The Minority Report: Targeting the Rare Oncogenes in NSCLC

Abstract

Lung cancer is still responsible for the highest number of cancer deaths worldwide. Despite this fact, significant progress has been made in the treatment of non-small cell lung cancer (NSCLC). Specifically, efforts to identify and treat genetic alterations (gene mutations, gene fusions, gene amplification events, etc.) that result in oncogenic drivers are now standard of care (EGFR and ALK) or an intense area of research. The most prevalent oncogenic drivers have likely already been identified; thus, there is now a focus on subgroups of tumors with less common genetic alterations. Interestingly, as we explore these less common mutations, we are discovering that many occur across other tumor types (i.e., non-lung cancer), further justifying their study. Furthermore, many studies have demonstrated that by searching broadly for multiple genetic alterations in large subsets of patients they are able to identify potentially targetable alterations in the majority of patients. Although individually, the rare oncogenic drivers subgroups may seem to occur too infrequently to justify their exploration, the fact that the majority of patients with NSCLC harbor a potentially actionable driver mutation within their tumors and the fact that different types of cancers often have the same oncogenic driver justifies this approach.

Introduction

There are expected to be ~224,000 new cases of lung cancer diagnosed in the United States this year and ~160,000 deaths. Although lung cancer is still responsible for the highest number of deaths across all cancers, the prevention and treatment strategies that have emerged over the past decade have begun to make significant advances in the overall survival of this disease [1, 2]. One of the primary reasons for the improvements in survival observed in some subgroups of patients with lung cancer is the focused effort on a precision medicine approach. This paradigm matches patients whose tumors harbor specific genetic mutations with drugs that specifically target these mutations. This has been driven by the dramatic clinical benefit derived from targeting the tyrosine kinases that are activators of growth and proliferation pathways. Due to this success, mutation analysis for epidermal growth factor receptor (EGFR) and anaplastic lymphoma kinase (ALK) in lung adenocarcinoma is now recommended in the most recent NCCN guidelines for NSCLC (Version 4.2014). The next logical step is to identify additional subgroups of lung cancer that contain oncogenic drivers and pair these with targeted therapeutics that may provide clinical benefit to patients. In this article we will explore some of the more promising oncogenic drivers that are currently under clinical investigation.

ROS1 Fusions

ROS1 is receptor tyrosine kinase (RTK) encoded by the ROS proto-oncogene 1, receptor tyrosine kinase (ROS1) gene, which was discovered by its homology to the v-ros oncogene in the UR2 avian sarcoma virus [3]. The first oncogenic ROS1 fusion was unveiled in glioblastoma [4]. ROS1 fusions were later discovered in lung cancer by phosphoproteomic analysis of NSCLC cell lines [5]. When a gene rearrangement occurs in ROS1, the extracellular domain (and sometimes the transmembrane domain) is replaced by sequences from an unrelated gene leading to constitutive activation of the ROS1 kinase domain [6]. The HCC78 NSCLC cell line harbors an SLC34A2-ROS1 fusion and when treated with the ROS1/ALK/MET inhibitor crizotinib shows decreased cell viability [7]. When expressed ectopically in the basal ganglia of mice, it promotes the formation of tumors and a transgenic mouse model that expresses EZR-ROS1 induces lung adenocarcinomas in mice [8, 9]. Activation of downstream pathways (JAK/STAT, PI3K-AKT, RAS/MAPK) by FIG-ROS1 fusions, as well as response to treatment with kinase inhibitors, has been demonstrated in cholangiocarcinoma and glioblastoma [9, 10].

