with the number of mutations present in cancers, the latter leading to expression of neo-epitopes increasing the cancers’ visibility for the immune system. ⁷⁵ A high tumor mutational burden (TMB) is also associated with T cell infiltration and PD-L1 expression in tumors. ⁷⁶ The majority of colorectal cancers harbor only low to moderate numbers of mutations; however, there is a hypermutated subgroup of colorectal cancer encompassing roughly 5% of metastatic tumors. ⁷⁷ The majority of these hypermutated tumors are characterized by defects in the DNA mismatch-repair (MMR) system leading to MSI. ²² Le et al. ⁷⁸ recruited patients with colorectal cancers and other tumors with MMR deficiency (dMMR) and MMR proficient (pMMR) colorectal cancers in a phase II study to test clinical activity of the checkpoint inhibitor pembrolizumab. In this seminal study, they found an impressive 40% response rate and 78% 20-week PFS rates for dMMR tumors (vs 0% and 11% in pMMR colorectal cancers). After further studies in colorectal and non-colorectal dMMR tumors confirming durable response with this therapy, the Food and Drug Administration (FDA) issued morphology-agnostic approval of pembrolizumab in dMMR cancers in 2017.^79–82 In addition, nivolumab, another PD-1 antibody, has shown durable responses (response rate 31%, disease control 69%) and 73% 12-month OS in pretreated dMMR colorectal cancer patients according to the phase II CheckMate 142 trial. ⁸³ The combination of nivolumab with CTLA4 antibody ipilimumab further increased efficacy (response rate 55%, disease control rate 80%, 12-month OS 85%) in another arm of the same trial. ⁸⁴ These results prompted FDA approval of nivolumab (2017) and nivolumab/ipilimumab (2018) for pretreated dMMR metastatic colorectal cancer. Other checkpoint inhibitors, including PD-L1 antibodies atezolizumab, durvalumab, and avelumab, have also been studied in early clinical trials and have shown encouraging results in dMMR colorectal cancer.^85–87 Checkpoint inhibitors have been investigated in the first-line setting of metastatic dMMR/MSI colorectal cancer. The combination of nivolumab and ipilimumab showed 24-month PFS and OS of 74% and 79%, respectively, while median PFS and OS were not yet reached after median follow-up of 29 months. ⁸⁸ Pembrolizumab was tested in the first-line setting against investigators’ choice combination chemotherapy (FOLFOX/FOLFORI ± bevacizumab or cetuximab) in the Keynote-177 phase III trial (NCT02563002). ⁸⁹ In the updated analysis, the pembrolizumab group had a lower risk of death, although significance was not reached (HR 0.74). Median PFS was 16.5 months in the pembrolizumab group versus 8.2 months in the chemotherapy group, while the overall response rate (ORR) was also higher in the pembrolizumab group (45.1% vs 33.1%). In addition, adverse events were significantly lower in the pembrolizumab group. ⁹⁰ Hence, checkpoint inhibitor therapy is becoming the standard of care in first-line setting of dMMR colorectal cancer and dMMR/MSI testing must be done in every colorectal cancer patient.

Immunotherapy for non-dMMR cancers?

While immunotherapy is becoming standard of care in dMMR colorectal cancer, the majority (approx. 95%) of metastatic colorectal cancers are characterized by pMMR/microsatellite stable (MSS) status. In these patients, efficacy of immunotherapy as monotherapy has been disappointing. In addition, not all patients with dMMR disease respond to checkpoint inhibitors. Therefore, current studies aim to identify predictive markers to improve patient stratification and to establish combination strategies for improving efficacy, especially in patients with pMMR tumors. Several studies are testing combination strategies of checkpoint inhibitors with chemotherapy, radiation, anti-VEGF antibodies, anti-EGFR antibodies, inhibitors of mitogen-activated protein kinase (MAPK) signaling, or multi-tyrosine kinase inhibitors aiming to turn immunologically ‘cold’ tumors into ‘hot’ tumors and thereby making them susceptible to immunotherapies (reviewed in Pecci et al. ⁹¹ and Hirano et al., ⁸⁰ Figure 1).

Figure 1.

Combination strategies to enhance immune checkpoint inhibitor efficacy in pMMR colorectal cancers tested in current clinical trials.

