The MET protein is a transmembrane receptor tyrosine kinase.¹ Its ligand, HGF, binds to the extracellular domain, leading to receptor dimerization and phosphorylation of the intracellular kinase domain within the cytoplasm. Phosphorylation promotes signaling through the PI3K and MAP kinase pathways, which are important pathways for proliferation, survival, motility, and invasion. Wild-type MET is expressed in most epithelial cells and has a key role in tissue regeneration, healing of the skin, and embryonic development. 

There are several ways the MET pathway can be dysregulated, including overexpression of the receptor or of its ligand, HGF, MET structural aberrations that drive oncogene addiction, including rare MET fusions, or MET mutations, typically exon 14 mutations.^2-4 MET gene amplification as a means of MET overexpression can be a primary event or a secondary event in the context of resistance to treatment.

The MET gene is translated as a pre‑mRNA that goes through a maturation step before being becoming a mature messenger RNA that is finally translated into protein. In some cases, errant splicing may occur during maturation as a result of a mutation, causing the loss of exon 14.^5,6

Exon 14 lies in the juxtamembrane region and includes tyrosine 1003, which is the binding site for the E3 ligase. The E3 ligase is responsible for ubiquitination and subsequent degradation of the MET receptor by the proteasome. METex14-skipping mutations result in a MET receptor with increased stability and kinase activity, due to the lack of degradation.⁷

METex14-skipping mutations result in MET functioning as an oncogenic driver and therapeutically relevant target, which when inhibited offers patients benefit. The true biological implications of the different METex14 variants are not fully understood at present, and further research is warranted.

Lung cancer is seen as the paradigm of personalized medicine. Ideally, every patient with advanced nonsquamous (and in some cases, squamous) NSCLC should be assessed for various molecular abnormalities including all those with an FDA-approved targeted therapy: specifically EGFR and BRAF mutations, ALK, ROS1, NTRK, and RET translocations. Now METex14-skipping mutations should be included in testing.^8-21

Clinicians should not assume that a routine testing panel includes all actionable mutations and they should work with their pathology team to ensure that all targetable mutations are examined efficiently and cost-effectively. With 7 molecularly specific analyses, tissue and economic efficiencies strongly support panel-based next-generation sequencing over single gene assays. Despite recommendations for testing, real-world data suggest that numerous patients are not getting the recommended testing.^22,23

At the present time, we have several therapeutic options for treating tumors with METex14 alterations. Crizotinib is approved as an ALK and ROS1 inhibitor, but it also inhibits MET and is available as an off‑label option.²⁴

Tepotinib, a selective MET TKI, is approved by the Japanese Ministry of Health Agency for the treatment of advanced NSCLC with METex14-skipping mutations. ²⁵

Capmatinib, another MET TKI, is approved by the FDA for the treatment of patients with METex14-positive advanced NSCLC.¹⁰

Now that we have targeted agents effective against METex14-positive disease, it is imperative that this oncogenic driver be included in our biomarker testing panels.

The optimal technology for METex14 testing is RNA NGS, which is the most sensitive way to detect METex14-skipping mutations.^6,26-29 DNA NGS is design dependent and less sensitive than RNA NGS, possibly missing some relevant exon 14 mutations. PCR testing is also design dependent. IHC and FISH are not recommended.

The number of cases detected by RNA NGS is nearly double those detected by DNA‑based technologies.³⁰ DNA NGS may miss some types of METex14 variants.²⁹ The flanking regions of exon 14 should be included for DNA NGS and a fragment length analysis should be performed. 

There is evidence that liquid biopsies to isolate cell‑free DNA is a reasonable method to detect 70% to 80% of the METex14-skipping mutations found through DNA-based testing of tissue in advanced NSCLC,^3,31 and should be used if available tissue is limited for molecular testing. As with any liquid biopsy, although a positive result is valid, a negative result without a reasonable alternative oncogene being identified does not preclude a false-negative result.

As we will discuss below, the validity of liquid biopsy for METex14 mutation testing is supported by data from the VISION trial, which examined the role of tepotinib in METex14-positive NSCLC.³² Furthermore, they used liquid biopsies to effectively increase trial enrollment. After liquid biopsy screening is incorporated there is a dramatic shift in the patient enrollment curve.

To help us measure the educational impact of this activity, please answer the following question, which was presented to you at the beginning of the activity. As a reminder, your responses will be aggregated for analysis, and your specific responses will not be shared.

The overall incidence of METex14 mutations in NSCLC is 3% to 4%.^6,33-44  This incidence is higher in adenocarcinoma at 4% compared with 1.5% to 2.0% for squamous histology. Sarcomatoid carcinomas have the highest incidence, in the range of 20%. METex14 mutations occur more frequently in white patients than Asians, in more women than men, and in more smokers or former smokers, but one third of the patients are still never smokers. MET mutations are frequently the only driver alteration.