ROS1 fusions are most commonly identified in patient samples using fluorescence in situ hybridization (FISH) to demonstrate the presence of a chromosomal rearrangement within the ROS1 gene locus and/or polymerase chain reaction (PCR) or next-generation sequencing (NGS) to identify the translocation partner. Multiple 5’ gene partners have been identified for ROS1 fusions, including TPM3, SDC4, LRIG3, SLC34A2, CD74, EZR, and others [6, 11, 12]. In a case series of 13 Japanese patients, all fusions were associated with adenocarcinomas and were negative for EGFR and KRAS mutations [11]. Additionally, the presence of ROS1 fusions has been associated with young age and minimal tobacco history [11]. In a study of 428 NSCLC tumor samples, ROS1 fusion events were identified in 1.2% of tumors [7]. Similar to Takeuchi et al., the patients in this cohort with ROS1 fusions have low tobacco use histories. Interestingly the cohort had two patients with squamous cell histology, suggesting that this alteration, like ALK and EGFR, may not be limited to adenocarcinoma [13•]. A third cohort of 18 patients (~2% of screened patients) reported by Bergethon et al. also demonstrated a younger median age and never-smoker status and all patients demonstrated adenocarcinoma histology [14]. It is notable, however, that some sample groups were small, which makes extrapolation to the larger population of NSCLC difficult [11, 14]. Large mutation surveys have since demonstrated similar findings to these smaller studies; the Cancer Genome Atlas Research Network identified ROS1 fusion events in 4 of 230 (1.7%) tumors and Pan et al. identified 11 ROS1-positive tumors in their Asian cohort of 1,139 patients [15••, 16].

The kinase domain of ROS1 shares significant homology to ALK. As noted above, some ALK inhibitors also have demonstrated activity in cell lines with ROS1 fusion proteins. Bergethon et al. demonstrated a dramatic tumor response following crizotinib treatment of a patient with a ROS1 fusion (and importantly no ALK fusion) [14]. Ou et al. evaluated the safety and efficacy of crizotinib in patients with advanced ROS1 fusion-positive NSCLC. Although this trial is still ongoing (NCT00585195), data from the first 25 evaluable patients demonstrated an objective response rate of 56%; progression-free survival and overall survival have not yet been reported [17]. At this time, multiple, additional studies are evaluating crizotinib as well as other drugs in treatment of patients with ROS1 fusion-positive NSCLC (Table 1). Currently, there is no U.S. Food and Drug Administration (FDA)-approved treatment for ROS1 rearrangements in NSCLC.

Table 1.

Characteristics of targetable mutations in NSCLC and active clinical trials

Gene	Frequency (%)*	Predominant histology	Mechanism	Detection method	Clinical trials
ROS1 [7,4,15,16]	1–2%	AdenoCA	Fusion	FISH, NGS, PCR	NCT01945021 NCT02186821 NCT02183870 NCT02097810 NCT01970865 NCT01964157 NCT02034981 NCT01922583
RET [11,16,22]	1.2–2%	AdenoCA	Fusion	FISH, NGS, PCR	NCT01813734 NCT01639508 NCT01823068 NCT01829217 NCT01877083 NCT01831726 NCT01582191 NCT01922583 NCT02029001
NTRK1 [28]	<0.5%	AdenoCA	Fusion	FISH, NGS, PCR	NCT02048488 NCT02097810 NCT01804530
MET [15,32,33]	1–4%	AdenoCA	Gene amplification, (exon 14 splicing)	IHC, PCR, FISH	NCT02132598 NCT01610336 NCT01911507 NCT01324479 NCT02154490 NCT01039948 NCT00585195 NCT01014936 NCT02055066 NCT01472016 NCT01253707 NCT01468922 NCT01773018
FGFR [36,37]	10–20%	SCC	Gene amplification, (Fusion)	SNP array, FISH, IHC, qPCR	NCT01831726 NCT01935336 NCT02160041 NCT01004224 NCT01948297 NCT01795768 NCT01283945 NCT02109016 NCT01976741 NCT02052778 NCT01868022 NCT02154490 NCT01676714 NCT02133157 NCT02117167
HER2 [33,49–51]	1–3%	AdenoCA	Insertion/deletion	PCR, direct sequencing, mass spectrometry	NCT01922583 NCT02029001 NCT01827267 NCT01542437 NCT02183883 NCT01148849 NCT01526473 NCT02117167 NCT01306045
BRAF [33,55]	1–3%	AdenoCA	Point mutation (Fusion)	PCR, direct sequencing, MALDI-TOF, mass spec, SSCA, HRMA	NCT01531361 NCT01306045 NCT01922583 NCT02029001 NCT01336634 NCT02012231
NRG1 [62]	3.9%	Mucinous AdenoCA	Fusion	RT-PCR, FISH, gene copy number, transcriptome sequencing	None

