Patients with colorectal cancer could benefit from the COLMATE platform as it allows for opportunities in clinical trial enrollment.
Colorectal cancer (CRC) is estimated to be the third most prevalent cause of cancer death in both men and women in the United States in 2021.1 While some patients are diagnosed with localized disease and cured with appropriate treatments, most patients who are diagnosed with de novo or recurrent advanced disease face a life-limiting illness. Median overall survival (OS) in metastatic CRC (mCRC) continues to increase over time, largely due to improving systemic therapies, but more therapies are still needed. For the majority of patients with mCRC, even initially effective systemic therapies become intolerable or cancer resistance to chemotherapy develops.
In the last decade, the identification of actionable genomic alterations in mCRC has led to the FDA approval of several effective therapeutic regimens targeting these alterations (see below). Still, even guideline-aligned biomarker testing for alterations such as RAS, BRAF, and microsatellite instability high (MSI-H)/mismatch repair deficiency (dMMR) in mCRC remains suboptimal,2 highlighting the need for more streamlined testing. Once therapeutic resistance develops, recognition of this is necessary and new therapeutic options are needed. This article will describe currently established and actionable genomic alterations in mCRC, emerging targets, and a novel clinical trial platform designed to optimize therapeutic selection for patients with mCRC.
The identification of actionable genomic alterations in tumors such as mCRC was once performed by Sanger DNA sequencing of tumor DNA that was extracted from fixed paraffin-embedded tumor tissue, but this has now been replaced by next-generation sequencing (NGS), which allows for larger-scale and automated genome sequencing. Multiple commercial vendors, such as Foundation One, Caris, and Tempus, among others, provide comprehensive tumor molecular profiling. However, utilizing NGS data from a tumor biopsy of 1 anatomic location at 1 time point may be insufficient to treat patients with mCRC, as we recognize that there can be spatial tumor heterogeneity (between the primary tumor and metastatic lesions) and temporal tumor heterogeneity (when targeted therapies exert pressure for certain mutations to disappear and new mutations to evolve).3,4 The invasiveness of serial tissue biopsies and the turnaround time of 2 to 4 weeks for NGS analysis can both be limitations for clinical use. Thus, the development of blood-based NGS testing as a method to analyze circulating tumor DNA (ctDNA) has been a welcome one. This allows clinicians to identify the prominent actionable genomic alterations from the patient’s cancer as a whole, and to monitor the dynamic changes in the molecular profile serially as the patient progresses from one treatment to another.5
The FoundationOne Liquid CDX test (Foundation Medicine, Inc, Cambridge, MA) is now FDA approved as a companion test for multiple biomarkers, including BRCA1 and BRCA2 genes in ovarian cancer for rucaparib (Rubraca) treatment, ALK rearrangements in non–small cell lung cancer (NSCLC) for alectinib (Alecensa) treatment, PIK3CA mutations in breast cancer for alpelisib (Piqray) treatment, and BRCA1 and BRCA2 and ATM mutations in prostate cancer for olaparib (Lynparza) treatment. The Guardant360 CDX (Guardant Health, Redwood City, CA) assay has been FDA approved as a companion test for EGFR mutation for the use of osimertinib (Tagrisso) for patients with NSCLC. Thus, blood-based NGS should be considered an additional test that clinicians can utilize in the care of patients with mCRC.
Established Actionable Genomic Targets in Metastatic Colorectal Cancer
Several genomic alterations are now established and actionable in mCRC, including KRAS/NRAS mutations, BRAF V600E mutations, HER2 amplification, MSI-H/dMMR, and NTRK fusions.6 We will briefly review the current clinical trial data behind these actionable alterations.