The Profile 1001 trial was an early-phase trial that examined crizotinib in patients with either METex14 alterations, or ALK or ROS1 rearrangements. Patients were screened for METex14 mutations and MET amplification.²⁴ The METex14 cohort included 69 patients and the response rate was 32% among 52 evaluable patients, with a duration of response of 9 months and a PFS of approximately 7 months. No correlation was seen between different MET alterations and response. 

Tepotinib is a highly selective, ATP‑competitive, reversible MET TKI that may penetrate the CNS.^31,48-50

The single‑arm, phase II VISION trial evaluating tepotinib for MET-altered NSCLC enrolled patients (n = 152) with both newly diagnosed disease or those previously treated with cytotoxic therapy or immunotherapy.³² Both patients with METex14 or MET amplification were included. Patients previously treated with MET inhibitors and those with active brain metastases were excluded. Both tissue biopsies and liquid biopsies were used to identify MET alterations. Patients received tepotinib at 500 mg daily until progression. The primary endpoint of the study was ORR. PFS, OS, and safety were secondary endpoints.

Patients with METex14-skipping mutations comprised cohort A.³² The ORR was 46% for the cohort; the waterfall plot shows the reduction in tumor size. The ORR by biopsy type was 48% for liquid biopsy and 50% for tissue biopsy. Regarding response, liquid biopsy seems just as efficient as tissue biopsy for METex14.

PFS was 8.5 months for the overall population and for the subset of patients who were tested by liquid biopsy compared with 11.0 months for patients who were tested by tissue biopsy.³² Thus, tepotinib is an active drug that targets a driver mutation. Furthermore, as introduced above, this study showed the response is similar regardless of whether the METex14 mutation was detected by tissue biopsy or liquid biopsy, which is important in this older patient population in which obtaining additional tissue samples for testing may be difficult.

Summarizing the independent- and investigator-assessed outcomes by biopsy type shows an ORR range from 48.5% to 61.7%.³² The duration of response was 9.9-16.4 months and the PFS was 8.5-11.0 months. The OS was 15.8‑22.3 months with a short follow-up time. There were 11 patients with brain metastases who experienced comparable outcomes to those without brain metastases, and intracranial ORR was not reported. There is evidence of brain activity for tepotinib from a single compassionate-access patient with a MET fusion who had a brain response.³¹ This is early evidence of tepotinib activity in the brain. 

Most of the MET type I inhibitors have a similar toxicity profile, with peripheral edema being the most common class effect. In VISION, peripheral edema, generally grade 1/2, was the most common adverse event. Nausea, diarrhea, and some laboratory abnormalities occurred as well.³² Increased amylase and lipase, mostly grade 1/2 and often asymptomatic, was seen. Patients receiving tepotinib should be monitored for these adverse events. Overall, tepotinib was reasonably well tolerated but 11% of patients discontinued treatment due to adverse events. Overall, patients’ quality of life was maintained while receiving tepotinib.

Capmatinib is one of the more potent MET inhibitors in vitro based on a pure isolated kinase assay with an IC₅₀ of 0.6 nM.^31,48-50

In the GEOMETRY mono-1 study, capmatinib was examined in patients with advanced NSCLC and a METex14-skipping mutation.^50,51 Cohort 5b consisted of patients who had not received previous therapy (n = 28), whereas patients in cohort 4 had received other chemotherapies but not a MET inhibitor (n = 69). Capmatinib 400 mg was given twice daily. Capmatinib was initially given under fasting conditions but the food restriction was later dropped.

In the cohort of previously treated patients, the ORR was 40.6% by independent review and 42.0% by investigator assessment.⁵⁰

In the treatment‑naive cohort, there was an apparent line-of-therapy effect. Rather than a 40% response rate as seen in previously treated patients, the ORR in treatment-naive patients was 60.7% to 67.9%. Of interest, the independent review committee (IRC)–assessed ORR was higher at 67.9% vs 61.0% for the investigator-assessed ORR.⁵⁰ It is unclear whether the line of therapy effect was due to the drug or the characteristics of the patients in this small cohort as the line of therapy effect seems more pronounced for capmatinib than for other MET TKIs in cross-trial comparisons. 

PFS in the pretreated cohort was 4.8‑5.4 months and 9.7‑11.1 months in the treatment‑naive cohort, again varying by independent vs investigator assessment.⁵⁰

The duration of response in the pretreated cohort was 8.3‑9.7 months and was 11‑14 months in the treatment‑naive cohort.⁵⁰ There is a disconnect between PFS and duration of response, assuming a similar follow-up time, which suggests that maybe this is not a uniform population. There are responders who are clearly addicted and there are nonresponders, some of whom may be deriving benefit and some of whom may not.