AdenoCA, adenocarcinoma; FISH, fluorescence in situ hybridization; NGS, next generation sequencing; PCR, polymerase chain reaction; IHC, immunohistochemistry; SNP, single nucleotide polymorphism; qPCR, quantitative PCR; MALDI-TOF, matrix-assisted laser desorption/ionization time of flight; Mass spec, mass spectrometry; SSCA, single strand conformational analysis; HRMA, high-resolution melting analysis; RT-PCR, reverse transcriptase PCR.

^*Frequencies listed are derived from individual studies with different cohort characteristics and therefore are not comparable across mutation type.

Similar to ALK+ patients (and EGFR mutation-positive patients), acquired resistance to inhibitors of ROS1 has been demonstrated to occur through a secondary kinase domain mutation, G2032R, in ROS1. This ROS1 mutation results in decreased crizotinib binding [18]. Preclinical models also suggest a role for wild-type EGFR signaling as another mechanism of acquired resistance [19].

RET Fusions

RET gene fusions have long been described in papillary thyroid carcinomas and radiation-associated thyroid cancers where inversions of chromosome 10 lead to oncogenic activity [20]. RET fusions also have been described in chronic myelomonocytic leukemia as the drivers of hematopoietic differentiation to monocytic/macrophage lineage and act in the RAS pathway [21]. In lung cancer, RET fusions first came to attention as a potential therapeutic target in a study of 1,528 surgical specimens [11]. RET fusions were identified by a KIF5B split FISH assays to identify alternative fusion partners of KIF5B, which is a known fusion partner of ALK. After identifying RET fusions in 12 samples. Additional investigation led to the identification of another RET partner, CCDC6, in two additional patients. In total, the frequency of RET rearrangements was found to be 0.9% in NSCLCs and 1.2% in the adenocarcinoma subgroup in this cohort [11]. Lipson et al. used NGS to identify clinically actionable genomic alterations in their study. After first identifying a KIF5B-RET fusion in their initial cohort, they screened 561 lung adenocarcinomas and identified 11 (2%) additional KIFB-RET gene fusions [22]. Both of these studies focused on populations of light or never smokers. Additionally, Lipson et al. found a higher frequency of RET fusions in their cohort of 405 Asian patients using NGS, compared with European patients [16]. Another study of 1,139 lung adenocarcinomas from a predominantly Asian population determined the frequency of RET fusions to be 1.3% by qRT-PCR; these were found predominantly in younger patients compared with non-RET fusion containing samples [16].

The downstream targets of RET activation are thought to be the Ras/Raf/MEK and JAK/STAT pathways [23]. There are at least four FDA-approved tyrosine kinase inhibitors that have activity on the RET tyrosine kinase, ponatinib, sunitinib, vandetanib, and cabozantinib [24–26]. Clinical trials are ongoing to determine the efficacy of these agents in populations of patients with RET fusions (Table 1); preliminary results from these trials are encouraging. Cabozantinib is a multikinase inhibitor and with potent activity against RET. Preliminary results from the first three patients of an ongoing clinical trial (NCT01639508) have demonstrated partial responses in two patients and stable disease in a third patient [26]. Vandetanib is an inhibitor of RET, vascular endothelial growth factor, and EGFR tyrosine kinases, which have demonstrated efficacy in treatment of thyroid cancers. Vandetanib has been shown to be effective in the treatment of RET positive lung cancer and is currently under evaluation in a phase 2 clinical trial for treatment of patients with advanced non-small cell lung cancer (NCT01823068) [27]. Additional clinical trials are currently ongoing for patients with RET gene fusions utilizing other multikinase inhibitors, such as lenvantinib (NCT01877083), dovitinib (NCT01831726), sunitinib (NCT01829217), and ponatinib (NCT01813734). No preclinical or clinical resistance mechanisms to RET inhibitors have yet been described.

