Mechanisms of acquired resistance to first- and second-generation EGFR tyrosine kinase inhibitors

Abstract

Patients with non-small-cell lung cancer (NSCLC) whose tumours harbour activating mutations within the epidermal growth factor receptor (EGFR) frequently derive significant clinical and radiographic benefits from treatment with EGFR tyrosine kinase inhibitors (TKIs). As such, prospective identification of EGFR mutations is now the standard of care worldwide. However, acquired therapeutic resistance to these agents invariably develops. Over the past 10 years, great strides have been made in defining the molecular mechanisms of EGFR TKI resistance in an effort to design rational strategies to overcome this acquired drug resistance. Approximately 60% of patients with acquired resistance to the EGFR TKIs (erlotinib, gefitinib, and afatinib) develop a new mutation within the drug target. This mutation—T790M—has been shown to alter drug binding and enzymatic activity of the mutant EGF receptor. Less common mechanisms of acquired resistance include MET amplification, ERBB2 amplification, transformation to small-cell lung cancer, and others. Here, we present a condensed overview of the literature on EGFR-mutant NSCLC, paying particular attention to mechanisms of drug resistance, recent clinical trial results, and novel strategies for identifying and confronting drug resistance, while also striving to identify gaps in current knowledge. These advances are rapidly altering the treatment landscape for EGFR-mutant NSCLC, expanding the armamentarium of available therapies to maximize patient benefit.

Introduction

Activating mutations within the tyrosine kinase domain of the epidermal growth factor receptor (EGFR) are found in ∼10%–15% of lung adenocarcinomas in the United States and Europe and in ∼40%–50% of cases in Asia [1–5]. The most common alterations include exon 19 deletions and the L858R point mutation in exon 21, collectively accounting for up to 90% of EGFR mutations in the clinic. The presence of these mutations is associated with sensitivity to EGFR tyrosine kinase inhibitors (TKIs), several of which have been developed to date (Table 1). The remaining 10% of EGFR mutations show variable TKI sensitivity [6], with some being sensitive (e.g. G719X, L861Q, S768I, EGFR fusions, and kinase domain duplications) [6–9], while others show comparatively poor response to first- and second-generation EGFR TKIs (e.g. most EGFR exon 20 insertions) [6, 10, 11].

Table 1.

EGFR TKIs with regulatory approval for the treatment of EGFR-mutant NSCLC

Generation	Name	Selectivity	Reversible/Irreversible	Approval status
1st	Gefitinib	WT EGFR	Reversible	FDA, EMA
	Erlotinib	WT EGFR	Reversible	FDA, EMA
	Icotinib	WT EGFR	Reversible	CFDA
2nd	Afatinib	WT EGFR	Irreversible	FDA, EMA, CFDA
3rd	Osimertinib	Mutant EGFR	Irreversible	FDA, EMA
	Olmutinib	Mutant EGFR	Irreversible	KFDA (conditional)

WT, wild-type; FDA, US Food and Drug Administration; EMA, European Medicines Agency; CFDA, China Food and Drug Administration; KFDA, Korea Food & Drug Administration.

EGFR-mutant lung cancers display unique biological and clinical phenotypes, being one of the paradigmatic examples of oncogene-driven tumours. Among the major EGFR downstream effectors (i.e. RAS-RAF-MEK-ERK1/2, PI3K-AKT, and JAK2-STAT3/5), EGFR-mutated cells activate essential anti-apoptotic AKT and STAT signalling pathways, on which they have survival dependency [12]. From a clinical standpoint, patients with EGFR-mutant non-small-cell lung cancer (NSCLC) are often never smokers or light former smokers [1–4]. In addition, EGFR-mutant lung cancers may have a predilection for metastatic spread to the liver [13] and to the central nervous system (CNS) [14, 15]. Although a higher CNS tropism of EGFR-mutated disease has not been definitely established, it is remarkable that ∼25% of advanced-stage patients have baseline brain metastasis, and up to 50% will develop CNS metastasis during the course of their disease [14, 16].