RAS mutations in mCRC have long been recognized as a negative predictive biomarker, identifying patients who would not benefit from anti-EGFR antibody treatments such as cetuximab (Erbitux) or panitumumab (Vectibix).7-9 While KRAS exon 2 mutations were initially thought to be the sole predictive biomarkers in mCRC, subsequent data have shown that KRAS/NRAS exon 2, 3, and 4 mutations are all predictive. As such, guideline-aligned biomarker testing now includes extended RAS testing for completeness. For patients with KRAS/NRAS wild-type mCRC, treatment with cetuximab or panitumumab, either in combination with other cytotoxic chemotherapies or as monotherapy, can be an effective treatment strategy. For these patients, primary tumor sidedness is also prognostic, with left-sided primary tumors being associated with better overall prognosis independent of therapy10; left-sidedness is predictive of response to anti-EGFR therapy as well.11,12
In addition to its known role as a negative predictive marker, KRAS is now also a positive predictive biomarker. Long thought to be the “undruggable” target due to its high affinity for GTP and its lack of large binding pocket for allosteric inhibitors,13,14 the discovery of small molecules that can bind to KRAS G12C led to rapid drug development.15 A phase 1 trial (NCT03600883) investigated sotorasib (Lumakras), an irreversible KRAS G12C inhibitor, in patients with advanced solid tumors harboring the KRAS p.G12C mutation.16 Overall, 7.1% of the 42 mCRC patients treated on study had a confirmed objective response, and 73.8% had disease control. However, median progression-free survival was only 4 months. Other KRASG12C inhibitors in development such as MRTX849 have also shown encouraging preclinical and early clinical data,17 although mechanisms of resistance—such as MAPK reactivation, stimulation of CDK4/6-dependent cell-cycle transition, and immune defects to KRAS G12C inhibition—are now being identified.18 Additionally, since KRAS G12C mutations are present in only approximately 3% of CRC cases, just a small minority of patients are eligible for this targeted approach. However, the fact that a previously undruggable target may now be druggable in mCRC has been an exciting development and holds promise for future drug development and discovery.
BRAF V600E mutations are both prognostic and predictive biomarkers in mCRC and are present in approximately 8% to 10% of mCRC cases.19,20 Interestingly, non-V600E mutations do not confer a prognosis that is as poor,21 but they are also targets of interest in drug discovery. For patients with previously treated BRAF V600E–mutant mCRC, the phase 3 BEACON CRC study (NCT02928224) found that combined inhibition of BRAF and EGFR with encorafenib (Braftovi) and cetuximab, respectively, led to improvements in OS, objective response rate (ORR), and progression-free survival (PFS) when compared with treatment with standard chemotherapy.22 Median OS, ORR, and PFS for the doublet arm were 9.3 months, 19.5%, and 4.3 months, compared with 5.9 months, 1.8%, and 1.5 months, respectively, for the control arm. These results led to combined BRAF and EGFR inhibition in previously treated patients with BRAF V600E mutant mCRC becoming a new standard of care.
Combined BRAF/MEK/EGFR inhibition in patients with BRAF V600E– mutant mCRC has also been investigated in the first-line setting in a single-arm phase 2 study.23 In the first stage of this trial, which incorporated results from the first 40 patients, ORR was 50%, with a disease control rate of 85%; PFS was only 4.9 months. Final results from this trial are awaited, but, given the overall poor prognosis of this disease, more treatment options are needed.
HER2 amplification has rapidly become an actionable target in mCRC over the past few years, and this target is found in approximately 5% of patients with RAS wild-type mCRC. Multiple small studies targeting HER2 have been encouraging despite it being a relatively rare genomic finding in mCRC. The phase 2 HERACLES study (NCT03225937) of trastuzumab and lapatinib demonstrated an ORR of 30% in patients with refractory KRAS exon 2 wild-type, HER2-amplified mCRC,24 with responding tumors harboring high HER2 gene copy numbers. The MyPathway phase 2 study utilized trastuzumab and pertuzumab (Perjeta) in a similar patient population, generating an ORR of 32%25; however, patients with KRAS-mutant disease were not excluded and tended to have lower response rates to dual anti-HER2 blockade. The phase 2 TRIUMPH study also tested the combination of trastuzumab plus pertuzumab but utilized both tissue and ctDNA for HER2 testing26; response rates were similar in both tissue-positive and ctDNA positive groups (ORR, 35% vs 33%, respectively). Initial results from the ongoing MOUNTAINEER study of tucatinib (Tukysa) plus trastuzumab demonstrated an impressive ORR of 52%, with a median PFS of 8.1 months and median OS of 18.7 months.27
The recently published results of the phase 2 DESTINY-CRC01 (NCT03384940) trial utilized a different approach in targeting HER2-positive mCRC: The investigators studied trastuzumab deruxtecan, an antibody-drug conjugate of a humanized anti-HER2 antibody with a topoisomerase inhibitor payload.28 Despite 30% of patients having received prior anti-HER2 therapy, the objective response rate with trastuzumab deruxtecan was 45.3%; 6% of patients did develop interstitial lung disease or pneumonitis, including 2 grade 5 events.
