Non-small cell lung cancer (NSCLC), which accounts for ~85% of lung cancer diagnoses, is a heterogeneous disease consisting of multiple histologies and many known driver mutations (1–4). MET exon 14 skipping mutation (METex14) is observed in ~3% (95% CI, 1%-4%) of NSCLC cases; is more likely to occur in females, elderly patients (median age ranging from 65 to 76 years old), and nonsmokers (5); and typically occurs in the absence of other driver mutations (2, 3, 6). In retrospective studies of patients with resected NSCLC, METex14 has been shown to be an independent poor prognostic factor and is associated with shorter overall survival (7, 8), underscoring the need for identification of this biomarker in patients with NSCLC and for targeted therapies (9).

Capmatinib was approved by the US Food and Drug Administration (FDA) on May 6, 2020, for the treatment of adult patients with metastatic NSCLC whose tumors have a mutation that leads to METex14. Approval was based on results from the phase 2 GEOMETRY mono-1 trial (NCT02414139) (10, 11). In the trial, a blinded independent review committee confirmed an overall response rate (ORR) of 68% (95% CI, 48%-84%) and median duration of response (DOR) of 12.6 months (95% CI, 5.6 months-not estimable [NE]) among treatment-naive patients (N=28). An ORR of 41% (95% CI, 29%-53%) and median DOR of 9.7 months (95% CI, 5.6-13.0 months) were established among previously treated patients (N=69) (11).

Additionally, capmatinib has shown clinical evidence of intracranial activity. In the GEOMETRY mono-1 trial, among 13 patients who had data evaluable by an independent neuroradiologic review committee, 12 (92%) had intracranial disease control and 7 (54%) had an intracranial response. Of the patients with an intracranial response, 4 (57%) had a complete response (11).

Tepotinib, another MET inhibitor, was approved by the FDA on February 3, 2021, for the treatment of adult patients with metastatic NSCLC harboring METex14 mutations. Approval was based on results from the phase 2 VISION trial (NCT02864992) (12, 13). Among patients with METex14 NSCLC treated with tepotinib, the ORR was 46% (95% CI, 36%-57%) and the median DOR was 11.1 months (95% CI, 7.2 months-NE), according to independent review (12). In the 11 patients with brain metastases, the response rate by independent review was 55% (95% CI, 23%-83%), with a median DOR of 9.5 months (95% CI, 6.6 months-NE) (12).

Prior to the FDA approval of capmatinib, there were no US-approved therapies that specifically targeted METex14 in NSCLC (4, 10, 14), but several other targeted therapies were given off-label to patients with METex14 NSCLC (15). These included crizotinib, a tyrosine kinase inhibitor with activity against MET, ALK, and ROS1; and cabozantinib, a tyrosine kinase inhibitor with activity against a broad range of targets, including MET, VEGFR2, RET, FLT3, and KIT (16). Reports have shown promising clinical activity for these agents in patients with METex14 NSCLC (2, 16, 17).

For lung cancer patients with METex14, broad molecular profiling, such as next-generation sequencing (NGS) testing, can be successfully used to guide treatment decisions (2, 16, 17). Studies have shown that broad molecular profiling, which allows patients to match with an appropriate therapy, can result in improved outcomes (18–20). NGS testing also saves personnel working time and consumables related to biomarker tests and reduces the overall cost of testing per patient (21, 22).

A potential barrier to use of targeted agents may be the time required for NGS testing to identify METex14 as well as other actionable biomarkers in patients with NSCLC (23), which may result in some patients initiating treatment with systemic therapies prior to receiving broad molecular testing results. Over the years, testing companies have tried to expedite the processing time for NGS testing, but the time needed to receive testing results can still be multiple weeks (23). Although NGS panels have longer testing turnaround times compared with single tests, NGS can reliably cover a broader spectrum of alterations (24). Patients with METex14 NSCLC have commonly been treated with chemotherapy or immuno-oncology (IO) therapy (9). Programmed death-ligand 1 (PD-L1) expression ≥50% has been detected in nearly half of patients with METex14 (25); however, as seen in patients with many other targetable oncogenic drivers, patients with METex14 tend to have generally poor responses to immunotherapy (25, 26), and the evidence of the clinical effectiveness of these treatments, particularly IO therapy, on patients with METex14 has been mixed, potentially due to differences in study populations, designs, and outcomes assessments (9, 27–29). Insights into real-world testing patterns and clinical outcomes associated with current treatments in the METex14 advanced NSCLC (aNSCLC) population may help physicians to make appropriate treatment choices for these patients. This study used a nationwide US de-identified clinico-genomic database (CGDB) to describe real-world biomarker testing patterns and assess the clinical outcomes of the most frequently used first-line (1L) and second-line (2L) therapies (IO monotherapy or chemotherapy) in patients with METex14 aNSCLC.