First- and second-generation EGFR TKIs (Table 1), which were designed against the wild-type EGF receptor, yield robust and relatively durable responses in the majority of lung cancer patients whose tumours harbour EGFR sensitizing mutations. The first-generation drugs, including erlotinib and gefitinib, are reversible inhibitors. The second-generation inhibitors, including afatinib and dacomitinib, are irreversible inhibitors which covalently bind to EGFR. Multiple phase 3 clinical trials have shown that patients with EGFR-mutant tumours display >70% radiographic response rates (RRs) and a statistically significant improved progression-free survival (PFS) when treated with first-generation (erlotinib and gefitinib) or second-generation (afatinib) EGFR TKIs (Table 1) compared with platinum-based chemotherapy [17–20]. Consequently, these three drugs are globally approved without distinctions for these patients (Table 1). In addition, the first-generation TKI icotinib is approved in China (Table 1). Based on recently presented phase III data, the second-generation TKI dacomitinib may also be considered for regulatory approval [21].

Unfortunately, these drugs do not cure patients. Acquired resistance, defined as systemic progression of disease (based on RECIST or WHO criteria) while on continuous treatment with an EGFR TKI [22], remains a significant barrier to overcome to deliver the most effective treatments to patients with EGFR-mutant NSCLC. Typically, acquired resistance develops after a median of 9.2–14.7 months of TKI therapy [17, 18, 20, 21, 23–26]. Extensive research over the past few years has characterized some of the major resistance mechanisms [27, 28], allowing the development of effective therapies to overcome acquired therapeutic resistance.

In this article, we summarize the latest evidence for practices which maximize the benefit of first- and second-generation EGFR inhibitors. We also provide an update on the molecular subtypes of acquired resistance. Finally, we comprehensively review current data on methods for identifying T790M-mediated versus non–T790M-mediated resistance, and we suggest a treatment algorithm considering clinical progression patterns together with molecular resistance mechanisms.

Maximizing the efficacy of first-line EGFR TKI therapy

In experimental cell-line based assays and xenograft tumour models, the second-generation EGFR inhibitors showed a higher potency than erlotinib or gefitinib against both EGFR sensitizing mutations (i.e. exon 19 deletion and L858R) and the EGFR T790M resistance mutation [27–30]. These findings, together with a broader pan-ErbB inhibition offered by second-generation EGFR TKIs [29, 30], served as the rationale to compare the efficacy of first- versus second-generation inhibitors in two randomized clinical trials. In the phase IIb LUX-Lung 7 trial, 319 patients were randomly assigned to receive first-line afatinib or gefitinib. There was a statistically significant increase in RR (70% versus 56%; P = 0.0083) and PFS (primary end point: hazard ratio (HR) 0.73, 95% confidence interval (CI) 0.58–0.92; P = 0.0073) in favour of afatinib [23], but overall survival (OS) differences (27.9 versus 24.5 months) did not reach statistical significance (HR 0.86, 95% CI 0.66–1.12; P = 0.258) [31]. Of note, there was no incremental benefit favouring afatinib in tumours with exon 19 deletions compared with L858R mutations [31], as was initially suggested from a retrospective pooled analysis of the LUX-Lung 3 and LUX-Lung 6 trials [6]. In the ARCHER 1050 trial, a randomized phase III study that enrolled 452 patients without CNS metastasis, dacomitinib showed a more robust PFS compared with gefitinib (14.7 versus 9.2 months; HR 0.59, 95% CI 0.47–0.74; P < 0.0001) [21]. Data on OS are still immature and have not been reported for this trial. Thus, both studies consistently suggest that more potent front-line pan-ErbB inhibition delays the emergence of resistance and disease progression. However, the adverse event profile was more pronounced with both afatinib [23] and dacomitinib [21] compared with gefitinib, predominantly due to an increase in skin and gastrointestinal toxicities. Therefore, this benefit to risk ratio should be individually evaluated when deciding between first- and second-generation TKIs for the front-line treatment of patients with EGFR-mutant lung cancer.