According to preliminary results, combinations of checkpoint inhibitors with VEGF antibodies or multi-tyrosine kinase inhibitors (also targeting VEGF receptors) appear promising. Preclinical studies in different tumor types have reported on endothelial-mediated (or VEGF-mediated) immunosuppression within tumors, which may be overcome with this dual strategy.^92,93 In addition, targeting VEGF may improve T cell infiltration into the tumor microenvironment.^94,95 Accordingly, combination strategies have been successful in hepatocellular carcinoma, renal cell cancer, or lung cancer, also in combination with chemotherapy.^96–98 In colorectal cancer, the BACCI phase II trial assessed atezolizumab in combination with bevacizumab and capecitabine versus placebo with bevacizumab and capecitabine in 133 pretreated metastatic colorectal cancer patients. ⁹⁹ The combination showed a modest (but statistically significant) benefit in PFS of 4.4 months versus 3.3 months, thereby reaching its prespecified primary end point. The phase I REGONIVO trial assessed regorafenib with nivolumab in 50 pretreated patients with colorectal or gastric cancers (25 patients each). In the colorectal cancer cohort, 36% objective response and a PFS of 7.9 months was observed with a manageable safety profile. This was considered an encouraging result warranting further study in larger trials. ¹⁰⁰ In contrast, the REGOMUNE trial, testing regorafenib and avelumab in pMMR colorectal cancers reported no objective responses (stable disease in 54% of 48 patients), but the authors noted recruitment of antitumor immunity in a subset of patients, supporting the general concept. ¹⁰¹ Currently ongoing studies also test checkpoint inhibitors together with bevacizumab and chemotherapy in first-line setting. The NIVACOR ¹⁰² and ATEZOTRIBE (NCT03721653) trials, for instance, investigate the combination of intensified chemotherapy with FOLFOXIRI together with bevacizumab and checkpoint inhibition (nivolumab or atezolizumab, respectively). According to preliminary results from NIVACOR, the combination was generally well tolerated, and efficacy data are pending.

The importance of MAPK signaling in colorectal cancer therapy is discussed in detail below. With respect to immunotherapy, the pathway has been implicated in tumor–immune interaction in multiple ways, including regulation of immunosuppression through cytokines and growth factors, regulation of human leukocyte antigen (HLA) expression and thereby immune cell evasion from the tumor microenvironment. ⁹¹ The combination of atezolizumab with MEK inhibition using cobimetinib, however, led to disappointing efficacy in the phase III IMblaze370 trial, analyzing 363 pretreated patients: the combination led to an OS of 8.9 months compared with 7.1 months for atezolizumab alone and 8.5 months with the control treatment regorafenib. ¹⁰³ Clinical trials combining checkpoint inhibitors with EGFR antibodies cetuximab and panitumumab with and without chemotherapy are ongoing. ⁸⁰ Cetuximab has previously been shown to induce antibody-dependent cell-mediated cytotoxicity, which could synergize with checkpoint inhibitors. ¹⁰⁴ For instance, the AVETUX phase II trial tested the combination of avelumab and cetuximab together with FOLFOX in the first-line setting. Overall, 39 patients of the intention-to-treat (ITT) cohort reached a 79.5% ORR and 92.3% disease control rate, thus a randomized trial appears feasible. ¹⁰⁵

In summary, while immune checkpoint inhibitor therapy has become standard of care in dMMR/MSI colorectal cancer, finding combination therapies for enhancing efficacy in dMMR cases remains challenging. Results from larger studies pertaining this group are awaited in the next months and years. Higher-order combinations of checkpoint inhibitors with targeted therapies and chemotherapeutics increase the chance of efficacy by elevating the probability that combinations include an effective drug. ¹⁰⁶ This strategy, however, may lead to enhanced side effects, demanding for more extensive clinical and preclinical biomarker discovery for better stratification of patients.

Targeting the MAPK pathway

The EGF/MAPK pathway is an intercellular and intracellular signal transduction cascade that regulates a plethora of processes, most importantly not only proliferation, cell growth, and apoptosis but also metabolism or migration of cells and others (Figure 2). The ligands, such as EGF, bind to the human EGF 1–4 (HER1–4 or ERBB1–4) family of transmembrane receptors. HER1 is also referred to as EGFR. Ligand binding leads to dimerization of the receptor, either as homo- or heterodimer (e.g. HER1 with HER2), followed by a cross-phosphorylation of the intracellular receptor domains. Prompted by this phosphorylation, Src homology and collagen (SHC) binds to the receptor and associates growth factor receptor binding protein 2 (GRB2) that in turn recruits son of sevenless (SOS) from the cytosol.^107–109 SOS is a guanine exchange factor that catalyzes the exchange from RAS-GDP to RAS-GTP, thereby activating the RAS protein.^110–112 Activated RAS-GTP in turn directly activates downstream components of the MAPK pathway, for example, RAF, PI3K, and several other effectors. ¹¹³ The rapidly accelerated fibrosarcoma (RAF) protein family includes BRAF ¹¹⁴ and is activated by RAS-GTP through direct interaction. ¹¹⁵ Active RAF phosphorylates and thereby activates the dual-specificity protein kinases MEK1 and MEK2. ¹¹⁶ They again phosphorylate and activate their substrates ERK1 and ERK2 (extracellular signal-regulated kinase), which phosphorylate and activate several cytosolic and nuclear substrates, activating transcription of proteins that enhance and promote proliferation, growth, evasion of apoptosis and affecting various cellular functions, including metabolism, migration, angiogenesis, and immune regulation. ¹¹⁷ Based on these physiological functions, the MAPK pathway plays a prominent role in the development of cancer.

Figure 2.

Targeting the MAPK pathway in colorectal cancer. Druggable receptors and intracellular signaling components of the pathway are depicted. Drugs targeting the pathway discussed in the text are highlighted.