Approximately 3% to 4% of patients with mCRC will be found to have MSI-H or dMMR, a highly actionable predictive biomarker of response to immunotherapy. Once a phase 1 study of the PD-L1 inhibitor MK-3475 (pembrolizumab [Keytruda]) identified CRC as a potentially sensitive tumor type,29 Le et al then investigated the efficacy of pembrolizumab monotherapy in patients with MSI-H/dMMR CRC, microsatellite stable (MSS)/proficient mismatch repair CRC, and MSI-H/dMMR malignancies other than CRC.30,31 Objective radiographic responses were seen in 53% of patients with MSI-H/dMMR tumors, including complete responses in 21%, and duration of response was significant. Similarly, treatment of MSI-H/dMMR mCRC with the PD-1 inhibitor nivolumab (Opdivo) demonstrated a response rate of 31.1%, with 8 patients at the time of publication having responses lasting 12 months or longer.32 The addition of the CTLA-4 inhibitor ipilimumab to nivolumab therapy in CheckMate 142 led to response rates of 55%, with median duration of response not reached by the time of publication.
In addition to being used to as monotherapy to treat refractory MSI-H/dMMR mCRC, pembrolizumab has been investigated in the frontline setting with very encouraging results. The phase 3 KEYNOTE-177 trial explored the use of pembrolizumab vs chemotherapy (5-fluorouracil-based therapy with or without bevacizumab [Avastin] or cetuximab) in MSI-H/dMMR mCRC.33 Pembrolizumab was found to be superior to chemotherapy with respect to PFS (16.5 vs 8.2 months; hazard ratio, 0.60; 95% CI, 0.45-0.80; P = .0002) and ORR (43.8% vs 33.1%). Among patients with a response, 83% in the pembrolizumab group had ongoing responses at 24 months, compared with 35% in the chemotherapy group. While an OS benefit for pembrolizumab has not been demonstrated in this study, crossover occurred in 60% of patients.34 Given these practice-changing results, it has become imperative to perform molecular testing in mCRC immediately at diagnosis.
While immunotherapy has been revolutionary for many patients with MSI-H/dMMR mCRC, some caveats exist. For instance, some patients have primary resistance to PD-1 therapy, and some develop acquired resistance for reasons that are unclear. Furthermore, given that less than 5% of patients with mCRC have MSI-H/dMMR disease, the vast majority of patients with mCRC will not be eligible for this therapy. Further investigation into developing better immunotherapy approaches for patients with MSS or refractory MSI-H/dMMR disease are urgently needed.
Fusions involving 1 of 3 TRK proteins encoded by the genes NTRK1, NTRK2, and NTRK3 are sometimes found in mCRC,35,36 although very uncommonly and often in MSI-H/dMMR disease.37 A tumor-agnostic screening program of 26,000 patients showed the prevalence of TRK fusions in CRC to be 8/2306, or 0.35%.38 In a pooled analysis of 3 phase 1/2 clinical trials, patients with TRK fusions were treated with the selective TRK inhibitor larotrectinib (Vitrakvi), with an ORR of 79%; 8 of the 153 patients had colon cancer, and 4 (50%) of these had an objective response to larotrectinib.39 Similarly, treatment with entrectinib led to an ORR of 57% overall, with 2 of 4 (50%) patients with CRC achieving an objective response to therapy.40 Thus, despite their relative rarity in mCRC, TRK fusions are actionable in this disease.