Combinatorial therapies with EGFR TKIs have also been investigated as an alternative to prolong disease control. Among them, anti-angiogenics have proven effective. Increased vascular endothelial growth factor (VEGF) levels have been detected in xenograft models of EGFR TKI-resistant lung cancer [32]. The combination of EGFR TKI plus anti-VEGF therapy delayed the onset of T790M-mediated resistance compared with single-agent EGFR TKI and also resulted in tumour regression in TKI-resistant EGFR L858R/T790M-positive tumours [32]. Translating these results to the clinic, a randomized phase II trial conducted in Japan (n = 152) showed that the combination treatment of erlotinib plus the anti-VEGF monoclonal antibody bevacizumab significantly prolonged PFS over erlotinib alone (16 versus 9.7 months: HR 0.54, 95% CI 0.36–0.79; P = 0.0015), without significant differences in RR (69% versus 64%) [33]. Typical bevacizumab-related toxicities were significantly increased in the combination arm (> 40% higher incidences of hypertension, proteinuria, and bleeding events), but no significant differences in erlotinib-related adverse events were noted [33]. Encouraging efficacy data have been recently reported in single-arm trials conducted with gefitinib in Japan (n = 42, median PFS = 14.4 months) [34] and erlotinib in Europe (n = 109, median PFS = 13.8 months) [35]. In the European study, and consistent with the preclinical data, the benefits of the combination seemed higher in patients with baseline EGFR T790M-positive tumours (n = 37, 34%), who reached a median PFS of 16 months and a 1-year survival rate of 72.4% [35]. Bevacizumab has regulatory approval in Europe, United States, and Asia for the first-line treatment of EGFR-mutant NSCLC. A confirmatory phase III trial is currently ongoing (NCT02633189). Ramucirumab is also being tested in combination with erlotinib in another randomized phase III study in this setting (NCT02411448).

Another consideration when deciding among the optimal front-line treatment options is which, if any, of these agents maximizes CNS disease control. The CNS is considered a ‘sanctuary site’, in that the amount of drug which effectively crosses the blood–brain barrier is an obstacle to achieve maximal antitumour effect in the CNS. Studies suggest that <2% of plasma drug concentration at steady state is detected in cerebrospinal fluids [36, 37], and it is likely that the concentration needed to inhibit EGFR above the IC₅₀ is only maintained for a short time in the CNS at conventional doses [38]. Clinically, no significant differences in terms of CNS responses between these compounds have been reported in small series [39–41]. However, some CNS responses have been reported with afatinib in patients previously treated with erlotinib or gefitinib [36]. On the other hand, bevacizumab has been shown to prevent CNS metastasis in mouse models of lung adenocarcinoma [42], prompting speculation that the combination of an EGFR TKI plus bevacizumab could prevent or delay CNS metastasis in humans, although this has not yet been clinically evaluated.

Molecular mechanisms of resistance to EGFR TKIs: target-dependent mechanisms

For proper disease management, clinical indicators of progression must be analysed alongside biological information about underlying mechanisms of drug resistance. While multiple mechanisms can lead to TKI resistance, the EGFR T790M mutation is the most common mechanism of acquired resistance to first- and second-generation EGFR TKIs, being present in 50%–60% of the cases (Figure 1) [27, 28, 43]. The T790M mutation was first reported in 2004 as a rare variant co-occurring with EGFR L858R in a pulmonary resection from a TKI-naive patient [44]. In 2005, two groups independently identified EGFR T790M as a secondary mutation occurring in patients who had had progressed on gefitinib after an initial response [45, 46]. The T790 residue is located within the ATP-binding pocket of the EGFR protein and mediates TKI resistance by increasing protein affinity for ATP [47]. When it occurs in conjunction with activating mutations, T790M lowers K_m[ATP], the concentration of ATP necessary to achieve a half-maximal reaction rate [47]. These biochemical changes return ATP affinity to a level that more closely resembles wild-type EGFR, thereby decreasing the efficacy of first- and second-generation EGFR TKIs [47]. The accompanying decrease in k_cat, however, leads to decreased ATP throughput and lowering enzymatic turnover. This likely explains why, in the absence of EGFR TKIs, the T790M mutation confers a growth disadvantage in cells with canonical EGFR activating mutations [48].

Figure 1.

Mechanisms of acquired resistance to first- and second-generation EGFR TKIs. Frequencies (given in parentheses) are approximations. Only mechanisms that have been detected in patients and reported in at least two independent data series are included.