In colorectal cancer, MAPK pathway alterations are very common (59% of non-hypermutated and 80% of hypermutated cases) and most prominently affect KRAS, NRAS, or BRAF. ²² Several antibodies and small molecule inhibitors of the pathway have been developed that are being tested in clinical trials or entered clinical use (Figure 2). As previously discussed, EGFR antibodies, cetuximab and panitumumab, are included in standard-of-care first-line therapy regimens for RAS WT, left-sided colorectal cancer. Small molecules targeting EGFR are also available, but do not play a role in clinical use as they were found to be less effective while having enhanced toxicities in colorectal cancer. In the next paragraph, we will highlight the most important developments for clinical management of cancers with alterations in HER2, BRAF, and KRAS, for which direct targeting has become a clinically meaningful option.

HER2-targeted therapy

The HER2 (ERBB2) receptor is similar to EGFR in its structure and function, inducing downstream signaling particularly through MEK/ERK and PI3K/AKT, which leads (among other functions) to mitogenic stimuli, cell growth, and survival. ¹¹⁸ Inhibiting this tumorigenic function with antibodies and small molecules has become a standard treatment in breast cancer and gastric cancer. Its broad activation in several other cancers, including that of the colorectum, has also made it an emerging treatment option for these diseases. ¹¹⁹ The activation of HER2 signaling within the oncogenic process depends on overexpression due to gene amplification in the majority of cases.^120,121 Specifically, this leads to the formation of homodimers or heterodimers of the receptor with other HER family members and subsequent activation of downstream pathways. ¹²² While several mutations in HER2 have been found in large-scale sequencing projects, ¹²³ not all of them seem to be activating.^119,120,124 In histopathological assessment, usually a strong expression (3+) in immunohistochemistry (IHC) or a 2 + expression together with a detection of amplification in in situ hybridization (ISH) is considered as HER2-high and predictive of response to HER2 therapies. The association of lower HER2 expression levels (2 + expression in IHC without amplification in ISH or 1 + expression) or HER2 mutations with treatment response is less clear. ¹²⁵ Trastuzumab was the first HER2 antibody clinically approved after showing significantly improved OS in combination with chemotherapy in HER2 positive breast cancer and has since then emerged as a standard-of-care in this disease. ¹²⁶ The results of the TOGA trial ¹²⁷ also led to approval of trastuzumab in combination with chemotherapy in HER2 positive metastatic gastric and gastroesophageal junction cancers. Other drugs targeting HER2, which are approved in breast cancer and are being studied in other solid tumors, include pertuzumab (an antibody with a different binding site than trastuzumab), and small molecules lapatinib (targeting both HER2 and EGFR) and neratinib (inhibitor of HER1, 2, and 4).^119,120 While the frequency of HER2 overexpression is reported to be 20–25% in breast cancer and 10–15% in gastroesophageal adenocarcinoma, only about 3–5% of colorectal cancers harbor HER2 amplifications.^119,128 Nevertheless, effective targeting of HER2 in breast and gastroesophageal cancer has led to efforts to determine if it is exploitable as a target in colorectal cancer. ¹²⁹ Interestingly, RAS WT HER2 overexpressing colorectal cancer has been associated with poor prognosis and resistance toward EGFR antibody treatment. ¹³⁰ To this end, mouse xenograft studies suggested a compensatory upregulation of HER2 on EGFR antibody treatment, revealing an acquired resistance mechanism toward EGFR-targeted therapy through HER2. ¹³¹ This suggested a potential salvage therapy option for advanced colorectal cancer patients refractory to EGFR antibodies. Accordingly, the HERACLES phase II trial tested a combination of trastuzumab and lapatinib in HER2 positive metastatic colorectal cancer refractory to standard of care (including cetuximab or panitumumab). ¹³² However, 27 of 914 screened patients were eligible for trial and 8 of those (30%) had an objective response, while 12 more patients (44%) had stable disease (disease control rate 74%). Median PFS was 5.2 months, OS was 11.5 months, and the treatment was well tolerated in this heavily pretreated group (median of five previous therapy lines). Consequently, these results were rightfully interpreted as very promising. Further substantiating these results, the subgroup of colorectal cancers with HER2 activation (amplification, overexpression, or activating mutation) included in the MyPathway basket trial also showed a notable objective response rate of 38% (14 of 37 patients) and a PFS of 5.6 months. ¹³³ More recently, HER2 antibody drug conjugates, such as trastuzumab emtansine^134,135 or trastuzumab deruxtecan, ¹³⁶ have been introduced. Early basket clinical trials for different pretreated tumor entities, including colorectal cancer, had promising results.^137,138 In colorectal cancer, however, the HERACLES-B trial, testing trastuzumab emtansine with pertuzumab in 31 heavily pretreated patients, did not reach the predefined efficacy end point of ORR (9.7%, estimated 30%), but led to stable disease in 67.7% of patients and a PFS of 4.1 months. Nausea and fatigue were the most frequent adverse events. Most recently, data from the phase II DESTINY-CRC01 trial, evaluating trastuzumab deruxtecan in pretreated metastatic colorectal cancer patients, were published. ¹³⁹ Overall, 78 patients were enrolled in this study, 58 of those were with HER2 3 + or 2 + and ISH positive cancers. In these patients with the highest HER2 expression status, an objective response was observed in 24 patients (45.3%). Neutropenia and anemia were the most frequent adverse events, interstitial lung disease, or pneumonitis occurred in 5% of patients. Thus, especially trastuzumab deruxtecan appears to be a promising agent for HER2 amplified colorectal cancer. It will be interesting to see more survival and quality of life data with this drug soon. In addition, further studies should test the impact of different HER2 mutations on efficacy of HER2-targeted therapy.