The FGFR family consists of 4 receptors—FGFR1, FGFR2, FGFR3, and FGFR4—each of which has an extracellular ligand binding domain, transmembrane domain, and intracellular tyrosine kinase domain.41 Eighteen ligands are identified in the FGFR family, and subsequent signal transduction leads to the activation of both the RAS-MAPK and PI3K pathways. In addition, FGFR1 alterations have been reported to contribute to resistance to anti-EGFR therapy in CRC.42 Approximately 5.3% of patients with CRC were identified as harboring tumors with FGFR1 amplification in a study analyzing 454 primary tumors.43 A separate study observed 3.5% of samples having FGFR1/2/3 amplification.44 Pemigatinib (Pemazyre) is an oral inhibitor of FGFR1, FGFR2, and FGFR3, which has now been FDA approved for the treatment of patients with metastatic cholangiocarcinoma who have tumors that harbor FGFR2 fusions. That approval was based on the results from the open-label phase 2 FIGHT-202 trial, which demonstrated that 35.5% of patients with metastatic cholangiocarcinoma achieved an objective response with pemigatinib.45 In order to investigate whether this is a tumor-specific effect or not, a phase 2 study of pemigatinib is ongoing in patients with mCRC whose tumors harbor FGFR alterations, including mutations, translocations, and amplifications (NCT04096417).
Phosphatidylinositol 3-kinases (PI3Ks) are lipid kinases that are crucial to the PI3K/AKT1/MTOR signaling pathway in the proliferation, growth, and survival of cancer cells.46 PIK3CA mutations in exon 9 and/or exon 20 occur in approximately 15% to 20% of CRC tumors.47 The prognostic role that the presence of PIK3CA mutations has on patients with CRC is unclear; a meta-analysis examining 28 studies concluded neutral prognostic effects on OS and PFS.48 Some preclinical data suggest that the inhibition of the COX-2 pathway by aspirin downregulates the PIK3CA pathway. Results of the Nurses’ Health Study and the Health Professionals Follow-up Study showed that in patients with a history of PIK3CA-mutant CRC, the patients who had regular intake of aspirin had superior CRC-specific survival and OS.49 Patients with PIK3CA wild-type CRC who took aspirin did not have the same benefit. Clinical trials for patients with mCRC whose tumors harbor PIK3CA mutations include ALCAP, a phase 1b/2 study of alpelisib (an oral alpha PIK3CA inhibitor) and capecitabine (NCT03592641), and C-PRECISE-01, a phase 1b/2 study of MEN1611 (a PI3K inhibitor) and cetuximab (NCT04495621).
Homologous recombination (HR) is a DNA repair mechanism for double-stranded breaks. Homologous recombination deficiency (HRD) was initially described in patients with germline BRCA1 or BRCA2 deficiency, who had increased risks of breast, ovarian, and pancreatic cancer.50 Other activating mutations that have been identified as involved in HR now include ATM, ATR, CHEK1/2, RAD51, and PALB1/2.51 PARP1 and PARP2 are nuclear enzymes that are activated by DNA single-strand breaks and replication fork damage, and they facilitate DNA repair. PARP inhibition leads to an increase in double-stranded DNA breaks; as such, the drugs are selectively active in cells that have HRD. ATM and BRCA1/2 are mutated in approximately 20% of patients with CRC.51 Although a phase 2 trial of olaparib in 33 patients with mCRC (20 with MSS tumors, 13 with MSI-H tumors) did not show complete or partial response to PARP inhibition, there was no selection for HRD genomic markers.52
Several PARP inhibitors are currently approved in prostate, breast, and gynecologic cancers; however, this drug class is largely still under investigation in gastrointestinal cancers. Patients with pancreatic cancer who have germline BRCA1/2 mutations are potentially eligible for olaparib maintenance therapy per the POLO study.53 Patients with mCRC were part of the Academic and Community Cancer Research United (ACCRU) 1603 phase 1/2 study of 5-fluorouracil, liposomal irinotecan, and rucaparib, which enrolled patients with gastrointestinal cancer and HRD biomarkers; results will indicate whether the patients with mCRC benefited from the combination of chemotherapy and PARP inhibition. The LODESTAR phase 2 study of rucaparib in patients with solid tumors with deleterious mutations in HRD genes (NCT04171700) is currently enrolling patients with mCRC.
