Targeted Molecular Treatments in Non-Small Cell Lung Cancer: A Clinical Guide for Oncologists

Targeted molecular treatments have changed the way non-small cell lung cancer (NSCLC) is managed. Epidermal growth factor receptor (EGFR), anaplastic lymphoma kinase (ALK), v-raf murine sarcoma viral oncogene homolog B1 (BRAF), and c-ros oncogene 1 (ROS1) mutations are now used to guide specific anti-cancer therapies to improve patient outcomes. New targeted molecular treatments are constantly being developed and evaluated as a means to improve efficacy, overcome resistance, or minimise toxicity. This review article summarises the current evidence for the efficacy, resistance mechanisms, and safety of targeted molecular treatments against specific mutations in NSCLC.


Introduction
Worldwide, lung cancer is the most commonly diagnosed cancer and the leading cause of cancer mortality. It accounted for 1.8 million new cases in 2012 (12.9% of all cancers) and caused 1.6 million deaths (19.4% of all cancers) [1]. Non-small cell lung cancer (NSCLC) accounts for 85% of all lung cancer cases, and adenocarcinoma is the most frequent histological subtype, accounting for nearly 40% of all lung cancer cases [2]. Prior to the use of targeted therapy and immunotherapy, patients with advanced lung cancer had a poor prognosis. Platinum doublet chemotherapy, which was the standard of care for all patients with incurable locally advanced or metastatic NSCLC [3], achieved a response rate of 19%, and a median overall survival of 7.9 months [4]. Modern trials with targeted treatments have resulted in significantly better outcomes, with median overall survival now extending to around, or even beyond, two years in selected populations [5][6][7][8].
Specific mutations have been found to be prevalent in lung adenocarcinomas, some of which are predictors for response to targeted treatment. The most common mutations occur in kirsten rat sarcoma virus (KRAS, found in 24% of cases), epidermal growth factor receptor (EGFR, 17%), anaplastic lymphoma kinase (ALK, 3%), v-raf murine sarcoma viral oncogene homolog B1 (BRAF, 2%), and c-ros oncogene 1 (ROS1, 1-2%) [9,10]. We will explore the current evidence for targeted therapy for different mutations in NSCLC, with the aim of providing clinical guidance for oncologists treating NSCLC.

EGFR Mutation-Positive NSCLC
EGFR mutations were first described in lung adenocarcinomas in 2004 [11] and were rapidly recognised as a predictor for response to EGFR tyrosine kinase inhibitors (TKIs). The frequency of mutations in this gene varies based on phenotypic characteristics of patients. They occur more frequently in Asian non-smoking women, with an incidence up to 40% [12]. The most common EGFR mutations are exon 19 deletions and exon 21 L858R point mutations [9]. Together, these two mutations account for 90% of all EGFR mutations [13]. The remaining EGFR mutations consist of a range of rarer mutations (which can either be sensitising or non-sensitising with respect to EGFR TKIs), and include exon 18 insertions, G719X point mutations in exon 18 (1-4%), exon 20 mutations (2-5%), and complex mutations (1-2%) [14,15].
EGFR TKIs have been developed to treat EGFR mutation-positive lung cancers, and a list of these are included in Table 1.