Evaluating EGFR T790M status is necessary in the setting of acquired resistance as it informs selection of subsequent therapies and is frequently required for access to clinical trials. Originally, re-biopsy of tumour tissue was the only option for acquiring material to assess T790M status [27, 28]. However, recently liquid biopsy genotyping has become an increasingly attractive alternative to tissue re-biopsy. Liquid biopsies offer a number of potential improvements over tissue biopsy, including decreased cost, improved patient safety, quicker turnaround time [49, 50], and a more ‘system-wide’ assessment to counteract heterogeneity between (or even within) different metastases [50, 51]. Liquid biopsy studies use either of two blood-based sources of genetic material: circulating tumour DNA (ctDNA) or circulating tumour cells (CTCs). Analysis of ctDNA from patients who were known to harbour T790M in their tumour was first reported in 2009 [52]. The first case in which longitudinal blood-based EGFR testing revealed acquisition of T790M was reported in 2013, using whole-exome sequencing of ctDNA [53]. Since then, more than a dozen studies have looked at the feasibility of liquid biopsies for identifying EGFR resistance mutations, utilizing ctDNA [54–63] as well as CTCs [57, 61]. Recently, a prospective study of 180 patients showed acceptable correlation between ctDNA and tissue re-biopsy [50]. Specifically, sensitivity of ctDNA-based detection of T790M via digital droplet PCR was calculated at 77.1%, specificity was 63.2%, and positive predictive value was 79% [50]. These findings are in line with retrospective analyses of plasma-based versus tissue-based detection of T790M [57, 64]. In one such retrospective analysis, the RR (62% plasma and 63% tumour) and median PFS (9.7 months for both plasma and tumour) were similar in patients with T790M-positive plasma or T790M-positive tumour [64]. However, based on these data, a negative plasma T790M test (i.e. T790M not detected in plasma) was insufficient to identify patients who could potentially benefit from osimertinib [64]. The RR in patients with T790M plasma-negative/tumour-positive was 69% compared with 25% in patients with T790M plasma-negative/tumour-negative results, indicating that plasma-based testing alone would have missed some patients who would have responded to osimertinib [64]. Based on these results, the National Cancer Comprehensive Network (NCCN) guidelines and the European Society for Medical Oncology (ESMO) guidelines now include plasma genotyping as an option at disease progression when tissue-based testing is not feasible [65, 66], although they both still encourage secondary re-biopsy to confirm a negative plasma assessment of T790M.

Aside from the T790M mutation, only a handful of other secondary EGFR mutations that desensitize tumours to erlotinib, gefitinib, and afatinib have been reported. Case reports with accompanying in vitro validation exist only for D761Y [67, 68], L747S [69, 70], and T854A [71]. EGFR L747S has also been reported as a rare variant in TKI-naive patients [72]. A more common event is EGFR amplification, which occurs in ∼10% of patients who develop acquired resistance to these therapies. However, EGFR amplification is always detected in the presence of EGFR T790M (Figure 1) [67, 73]. It remains an open question whether EGFR amplification is an early event that promotes drug tolerance in these cells, a late event that helps cells circumvent deleterious effects of T790M, or simply a passenger event.

Molecular mechanisms of resistance to EGFR TKIs: target-independent mechanisms

Fewer than half of the cases of acquired EGFR TKI resistance are attributed to non–EGFR-centric adaptations. Of these, many are often classified as ‘bypass’ resistance mechanisms, as they utilize alternative cellular pathways to activate the same key downstream effectors of tumour cell growth and survival as EGFR, namely ERK1/2 and AKT1. The most common ‘bypass’ mechanism for resistance to first- and second-generation EGFR TKIs is amplification of ERBB2, a gene that encodes for the ErbB family member HER2. Although most well studied in breast cancer, where ERBB2 amplification drives 15%–30% of tumours [74], ERBB2 has also been implicated in other tumour types, most notably endometrial and gastric cancer [75, 76]. ERBB2 amplification occurs in ∼10%–15% of patients with EGFR TKI resistance (Figure 1) [77]. To date, there are no HER2-specific targeted therapies approved for patients with EGFR-mutant cancer who develop resistance via ERBB2 amplification, nor are there enough data to suggest whether these patients would benefit from HER2-targeted therapies.