The time associated with biomarker testing in aNSCLC is known as a critical operational issue impacting treatment decision-making (26); however, few studies have provided up-to-date assessments on this topic. We found only two studies in this domain (33, 34), neither of which looked at testing time for NGS in recent years. Our study utilized the most recent data available to provide contemporary assessment on the timing associated with biomarker testing (both PD-L1 and NGS) from collection of tumor specimen to receipt of test results among patients with METex14 aNSCLC. For patients with advanced cancer, the time needed for receiving NGS testing results may affect real-world treatment decisions regarding the adoption of targeted therapy. Importantly, NGS testing requires about 2 weeks for results and costs more than individual single-gene tests; however, the overall cost is lower and the time to all results is shorter compared with multiple sequential single-gene tests for all potentially actionable oncogenic drivers, which makes NGS testing a preferred method for comprehensive molecular profiling (22). It should be noted that the assessment of time from biopsy to receiving testing results, in relation to the subsequent clinical benefit of patients receiving various therapies, has not been well studied. This study examined real-world testing patterns for PD-L1 and FMI NGS testing and assessed the clinical outcomes with chemotherapy and IO monotherapy in patients with METex14 aNSCLC prior to capmatinib US approval.

Our results show that the time from specimen collection to reported results was shorter for PD-L1 testing than for FMI NGS testing in 2019, with a median difference of 1 week. Whether this arose from differences in clinical decision-making, characteristics of the 2 forms of testing, or some combination of these factors could not be ascertained with the data available. Additionally, the duration between specimen collection and report date in this study should not be interpreted as turnaround time, which is defined as the time between test order and report date; note that specimen collection could have occurred before or after the test order in real-world practice. In this study, based on a primarily community oncology setting in the US, we found that twice as many patients with METex14 initiated 1L therapy prior to receiving their FMI NGS reports as initiated 1L therapy after receiving their reports. Although multiple lung cancer guidelines recommend early comprehensive genomic profiling prior to systemic treatment initiation (4, 35, 36), our study showed that testing was often postponed until after the treatment decisions were made, and therefore the optimal clinical outcomes were not achieved. Our study sample was restricted to those who received NGS testing; if otherwise assessed in a broader sample of patients with NSCLC, it is expected that comprehensive biomarker testing may be further delayed or not completed. One of the primary reasons for this delay or lack of NGS testing, excluding cost, is the consideration around the additional waiting or turnaround time to obtain testing results. Because NGS testing turnaround time may be shortened due to advancement of testing technology in recent years, we attempted to further explore the most recent data, which showed a difference of only about 1 week between NGS testing and PD-L1 testing. In our analysis, we saw a trend toward more patients receiving an NGS report prior to treatment initiation in recent years, and fewer patients having the opposite order ( Table 2 and Figure 2 ). The shortened turnaround time and increased awareness of the importance of early NGS testing could be reasons leading to the positive shift with regard to testing pattern shown in Table 2 and Figure 2 . The data we presented here will further help physicians to rationalize the best treatment options for patients, as well as communicate to patients the importance of waiting for NGS testing results to achieve the best clinical outcomes. Furthermore, a few studies also support the idea of waiting in order to ensure that the most appropriate treatment is initiated first (19, 37, 38). Importantly, earlier treatment initiation for advanced NSCLC may not be associated with longer survival (37, 38), and a large retrospective study has shown that in patients with actionable oncogenic drivers, receiving a targeted therapy is associated with longer survival (19).

Our analysis of measures of effectiveness for chemotherapy or IO monotherapy in patients with METex14 demonstrated a median 1L rw-PFS of 5.7 months and 2.4 months in patients receiving 1L chemotherapy and 1L IO monotherapy, respectively, and a median 2L rw-PFS of 3.5 months and 4.7 months in patients receiving 2L chemotherapy and 2L IO monotherapy, respectively. Patients who received IO monotherapy in 1L showed 3- and 6-month 1L rw-PFS rates of 33% and 17%, respectively, indicating that most patients experienced rapid disease progression even with these treatments. Our analysis also found a median OS of 20.0 months in patients receiving 1L chemotherapy but not reached in patients receiving 1L IO monotherapy, respectively, and a median OS of 15.3 months and 19.3 months in patients receiving 2L chemotherapy and 2L IO monotherapy, respectively. These results are consistent with previous reports showing lower overall survival in patients with METex14 NSCLC receiving nontargeted therapies (chemotherapy and IO) relative to those receiving targeted therapies (9, 39). In a retrospective chart review study, patients who received treatment with a MET inhibitor had a median overall survival of 25.3 months, whereas patients who did not receive a MET inhibitor had a median overall survival of 10.9 months (39). In another retrospective review of stage IV NSCLC patients with METex14 comparing those who received a MET tyrosine kinase inhibitor in 1 or more lines of therapy (n=27) with those who did not (n=34), the median overall survival was 24.6 months and 8.1 months, respectively. Among patients who received crizotinib as their first MET inhibitor, the median PFS was 7.4 months (9). Patients in these studies could have received a MET inhibitor in any line of treatment, not just in 1L. A report from the Sarah Cannon Research Institute showed that patients with METex14 NSCLC were responsive to MET inhibitor therapy after receiving standard-of-care therapy in 1L (40).