Efficacy
Early studies investigated the efficacy of EGFR TKIs in a pre-treated population of unselected patients with advanced NSCLC. The BR.21 trial found a 30% improvement in overall survival in patients treated with erlotinib compared to placebo, with an absolute survival benefit of two months [16]. Other studies in unselected populations have not demonstrated a statistically significant overall survival benefit of using EGFR TKIs when compared to chemotherapy or placebo [17][18][19][20], in combination with chemotherapy [21,22], or as maintenance therapy after chemotherapy [23][24][25][26][27][28]. The role of erlotinib in EGFR wild-type NSCLC as maintenance therapy was most recently discounted in a phase III trial when erlotinib as maintenance treatment resulted in a median overall survival of 9.7 months compared to a median overall survival of 9.5 months when erlotinib was used on progression [29].
Although designed as a study to select patients based on phenotypic characteristics (ethnicity and smoking history), the Iressa Pan-Asia Study (IPASS) study was the first to demonstrate differential outcomes for patients treated with an EGFR TKI (gefitinib) based on the presence or absence of an activating EGFR mutation. These data were based on a subset analysis of patients, which demonstrated that the benefit of EGFR TKIs was exclusive to patients with an EGFR mutation [30]. Subsequently, trials have been performed investigating gefitinib, erlotinib, or icotinib in treatment-naive patients selected for the presence of an activating EGFR mutation. The results of these trials are summarised in Table 2. Treatment with an EGFR TKI typically resulted in superior median progression-free survival (PFS) of 9-13 months when compared to platinum doublet chemotherapy, which had median PFS in the range of 4-6 months. Furthermore, the response rates were as high as 83% in patients on an EGFR TKI, compared to 36% in patients who received chemotherapy. Due to crossover between the study arms, none of these trials demonstrated a statistically significant improvement in overall survival, which can extend up to 38 months [5,6,[30][31][32][33][34][35][36][37]. The question of whether there are meaningful differences between first-generation TKIs has been addressed in three small studies. In these studies, no significant differences between erlotinib and gefitinib were observed, although there were some differences in the pattern of side effects [47][48][49]. The benefits of EGFR TKIs observed with EGFR mutation-positive cancers does not translate to patients who have high EGFR expression identified using immunohistochemistry or increased EGFR copy number detected by fluorescence in situ hybridization [31,50].
In an effort to improve further outcomes for these patients, second-generation EGFR TKIs have been developed. Afatinib and dacomitinib were both designed to bind covalently to the mutated EGFR protein. Additionally, these agents are pan-HER inhibitors and block activation of other members of the EGFR/HER family. These agents result in superior PFS compared to chemotherapy in treatment-naïve patients with EGFR-mutated tumours [7,39,40,43]. There are only limited data comparing secondand first-generation agents to each other. In a randomised phase II study, Lux Lung 7, afatinib and gefitinib were compared as first-line therapy for treatment-naïve patients. Afatinib was found to have a statistically significant PFS benefit; however, the absolute benefit was small (0.1 months) [41,42] and there seems to be little meaningful efficacy difference between the agents. The recently published overall survival data for the phase III trial of dacomitinib compared to gefitinib did show a statistically significant improvement in median overall survival (34.1 vs. 26.8 months). Despite these results, the clinical use of dacomitinib is likely to be limited by the FLAURA trial (discussed in the next paragraph), especially given the toxicity profile of dacomitinib [44].
Osimertinib, a third-generation irreversible EGFR TKI, has greater efficacy than the first-and second-generation agents. FLAURA, a randomised study comparing first-line osimertinib to erlotinib or gefitinib, showed that, despite similar response rates, patients treated with osimertinib had better PFS (18.9 months vs. 10.2 months). In patients who had stable central nervous system (CNS) metastases at time of trial enrolment, osimertinib also had a superior PFS to the first-generation EGFR TKIs (15.2 months vs. 9.6 months). Overall survival data from this study are currently immature [45].
There are no prospective studies investigating the efficacy of EGFR TKIs in patients with uncommon EGFR mutations. Observational data with small sample sizes do indicate activity of first-generation EGFR TKIs in some of the rarer EGFR mutations; however, the response rates may be lower compared to patients with common EGFR mutations [14]. In vitro data has demonstrated that cells with exon 18 mutations had better responses to second-generation EGFR TKIs such as afatinib and neratinib compared to first-or third-generation EGFR TKIs [51]. Ad hoc analyses of trial data showed a greater benefit of afatinib in patients with point mutations and duplications in exons 18-21, with a disease control rate of 84%, median PFS of 10.7 months, and median overall survival of 19 months. Meanwhile, patients who had de novo T790M mutations or exon 20 insertions had lower response rates (15% and 9%, respectively), shorter median PFS (2.9 months and 2.7 months), and shorter overall survival (14.9 months and 9.2 months) [52]. The resistance to EGFR TKIs and the poorer prognosis associated with exon 20 mutations was also seen in a retrospective analysis of 20 patients by Noronha et al. [53]. A phase II study of poziotinib in patients with EGFR exon 20 mutant advanced NSCLC is currently recruiting, with early results suggesting activity [54]. Without phase III evidence to support a different approach, EGFR TKIs are still the recommended first-line option for patients with uncommon but activating EGFR mutations.
Although several studies have been conducted in the adjuvant setting, only one trial has been completed where patient selection was prospectively based on the presence of an activating EGFR mutation. Consequently, interpretation of results is difficult. Based on the available data, EGFR TKIs may improve PFS, though the data for overall survival remains immature [55][56][57]. A phase III trial of adjuvant osimertinib in EGFR mutation-positive patients is currently recruiting, with results expected in late 2021 [58]. EGFR TKIs have yet to be implemented into routine clinical practice in this setting.
Continuation of a first-or second-generation EGFR TKI after progression on these agents has not been shown to improve outcomes [65,66]. However, osimertinib has demonstrated activity against the exon 20 T790M mutation, which occurs when threonine at position 790 is replaced by methionine [67]. This substitution leads to increased ATP affinity, cell proliferation and survival, and, ultimately, resistance to first generation TKIs [68,69]. AURA3 compared treatment with osimertinib to standard platinum doublet chemotherapy in patients with secondary T790M mutation after progression with a first-generation EGFR TKI. Osimertinib resulted in better PFS (10.1 vs. 4.4 months) and higher response rates (70% vs. 30%). A PFS benefit in patients with stable CNS metastases was also seen with osimertinib (8.5 months vs. 4.2 months) [46].