MET is a proto-oncogene that encodes a tyrosine kinase receptor that binds the ligand hepatic growth factor. Activation of MET leads to the downstream activation of the MAPK and PI3K/AKT pathways. MET amplification has been described as an escape mechanism pathway for CRC that is targeted by anti-EGFR therapy, due to the crosstalk between MET and MAPK pathways. MET amplification occurs in 12.5% of patients with CRC who do not have alterations in KRAS, NRAS, PIK3CA, and HER2 genes and also are not responsive to anti-EGFR therapy; incidence of MET amplification is likely higher in patients who have acquired resistance to anti-EGFR therapy.54 Tepotinib (Tepmetko) and capmatinib (Tabrecta) are oral MET inhibitors that are FDA approved for patients with metastatic NSCLC with tumors that harbor MET exon 14 skipping alterations, but they are investigational in mCRC. Ongoing trials for certain patients with mCRC—those with RAS/BRAF wild-type tumors who have acquired resistance to anti-EGFR therapy due to MET amplification—are the PERSPECTIVE phase 2 study of tepotinib and cetuximab (Erbitux) (NCT04515394) and a phase 2 study of savolitinib (NCT03592641).
EGFR mutations have been identified as a mechanism of resistance to anti-EGFR therapy in both lung cancer and CRC. Bertotti et al examined genomic alterations that affected anti-EGFR responses in KRAS wild-type CRC. They identified a sequence alteration in the kinase domain of EGFR (V843I) and evidence of tumor growth in the presence of cetuximab.42 When the EGFR-mutated tumor cells were treated with afatinib (Gilotrif), a tyrosine kinase inhibitor that targets EGFR mutations, and cetuximab, tumor growth inhibition was observed. The UCGI 25: UNICANCER phase 2 trial treated patients with KRAS wild-type tumors (no selection for EGFR mutations) with cetuximab and afatinib vs cetuximab alone, and results showed that the ORR was 8.3% in the monotherapy arm compared with 26% in the combination arm.55
CHRONOS was a phase 2 study of anti-EGFR rechallenge with panitumumab for patients with RAS/BRAF wild-type mCRC.56 The rationale was that patients develop mutant RAS and EGFR ectodomain clones when they are treated with anti-EGFR therapy. However, these clones decay when there is withdrawal of the anti-EGFR antibody, and thus there are patients who are candidates for anti-EGFR rechallenge. Patients who had received prior anti-EGFR therapy and achieved an objective response could be eligible for rechallenge with panitumumab after they had been on a non–anti-EGFR intervening line of therapy. Circulating DNA was analyzed for RAS, BRAF, and EGFR ectodomain mutations; if wild type, patients would then undergo panitumumab rechallenge. A total of 69% of the patients were eligible for anti-EGFR rechallenge based on their ctDNA analysis, and the disease control rate for at least 4 months with panitumumab rechallenge was 63%. Currently, a phase 2 study of pemetrexed and erlotinib (Tarceva) for patients with EGFR-overexpressed mCRC is enrolling in Korea (NCT03086538); another phase 2 study, PULSE, is examining panitumumab rechallenge in patients with mCRC with RAS/BRAF wild-type tumors who have had prior response to anti-EGFR therapy (NCT03992456).
Given the increasing prevalence of actionable genomic alterations in advanced CRC, an efficient mechanism to identify patients with actionable alterations and the ability to match these patients with molecularly assigned clinical trials was needed. The ACCRU consortium developed a screening umbrella platform for this purpose entitled COlorectal Cancer and Liquid BiOpsy Screening Protocol for Molecularly Assigned ThErapy, or COLOMATE (NCT03765736).57 This platform utilizes molecular profiling through the Guardant360 liquid NGS panel to identify potential patients for the COLOMATE companion study arms. Patients eligible to enroll on the COLOMATE study are those with metastatic colorectal adenocarcinoma whose disease has progressed on, who have been intolerant to, or who have a contraindication to a fluoropyrimidine, oxaliplatin, irinotecan, anti-VEGF monoclonal antibodies (mAbs), anti-EGFR mAbs, and anti–PD-1 mAbs.
Once enrolled on the COLOMATE platform, patients undergo liquid biopsy NGS with the Guardant360 assay to identify current actionable genomic alterations, even if they have previously undergone such testing, as long as prior testing was completed at least 60 days prior to trial registration. Rationale for this stems from the knowledge that molecular profiles in mCRC can be dynamic; for instance, patients with RAS wild-type mCRC can develop RAS mutant subclones in the face of selective pressure from prior treatment with anti-EGFR mAbs.5 With a short turnaround time for Guardant360 testing, treating physicians and the COLOMATE Steering Committee can quickly identify (and prioritize, if multiple) molecularly assigned trial options for patients. Of note, patients can reenroll in COLOMATE up to 3 times, maximizing their trial options within this platform.