Amplification of the MET gene, which encodes the MET tyrosine kinase receptor, is reported in ∼5% of patients with acquired resistance to first- and second-generation EGFR TKIs (Figure 1) [27, 28, 78]. MET is activated by hepatocyte growth factor, thereby potentiating growth, survival, and invasion pathways via Src, PI3K, and RAS family members [79]. Although there are no targeted therapies specifically approved for patients with EGFR-mutant NSCLC who progress on EGFR TKI due to MET amplification, there are case reports of clinical benefit when combining a MET inhibitor with an EGFR TKI in EGFR-mutant lung cancer patients whose tumours acquire MET amplification [80, 81]. Two phase Ib/II trials of the MET inhibitor capmatinib (INC280) alone or in combination with erlotinib (NCT02468661) and gefitinib (NCT01610336) are currently under way.

Other ‘bypass’ mutations have been reported at a lower frequency (Figure 1). These include mutations in BRAF [82], PIK3CA [27], KRAS [63], PTEN loss [83], NF-1 loss [84], and CRKL amplification [85]. Furthermore, resistance mediated by non-genetic changes has also been described, including up-regulation of IGF1R [86], FGFR [87], hepatocyte growth factor, and the ligand for MET [88, 89]. Likewise, activation of AXL kinase is associated with resistance to EGFR TKIs [90].

These rare events, along with the more common ERBB2 and MET amplifications, sometimes co-occur with EGFR T790M [28, 63]. This is likely a result of metastatic heterogeneity [28, 63]. This phenomenon, which remains understudied, is becoming increasingly clinically relevant. Although binary designation of acquired resistance as T790M positive or T790M negative is useful for making therapeutic decisions, there is evidence that up to 46% of EGFR-mutant tumours with acquired resistance to first- or second-generation EGFR TKIs demonstrate co-occurring mechanisms of resistance [63]. These co-occurring mutations likely influence treatment outcomes [63].

Lastly, changes in tumour phenotype at time of disease progression are also well documented. Up to 10% of tumours from patients who fail EGFR TKI therapy show transformation to small-cell lung cancer morphology (Figure 1) [27, 28]. These transformed tumours display molecular profiles similar to classical small-cell lung cancer, such as loss of RB activity [91, 92] and TP53 activity [92]. While the tumours retain the original EGFR mutation, EGFR expression is drastically reduced [27, 91], suggesting that these tumours are no longer driven by EGFR signalling [91]. Epithelial-to-mesenchymal transition is also reported at the time of disease progression, although this is rare (1%–2% of the cases) [27, 90, 93]. There is also a recent report of two patients with adeno-to-squamous transformation concurrent with T790M acquisition [94], but this phenomenon has yet to be reported more broadly. It is important to note that any histological change at the time of disease progression on EGFR TKI therapy will be missed if one relies on liquid biopsies alone.

Potential pharmacologic mediators of patient response to EGFR TKIs

As noted previously, the blood–brain barrier is perhaps the most frequently encountered mediator of intra-patient differences in EGFR TKI bioavailability and efficacy [36–41]. Perhaps the most striking effect of poor CNS penetrance is differential disease evolution in response to TKI therapy, with brain metastases frequently showing different biological mechanisms of acquired drug resistance [27, 95]. As such, poor penetration of EGFR TKIs into the CNS is likely a factor that contributes to disease progression in otherwise TKI-responsive patients.

Other mediators of drug bioavailability have also been studied. There are limited reports detailing non-biologic mediators of EGFR TKI exposure, although the extent to which these factors contribute to patient response may not be fully elucidated. Smoking, for example, results in decreased erlotinib plasma concentrations [96]. Indeed, patients with EGFR-mutant lung cancer who have a history of smoking have worse PFS with EGFR TKIs than non-smokers [97–99], although reduced plasma concentration was never shown to be causative. On the contrary, the phase III CurrentS study demonstrated that achieving higher plasma concentrations by increasing erlotinib dosage from 150 mg daily (the standard) to 300 mg did not result in increased PFS or OS for smokers [100].