Some data indicate that the new selective MET TKIs are active in the brain. There were 13 patients with measurable CNS metastases on the GEOMETRY mono-1 trial and the intracranial ORR was 54%.⁵⁰  

As mentioned above, MET TKIs have a class effect of peripheral edema. In GEOMETRY mono-1, 7.5% of patients had grade 3/4 peripheral edema with capmatinib.⁵⁰  

Many of the MET inhibitors, including crizotinib, tepotinib, and capmatinib, increase creatinine levels but this is not a nephrotoxic effect.^32,50,52 Instead, these MET TKIs interfere with the secretion of creatinine and functioning similar to cimetidine or trimethoprim.⁵⁰ Creatinine levels rise, and the apparent glomerular filtration rate (GFR) drops by approximately 20% to 25%, but the MET TKI is actually just interfering with the validity of creatinine as a measure of your GFR. The classic response is that creatinine will rise and then plateau. Because of this, if consideration to altering dosing of the MET TKI or other drugs is being given based on creatinine, it is recommended that a noncreatinine‑based assessment of renal function be performed to verify GFR test results.

A phase II study of the selective class I MET TKI savolitinib preferentially enrolled patients with sarcomatoid lung cancer, but other subtypes were eligible to enroll.^53,54 In the resulting study population (n = 70), 57% of patients had adenocarcinoma and 35% had sarcomatoid lung cancer. The median age of the patient population was older at 68.7 years. Treatment with savolitinib achieved an ORR of 47.5% and median PFS of 6.8 months.

Again, the most common adverse event with savolitinib was peripheral edema, followed by nausea, vomiting, and some laboratory abnormalities.^53,54 There were approximately 40% grade ≥ 3 adverse events and approximately 14% of patients discontinued therapy due to treatment‑related adverse events.

When comparing outcomes for the key selective MET TKIs, the response rate for previously treated patients is similar at approximately 40%.^{24,31,50,51,54-57} The duration of response and PFS with capmatinib are slightly less than tepotinib.^31,50 The exaggerated ORR with capmatinib as frontline therapy is out of keeping with the other MET TKIs, but this is a small cohort of patients. The question of whether capmatinib is really more affected by line of therapy than other MET TKIs or whether the small sample size skews the data remains to be determined.

To help us measure the educational impact of this activity, please answer the following question, which was presented to you at the beginning of the activity. As a reminder, your responses will be aggregated for analysis, and your specific responses will not be shared.

None of these MET TKIs offer a cure. Acquired resistance can develop to any targeted therapy in several different ways.⁵⁸ From a pharmacologic perspective, the most common type of resistance is CNS progression due to failure of the drug to cross the blood–brain barrier.  

Biological mechanisms of resistance include on‑target resistance mechanisms and secondary driver mutations. On‑target resistance mechanisms are usually due to point mutations in MET. Although MET amplification can occur, it does not appear to have a large impact on resistance to MET inhibitors. Secondary pathways may also emerge as a mechanism of resistance.

The variety of mutations that confer resistance to type I and type II MET inhibitors overlap; for example, mutations at D1228 are common to both types I and II,^59,60 but there are differences, as can be seen in the figure. This variability suggests that switching to a different class of inhibitor might provide a benefit to some patients depending on the type of resistance mutation.

Acquired bypass pathways need to be distinguished from cases with potentially 2 driver pathways coexisting pretreatment, which may reflect a true de novo codriven state, or failure of the MET pathway to manifest as a true driver (failed driver). A second potential driver appears in approximately 5% of de novo cases, mostly KRAS.^60-62

In the approximately 5% of cases where METex14 is not the only apparent potential driver of oncogenesis, the presence of other drivers could explain why the response rate to MET TKIs is low compared with therapies for other driver oncogenes.^24,31,50,55 Such cases may reflect a true de novo codriven state, or failure of the MET pathway to manifest as a true driver (failed driver). Evidence that in some of these cases no identifiable MET protein exists suggests that MET as a failed driver, despite the presence of a METex14-skipping mutation, can occur.

Immuno-oncology and antibody–drug conjugate approaches to METex14-positive NSCLC are also being examined. ABBV-399 is a MET antibody–drug conjugate that targets a toxin to MET and downregulates MET signaling.⁶³ In a general NSCLC population—not specifically in METex14—the response rate was approximately 20% to 40%, varying by level of c-MET expression and drug dosing schedule. The response rate is similar in EGFR‑mutant tumors, where MET may have a specific signaling role.

Sym015 is a mixture of 2 MET antibodies that block HGF ligand binding, produce antibody-dependent cellular cytotoxicity, and downregulate MET signaling.⁶⁴ In METex14-positive NSCLC, Sym015 therapy resulted in responses, including in some patients with previous MET TKI exposure.

Using checkpoint inhibition for METex14 lung cancer, there was some activity including a change from baseline of approximately 70% in one case.⁶⁵ The responses did not correlate with the level of PD-L1 expression or with high levels of tumor mutational burden. The importance of a smoking history in determining METex14-positive NSCLC responsivity to checkpoint inhibition continues to be explored.

JNJ‑372 is a bispecific antibody targeting EGFR and cMET.⁶⁶ There is crosstalk between EGFR and MET, but it is unknown whether the same crosstalk occurs with METex14. Because MET inhibition with the bispecific occurs preferentially where there is also EGFR expression, the level of toxicity from normal tissue exposure may be lower. In NSCLC with EGFR exon 20 insertions, the response rate with JNJ-372 is approximately 30%.⁶⁷