Sequencing Therapy for Genetically Defined Subgroups of Non–Small Cell Lung Cancer

OVERVIEW

The practice of precision medicine for patients with metastatic non–small cell lung cancer (NSCLC), particularly those patients with adenocarcinoma histology (the predominant subtype of NSCLC), has become the accepted standard of care worldwide. Implementation of prospective tumor molecular profiling and rational therapeutic decision-making based on the presence of recurrently detected oncogenic “driver” alterations in the tumor genome has revolutionized the way that lung cancer is diagnosed and treated in the clinic. Over the past two decades, there has been a deluge of therapeutically actionable driver alterations and accompanying small molecule inhibitors to target these drivers. Herein, we synthesize a large and rapidly growing body of literature regarding therapeutic inhibition of driver mutations. We focus on established targets, including EGFR, anaplastic lymphoma kinase (ALK), ROS1, BRAF, RET, MET, HER2, and neurotrophic tyrosine kinase receptor (NTRK), with a particular emphasis on the sequencing of small molecule inhibitors in these genetically defined cohorts of patients with lung cancer.

SEQUENCING AGENTS IN THE TREATMENT OF PATIENTS WITH EGFR-MUTANT LUNG CANCER

EGFR-mutant lung cancers represent 12% to 17% of all lung adenocarcinomas.^1,2 EGFR mutations—most commonly small deletions in exon 19 (19del) or point mutations in exon 21 (L858R)—identify a subset of patients who are most appropriately treated with first-line EGFR TKIs. Many prospective studies have compared first-line EGFR TKIs with standard cytotoxic platinum doublet chemotherapy and have confirmed the superior response rates and progression-free survival with EGFR TKIs compared with chemotherapy in patients with EGFR-mutant lung cancers.^3–5 First-generation (erlotinib, gefitinib) and second-generation (afatinib) EGFR TKIs have multiple global approvals for first-line treatment of patients with metastatic EGFR-mutant lung cancers (Table 1). Dacomitinib (a second-generation EGFR TKI) has recently been compared head to head with gefitinib, revealing an increase in median progression-free survival (mPFS; 14.7 vs. 9.2 months, respectively).⁶ The results on overall survival are expected this year and could lead to a new regulatory approval.

TABLE 1.

Targeted Therapies Currently Approved and Under Development for Lung Cancer

Mutation and Drug	Category/Description	Approval Status
		United States	European Union
EGFR mutations
Erlotinib (OSI744)	First-generation EGFR TKI	X	X
Gefitinib (ZD1839)	First-generation EGFR TKI	X	X
Afatinib (BIBW2992)	Second-generation pan-HER TKI	X	X
Dacomitinib (PF299804)	Second-generation pan-HER TKI
Osimertinib (AZD9291)	Third-generation EGFR TKI	X	X
Nazartinib (EGF816)	Third-generation EGFR TKI
ALK fusion
Crizotinib (PF2341066)	First-generation ALK, MET, ROS1, multikinase inhibitor	X	X
Ceritinib (LDK378)	First-generation ALK, ROS1, multikinase inhibitor	X	X
Alectinib (CH5424802)	Second-generation ALK TKI	X	X
Brigatinib (AP26113)	Second-generation ALK, mutant EGFR, ROS1, multikinase inhibitor	X
Ensartinib (X396)	Second-generation ALK TKI
Lorlatinib (PF6463922)	Third-generation ALK, ROS1, multikinase inhibitor
ROS1 fusion
Crizotinib	ROS1, ALK, MET, multikinase inhibitor	X	X
Ceritinib	ROS1, ALK, multikinase inhibitor
Brigatinib	ROS1, ALK, mutant EGFR, multikinase inhibitor
Lorlatinib	ROS1, ALK, multikinase inhibitor
Entrectinib	ROS1, ALK, TrkA/B/C (NTRK1,2,3), multikinase inhibitor
DS-6051b	ROS1, TrkA/B/C (NTRK1,2,3), multikinase inhibitor
BRAF V600E mutation
Vemurafenib	BRAF V600E selective inhibitor (serine/threonine protein kinase inhibitor)
Dabrafenib	BRAF V600E selective inhibitor (serine/threonine protein kinase inhibitor)	X	X
Trametinib	MEK1-2 inhibitor	X	X
RET fusion
Cabozantinib	RET, MET, VEGFR2, AXL, c-KIT, FLT3, multikinase inhibitor
Vandetanib	RET, EGFR, VEGFR multikinase inhibitor
Sunitinib	RET, PDGFR, VEGFR1-3, c-KIT, FLT3 multikinase inhibitor TKI
Sorafenib	RET, VEGFR1-3, PDGFR, RAF, c-KIT, FLT3 multikinase inhibitor
Alectinib	RET, ALK, FLT3 multikinase inhibitor
Lenvatinib	RET, VEGFR1-3, FGFR1-4, PDGFR multikinase inhibitor
Nintedanib	RET, VEGFR, FGFR, PDGFR, Flt3, multikinase inhibitor
Ponatinib	BCR-ABL+, FLT3, SRC, c-KIT, FGFR, VEGFR, RET multikinase inhibitor
Regorafenib	RET, VEGFR, PDGFR, FGFR, KIT, RAF multikinase inhibitor
MET amplification
Crizotinib	MET, ALK, ROS1 multikinase inhibitor
Onartuzumab	Anti-MET monoclonal antibody
Tivantinib	MET selective inhibitor
Savolitinib	MET selective inhibitor
Glesatinib	MET, AXL, multikinase inhibitor
Capmatinib	MET selective inhibitor
Cabozantinib	MET, VEGFR2, AXL, c-KIT, FLT3, RET multikinase inhibitor
MET exon 14 skipping mutation
Crizotinib	MET, ALK, ROS1 multikinase inhibitor
Capmatinib	MET selective inhibitor
NTRK fusion
Entrectinib	TrkA/B/C (NTRK1/2/3), ALK, ROS1 multikinase inhibitor
Loxo-101	TrkA/B/C (NTRK1/2/3) TKI
DS-6051b	TrkA/B/C (NTRK1/2/3), ROS1 multikinase inhibitor
HER2 amplification
Trastuzumab	HER2 monoclonal antibody
Ado-trastuzumab emtansine	HER2 monoclonal antibody
HER mutation
Afatinib	Pan-HER TKI
Dacomitinib	EGFR, HER2 TKI
Lapatinib	EGFR, HER2 TKI
Neratinib	EGFR, HER2 TKI
Trastuzumab	HER2 monoclonal antibody
Ado-trastuzumab emtansine	HER2 monoclonal antibody