Likewise, antacid use is associated with altered absorption of erlotinib and perhaps gefitinib. Since these drugs are weakly basic, having a less-acidic stomach environment pushes them toward the non-ionized form, which is thought to decrease absorption [101]. One study has shown retrospectively that concomitant use of proton pump inhibitors (PPIs) correlated with worse survival outcomes for patients who had taken erlotinib [102], although a separate, smaller study of gefitinib therapy concomitant with PPI use showed no such decrease in survival (507 patients in erlotinib study; 43 patients in gefitinib study) [103]. As such, the development of the third-generation EGFR TKI osimertinib included a trial assessing pharmacokinetic changes in the presence of PPIs (NCT02224053). To further investigate this potential adverse drug–drug interaction, one prospective study investigated the use of cola, an acidic beverage, to counteract the effects of PPIs. Patients taking erlotinib plus PPI had increased erlotinib bioavailability when erlotinib was administered with cola, though this study did not assess whether there were changes in drug efficacy [104].

Clinical identifiers of and treatment options for EGFR TKI resistance

As described above, the Jackman criteria [22] are used to define acquired therapeutic resistance. This definition is usually considered as a standardized criterion for clinical trial selection, but substantial heterogeneity exists regarding clinical progression patterns. Acquired resistance can be categorized according to the presence or absence of symptoms derived from progressive disease, kinetics of tumour growth, and number of progressive metastases. Three clinical subtypes of acquired resistance according to the extent and sites of progressive disease are generally accepted: (i) systemic or multi-site progression, (ii) oligo-progression (three or less progressing locations), and (iii) isolated CNS progression [105]. Data from retrospective series have shown that most patients (∼60%–70%) usually experience systemic progression, commonly (50%) at original metastatic sites [106–110]. Fewer patients fall into the oligo-progressive (∼20%–25%) [106, 108, 109] and CNS-only (∼15%) [111, 112] categories. Notably, the patterns of clinical progression have been shown to impact prognosis [106, 108] and therefore should be considered when dealing with resistance in the clinic.

When a patient develops symptomatic multisite progression, second-line systemic treatment is recommended. Currently, the choice of second-line systemic therapy is primarily based on the presence or absence of the EGFR T790M mutation, detected either in plasma or tumour tissue. For the ∼50%–60% of patients who develop the T790M mutation on progression with erlotinib, gefitinib, or afatinib, the current standard therapy is osimertinib, a mutant-selective EGFR TKI which was designed to overcome the T790M resistance mutation [113]. Based on convincing clinical trial data with this agent [114, 115], osimertinib is approved in the United States, Europe, and Asia for the treatment of EGFR-mutant NSCLC patients with T790M-mediated acquired resistance. Besides osimertinib, other mutant-selective EGFR TKIs are being developed (Table 1). They will be discussed in a separate review in this supplement of the journal.

For those patients with T790M-negative tumours at the time of acquired resistance, standard treatment is platinum-based chemotherapy (Figure 2). When platinum-based chemotherapy is considered, the IMPRESS trial showed that maintaining the EGFR TKI did not further improve survival outcomes [116]. In contrast, in the context of indolent systemic progression, it is safe to continue the EGFR TKI and delay the initiation of a new systemic treatment (Figure 2). Thus, in the ASPIRATION trial, 54% of the patients were deemed to be candidates to continue first-line erlotinib beyond extracranial RECIST progression at physician discretion, thereby delaying salvage therapy for a median of 3.1 months in these subjects [117]. Of note, up to 20% of the patients that are candidates to continue post-progression erlotinib do so for >12 months before initiating a new systemic therapy [109]. Patients with longer time from best overall response to disease relapse, presence of slow-growing disease, and progression without extrathoracic metastasis are the better candidates for this approach [109, 117]. Some studies have suggested a survival benefit for continuing EGFR TKI beyond progression [106, 107, 117, 118], but this is likely a result of selection bias. Continuing EGFR TKI beyond RECIST progression in asymptomatic progressors could be an alternative to second-line platinum-based chemotherapy in patients with T790M-negative tumours (Figure 2).

Figure 2.