We can conclude that targeting of HER2 in RAS-WT colorectal cancer refractory to standard therapy is a viable treatment option on basis of the HERACLES and the MyPathway trial results. We therefore recommend HER2 analysis in a fresh tumor biopsy and subsequent treatment of HER2 positive patients, ideally within a clinical trial.

Treatment of BRAF mutant cancers

BRAF mutations can be found in 5–10% of colorectal cancers, predominantly the V600E type. ¹⁴⁰ This mutation increases the catalytic activity of BRAF and reduces autoinhibition, thereby leading to a constitutive activation of the MAPK pathway. ¹⁴¹ The BRAF V600E mutation is associated with poor prognosis and generally a poor response to therapy. ¹⁴² Further typical features of BRAF V600E-mutated colorectal cancers are right-sided tumor location, poor tumor differentiation, peritoneal metastases,^143,144 and hypermethylation, MSI, and CMS subtype 1 (compare below). ¹⁴⁵ Surprisingly, BRAF mutations other than V600E might be associated with better survival. ¹⁴⁶

Due to the association with poor response and prognosis, treatment of BRAF V600E-mutated colorectal cancer is challenging. Resection of metastases is controversially discussed in this subgroup and should only be performed after careful patient selection.^147,148 An intense regimen is often recommended for fit patients in first line, usually a cytotoxic triplet (FOLFOXIRI) in combination with bevacizumab. ²⁰ This recommendation is mainly based on the results of the phase III TRIBE trial comparing an irinotecan-based cytotoxic doublet (FOLFIRI) plus bevacizumab to a triplet (FOLFOXIRI) plus bevacizumab. In a subgroup analysis of 28 BRAF V600E cases, both OS (19.0 vs 10.7 months) and PFS (7.5 vs 5.5 months), were better in the FOLFOXIRI arm. Significance was not reached in both cases, potentially due to the small sample size. ⁴⁷ In the following phase III TRIBE-2 study, the experimental group received FOLFOXIRI plus bevacizumab as first-line induction therapy, while the control arm received FOLFOX plus bevacizumab. Maintenance treatment consisted of 5-FU plus bevacizumab in both groups. After progression, the experimental group received FOLFOXIRI plus bevacizumab again, while the control group was treated with FOLFIRI plus bevacizumab. In a post hoc subgroup analysis, the strong benefit from the TRIBE study could not be confirmed, which was explained by the different treatment regimen in the control group. ¹⁴⁹ In an individual patient data, meta-analysis of five trials comparing FOLFOXIRI plus bevacizumab with doublet plus bevacizumab, no benefit for the triplet in the BRAF-mutated subgroup was observed. This was assumed to be caused by the different doublet therapy regimen in TRIBE (FOLFIRI) versus all other trials (FOLFOX). It can be concluded that FOLFORIXI plus bevacizumab should no longer be the first choice for BRAF V600E-mutated patients, but FOLFOX plus bevacizumab indeed seems to be preferrable. ¹⁵⁰

In second line, due to the association of BRAF mutations with MSI, some patients benefit from immune checkpoint inhibitors, so that MMR and MSI should always be tested. ⁸⁹ Targeted therapy has become another option for the remaining majority of patients recently. Over the last decade, several BRAF-inhibitors have been established, including vemurafenib, dabrafenib, or encorafenib. These inhibitors showed encouraging response rates in advanced BRAF V600E-mutated non-small-cell lung cancers^151–153 and improved survival of patients with advanced melanoma bearing a BRAF V600E mutation.^154–156 Attempts to target BRAF V600E in metastatic colorectal cancer as monotherapy have been largely unsuccessful, though. In a phase I dose expansion study with encorafenib, about two-thirds of 18 patients showed stable disease but no patient responded. ¹⁵⁷ In a phase II study of 21 patients treated with vemurafenib, 1 patient had a partial response, while one-third of patients had a stable disease. ¹⁵⁸ Similar results were obtained in a basket trial of vemurafenib treatment in BRAF-mutated non-melanoma cancers: no objective response was detected in 10 colorectal cancer patients, 50% had stable disease. These rather disappointing results can be explained by a feedback activation of EGFR as response to BRAF V600E inhibition, which, however, could be overcome by EGFR inhibition in preclinical models.^159,160 This combination of vemurafenib and cetuximab was also tested in the basket trial mentioned above. An increase in response rate to 4% and disease control rate to 73% was noted. ¹⁵¹ Further drugs have been added to the combination of BRAF and EGFR inhibition in clinical trials to improve efficacy. The cytotoxic agent irinotecan was tested in combination with cetuximab and vemurafenib in a phase I study leading to a response rate of 35% and stable disease in 53% of patients. ¹⁶¹ The combination of encorafenib and cetuximab was tested with and without the PI3K inhibitor alpelisib in another phase I study. ORR was 19.2% in the group without alpelisib and 17.9% with alpelisib, while stable disease was noted in 57.7% and 75.0%, respectively, leading to a disease control rate of 76.9% versus 92.9%. ¹⁶² Another study compared different combinations of the MEK inhibitor trametinib (T), the EGFR antibody panitumumab (P), and the BRAF inhibitor dabrafenib (D). Response rates were 0% for the T + P group, 10% for the D + P group, and 21% for the triple combination of T + P + D, underlining the stronger effect of combinations of inhibitors of the MAPK pathway. ¹⁶³