NTRK1 fusions

Neurotrophic tyrosine kinase, receptor, type 1 (NTRK1) fusions are another class of gene rearrangements recently identified in NSCLC that shows potential as a therapeutic target. The NTRK1 gene encodes the high affinity nerve growth factor receptor or TRKA protein RTK. Currently, two 5’ gene fusion partners of NTRK1 have been identified in tumor samples from NSCLC patients: MPRIP and CD74, were identified by NGS and FISH assays on tumor tissue from patients who were negative for EGFR, KRAS, ALK, and ROS1 oncogenic alterations [28]. Of the 91 individual lung adenocarcinoma samples evaluated, 3 (3%) demonstrated evidence of an NTRK1 gene rearrangement. Vaishnavi et al. reported that these fusion genes demonstrated constitutive TRKA kinase activity and induced tumor formation in nude mice. Fusions of NTRK1 with other partners have been reported in both colorectal and thyroid cancers [29]. One lung adenocarcinoma patient with an MPRIP-NTRK1 fusion was treated with crizotinib but only experienced a minor radiographic response and ultimately progressed after 3 months. The lack of a significant response in this patient was consistent with the modest activity of crizotinib as a TRKA inhibitor [28]. The Cancer Genome Atlas project failed to identify NTRK1 fusions among a group of 230 lung adenocarcinomas [30]. Currently, there are at least three clinical trials of TRK inhibitors in NSCLC (Table 1). Importantly, one of these drugs, RXDX-101 demonstrated an objective tumor response in a colon cancer patient harboring a TPM3-NTRK1 fusion [31].

MET

MET encodes the cMET RTK, which is typically activated by its cognate ligand, hepatocyte growth factor (HGF). MET gene amplification is estimated to occur in approximately 1–4% of NSCLC as measured by FISH analysis [15••, 32, 33••]. The exact percentage of patients may vary depending on the cutoff used to define gene amplification. Indeed, the precise definition of MET gene amplification appears to have implications for its use as a predictive biomarker in cancer; recent results from the phase I trial of crizotinib in a cohort of NSCLC patients with MET gene amplification demonstrated a higher likelihood of benefit in patients with higher gene amplification [34].

As demonstrated in this review, there exist multiple mechanisms of activation of a proto-oncogene (fusions, point mutations, insertions/deletions, overexpression). Seo et al. identified MET exon 14 skipping in 3 samples from a cohort of 87 (3.4%) lung adenocarcinoma patients [30]. Similarly, The Cancer Genome Atlas project identified 10 of 230 lung adenocarcinoma (4.3%) samples with evidence of MET exon 14 skipping [15••]. In both cohorts, the MET exon 14 alteration occurred only in samples without other known oncogenes. Exon 14 of MET encodes the juxtamembrane domain, which regulates endocytosis and degradation of the receptor. Absence of this domain increases expression of MET, similar to MET gene amplification, leading to ligand-independent activation to downstream signaling and its oncogenic properties [35]. Treatment of cells expressing this abnormal MET receptor inhibits MAPK signaling and decreases cell viability. Thus, MET gene amplification and MET exon 14 skipping may collectively represent 7–8% of NSCLC and are a potential target for MET inhibitors.