Established Treatments

These first- and second-generation EGFR TKIs are effective therapies, but the majority of patients develop disease progression on these agents after 8 to 10 months.^3–5,7 It has become standard of care to rebiopsy patients at the time of clinical progression, and the acquired molecular alterations identified serve as the molecular mechanisms of resistance to EGFR TKI treatment.^7,8 Another emerging option is liquid biopsy where tumor cfDNA within plasma is utilized for mutation testing. This has emerged as a viable alternative to tumor rebiopsy. The most common acquired mutation is EGFR T790M, but other acquired alterations include MET amplification, HER2 amplification, PIK3CA mutations, small cell histologic transformation, and epithelial to mesenchymal transition.^7,8 To address these resistance mechanisms to EGFR TKIs, multiple combination treatments using first- and second-generation EGFR TKIs have been assessed. EGFR TKIs have been combined with EGFR antibodies,^9,10 mTOR inhibitors,¹¹ HDAC inhibitors,¹² HSP90 inhibitors,¹³ MET inhibitors,¹⁴ dasatinib,¹⁵ cabozantinib,¹⁶ and other agents with limited efficacy seen with the combinations.

Osimertinib

Osimertinib is a third-generation, mutant-selective, covalent EGFR inhibitor (Table 1) that targets both the sensitizing EGFR mutations as well as EGFR T790M. Its initial approval in many countries is for patients with EGFR-mutant lung cancers who were previously treated with an EGFR TKI and have acquired EGFR T790M. This approval was based on the phase I AURA study¹⁷ and confirmed by a randomized study of osimertinib versus platinum doublet chemotherapy in this clinical setting.¹⁸ Patients treated with osimertinib as second-line EGFR TKI had an overall response rate (ORR) of 71% and an mPFS of 10.1 months. Based on the efficacy in the later-line setting, osimertinib was subsequently assessed prospectively in a randomized phase III study of osimertinib or standard of care EGFR TKI (erlotinib or gefitinib) as first-line treatment of patients with metastatic lung adenocarcinoma in the FLAURA study.¹⁹ In the first-line setting, the mPFS on osimertinib was 18.9 months compared with 10.2 months with standard EGFR TKIs. Based on these data, osimertinib is expected to receive global approval as a first-line treatment option for patients with metastatic EGFR-mutant lung cancers.

Resistance to Osimertinib and New Combination Therapies

Patients treated with osimertinib also develop resistance attributable to acquired molecular alterations. There are less robust data to suggest which mechanisms of acquired resistance occur frequently and are clinically meaningful, and we do not yet have any data on mechanisms of resistance to first-line treatment with osimertinib. Resistance mechanisms identified to third-generation EGFR TKIs include acquired EGFR C797S mutation (C797 is the site at which osimertinib binds to the EGFR kinase domain),^20,21 loss of EGFR T790M,²² MET and HER2 amplification,^23,24 YES1 amplification,^25,26 and acquired mutations, including KRAS, PIK3CA, and HER2.^24,27

New combinations will presumably be assessed both in the first-line setting and after progression on osimertinib to attempt to prevent and reverse resistance to osimertinib, respectively. The combination of osimertinib and savolitinib, an MET inhibitor, has shown activity in patients with MET amplification after EGFR TKI therapy with erlotinib, afatinib, gefitinib, or osimertinib.²⁸ Other combinations that are being assessed include osimertinib and bevacizumab (NCT02803203, NCT03133546, and NCT02971501), osimertinib and selumetinib (NCT03392246), osimertinib and dasatinib (NCT02954523), osimertinib and the JAK inhibitor INCB039110 (NCT02917993), osimertinib and navitoclax (NCT02520778), and osimertinib and an mTOR inhibitor MLN0128 (NCT02503722) among others. With the milder toxicity profile of osimertinib compared with earlier-generation EGFR inhibitors, combination studies may prove to be more efficacious by reaching optimal doses of both drugs before being limited by toxicity.

Sequencing of EGFR Inhibitors

As a general practice, our most effective treatments should be used first, because there is a clear minority of patients who progress quickly with declining functional status and do not receive second-line therapy—as high as 50% in historical data sets.²⁹ Other factors that should be considered when choosing a first-line treatment of this population include toxicity profile and central nervous system (CNS) efficacy. In the FLAURA study, there were fewer patients with greater than or equal to grade 3 toxicities, fewer fatal adverse events, and a lower rate of adverse events leading to permanent discontinuation with osimertinib compared with standard EGFR TKI.¹⁹ In addition, there were fewer events of CNS progression with osimertinib compared with standard EGFR TKI (6% vs. 15%). Osimertinib is well positioned to be the new first-line treatment of choice because of a marked improvement in efficacy and superior CNS penetration/efficacy while also being better tolerated by patients.