Based on the promising results of double and triple combinations, the open-label, randomized phase III BEACON trial recruited 665 patients with BRAF V600E metastatic colorectal cancer that had previously received one or two therapy regimens. ¹⁶⁴ Patients received either a triple combination of encorafenib (BRAF), cetuximab (EGFR), and binimetinib (MEK), or a doublet combination of encorafenib and cetuximab. The control group received either irinotecan or FOLFIRI combined with cetuximab according to investigators’ choice. Median OS (primary end point) was 9.0 months within the triplet group and 8.4 months in the doublet group versus 5.4 months in the control group. Compatible results were obtained regarding the response rates: objective response rate was 26% for the triplet group, 20% for the doublet group, and 2% for the control group. Interestingly, the experimental arms were also superior regarding adverse events of grade 3 or higher: 58% triplet, 50% doublet, and 61% control. These results led to FDA and European Medicines Agency (EMA) approval for the doublet combination of encorafenib and cetuximab as second-line treatment regimen for metastatic BRAF V600E-mutated colorectal cancer, thereby changing current clinical practice. Based on the results of the BEACON trial, the single-armed phase II ANCHOR trial aims to explore encorafenib, binimetinib, and cetuximab as a first-line therapy regimen. During the phase I stage of this trial, an objective response rate of 50% and a median PFS of 4.9 months could be observed, so that phase II within this trial was initiated and results are expected soon. ¹⁶⁵ The observed PFS is in the range of current standard therapy FOLFOXIRI + bevacizumab. ⁴⁷ If efficacy and safety prove to be encouraging, a phase III study will be needed to establish the targeted regimen as first-line treatment. The phase III BREAKWATER (NCT04607421) trial is currently evaluating encorafenib and cetuximab ± cytotoxic chemotherapy as first-line therapy for BRAF V600E metastatic colorectal cancer patients. In the safety lead-in stage, 60 patients will receive either encorafenib with cetuximab and FOLFIRI or encorafenib with cetuximab and FOLFOX. In the phase III part, encorafenib and cetuximab or the combination of encorafenib with cetuximab and either FOLFIRI or FOLFOX will be administered in experimental arms. The control group will be treated with a chemotherapy doublet or triplet with or without bevacizumab according to investigators’ choice. Enrollment started in January 2021, and the study is expected to be completed in 2026, hopefully enlightening the question of the best first-line therapy for BRAF V600E positive colorectal cancer patients. ¹⁶⁶

In conclusion, novel treatment options for BRAF V600E metastatic colorectal cancer patients have improved survival and response rates. We recommend testing of dMMR/MSI in all patients to evaluate the option of checkpoint inhibitor treatment. For other patients, FOLFOX with bevacizumab is the current first-line therapy of choice, followed by encorafenib and cetuximab in second line. The addition of established cytotoxic agents could further improve this therapy strategy and may move it to the first-line setting.

KRAS inhibitors – drugging the undruggable?