FGFR1

The majority of targetable oncogenic drivers identified to date are found predominantly in tumors with adenocarcinoma histology. In an attempt to identify oncogenic drivers in squamous cell lung cancer, Weiss et al. screened 155 squamous cell lung cancers with single nucleotide polymorphism (SNP) arrays to uncover additional targetable genomic alterations. They identified fibroblast growth factor receptor 1 (FGFR1) gene amplification in 15 of 155 squamous cell lung cancers, but no FGFR1 gene amplification in 77 lung adenocarcinomas [36]. Using FISH on an independent dataset the authors identified FGFR1 amplification in 34 of 153 patients with squamous cell carcinoma. In both cohorts, all patients were current or former smokers. Heist et al. had similar results when they looked retrospectively at a cohort of 226 squamous cell lung cancer patients and found that 16% of them had amplifications in gene copy number to a ratio of 2.2 or greater [37].

FGFR1 is a member of a family of RTKs, including FGFR2, FGFR3, and FGFR4, and activates the phospholipase-Cy, RAF-MAPK, and PI3K/AKT pathways when triggered by ligand binding [38]. The oncogenic potential of this amplification also was demonstrated by two groups who showed that inhibition of FGFR1 in NSCLC-FGFR1 amplified cells decreased viability significantly [36, 39]. Further work has demonstrated that mRNA and protein levels may be better predictors of oncogenic potential than gene copy number as increased gene copy number did not always correlate with high expression or drug sensitivity [40•]. Clinical trials of several multikinase inhibitors with activity against FGFR1 are underway. Clinical trials of dovitinib, an FGFR1/multikinase inhibitor, have demonstrated promising results in breast cancer and renal cell carcinoma [40•, 41]. BGJ398, a pan FGFR inhibitor, has recently demonstrated promising results in a phase I trial: of 17 evaluable NSCLC patients with amplifications of FGFR1, the ORR was 45% at the data cutoff [42]. The final results from these and other studies of FGFR1 inhibitors are eagerly anticipated. In addition to FGFR1 gene amplification, oncogenic fusions involving both FGFR1 and FGFR3 have been identified in lung cancer and also may predict for susceptibility to FGFR targeted therapies [43, 44]. Table 1 lists active studies of FGFR1 inhibitors.

HER2 Mutations

The human epidermal growth factor receptor 2 (HER2, ERBB2) first gained recognition for its importance in breast cancer as a prognostic and predictive marker [45, 46]. The HER2 gene encodes a transmembrane protein that is a receptor tyrosine kinase in the ERBB family that, when activated, promote cellular proliferation, differentiation, migration, and apoptosis. Despite the demonstrated utility of HER2 protein expression-based treatment in breast and gastric cancers, HER2 targeting in lung cancer has not been proven effective when treatment is based on predictive biomarkers that measure protein expression [46–48]. Importantly, activating mutations within the tyrosine kinase domain of the HER2 gene were discovered in a small subset of NSCLC patients. Arcila et al. evaluated 1,478 lung adenocarcinoma samples using mass spectrometry analysis (Sequenome) and found that of the group of patients negative for EGFR and KRAS mutations, 25 of 507 (4.9%) patients demonstrated HER2 mutations [49]. The overall prevalence in the cohort of 1,478 was 1.8% and in-frame insertions in exon 20 accounted for the majority (96%) of these mutations [49]. Clinically, patients with HER2 mutations were slightly younger and never smokers, but there were no differences between genders or in the prevalence between Asian and Caucasian patients. They also evaluated 104 squamous cell and six small cell carcinoma samples but found no HER2 mutations. Buttitta et al. studied a European cohort of 407 lung adenocarcinomas for HER2 mutations by PCR and direct sequencing methods and identified 9 (2.2%) mutations. They found a predominance of HER2 mutations in patients whose adenocarcinoma had bronchoalveolar features; however, Arcila et al. did not observe this association. They did find a trend toward younger patients who were never smokers and a slightly higher prevalence in women, although none of these findings were significant [50]. Li et al. utilized direct sequencing in a cohort of 224 Asian lung adenocarcinoma patients and found 8 (3.6%) with HER2 mutations. In this cohort, HER2 mutations were associated with women and never smokers [51]. Overall, the group of patients harboring HER2 mutations comprised <4% of patients with lung adenocarcinoma in these and other studies [33••, 49–51].