One potential criticism of first-line osimertinib is that there are no approved EGFR-directed treatments after progression on osimertinib. However, only 50% of patients on standard EGFR TKI acquire EGFR T790M and are eligible for second-line osimertinib, and the additive time on treatment with a standard EGFR TKI followed by osimertinib is essentially equivalent to the time on first-line osimertinib. After clinical progression on first-line osimertinib in the setting of C797S and the absence of EGFR T790M, there are preclinical data to suggest efficacy of a standard EGFR TKI, such as gefitinib, in that setting.³⁰ We await clinical data that such sequencing is a viable option.

Other Strategies to Improve Outcomes for Patients With EGFR-Mutant Lung Cancers

There is substantial heterogeneity in the clinical course of patients with EGFR-mutant lung cancers. As next-generation sequencing of tumors becomes the standard of clinical care, we will be able to discern the impact of concurrent genetic alterations in addition to activating EGFR mutations within lung cancers. Several reports have shown concurrent TP53 alterations to be a negative prognostic factor^31,32 associated with shorter overall survival. In addition, several pretreatment concurrent alterations, including HER2 amplification, MET amplification, and TP53 mutations, seem to be associated with shorter time to progression on EGFR TKIs.³³ Highlighting concurrent alterations that may have prognostic significance is important, because it identifies a subset of patients with poorer outcomes on whom new therapeutic options should be focused.

Data from the ASPIRATION study and others have shown the utility of treatment with EGFR TKIs beyond radiographic progression in the setting of more indolent disease and in the absence of symptoms.³⁴ Treatment beyond progression allows for additional clinical benefit and delays the time before new treatments are required. Cancers can be heterogeneous in their response and resistance to therapy, such as a scenario where the majority of the target lesions continue to respond to treatment while one lesion has begun to grow. In this situation, local therapy to the oligoprogressive metastasis followed by continuing previous systemic therapy is another means to prolong time on treatment that is largely continuing to benefit a patient.³⁵

SEQUENCING AGENTS IN THE TREATMENT OF PATIENTS WITH ALK-REARRANGED LUNG CANCER

ALK-Rearranged Lung Cancer

Rearrangements in the gene encoding the ALK on chromosome 2p were first discovered as oncogenic driver alterations in NSCLC in 2007.³⁶ These chromosomal rearrangements result in the production of a chimeric fusion protein—most commonly EML4-ALK, although several other ALK fusions have been described. ALK rearrangements are detected in approximately 4% to 8% of NSCLCs^2,37 and can be detected in tumor samples by several diagnostic measures, including immunohistochemistry, fluorescence in situ hybridization, and next-generation sequencing.

Overview of ALK TKIs in Clinical Use

Now, approximately 10 years after the initial discovery of ALK as a lung cancer driver, there have already been numerous large, international, prospective clinical trials testing the efficacy of ALK TKIs in this patient population. Remarkably, five ALK TKIs have already gained regulatory approval (Table 1). These ALK TKIs can be broken down into three “generations” of inhibitors defined by increasing “on-target” efficacy toward ALK. Crizotinib was the first ALK TKI developed in the clinic and the first ALK TKI to obtain regulatory approval. Crizotinib also targets MET and ROS1 (described below). Therefore, the development of more potent and more specific second- and third-generation inhibitors was needed (Table 1). Below, we summarize a large amount of clinical trial data discussing the sequencing of ALK TKI therapies.

First-line Therapy for ALK-Rearranged Lung Cancer

Crizotinib was the first ALK TKI to be approved for first-line treatment of ALK+ lung cancer based on the PROFILE 1014 study.³⁸ This phase III trial enrolled 343 patients who were treatment naïve and randomized them to crizotinib or chemotherapy (platinum/pemetrexed). Cross over to crizotinib treatment after disease progression was permitted for patients receiving chemotherapy. The ORR was 74%, mPFS was 10.9 months, and hazard ratio (HR) for progression or death in the crizotinib group was 0.45 (95% CI, 0.35–0.60). Crizotinib is approved by the U.S. Food and Drug Administration (FDA) and the European Medicines Agency for first-line therapy of ALK+ NSCLC.

More recently, ceritinib, a second-generation ALK TKI, has also been approved for first-line treatment of ALK+ lung cancer based on the ASCEND-4 study.³⁹ This study enrolled patients with metastatic ALK+ NSCLC who were treatment naïve and had asymptomatic, neurologically stable brain metastases; 189 patients were randomly selected to receive ceritinib (750 mg orally once a day), and 187 patients were randomly selected to receive platinum/pemetrexed for four cycles (maintenance pemetrexed was allowed). Crossover to ceritinib was permitted after disease progression for patients receiving chemotherapy. mPFS was 16.6 months (95% CI, 12.6–27.2) in the ceritinib group and 8.1 months (95% CI, 5.8–11.1) in the chemotherapy group (HR 0.55; 95% CI, 0.42–0.73]; p < .00001).

In patients with measurable CNS lesions at baseline, the confirmed overall intracranial response rate was 57% (95% CI, 37%–76%) in the ceritinib arm and 22% (95% CI, 9%– 42%) in the chemotherapy arm. The most common adverse events in the ceritinib arm were diarrhea (85% all grades, 5% grade 3/4), nausea (69% all grades, 3% grade 3/4), vomiting (66% all grades, 5% grade 3/4), and an increase in liver function enzymes. The ASCEND-8 trial⁴⁰ evaluated a lower dose of ceritinib (450 mg) in patients and found it to be similarly efficacious; the mPFS in the 450-mg dose arm was 17.6 months compared with 10.9 months in the 750-mg dose arm. There were also improvements with the 450- versus 750-mg dose arm for grade 3/4 diarrhea (1.1% vs. 7.8%), nausea (0% vs. 5.6%), and vomiting (0% vs. 4.4%).