While BRAF mutations are found in only around 5% of metastatic colorectal cancers, RAS mutations are present in more than 40%, and in up to 20% of all cancer types. ¹⁶⁷ Therefore, targeting this driver gene could potentially benefit a significantly larger number of patients. Of the three different isoforms of RAS, KRAS mutations are predominant in colorectal cancer with around 40%, NRAS mutations occur in approximately 5%, and HRAS plays an insignificant role in colorectal cancer. ¹⁶⁸ Oncogenic driver mutations occur most often in codons 12, 13, and 61, leading to a lower rate of guanosine triphosphate (GTP) hydrolysis and thereby constitutively activating the pathway. Besides decreased GTPase activity, a change in the affinity to downstream targets, which differs between the mutations, can also play a role in the activity of the mutation. ¹⁶⁹ KRAS mutations are associated with worse OS and inferior response toward EGFR antibodies (compare above). Directly targeting the RAS protein, however, is difficult given its high affinity to GTP and its lack of hydrophobic pockets, limiting the binding capacity for small molecules. RAS has therefore been considered an undruggable gene for decades. ¹⁷⁰ The most promising attempt was the development of allele-specific inhibitors of the G12 C mutation in the last years. This is based on the relatively strong GTPase activity in G12 C-mutated KRAS compared with other KRAS mutations. ¹⁶⁹ In a fragment-based screen, a pocket was found in which a compound could bind to a reactive cysteine, stabilizing the guanosine diphosphate (GDP) bound (inactive) state of KRAS and thereby leading to decreased KRAS activity. ¹⁷¹ One big advantage of this mode of action is the selectivity toward mutated RAS because only the mutated protein contains a cysteine. ¹⁷¹ Two inhibitors of the G12 C-mutated KRAS have been investigated in clinical trials: Sotorasib (AMG510) was the first G12 C-specific KRAS inhibitor tested in a basket trial of 129 patients, 42 of them with colorectal cancer. In the colorectal cancer subgroup, 7.1% had a confirmed response while the disease control rate was 73.8%. ¹⁷² Another inhibitor, adagrasib, was tested in a phase I/II trial as monotherapy or in combination with cetuximab. In 45 colorectal cancer patients evaluable for analysis with adagrasib monotherapy, 22% had a confirmed response while the disease control rate was 87% and PFS was 5.6 months. The combination therapy with cetuximab led to 43% response rate and 100% disease control rate in 28 patients^173,174 Thus, both studies showed promising activity in pretreated patients. Further G12 C inhibitors JNJ-74699157, GDC-6036, and JDQ433 are currently being tested in similar trials with colorectal cancer patients (NCT04006301, NCT04449874, and NCT04699188). In some of these trials, combinations with EGFR antibodies or immune checkpoint inhibitors are evaluated since in vitro data have shown that G12 C inhibitors are able to increase the effect of targeted therapy and immunotherapy. ¹⁷⁵

Since the G12 C mutation does only occur in about 3% of colorectal cancers, there is still a lack of therapies for more common KRAS mutations.^175,176 A possible solution could be inhibitors that prevent the interaction between the KRAS molecule and SOS, which is mainly responsible for RAS activation. This principle was shown to be effective in vitro in fragment-based screens. By binding into a certain pocket in the KRAS protein, the interaction with SOS is prevented, leading to significant decrease in KRAS activity.^177,178 Similar results were obtained for the KRAS inhibitor BI-2852 that could bind to KRAS in nanomolar levels. ¹⁷⁹ The compound BAY-293 was found to specifically inhibit the interaction of KRAS and SOS1 at picomolar concentrations, suggesting it as a promising candidate for further investigation. ¹⁸⁰ Another SOS1 interaction inhibitor, BI1701963, is currently under investigation in two clinical trials. First, in combination with the MEK inhibitor trametinib in different solid tumors with KRAS mutation (NCT04111458), and second, in combination with irinotecan specifically in colorectal cancer patients (NCT04627142). The most important challenge of this approach is the non-specificity of the compounds toward mutant KRAS.^177,179,180 BAY-293, for example, inhibits the proliferation of KRAS WT cells in lower concentrations than in KRAS mutant cells, ¹⁸⁰ therefore the toxicity profiles of the compounds in ongoing clinical trials are awaited with great interest.

Targeting WEE1 in KRAS- and TP53-mutated tumors

Adavosertib is an inhibitor of WEE1, a tyrosine kinase that is involved in cell-cycle regulation. ¹⁸¹ Based on preclinical considerations of a potentially increased sensitivity of tumors with TP53 and KRAS mutations due to DNA replication aberrations, the drug was tested as maintenance therapy in Arm C of the FOCUS4 platform trial. This setting was likely inspired by trials of other DNA damage response targeting agents, such as PARP-inhibitors, that have, for instance, recently been approved as maintenance therapy after platinum-based induction in BRCA-mutated pancreatic cancer. ¹⁸² Patients with stable disease or response to first-line induction chemotherapy (approx. 2/3 platinum-based) were randomized to receive adavosertib (N = 44) or active monitoring (N = 25). Adavosertib improved the primary end-point PFS (median 3.61 vs 1.87 months), thereby making it an interesting candidate for a larger phase III trial, even though no improvement of OS was observed in this cohort. Other efforts to therapeutically exploit DNA damage response in colorectal cancer have been investigated in preclinical studies and in a few trials testing PAPR inhibitors (olaparib, veliparib) as mono- or in combination therapies ¹⁸³ No meaningful clinical benefit has been reported in patient cohorts so far. These were, however, mainly unstratified with respect to homologous recombination defects. Further studies are needed to find out if translation of targeting DNA damage response into the clinical setting of colorectal cancer treatment will be successful.