Evaluation of HER2-targeted agents in patients both within and outside of clinical trial has yielded significant findings. Interestingly, the positive studies have focused on HER2 mutation analysis as a predictive biomarker. Afatinib, a potent ERBB family blocker, demonstrated an objective response in all three patients with HER2 mutations [52]. In a retrospective analysis of 65 patients with stage IV NSCLC with HER2 mutations, there was an overall response rate of 50% and disease control rate (DCR) of 82% using a variety of HER2-targeted therapies. Trastuzumab combined with chemotherapy or afatinib monotherapy were particularly effective (DCR of 96% and 100%, respectively), but progressive disease was observed in patients treated with lapatinib or masatinib [53]. Currently, there are multiple trials ongoing to identify effective agents that target HER2 mutations in NSCLC (Table 1).

BRAF Mutations and Fusions

Somatic BRAF (B1 homolog of the v-raf murine sarcoma viral oncogene) mutations are clinical important components in many types of cancers including thyroid, melanoma, and colorectal [54]. BRAF belongs to the RAF kinase family and is in the group of serine-threonine kinases that is important in the RAS/RAF/MAPK signaling pathway. Activating point mutations in BRAF result in constitutive activation of this pathway leading to cell growth and proliferation [52]. In NSCLC, the BRAF mutation rate is approximately 1–3% [33••, 55]. Chen et al. performed a meta-analysis of 10 studies and found BRAF mutations in 98 of 2,224 adenocarcinoma samples (4.04%) and 6 of 1,037 nonadenocarcinomas (0.58%) [56]. Across the studies, a number of different detection techniques were utilized to identify mutations. Within this study, the authors focused on the most common BRAF mutation: V600E (53.6%). They found that patients were more likely to be never smokers and that patients were more likely to have adenocarcinoma histology.

Early data on treatment of BRAF positive patients with BRAF inhibitors has demonstrated strong potential as a therapeutic strategy. A patient with stage IV NSCLC harboring a BRAF V600E demonstrated a complete response by PET to vemurafenib [57]. Early results from a phase II study of dabrafenib demonstrated an objective response rate of 54% among 17 NSCLC patients with a BRAF V600E mutation; progression-free survival and overall survival have not yet been reported [58]. Based on successful combination of BRAF and MEK inhibitors in melanoma patients with BRAF V600E, this study has a second arm that is currently still in process to determine the safety and efficacy of dabrafenib in combination with trametinib, a MEK inhibitor [58, 59].

A group investigating mucinous adenocarcinomas recently identified another genetic alteration involving BRAF. They identified a constitutively active BRAF fusion protein (TRIM24-BRAF), similar to the BRAF fusions seen in prostate cancer, gastric cancer and melanoma [60, 61]. They demonstrated not only that this aberrant protein activates the BRAF kinase but also that it is responsive to sorafenib, a commercially available RAF kinase inhibitor [60]. Clearly, these studies provide compelling results and additional clinical trials are ongoing to determine the clinical benefit of BRAF inhibitors in lung cancer patients (Table 1).

NRG1 Gene Fusions

The NRG1 gene encodes the neuregulin protein, a ligand for ERBB3 (HER3) and ERBB4 (HER4) RTKs. Fernandez-Cuesta et al. utilized gene copy number and transcriptome sequencing analysis to evaluate 25 lung adenocarcinomas from never smokers without EGFR or KRAS mutations. This analysis yielded a novel gene fusion containing the 5’ end of the CD74 gene in frame with the 3’ end of the NRG1 gene, encoding the EGF-like domain [62]. Follow-up analysis using RT-PCR and FISH in a validation cohort of 102 never-smoker patients who were negative for other identified an additional four cases (3.9%) with CD74-NRG1 fusions. In sum, all five cases were women with invasive mucinous adenocarcinoma. Another study focusing on invasive mucinous adenocarcinoma used whole transcriptome sequencing to identify 6 of 90 cases with NRG1 fusions (one SLC3A2- and five CD74-NRG1) [60]. As outlined above, this group also identified BRAF fusions and another potentially actionable fusion, ERBB4 (HER4). NRG1 gene rearrangements were previously identified in breast cancer with a reported prevalence of 4.7–6% of breast cancers and in head and neck cancers and may be a predictive biomarker for lapatinib sensitivity [63–65].