Alectinib, another second-generation ALK TKI, has also been tested in the first-line setting. In the global phase III ALEX study, 286 patients with advanced/metastatic ALK+ NSCLC who were treatment naïve were enrolled and randomized to receive alectinib (600 mg orally twice a day; 152 patients) or crizotinib (250 mg orally twice a day; 151 patients).⁴¹ Platinum-based chemotherapy was not the comparator in the ALEX study. Asymptomatic brain metastases were allowed. mPFS (investigator assessment) was not reached (17.7 months to not reached) for alectinib and 11.1 months (9.1–13.1) for crizotinib, with an HR of 0.47 (95% CI, 0.34–0.65; p < .0001). For patients with baseline CNS metastases, mPFS was not reached (9.2 months to not reached) for alectinib and 7.4 months (6.6–9.6) for crizotinib, with an HR of 0.40 (95% CI, 0.25–0.64). For patients without baseline CNS metastases, mPFS was not reached for alectinib and 14.8 months (10.8–20.3) for crizotinib, with an HR of 0.51 (95% CI, 0.33–0.80). Compared with the crizotinib arm, patients in the alectinib arm experienced more myalgias (16% all grades), weight gain (10% all grades), and laboratory abnormalities, including elevated bilirubin (15% all grades) and anemia (20% all grades). Alectinib is approved by the FDA and the European Medicines Agency for first-line therapy of advanced/metastatic ALK+ NSCLC.

Several other ALK TKIs are being tested in the first-line setting. The ALTA-1L study (NCT02737501) comparing the efficacy of brigatinib versus crizotinib has completed accrual. Ensartinib (eXalt3; NCT02767804) and lorlatinib (NCT03052608) are also being tested against crizotinib in patients with treatment-naïve ALK+ lung cancer. The issue with many of these studies is that they began enrollment before the global ALEX trial was reported.

Second-Line ALK Inhibitor Therapy and Beyond

Since crizotinib was the first ALK TKI studied and approved, most of the available data to date focus on the efficacy of second- and third-generation ALK inhibitors (Table 1) in patients with acquired resistance to crizotinib. This topic has been reviewed extensively in the literature^42,43 and for sake of space constraints and also relevance (with the emergence of first-line ceritinib and alectinib), will only be reviewed briefly here. In general, the factors that are important for selection of a second-line ALK TKI for a patient with acquired resistance to crizotinib are systemic activity, CNS activity, safety/adverse events profile, and resistance profiles. At present, ceritinib,^44–46 alectinib,⁴⁷ and brigatinib^48,49 are all FDA approved for patients who are intolerant of or have experienced disease progression with crizotinib. Response rates range from approximately 40% to 60% across the studies, with mPFS of approximately 5 to 15 months depending on the study. Each of these inhibitors has documented CNS activity as well.

With the emergence of second-generation inhibitors, such as ceritinib and alectinib, in the first-line setting of metastatic ALK+ NSCLC, it is not precisely clear what standard of care will emerge for second-line treatment after acquired resistance to one of these more potent ALK TKIs. It is beneficial to understand acquired resistance to these inhibitors to develop rational therapeutic strategies at the time of disease progression. As seen with other TKIs in clinical use, resistance mechanisms encompass two broad categories—“on-target” mechanisms (such as kinase domain mutations and genomic amplification of the ALK fusion) and “off-target” mechanisms (predominantly “bypass” signaling pathways). Here, we will focus on overcoming “on-target” resistance mechanisms (“off-target” resistance mechanisms will be explored in greater detail below). Although current data sets are limited, at the time of acquired resistance to second-generation ALK TKIs, ALK kinase domain mutations will be detected in approximately 50% of drug-resistant tumors. In contrast to the experience with EGFR-mutant lung cancer, where T790M is the dominant (more than 60%) resistance mutation,^7,8 many different ALK resistance mutations (such as C1156Y, I1171T/N/S, F1174L/C, V1180L, R1192P, L1196M, G1202R, D1203N, S1206Y/C, E1210K, A1280V, G1269A, and others) have been described.^50,51 Of these mutations, the G1202R solvent front mutation is thought to be the most recalcitrant mutation, with increasing frequency of this mutation detected with increased “on-target” potency of the ALK inhibitor.

Lorlatinib is a third-generation ALK TKI, which was designed to be increasingly selective/potent against ALK and have increased brain penetration (lorlatinib is also an ROS1 inhibitor as described in below). Lorlatinib has activity against most known ALK kinase domain mutations, including G1202R.⁵² Initial reports from the ongoing international phase I trial of lorlatinib (NCT01970865) have recently been published.⁵³ Forty-one patients with ALK+ NSCLC received at least one dose of lorlatinib, 52% of whom had received two or more prior TKIs and 72% had documented CNS metastases. The ORR was 46% (95% CI, 31%–63%) for all patients who were ALK+ and 42% for those patients who had received two or more prior ALK TKIs. Adverse effects of lorlatinib included hypercholesterolemia (72% across the entire study), hypertriglyceridemia (39%), peripheral edema (39%), and peripheral neuropathy (39%). Neurocognitive adverse events (slowed speech, slowed mentation, and word-finding difficulty) were also observed, although the precise frequency was not defined.