Toxicity
The adverse events of the first-generation EGFR TKIs are often Grade 1 or Grade 2 and are less likely to cause dose reductions (~20%) or drug discontinuation (<6%) compared to chemotherapy. The most common adverse events of any grade are rash or acne (66-80%), diarrhoea (25-55%) and elevated liver transaminases, particularly of alanine aminotransferase (37-55%). Most of the Grade 3 to Grade 5 adverse events occur in <6% of patients, with the exception of elevated transaminases which can occur in up to 26% of patients. The rates of pneumonitis are uncommon, in order of 1-5% [6,30,33,34,36]. There are subtle differences in the pattern of toxicity between erlotinib and gefitinib. Erlotinib is more likely to cause rash or diarrhoea, while gefitinib is more likely to cause liver function abnormalities [49]. However, these differences have little meaningful clinical impact.
Second-generation EGFR TKIs such as afatinib are more likely to result in adverse events when compared to chemotherapy or first-generation EGFR TKIs. Adverse events that are Grade 3 or higher occur in up to 49% of patients and include severe diarrhoea (up to 15%) and rash (up to 16%) [7,41]. In most studies, these have been successfully managed with dose reductions, and consequently the rate of discontinuation of drug as a result of toxicity is comparable to first-generation EGFR TKIs (6-10%) [41,43].
Meanwhile, osimertinib has been shown to have a better toxicity compared to first-generation EGFR TKIs, with a lower frequency and lower severity of adverse events of rash and transaminase elevations [45].

ALK-Positive NSCLC
ALK rearrangements were first identified in 2007 [70]. Although a single rearrangement with echinoderm microtubule-associated protein-like 4 (EML4) was initially identified, it has become apparent that there are several variants based on the location of the rearrangement. These variants may have prognostic significance with differences in outcome noted between them [71].
ALK rearrangements are more common in younger patients who have never smoked, or who have a light smoking history [72]. The commonly used ALK TKIs are listed in Table 1, and the pivotal randomised controlled trials for ALK TKIs are listed in Table 3.

Efficacy
Crizotinib was initially developed as a mesenchymal-to-epithelial transition (MET) inhibitor. However, during phase I trials it became apparent that it had substantial activity against ALK-rearranged tumours [73]. It was the first ALK TKI to show a clinically significant benefit in ALK-positive NSCLC, when it was evaluated in the second-line setting compared to chemotherapy. Patients treated with crizotinib had a median PFS of 7.7 months compared to 3.0 months with chemotherapy [8]. Data from this study resulted in crizotinib becoming a standard treatment in this group of patients.
Subsequently, ALK TKIs have been evaluated in the first-line setting. Crizotinib was compared to standard platinum doublet chemotherapy in the PROFILE 1014 study, with better PFS and higher response rates observed for crizotinib than for chemotherapy [74]. A study with the same design but using ceritinib resulted in similar outcomes [75]. Both crizotinib and ceritinib have shown activity in stable CNS metastases compared to chemotherapy, with median PFS of 9.0 months with crizotinib (vs. 4.0 months) and 10.7 months with ceritinib (vs. 6.7 months) [75,76].
Most recently, alectinib has been compared to crizotinib in the first-line setting in the ALEX study. In this study, alectinib had superior efficacy, with the median PFS not reached in the alectinib arm and 11.1 months in the crizotinib arm. Alectinib does have meaningful CNS activity compared to crizotinib. In patients treated with alectinib, the cumulative rate of CNS metastases was markedly reduced (9.4% vs. 41.4%), and, in patients with known CNS metastases at time of trial enrolment, the response rates were higher (59% vs. 26%) [77].
Ensartinib, a third-generation ALK TKI, had response rates of 66% and a median PFS 9.2 months in an early phase I/II trial. Recruitment is currently ongoing for a first-line phase III trial of ensartinib compared to crizotinib in ALK-positive NSCLC [78].
In parallel with trials of EGFR TKIs, these studies have all allowed crossover of patients. Consequently, it has not been possible to demonstrate improvements in overall survival.

Resistance
Acquired resistance usually occurs within the first two years of ALK TKI treatment and can occur due to acquired point mutations in ALK, or due to bypass track activation via activation of EGFR or amplification of CKIT. The most common resistance mutation is L1196M, though there are often multiple resistance mutations that occur concurrently [72].
Alectinib and ceritinib have been studied in phase III trials in the second line setting after resistance to crizotinib. They had superior efficacy when compared to chemotherapy with response rates between 35-40%, a median PFS of approximately 9 months, and evidence of efficacy in patients with known CNS metastases [79,80]. Phase II trials investigating brigatinib and lorlatinib show response rates up to 65% in patients with and without CNS metastases [81,82]. A phase II trial for entrectinib, a pan-tropomyosin receptor kinase (TRK), -ROS1, and -ALK TKI, is currently underway for mutation-positive patients with solid tumours after promising phase I data [83].
Using in vitro data on cell lines, Gainor et al. demonstrated that the activity of second-line therapy is influenced by the specific resistant mutation that occurs, as shown in Table 4 [84]. Sequential ALK TKI can also lead to the development of compound ALK mutations that may suggest resistance to specific ALK TKIs [85], thus highlighting the importance of patient selection and treatment sequencing when determining the optimal management pathway for each patient.