Targeting rare cancer drivers

Recently, molecular characterization of tumors by whole exome and even whole genome sequencing has greatly increased our knowledge of recurrent mutations and structural aberrations, also in colorectal cancer.^{22–24,184,185} Accordingly, several studies are testing the efficacy of specific inhibitors in mainly tumor-type agnostic ‘precision medicine’ trials and molecular tumor boards.^186–192 However, clinical benefits of this genomics-based therapy stratification remain unproven. To date, only a small fraction of oncogenic driver genes with low prevalence can be targeted by specific inhibitors; hence, only few patients with tumors harboring these alterations can benefit. Table 2 summarizes currently available predictive markers for precision therapy of colorectal cancer. As discussed above, RAS mutations and right-sided primary tumor location represent negative predictive factors for EGFR antibody treatment, while MSI-high, HER2 amplification, and BRAF V600E mutations are predictive for response toward specific treatments.

Table 2.

Therapy targets and molecular predictive markers in metastatic colorectal cancer.

Target/marker	Frequency (%)	Drug(s)	Type of marker
BRAF V600E	5	Encorafenib + cetuximab	Positive predictive
MSI/dMMR	5–10	Checkpoint inhibitors	Positive predictive
NTRK	<1	Entrectinib, larotrectinib	Positive predictive
HER2 amplification	2.5	Trastuzumab + pertuzumab, Trastuzumab deruxtecan, Trastuzumab emtansine	Positive predictive
KRAS/NRAS	40	Cetuximab, panitumumab	Negative predictive
Right colon	20	Cetuximab, panitumumab	Negative predictive

In addition, neurotrophic receptor tyrosine kinase (NTRK) gene fusions, very rare targetable alterations, can be found in colorectal cancer. This family of NTRK1–3 genes is normally involved in physiological regulation of the nervous system; however, fusion of the kinase domain of these genes with a variety of partners (including RET, MET, ERBB2, FGFR, and others) in chromosomal rearrangements can lead to potent and ligand-independent cancer drivers (reviewed here). ¹⁹³ In colorectal cancer, NTRK fusions occur in 1–2% of cases and can be detected by next-generation sequencing. Importantly, the efficacy of the available (and FDA/EMA-approved) inhibitors, larotrectinib and entrectinib, is very high according to recent data. ¹⁹³ Larotrectinib has been evaluated in 55 mainly pretreated patients with diverse tumor types (four of them colon cancers) with a response rate of 75% and a 1-year PFS of 55%. ¹⁹⁴ Furthermore, preliminary results of three ongoing trials evaluating entrectinib reported that 57% of 54 patients had an objective response with a median duration of 10 months. ¹⁹⁵ Thus, given the putative high efficacy of direct inhibition, patients who are refractory toward first- and second-line therapies should be offered testing for NTRK fusions, especially if no other drivers (RAS/RAF/HER2) have been detected.

Consensus molecular subtypes: do they have therapeutic implications?

Beyond analysis of mutations or structural genomic aberrations as predictive markers or therapy targets, transcriptome-based analyses aiming to improve therapeutic patient stratification by establishing new molecular classification systems of colorectal cancer have been published in recent years.^196–201 For better clinical standardization and translation, the classification systems were unified into the ‘consensus molecular subtypes’ (CMS) by an international consortium. ²⁰² CMS categorizes colorectal cancers into four subtypes with distinct tumor biology: CMS1 (MSI/immune), CMS2 (canonical), CMS3 (metabolic), and CMS4 (mesenchymal). ²⁰² The subtypes were related to previously known genomic, epigenetic, histological, and clinical features of colorectal cancer.^145,202 To this end, CMS1 largely overlapped with dMMR/MSI cancers, a high degree of immune cell infiltration, low tumor differentiation, a higher frequency of BRAF mutations, hypermethylation, and location of the primary tumor in the proximal colon. In contrast, CMS2 and CMS4 had many copy number variations, were pMMR, and low in methylation rate. CMS2, representing the ‘canonical subtype’, was also found to be associated with WNT and MYC activation, and left-sided location, while CMS4 was related to mesenchymal signatures, TGF-beta activation, and extracellular matrix remodeling, histologically characterized by a strong desmoplastic reaction.^145,202 Finally, CMS3 was characterized by marked metabolic reprogramming, including activation of lipogenesis, while MSI status was mixed and copy number variations and methylation status were low.^145,202,203 CMS1 is associated with the worst prognosis in Stage IV disease, while CMS2–4 have better overall outcome in this stage.^204,205