More recently, the results from the phase II lorlatinib study were presented.⁵⁴ This trial included five cohorts of patients with ALK+ NSCLC (and a sixth cohort for ROS1). In cohort 1 (30 patients who were ALK+ treatment naïve), the ORR was 90%, and the intracranial ORR was 75%. In cohorts 2 (27 patients) and 3A (32 patients; cohorts 2 and 3 included patients who were ALK+ and had previously received crizotinib only [cohort 2] or crizotinib with or without chemotherapy [cohort 3A]), the ORR was 69%, and the intracranial ORR was 68%. In cohort 3B (28 patients who were ALK+ and had previously received a noncrizotinib ALK inhibitor with or without chemotherapy), the ORR was 33%, and the intracranial ORR was 42%. In cohorts 4 (65 patients) and 5 (46 patients; cohorts 4 and 5 included patients who were ALK+ and had previously received two prior TKIs [cohort 4] or three prior TKIs [cohort 5] with/without chemotherapy), the ORR was 39%, and the intracranial ORR was 48%. These studied led the FDA to give lorlatinib breakthrough therapy designation for the treatment of patients with ALK+ metastatic NSCLC previously treated with one or more ALK inhibitors.

Rational Combination Strategies to Improve Outcomes for Patients With ALK-Rearranged Lung Cancers

Resistance to ALK TKI therapy also can be mediated by numerous ALK-independent mechanisms, most commonly thought to be activation of “bypass” signaling pathways that circumvent the inhibited ALK fusion protein. Bypass signaling pathways that have been shown to occur clinically include EGFR pathway activation,⁵⁵ IGF-1R pathway activation,⁵⁶ SRC signaling,⁵⁷ cKIT amplification,⁵⁵ and MAPK pathway activation (via KRAS copy number gain or loss of the phosphatase DUSP6⁵⁸). Histologic changes, such as epithelial to mesenchymal transition⁵⁰ and transition to small cell lung cancer,⁵⁹ have also been described. Several ongoing clinical trials hope to address the proper sequence of ALK TKIs and the role of rational combination therapies to maximize benefit and outcomes for patients with ALK+ NSCLC. For example, the combination of ceritinib and the allosteric MEK inhibitor, trametinib, is being tested in an ongoing phase I/II trial (NCT03087448). Ceritinib is also being evaluated in combination with the CDK4/6 inhibitor ribociclib/LEE011 (NCT02292550).⁶⁰ Finally, ALK TKIs are being tested in combination with immune checkpoint inhibitors; however, retrospective data suggest that the magnitude of benefit seen in this cohort of patients with ALK+ NSCLC treated with checkpoint inhibitors is small.⁶¹ Additional rational study designs to forestall or overcome ALK TKI resistance are urgently needed.

OTHER TARGETABLE GENOMIC ALTERATIONS IN PATIENTS WITH LUNG CANCER

ROS1 Rearrangements

Approximately 1% of lung adenocarcinomas are driven by oncogenic ROS1 rearrangements.⁶² The ROS1 and ALK kinase domains show considerable homology, explaining crizotinib’s high affinity for both.⁶³ In the phase I PROFILE 1001 study, among 50 patients with ROS1-rearranged NSCLC, the ORR was 72% with a disease control rate of 90%, and mPFS reached 19.2 months.⁶³ In a prospective phase II study and a retrospective EUROS1 study, mPFS times were 10 and 9.1 months respectively, with ORRs of 72% and 80%, respectively.^64,65 In a larger Asian phase II study, mPFS in 127 patients was 13.4 months.⁶⁶ These studies led to crizotinib approval by the FDA (March 2016) and the European Medicines Agency (August 2016) for treatment of advanced ROS1-rearranged NSCLC.

Ceritinib is a second-generation ALK TKI that also has efficacy against ROS1. In a Korean phase II study, among 32 patients with ROS1 rearrangement (all crizotinib naïve except for two patients), the ORR was 67%, and mPFS reached 19.3 months; however, clinical response was not observed in the two patients who had received crizotinib.⁶⁷ Other ALK TKIs—including brigatinib and lorlatinib—have shown potential anti-ROS1 activity in early development studies.⁶⁸ Among three patients treated with brigatinib, the patient who was crizotinib naive had a partial response (ongoing at 21.6 months), and in 12 patients treated with lorlatinib (seven were crizotinib pretreated), the ORR was 50% with an mPFS of 7.0 months.^48,53

Other potential ROS1 inhibitors include entrectinib (a potent inhibitor of ALK, ROS1 kinase, and TRK), which was evaluated in 14 patients who were crizotinib naive (13 with NSCLC and one with melanoma) in two phase I studies, giving an ORR of 86% (no responses in six patients who were pretreated with crizotinib) and an mPFS of 19 months.⁶⁹ Preliminary results from a Japanese phase I study of DS-6051b (a ROS1/TRK inhibitor) gave an ORR of 62.5% among 13 patients with NSCLC (no responses in three patients who were pretreated with crizotinib).⁷⁰

The most common mechanism of resistance to crizotinib in ROS1+ NSCLC is mediated by the ROS1 Gly2032Arg mutation, analogous to ALK Gly1202Arg.⁵⁰ Second-generation ALK TKIs with ROS1 activity (ceritinib, brigatinib, and entrectinib) are ineffective against ROS1 Gly2032Arg. Lorlatinib retains potent activity against ROS1 Gly2032Arg in vitro and in vivo.⁵² Research is needed to assess its efficacy in ROS1+ patients who have relapsed after treatment with available TKIs.