Proposed algorithm for the management and treatment of non-central nervous system progression. Red text refers to standard treatment options. CNS, central nervous system; TKI: tyrosine kinase inhibitor.

In patients who progress at a limited number of metastatic sites, the use of local ablative therapies (radiation or surgery) in progressing areas while continuing the EGFR TKI is another treatment alternative (Figure 2), based on the assumption of an ongoing benefit of the TKI within non-progressing lesions [119–121]. This method extended systemic disease control for >10 months in highly selected groups of individuals with non-CNS oligoprogressive disease [120, 121].

Finally, medical treatment of CNS progression is challenging and requires taking into consideration the presence of neurological symptoms, extent of CNS involvement, previous radiotherapy to the brain, and the status of systemic disease (controlled versus not controlled) (Figure 3). In cases with isolated CNS progression, the use of local ablative therapies to the CNS, such as stereotactic surgery, is considered a standard of care. The EGFR TKI can be continued after local therapy is completed [65, 66]. Indeed, there are several studies demonstrating that with this approach extracranial progression is delayed for a median of 6–10 months in this subgroup of patients [111, 112, 119]. However, some isolated reports and small case series have reported CNS responses in patients with progressive brain metastasis on standard EGFR TKI doses or leptomeningeal disease [95, 122, 123]. This approach has not been tested in clinical trials, and thus remains experimental.

Figure 3.

Proposed algorithm for the management and treatment of isolated central nervous system progression. Red text refers to standard treatment options. CNS, central nervous system; WBRT, whole brain radiotherapy; TKI, tyrosine kinase inhibitor.

Discussion

Conclusion

New compounds and tools for the diagnosis and treatment of EGFR-mutant NSCLC continue to be developed. Among these, blood-based monitoring of disease progression may offer a safer and faster way of informing treatment decisions compared with tissue biopsy. Likewise, novel compounds and drug combinations offer the potential to delay or overcome acquired resistance to EGFR TKIs. Rapid adoption of evidence-based treatment paradigms offer the best hope for maximizing the survival benefit gained from first-line treatment of patients with EGFR-mutant NSCLC. Finally, it should be noted that there remains an unmet need to identify the cause of acquired resistance for 10%–15% of patients who stop responding to erlotinib, gefitinib, or afatinib. Interesting, although mechanisms of resistance to third-generation inhibitors are also incompletely understood, early reports implicate many of the same mechanisms as seen with earlier-generation TKIs [63]. Elucidation of resistance mechanisms to first- and second-generation EGFR TKIs should continue to be pursued as these TKIs are sure to retain clinical utility, and newly identified mechanisms will inform development of next-generation EGFR TKIs.

Funding

DW was supported by a Ruth L. Kirschstein NRSA Fellowship (T32HL094296). JZ was funded by Instituto de Salud Carlos III (Rio Hortega, CM15/00196). BCC was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Science, ICT & Future Planning (2016R1A2B3016282). CML received support from the National Institutes of Health (NIH) and National Cancer Institute (NCI) R01CA121210 and P01-CA129243. CML was also supported by a Damon Runyon Clinical Investigator Award, a LUNGevity Career Development Award, a V Foundation Scholar-in-Training Award, an AACR-Genentech Career Development Award, and U10CA180864. LPA was funded by ISCIII: PI1401964, PIE15/00076, RTICC (R12/0036/0028) and CIBERONC (C16/12/00442), co-funded by FEDER from Regional Development European Funds (European Union) (no grant number applies).

AstraZeneca has provided a sponsorship grant towards this independent publication.

Disclosure

BCC received research funding from Novartis, Bayer, AstraZeneca, MOGAM Institute, Dong-A ST. BCC has also served as a consultant for Novartis, AstraZeneca, Boehringer-Ingelheim, Roche, BMS, Yuhan, Pfizer, and Eli Lilly. CML has served as a consultant for Pfizer, Novartis, AstraZeneca, Genoptix, Sequenom, and ARIAD and has been an invited speaker for Abbott and Qiagen. All remaining authors have declared no conflicts of interest.

Key Message

When dealing with EGFR TKI resistance in EGFR-mutant NSCLC, clinical indicators of progression must be analyzed alongside biological information underlying mechanisms of drug resistance.