Unlike ALK, ROS1, RET, and NTRK1 gene fusions where the kinase domain of the respective genes is activated by the fusion event, NRG1 does not encode a kinase but the ligand for an RTK. Expression of CD74-NRG1 in cell leads to activation of HER2 and HER3 in vitro; similarly, all tumor samples harboring these fusions showed high expression and phosphorylation of HER2 and HER3. Thus, this gene fusion leads to high expression of the ligand, inducing an autocrine signaling loop in cancer cells. Inhibition of a cell line model using lapatinib and afatinib inhibited HER2 and HER3 signaling suggesting a possible clinical strategy for treatment of patients harboring these gene fusions [60]. Other potential strategies to target these novel fusions might include the use of HER3 monoclonal antibodies.

Conclusions

The landscape of personalized medicine is constantly evolving. Initially, we found that targeting molecular alterations in signal transduction pathways can generate significant improvements in patient outcomes. With the dramatic success of EGFR TKIs in patients with activating EGFR mutations researchers have demonstrated the value of targeting specific molecular pathways [66]. Treatment of ALK+ patients with crizotinib has lead to similar clinical improvements [67]. As a result of this success, clinical testing of EGFR and ALK genes are considered part of the standard of care in NSCLC.

Additionally, although age and smoking status often are associated with oncogenic drivers, as noted in many of the examples in this review, this may be more of a reflection of the biases in the sample population than a reflection of the population as a whole. In a study of sequential NSCLC surgical specimens, 40% of those tumors with EGFR mutations were in former or current smokers and there was no significant difference in ages between EGFR positive or negative patients [68]. Wakelee et al. found similar results when studying ALK-positive NSCLC patients. In their survey of 273 ALK+ patients, they found that 33% of patients had a >10 pack year history, 33% were former smokers, and 33% were never smokers. They also found that 52% of ALK-positive patients were aged 65 years or older at diagnosis, indicating that the stereotype of a never-smoker young patient may be imperfect and therefore should not be used to determine which patients will be screened [69].

Currently, both public and private ventures are working to identify additional targetable genetic alterations that may yield similar clinic benefit for patients. The Lung Cancer Mutation Consortium has sought to make mutation detection more efficient by using multiplex analysis to concurrently evaluate samples for genetic alterations. They have recently reported that in a screen for 10 different oncogenic drivers, 64% of lung adenocarcinomas samples screened contained an actionable oncogenic driver [33••]. The Cancer Genome Atlas Research Network recently published their results on the molecular profiles of 230 resected lung adenocarcinomas. They used a pathway-based approach and were able to demonstrate that a striking 76% of lung adenocarcinomas had oncogenic activities that are drivers of the RTK/RAS/RAF activation pathway [15••]. Thus, although individual mutations might be rare, they add up to a significant number of patients who may benefit from targeted therapy. These studies compliment new clinical trial designs that are oncogene- or pathway-based, instead of cancer-type specific. Hopefully, by using multiplexed genetic analysis to pair patients with targeted therapeutic agents we will be able to maximize the evaluation of new drugs. This new approach supersedes traditional clinical trial design, which is difficult when considering mutations that occur less frequently due to the difficulties of recruiting a sufficient number of patients with a specific cancer type and specific genetic alteration. The unknown variable is whether targeting these mutations will always yield significant clinical benefit, because as we learned from the experience of BRAF mutations in melanoma and colon cancer, not all genetic alterations that are identifiable across cancer types respond similarly [70, 71].

Although the new landscape of personalized medicine challenges us to provide selective treatment to patients in smaller subgroups, we can meet this challenge by utilizing multiplexing analysis and new clinical trial designs that take advantage of common oncogenic pathways across tumor types.