BRAF Mutations

The most common BRAF mutation, V600E (Val600Glu), is observed in 1% to 2% of lung adenocarcinomas.^1,71 In the phase II VE-BASKET trial with vemurafenib including patients with various BRAFV600-mutant tumors, the ORR and the mPFS in an NSCLC cohort of 19 patients were 42% and 7.3 months, respectively.⁷² In a recent phase II study of dabrafenib as monotherapy or combined with trametinib in patients with BRAFV600-mutant metastatic NSCLC (BRF113928), dabrafenib monotherapy gave a 33% ORR in 78 pretreated patients with a median duration of response of 9.6 months and an mPFS of 5.5 months.^73,74 The BRAF-MEK TKI inhibitor combination doubled the clinical benefit in 57 pretreated patients, with an ORR of 63% and an mPFS of 10.2 months.⁷⁵ The combination showed similar benefits in 36 nonpretreated patients with BRAFV600E, with an ORR of 64% and an mPFS of 10.8 months. Median overall survival was prolonged in the two combination cohorts (18.2 and 24.6 months in pretreated and nonpretreated cohorts, respectively).⁷⁶ The European Medicines Agency (April 2017) and the FDA (June 2017) have approved dabrafenib in combination with trametinib for treatment of BRAFV600E-mutant advanced NSCLC.

For non–BRAFV600-mutant NSCLC, six patients received BRAF inhibitors in the retrospective study.⁷⁷ All tumors with non-BRAFV600 mutant located outside the activation segment of the BRAF kinase domain (codons 596–600) were refractory to BRAF inhibitors. However, one patient with a G596V mutation achieved a partial response to vemurafenib. Additional studies are needed to assess the benefit of immune checkpoint inhibitors in patients with BRAFV600E mutants and also assess if patients with non–BRAFV600-mutant tumors can benefit from targeted therapies and/or immune checkpoint inhibitors. The current recommendation is to treat patients with BRAFV600E mutant with a BRAF-MEK inhibitor combination.

RET Rearrangements

RET fusions are found in 1% to 2% of NSCLCs and tend to be mutually exclusive with other oncogenic drivers.⁷⁸ KIF5BRET is the most common, with at least 10 fusion variants. Although RET-selective TKIs have not yet been developed, several multitarget agents with anti-RET activity have been evaluated that might be restricted in their ability to inhibit RET relative to their other kinase targets. The activity of multikinase inhibitors (cabozantinib, vandetanib, sunitinib, sorafenib, alectinib, lenvatinib, nintedanib, ponatinib, and regorafenib) in RET-rearranged NSCLC (ORR = 16%–47% and mPFS = 2.3–7.3 months) is clearly inferior to that seen with selective TKIs in other oncogene-addicted NSCLC models.^79–81 The activity of cabozantinib in a phase II trial in RET-rearranged tumors was comparable with monotherapy BRAF TKIs in BRAFV600E mutant, with an ORR of 28% and an mPFS of 5.5 months, but clearly inferior to combination BRAF plus MEK TKIs.⁸¹ Other RET-specific TKIs are under development, and evaluation of alternative signaling pathways is needed for combination therapies to overcome RET resistance and determine the best strategies in this population.

MET Alterations

Dysregulation of the MET pathway occurs through protein overexpression, gene amplification, mutation, and rearrangement. Several agents (TKIs or monoclonal antibodies) have been developed to target MET or its ligand, hepatocyte growth factor. Early trials focused on targeting MET overexpression (15%–70% in unselected NSCLC) but without a consensus on the definition of MET positivity. MET overexpression has been particularly associated with blocking the EGFR pathway, leading to phase II/III combinations. Two randomized phase III trials failed to show any clinical benefit in OS of onartuzumab or tivantinib in association with erlotinib in unselected patients or MET protein–overexpressing NSCLC.^82,83 Somatic MET mutations are diverse and include exon 14 skipping. The resulting mutant receptor shows increased MET signaling (truncated MET receptor leading to decreased ubiquitination and degradation of the MET protein). The diversity of MET exon 14 alterations presents challenges for diagnostic testing. MET exon 14 alterations are detected in 3% to 4% of NSCLCs, more frequently in adenocarcinoma and sarcomatoid histologic subtypes.⁸⁴ Approximately 20% to 30% of sarcomatoid carcinomas harbor MET exon 14 alterations. Case reports and cohorts have shown dramatic and durable partial responses with MET-targeting TKIs including crizotinib, capmatinib, and cabozantinib in patients with MET exon 14. Preliminary data from the phase I trial of crizotinib (PROFILE 1001) evaluating patients with advanced lung cancer with MET exon 14 alterations showed an ORR of 44%, and global retrospective series showed an mPFS of 7 months.⁸⁵ A small series has shown few responses to immunotherapy, even in patients with PD-L1 greater than or equal to 50%.⁸⁶

MET amplification causes protein overexpression and constitutive kinase activation. MET copy number gains arise from two distinct processes: polysomy and amplification. They are identified by fluorescence in situ hybridization, showing an increase in the MET to CEP7 ratio, although no clear consensus on the definition of MET positivity based on gene copy number has been reached. MET amplification occurs via acquired EGFR TKI resistance, representing bypass track signaling (5%–20% of cases) or de novo (1%–5%).⁸⁷ Crizotinib and capmatinib (INC280) have shown a potential clinical benefit in small phase I/II trials on MET amplification (MET/CEP7 greater than or equal to 5), with ORRs of 67% and 47%, respectively, that must be confirmed.^88,89

Clinical trials focusing on combined MET and EGFR TKIs for patients with acquired resistance to EGFR TKIs are ongoing. Encouraging antitumor activity has been seen with combined osimertinib and savolitinib (a selective MET TKI) in EGFR-mutated and MET-amplified patients (confirmed centrally by fluorescence in situ hybridization; MET gene copy greater than or equal to five or MET to CEP7 ratio greater than or equal to two; the TATTON trial).²⁸ The ORR in patients who are T790M− was 53%, and duration of response was not reached.