CMS has also been studied as a predictive signature for therapy selection in metastatic cancers. Here, CMS1 is most likely associated with sensitivity to checkpoint inhibitor treatment due to its association with MSI/hypermutation. Published data are partially conflicting with respect to prediction of survival on EGFR and VEGF antibody combination therapies. Retrospective transcriptomic analysis for CMS classification of tumors from the phase III FIRE-3 and CALGB/SWOG 80405 clinical trials have been performed.^204,205 These studies had compared the addition of bevacizumab or cetuximab to doublet chemotherapy as first-line treatment. In the CALGB/SWOG 80405 analysis, CMS1 patients treated with bevacizumab had better OS than those treated with cetuximab, while the CMS2 patients benefited more from cetuximab therapy. According to the FIRE-3 study data, OS was comparable in CMS1 and CMS2 subgroups, independent of targeted therapy, while CSM3 and CSM4 both favored cetuximab with a longer OS. In addition, retrospective analysis of AGITG MAX trial data suggested CMS2 and possibly CMS3 tumors to benefit from bevacizumab in addition to capecitabine chemotherapy, compared with other CMS. ²⁰⁶ Other retrospective data suggested a worse outcome associated with anti-EGFR antibodies in CMS1 and a favorable outcome in CMS2. ²⁰⁷ Differences of CMS predictive values found in these studies have been attributed to different chemotherapy backbones used in the trial populations, interaction of chemotherapy with targeted therapies, the tumor microenvironment, and differences in therapy sequence. ²⁰⁸ Interestingly, in a different approach classifying molecular subtypes based on gene copy number variations instead of gene expression, Smeets et al. ²⁰⁹ found that tumors with a high or intermediate degree of chromosomal instability had improved outcome after bevacizumab combination therapy, while the subgroup with low degree of copy number variations (corresponding CMS1 or dMMR/MSI/hypermutated tumors) did not benefit from bevacizumab treatment. With respect to conventional chemotherapies, improved outcome of irinotecan- versus oxaliplatin-based combination treatment has been reported for CMS4 tumors. ²⁰⁷ Notably, studies in preclinical models of colorectal cancer suggest associations of CMS with response to other specific anticancer drugs, including oxaliplatin, ²¹⁰ HSP-90 inhibitors, ²¹¹ birinapant, ²¹² or YM-155, ²¹³ indicating that the CMS classification may have predictive potential for specific substances.

Thus, cetuximab might be beneficial in CMS2–4 and irinotecan specifically in CMS4 tumors, while the situation for bevacizumab seems less clear. Due to this rather preliminary data and relatively laborious methodology, CMS currently has no application in routine clinical practice. Further data, favorably from prospective trials, are needed to define the future predictive value of CMS in metastatic colorectal cancer.

Functional precision medicine – science fiction or realistic path to explore?

Due to the low prevalence of druggable mutations and molecular biomarkers for drug efficacy, testing of drugs ex vivo in suitable model systems for personalized predictions has been proposed to complement molecular genetic testing and referred to as ‘functional precision medicine’. ²¹⁴ Tumor models, such as patient-derived xenografts (PDX) or patient-derived organoids (PDO), are used for preclinical drug screenings, co-clinical trials, and also personalized therapy predictions.^215,216 PDX are mouse models, in which tumor cells or fragments obtained from tumors of cancer patients are implanted orthotopically or heterotopically (for instance, subcutaneously or into the renal capsule) into immunodeficient mice. ²¹⁷ These implants form tumors within several weeks and have been shown to resemble their origin, so that PDX can be used for drug testing in a personalized ‘mouse avatar’.^215,218 Some studies have demonstrated high correlations of treatment response between patient and PDX;^215,216,219 however, the technology is laborious, engraftment rates differ between patients and tumors, engraftment time can be too long for personalized testing, and ethics are controversial.

Organoids are stem cell-derived three-dimensional cell cultures that grow in extracellular matrix with the help of culture medium supplemented with stem cell niche factors.^220–222 Similar to PDX, PDOs resemble their origin, that is, tumors or healthy epithelial tissues, with respect to molecular and morphological features.^223–225 They can be established from colorectal cancers with high efficiency (approx. 70%),^226–228 kept in culture long term and can be expanded for biobanking or drug profiling studies.^225,229,230 In addition, recent studies have reported a high degree of correlation between patients’ and matched organoid drug response, when treated with the same substances.^226–228 However, the only completed interventional trial (SENSOR) testing organoid-predicted precision treatments in ‘last-line’ colorectal cancer patients failed to show a meaningful clinical benefit. ²³¹ Further development of the model, and standardization and benchmarking with clinical response might be necessary for PDOs to harvest clinical benefit. Nevertheless, both PDX and PDOs are invaluable tools depicting the diversity of colorectal cancer in the laboratory for preclinical development of novel treatments and for biomarker research.

Conclusion and outlook

In conclusion, the therapy of metastatic colorectal cancer has greatly improved in recent years. Some patients can be cured by resection of metastases, especially in combination with advanced chemotherapy and targeted therapy protocols. These treatments need to be carefully tailored to molecular (RAS/RAF/MSI) and clinical (sidedness, performance status) predictive markers. Beyond, specific subgroups, such as patients with dMMR/MSI tumors, tumors with HER2 amplification, BRAF V600E mutation, and NTRK fusion, benefit from immunotherapy or targeted therapies, respectively. Therefore, analyzing these molecular characteristics of tumors is necessary to allow optimal patient care. More detailed molecular characterization, eventually in combination with functional testing using advanced preclinical models, may indicate further treatment options for advanced colorectal cancer patients in the future. However, most oncogenic drivers of colorectal cancer are currently not druggable, so that precision treatments for the majority of patients are not foreseeable and would require breakthroughs in basic research. Further developments of immunotherapy appear inciting and promising areas in oncological research in general. In this regard, exploiting the immune system in yet immunologically cold tumors seems to be a holy grail of preclinical and translational research in coming years.