The COLOMATE screening platform contains several dynamic companion study arms on which patients with specific molecularly identified alterations can enroll. Examples of trials currently enrolling or under development through this mechanism include those for patients whose tumors are RAS/BRAF wild type, HER2 amplified, FGFR altered, MET amplified, RAS mutated, or BRAF mutated (either V600E mutant or non-V600E mutant), or without any known actionable alteration. As this trial mechanism is fluid, additional arms can be added based on scientific rationale, discovery, and clinical need.
Although the primary objective of the COLOMATE study is to perform blood-based genomic profiling on patients with treatment-refractory mCRC to facilitate accrual to molecularly assigned therapies via companion clinical trials, additional end points will also be explored. Specifically, patient-matched tumor tissue and cell-free DNA (cfDNA) from peripheral blood will be obtained to facilitate clinically annotated genomic analyses, explore mechanisms of acquired resistance to molecularly assigned therapy, and analyze the correlation between cfDNA mutational burden (allele frequency, copy number) and clinical outcomes such as objective response rate, progression-free survival, and OS.
Overall, given the rapidly evolving genomic landscape evident in CRC, platforms such as COLOMATE hold promise to nimbly facilitate molecularly assigned clinical trials and allow physicians to respond quickly to the molecular changes in mCRC as they evolve in relation to prior therapies. The COLOMATE approach maximizes both efficiency in trial enrollment and opportunities for patients while accounting for a dynamic tumor environment, serving as a model for clinical cancer research in the years to come.
KC has received research grants to her institution from Array, Bristol Myers Squibb, Incyte, Daiichi Sankyo, Merck, Nucana, and Pfizer/Calithera; she is a consultant for Array, Foundation Medicine, Merck, Natera, Pfizer, and Taiho. JJ has received consulting/educational speaking fees from AstraZeneca, Bayer, E.R. Squibb & Sons, and OncLive®. JS is a consultant for AbbVie, Amgen, AstraZeneca, Bayer, Inivata, Mereo Biopharma, Natera, Pfizer, Seagen, Silverback Therapeutics, and Viatris; he has received research grants to his institution from AbbVie, Amgen, AStar D3, Bayer, Curegenix, Daiichi-Sankyo, Gossamer Bio, Nektar, Roche/Genentech, Sanofi, and Seagen. TB-S has received research grants to his institution from Abgenomics, Agios, Amgen, Arcus, Arys, Atreca, Bayer, Boston Biomedical, Bristol Myers Squibb, Celgene, Clovis, Genentech, Incyte, Ipsen, Lilly, Merck, Merus, Mirati, Novartis, Pfizer, and Seattle Genetics. He has received consulting fees to his institution from Arcus, Array Biopharma, Bayer, Genentech, Incyte, Ipsen, Merck, Pfizer, and Seattle Genetics, and consulting fees to himself from AbbVie, AstraZeneca, Beigene, Boehringer Ingelheim, Celularity, Daiichi Sankyo, Eisai, Exact Science, Foundation Medicine, Janssen, Kanaph, Natera, Sobi, TreosBio, and Xilis. He serves on the IDMC or DSMB for AstraZeneca, Exelixis, Lilly, PanCan, and 1Globe and on the Scientific Advisory Board for Imugene, Immuneering, and Sun Biopharma. CW has received research grants to her institution from Boston Biomedical Inc, INHBRX, Lycera, Rapt Therapeutics, Seattle Genetics, Symphogen, and Vaccinex; she has received honoraria for consultant/educational speaking from Nova Research Company, Oncology Learning Network, and PrecisCA.
Affiliations: 1Division of Hematology/Oncology, Department of Internal Medicine, Vanderbilt University Medical Center, Nashville, TN, USA; 2Mayo Clinic, Jacksonville, FL, USA; 3Duke University Medical Center, Durham, North Carolina, USA; 4Mayo Clinic, Scottsdale, AZ, USA; 5Department of Hematology and Medical Oncology, Emory University School of Medicine, Atlanta, GA, USA