ERBB2/HER2 Alterations

The ERBB2 gene encoding HER2 is a major proliferative driver activating downstream signaling via the PI3K-AKT and MEK-ERK pathways. Aberrations in HER2 have emerged as oncogenic drivers and therapeutic targets in lung cancers, with HER2 mutations (exon 20) in 1% to 5% and HER2 amplifications in 2% to 5% of lung adenocarcinomas. Kinase domain mutations, mainly exon 20 insertions and point mutations, lead to constitutive HER2 kinase activation.⁹⁰ Although the clinical relevance in NSCLC is questionable given the lack of definition of HER2 positivity in this indication, several case reports have shown responses with HER2 TKIs in patients with an HER2 mutation, including afatinib, lapatinib, neratinib, and neratinib plus temsirolimus.^91–93 Clinical benefit is generally low: for example, only three of 26 patients with HER2 mutant had a response (ORR of 12%) with dacomitinib (pan-HER TKI), 21% had a response with combined neratinib (pan-HER TKI) and temsirolimus (mTOR inhibitor), 0% had a response with neratinib alone in phase II trials, and in a recent basket trial with neratinib (SUMMIT trial), only one objective response was observed among 26 patients with lung cancer.^93–95 A partial response was reported in a patient harboring both an HER2 exon 20 mutation and an HER2 amplification with combined trastuzumab and paclitaxel.⁹⁶ In a retrospective cohort study, ORRs of 50% and 18.2% were reported in patients with HER2 exon 20 insertions treated with combined trastuzumab-chemotherapy or afatinib, respectively, although a randomized phase II trial did not show activity with the addition of trastuzumab to chemotherapy in HER2-overexpressing lung cancer.^97,98 Targeting HER2 mutations with ado-trastuzumab emtansine-1 seems promising in mutant tumors with no copy number change, with a 44% ORR, although response was low in HER2-overexpressing patients (0% ORR in IHC2+ and 20% ORR in IHC3+).⁹⁹ Rarer HER2 variants include transmembrane domain mutations (e.g., V659 and G660) with sensitivity to afatinib and trastuzumab emtansine-1.^99,100

NTRK Rearrangements

Among recently discovered dominant oncogenic mutations, NTRK chromosome rearrangements (less than 1%) have been identified in several solid malignancies, including NSCLC. Similar to ALK and ROS1 rearrangements, recurrent gene fusions involving NTRK1, NTRK2, and NTRK3 are actionable drivers. The rarity of these fusions across differing cancer types has resulted in basket trial design for drug development with LOXO-101 (larotrectinib) and RXDX-101 (entrectinib).^69,101,102 Entrectinib is a highly potent oral ATP-competitive TKI with efficacy against TrKA (encoded by NTRK1), TrKB (encoded by NTRK2), TrKC (encoded by NTRK3), ROS1, and ALK. All three NTRK1/2/3-rearranged advanced solid tumors responded (including in one patient with NSCLC whose tumor harbored an SQSTM1–NTRK1 fusion).⁶⁹ Entrectinib seems to be active in intracranial lesions and ROS1- and ALK-rearranged tumors. A phase II basket study (STARTRK-2; NCT02568267) is currently accruing patients with NTRK-, ROS1-, and ALK-rearranged cancers with the intent of confirming the results. Promising clinical benefit has also been shown with larotrectinib, a selective pan-TRK inhibitor with an ORR of 76% (95% CI, 62%–87%) in 50 patients (7% lung cancer), regardless of type of tumor and NTRK rearrangement (NTRK1/2/3).¹⁰¹

DISCUSSION

The rapid development of genomic biomarkers that define various molecular subtypes of NSCLC (Fig. 1) and the spectrum of currently available targeted therapies against these targets (Table 1) have completely reshaped the treatment paradigm for oncogene-addicted cancers. Nonetheless, several major challenges still remain in this field.

FIGURE 1. Molecular Cohorts of Lung Cancer.

(A) In a large academic center, these are the actionable mutations prospectively identified on a next-generation sequencing mutation platform (430 genes) over a set time period.¹⁰³ (B) In a national molecular testing effort, these are the actionable mutations (six-gene panel) prospectively identified over a set time period.¹

Abbreviation: WT, wild-type.

A major challenge is making molecular testing available to the maximum number of patients. Successful implementation of personalized medicine requires widely accessible tumor molecular profiling in routine practice settings worldwide along with molecular centers for high-quality testing. Promising examples of large-scale routine molecular profiling on a national level have been reported as part of large cooperative networks in France and the United States, and they have included thousands of patients with NSCLC^1,2 (Fig. 1B). It was encouraging to see that survival was improved for patients treated with biomarker-directed targeted therapies, showing the benefit of this treatment approach. Targeted gene sequencing or whole-exome sequencing by next-generation sequencing assays is gradually being integrated into clinical practice. Their success depends on them being made broadly available to practicing clinicians, applicable to small tumor biopsies, and affordable to patients and/or the health care system with a turnaround time to obtain results that is short.

An emerging option is the use of plasma genotyping with sequencing circulating tumor DNA (including the detection of high tumor mutational burden as a potential biomarker for immunotherapy), which has the logistical advantage of being rapid, noninvasive, cheap, and nononerous for the patient.^104,105 It is expected to become a new standard in daily clinical practice in the near future but still needs standardization, especially for the use of a large panel of genes. By carrying out increasingly extensive molecular analyses, many uncommon or rare alterations are detected for which clinical significance assessment constitutes a real challenge. Nevertheless, the generation of massive volumes of data highlights that tools to support the medical interpretation and interaction between clinicians and scientists must be developed to ensure a thorough and rapid outcome. The implementation of molecular tumor boards both within hospitals and as national networks is increasingly widespread, offering an environment for discussion of all of these elements and dissemination of standardized practices and shared knowledge.