A phase 2 study of capmatinib in patients with aNSCLC and confirmed METex14 (GEOMETRY mono-1) demonstrated substantial antitumor activity regardless of the line of therapy in which capmatinib was used. The median progression-free survival (PFS) was 12.4 months (95% CI, 8.2 months-NE) and 5.4 months (95% CI, 4.2-7.0 months) among treatment-naive patients in the 1L setting and previously treated patients in the 2L or third-line (3L) setting, respectively (11). A post hoc analysis of patients treated with capmatinib in the 2L or 3L setting in the GEOMETRY mono-1 trial reviewed responses in patients who did or did not receive IO in the 1L or 2L setting. Among patients who received prior IO, the median PFS was 8.3 months (95% CI, 4.2-12.6 months; data not mature). Among patients without prior IO, the median PFS was 5.4 months (95% CI, 4.2-6.9 months) (41). Furthermore, a recent analysis reviewed treatment-naive patients with METex14 aNSCLC treated with 1L capmatinib in the GEOMETRY mono-1 trial compared with a matched real-world cohort treated with 1L chemotherapy and/or immunotherapy. The median PFS was 12.0 months (95% CI, 5.5-20.7 months) and 6.2 months (95% CI, 3.4-9.1 months) for patients treated with 1L capmatinib or 1L chemotherapy and/or immunotherapy, respectively. Additionally, after left truncation was accounted for in the real-world dataset, the median overall survival was 20.8 months (95% CI, 12.6 months-not reached) and 14.8 months (95% CI, 9.0 months-not reached) for patients treated with 1L capmatinib or 1L chemotherapy and/or immunotherapy, respectively (31). Taken together, both real-world and clinical trial evidence highlights the need to test all patients with NSCLC for oncogenic driver mutations, including METex14, and the importance of early testing.

A retrospective review of 24 patients with METex14 NSCLC receiving single-agent (anti–PD-L1) or combination (anti–PD-1 and anti–CTLA-4) IO reported an ORR of 17%, with short DOR and PFS (28). In the 21 assessable patients, the median PFS was 1.9 months (95% CI, 1.7-2.7 months) (28). Furthermore, in the retrospective IMMUNOTARGET registry study, 23 patients with advanced METex14 treated with immunotherapy had a median PFS of 4.7 months (95% CI, 1.8-7.8 months) (26). These results are consistent with case reports showing limited efficacy of IO in patients with METex14 NSCLC (29, 42). In each of these 4 studies or reports, responses were limited even in tumors with high PD-L1 expression (26, 28, 29, 42). A recent study with a small sample size (N=13) by Mayenga et al, however, showed prolonged responses to 2L IO treatment in 6 patients with METex14 NSCLC previously treated with chemotherapy (N=13), with 5 patients showing a partial or complete response within the first 4 months (27). This study finding contradicts the results of all other studies with larger sample sizes, although it indicates that the response to IO in patients with METex14 is still not fully understood; results based on only 13 patients pose the risk of overinterpretation of the finding (27, 28). The importance of broad molecular testing for driver mutations is further underscored by evidence showing that patients harboring driver mutations may not respond well to IO, in the case of genes such as EGFR and ALK (26, 43–45), or may have only modest responses, in the case of BRAF (26, 46). Accordingly, the National Comprehensive Cancer Network Clinical Practice Guidelines in Oncology for NSCLC now considers the presence of oncogenes that predict lack of benefit from single-agent IO or combination IO-chemotherapy a contraindication to the use of these agents (4). In particular, the guidelines note that patients with METex14 and high PD-L1 expression do not generally respond to IO (4).

Our study found that pembrolizumab and other PD-L1 inhibitors and chemotherapies were used frequently as 1L treatments in this patient population during the study time frame (January 1, 2011, to December 31, 2019). Although newly approved MET inhibitors will likely change the treatment landscape, the clinical outcomes associated with IO or chemotherapy in this study are still very much relevant and reflective of the current outcomes associated with these regimens. In our study, a small number of patients received IO plus chemotherapy or MET inhibitors as 1L treatments during the study period; however, due to the limited number of patients receiving these treatments, we are unable to provide data on the clinical outcomes with these treatment regimens.

The data for this study are available from Flatiron, but restrictions apply to the availability of these data, which were used under license for the current study and so are not publicly available. Data are, however, available from the authors upon reasonable request and with permission of Flatiron.

The studies involving human participants were reviewed and approved by WCG IRB. Written informed consent for participation was not required for this study in accordance with the national legislation and the institutional requirements.

FAZVK was involved in the conceptualization, investigation, formal analysis, and supervision. BC was involved in the conceptualization, investigation, formal analysis, and supervision. HK was involved in the investigation and formal analysis. JK was involved in the investigation and formal analysis. VD was involved in the conceptualization and formal analysis. SZ was involved in the formal analysis. MS was involved in the conceptualization and formal analysis. All authors were involved in the writing, review, and editing of the manuscript. All authors read and approved the final manuscript.

FAZVK is an employee of Novartis Pharma AG. BC, VD, SZ, and MS are employees of Novartis Pharmaceuticals Corporation. HK and JK are employees of Genesis Research.

The authors declare that this study received funding from Pharmaceuticals Corporation, East Hanover, NJ, USA. The funder had the following involvement with the study: Employees of the company were involved in medical accuracy review.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.