ROS1-Positive NSCLC
ROS1 rearrangements are more likely to be present in younger, non-smoking Asian patients [10]. There are no TKIs that have been designed to specifically target ROS1. However, in early clinical trials, it became apparent that ALK TKIs had activity in ROS1-positive patients. Crizotinib had a response rate of 71.7% and a median PFS of 15.9 months in a phase II trial of 127 patients [87]. Ceritinib had a response rate of 62% and a median PFS of 9.3 months in a phase II trial of 32 patients, though the median PFS improved to 19.3 months in patients who were treatment-naïve [88]. Finally, a phase I trial of lorlatinib included 12 ROS1-positive patients and achieved a response rate of 50% with a median PFS of 7 months [82]. Importantly, the ALK inhibitor alectinib, which has a structure that is distinct from the other agents mentioned above, is not active in ROS1-mutated tumours [89].

BRAF Mutation-Positive NSCLC
BRAF V600E mutations have been noted in several tumour types, most notably melanoma. They may also occur in <5% of NSCLC and are often associated with poor response to platinum-based chemotherapy. The availability of RAF inhibitors has led to their evaluation in NSCLC, although the rarity of the mutation means that the data is limited to phase II trials. Dabrafenib in monotherapy resulted in a response rate of 33%, median PFS of 5.5 months and median overall survival of 12.7 months [90]. In keeping with the experience in melanoma, the addition of the Mitogen-activated protein kinase (MEK) inhibitor trametinib resulted in better outcomes (response rates of 64% and a median PFS of 10.9 months) [91]. These agents have regulatory approval for use in NSCLC from the Food and Drug Administration (FDA) and the European Medicines Agency.

Toxicity
The most common adverse event from dabrafenib and trametinib was pyrexia, which occurred in 64% of patients. Other toxicities such as nausea, diarrhoea, fatigue, peripheral oedema, vomiting, dry skin, anorexia, and headache each occurred in 25-36% patients [91].

KRAS Mutation-Positive NSCLC
KRAS mutations are the most common mutation found in NSCLC. They often occur at codon 12, and can rarely occur at codon 13 and 61 [92]. They are mutually exclusive with EGFR mutations and ALK translocations in almost all cases [93]. KRAS mutations are more common in smokers, and also convey a poorer prognosis [94,95].
There are no targeted treatments that have a clinically meaningful benefit in patients with KRAS mutations. While MEK inhibitors such as selumetinib and trametinib held promise in early research, benefit could not be demonstrated in larger trials in patients with advanced NSCLC. A phase III trial of second-line selumetinib and docetaxel compared to docetaxel alone showed no difference in PFS (3.9 months vs. 2.8 months) and no difference in median overall survival (8.7 months vs. 7.9 months) [96]. A phase II trial of second-line trametinib compared to docetaxel alone showed no difference in median PFS (12 weeks vs. 11 weeks) with a response rate of 12% in both arms [97].

Other Mutations in NSCLC
There are other less common mutations that have been investigated as potential drug targets. Mutations in the mesenchymal-to-epithelial transition (MET) gene that cause exon 14 skipping occur in 3% of non-squamous NSCLC, and are more likely in older patients [98]. Patients with MET mutations who never received MET inhibitor therapy had a poor prognosis (median overall survival 8.1 months), which was worse if there was concurrent MET amplification (median overall survival 5.2 months). Treatment with a MET inhibitor extended the median overall survival to 24.6 months [99]. Early trials have suggested an antitumour effect of crizotinib in patients with MET exon 14-altered NSCLC, and in patients with MET amplification [100,101]. Other MET TKIs, such as capmatinib, tepotinib, salvolitinib, cabozantinib, glesatinib, and merestinib, are currently being investigated for patients with MET mutations.

Conclusions
Targeted molecular therapies have revolutionised the management of advanced NSCLC and have become the international standard of care for patients with driver mutations. Individualised patient care has never been so important. The optimal sequencing of TKIs to provide the best outcomes for our patients is unknown, especially in the immunotherapy era of oncology. An improved understanding of molecular resistance will guide the development of new treatments and assist with decision-making about treatment selection. Funding: This research received no external funding.

Conflicts of Interest:
The authors declare no conflict of interest.