Another means of promoting wide access to genetic profiling is via innovative protocol designs, including molecular screening and both targeted therapy and immunotherapy arms for a single disease (umbrella trials) or a single targeted therapy for multiple diseases (basket trials). Examples of umbrella trials in advance NSCLC include the Lung-MAP (NCT02154490) for squamous cell carcinoma and the phase II trials Lung Matrix trial (NCT026649351), NCI-MATCH cooperative group trial (NCT01306045), and SAFIR02 lung trial (NCT02117167) for squamous and nonsquamous NSCLCs. In the SAFIR02 lung trial, high-throughput molecular analyses (comparative genomic hybridization array and next-generation sequencing) are used to evaluate whether treatment with guided targeted agents or immune checkpoint inhibitors improves clinical benefit compared with standard maintenance therapy in patients with metastatic NSCLC. For the basket trials, examples include the vemurafenib trial for BRAFV600-mutant patients, the French national AcSé trial (biomarker-driven access to crizotinib in ALK+, MET+, or ROS1+ malignancies in adults and children), the Larotrectinib (Loxo-101; a selective TRK inhibitor in patients with NTRK fusion cancer), and the recently published trial in HER2- and HER3-mutated patients.^64,72,94,101

Another major challenge is that—despite the high response rates—all targeted therapies remain effective for a finite period of time. Improvements in outcomes for our patients will build on what we understand about how cancers escape targeted therapies (Fig. 2). One strategy is to improve on target inhibition. As has been seen in ALK+ lung cancers and EGFR-mutant lung cancers, better inhibitors can and have been developed. We have seen marked improvements in progression-free survival when we compared first-line treatment with newer agents, such as alectinib and osimertinib compared with crizotinib and erlotinib/gefitinib.^19,106 Alternative dosing schedules, such as pulse dosing of targeted therapy, can be explored, especially in the setting of trying to improve CNS penetration and efficacy.¹⁰⁷ Dual target inhibition, such as the use of afatinib and cetuximab (an EGFR TKI and an EGFR antibody, respectively), can be used for maximum on target inhibition.^9,10

FIGURE 2. Mechanisms of Acquired Resistance to Targeted Therapies.

Rational combination treatments that address known mechanisms of resistance should be assessed at the time of disease progression or at initial treatment to attempt to reverse or prevent acquired resistance. Many combinations are being studied in each of the molecular subsets of lung cancer. To truly personalize medical care, it is essential to sample the tumor at the time of acquired resistance by either tumor biopsy or analysis of plasma-derived sequencing circulating tumor DNA to identify the relevant resistance mechanisms for a particular patient. Finally, we must understand how alternative treatments, such as standard cytotoxic chemotherapy and immunotherapy, fit in when sequencing treatments for patients. There are some data to suggest that some of the oncogene-driven lung cancers are less responsive to immunotherapies,⁶¹ which makes it even more important to identify in whom and when to use immunotherapy. We also can use information regarding concurrent molecular alterations to provide both prognostic and predictive information. Although the majority of patients appropriately selected have excellent responses to targeted therapies, there are clearly outliers who have primary progression or shorter limited responses to targeted therapies. If certain concurrent alterations can identify these poor responders, we can focus efforts on developing new strategies to improve outcomes for these patients.

Overall, the prospective identification and rational therapeutic targeting of oncogenic “driver” mutations have paved the way for implementing precision treatment strategies in NSCLC and are now the standard of care worldwide. Despite much success in this area, a large amount of work is still needed to optimize the effectiveness of these therapies for patients with lung cancer and understand how to best sequence our available treatments. Collaborative efforts and integration of mutational data with multiomic, functional, and clinic-pathologic data are critical steps for the future to advance our understanding of lung oncogenesis and hopefully, turn oncogene-driven lung cancer into a chronic manageable disease state.

PRACTICAL APPLICATIONS.

• Prospective tumor molecular profiling is now the standard of care for the treatment of patients with metastatic NSCLC. Molecular subtyping can be used to select patients for specific targeted therapeutic options.

• Many such molecular subtypes have been described in NSCLC—including EGFR, ALK, BRAF, ROS1, RET, MET, HER2, and NTRK—and small molecule inhibitors have been developed against each of these targets.

• The mutant-selective, third-generation EGFR TKI—osimertinib—has recently emerged as the new standard first-line therapy for metastatic EGFR-mutant NSCLC; however, it is unclear what the next best therapeutic option will be on development of osimertinib resistance. Several ongoing clinical trials are addressing combination therapies to forestall or overcome osimertinib resistance.

• “Second-generation” ALK TKIs, including alectinib and ceritinib, have shown superior clinical outcomes for the treatment of patients with metastatic ALK+ NSCLC compared with the first-generation ALK TKI, crizotinib. These more potent inhibitors are emerging as the new standard first-line therapy for patients with metastatic ALK+ NSCLC. However, it is not clear how best to sequence ALK TKI therapy or how to overcome acquired resistance that is not mediated by an ALK kinase domain mutation.

• Crizotinib has received regulatory approval for the management of ROS1+ NSCLCs, whereas the combination of a BRAF inhibitor (dabrafenib) and an MEK inhibitor (trametinib) has received regulatory approval for the management of BRAFV600E-mutant NSCLC.