You are here

Characteristics and Outcomes of Patients with Lung Cancer Harboring Multiple Molecular Alterations: Results from the IFCT Study Biomarkers France

Journal of Thoracic Oncology, Volume 12, Issue 6, June 2017, Pages 963 - 973



Little is known about the prevalence, prognosis, and response to treatment of advanced NSCLC harboring multiple genomic alterations.


The French Biomarkers France database, which includes 17,664 patients, was used. The prevalence of multiple alterations, their associations, their impact on prognosis (overall survival [OS]), and their response to targeted or conventional treatments (progression-free survival [PFS] and objective response rate) were assessed and compared with those of patients harboring single or no mutation.


We identified 162 patients (0.9%) with double alterations and three with triple mutations. Multiple molecular alterations preferentially involved KRAS (67.3%), phosphatidylinositol-4,5-bisphosphate 3-kinase catalytic subunit alpha gene (PIK3CA) (53.3%), and EGFR (42.4%). Patients with multiple alterations were more likely to be male (56.4%), be never-smokers (25.8 versus 34.7%, p < 0.001), and exhibit adenocarcinomas (83.6%). OS did not differ between single and multiple alterations. Patients with EGFR/KRAS and EGFR/PIK3CA mutations experienced worse PFS than did patients with only EGFR mutations (7.1 and 7.1 versus 14.9 months, p = 0.02 and 0.002, respectively). Concomitant mutations in patients harboring anaplastic lymphoma receptor tyrosine kinase gene (ALK) rearrangement bore little impact on OS (17.7 versus 20.3 months, p = 0.57) or PFS (10.3 versus 12.1 months, p = 0.93). Patients harboring KRAS mutations plus another alteration had an OS time (13.4 versus 11.2 months, p = 0.28), PFS time (6.4 versus 7.2 months, p = 0.78), and objective response rate under first-line chemotherapy (41.7% versus 37.2%) similar to those of patients harboring KRAS mutations only.


With almost 1% of patients harboring multiple alterations, the dogma of mutually exclusive mutations should be reconsidered. Although double mutations do not decrease OS, they do alter PFS under first-line treatment for patients with EGFR mutations. Among limited numbers of patients, therapies targeting the dominant oncogene seem to usually remain active.

Keywords: NSCLC, Single mutation, Multiple mutations, Biomarkers France, KRAS, EGFR.


In the last decade, scientists have characterized key molecular alterations that drive lung carcinogenesis, with many demonstrated to play a key role in lung cancer oncogenesis. Patients treated with therapies targeting these oncogenic drivers exhibit better prognoses than those without personalized therapies.

Activating mutations in EGFR have led to the identification of a distinct population of patients with NSCLC. Cancers that harbor EGFR mutations have been demonstrated to possess profound sensitivity to EGFR tyrosine kinase inhibitors (TKIs), giving them a unique biology and natural history.1 Echinoderm microtubule associated protein like 4 gene (EML4)–anaplastic lymphoma receptor tyrosine kinase gene (ALK) rearrangement is present in approximately 5% of adenocarcinomas and confers sensitivity to crizotinib and second- and third-generation drugs.2 Less frequently, mutations or rearrangements occur in other oncogenes, such as erb-b2 receptor tyrosine kinase 2 gene (HER2), BRAF, phosphatidylinositol-4,5-bisphosphate 3-kinase catalytic subunit alpha gene (PIK3CA), ROS1, MNNG HOS Transforming gene (MET), or ret proto-oncogene gene (RET). Therapeutics for these mutations are currently being evaluated in clinical trials.

The evolution of tumors bearing a molecular alteration is usually dependent on a single mechanism following the principle of oncogenic addiction, which has been described as the dependence of tumor cells on the specific activity of an activated oncogene.3 A single mutation or translocation is supposed to confer a survival advantage to the respective cells. Thus, it is commonly accepted that these molecular alterations are mutually exclusive.4 Nevertheless, case reports and limited series suggest that concomitant molecular alterations can occur in lung cancer and are, perhaps, underestimated. Their impact on response to antitumor agents has scarcely been reported.

In the era of personalized medicine and rapid development of next-generation sequencing, which provide access to the whole genome while increasing detection sensitivity, the probability of facing multiple molecular alterations is likely to increase in the near future.5 This study, which is based on a nationwide screening program conducted over 1 year,1 aimed to analyze the largest molecular database so far for concomitant mutations to better characterize the epidemiological characteristics of these patients and their responses to conventional and targeted therapies.

Materials and Methods


All consecutive patients with NSCLC who were routinely screened for molecular alterations over a 1-year period at one of the 28 certified molecular genetic centers in France were eligible for inclusion in this study. The data were recorded and monitored by the French Cooperative Thoracic Intergroup (IFCT), with the investigators having full access to the deidentified data and analyses for the current report. The database of this program collected the results of 18,679 molecular analyses conducted from April 2012 to April 2013. The prevalence of co-mutations, associated epidemiological characteristics, and impact on prognoses and treatment outcomes were compared with those of patients harboring a single genomic aberration (50.0%). Details are available in the seminal article.1

Molecular Analysis

The molecular analyses of EGFR, HER2, KRAS, BRAF, and phosphatidylinositol-4,5-bisphosphate 3-kinase catalytic subunit alpha gene (PIK3CA) mutations and ALK rearrangements were done on a routine basis at the molecular genetics centers. Briefly, each molecular genetics center used either the Sanger sequencing method or a more sensitive validated allele-specific technique (generally to be confirmed by Sanger sequencing) to assess EGFR (exons 18–21), HER2 (exon 20), BRAF (exon 15), KRAS (exon 2), and PIK3CA (exons 9 and 20) mutations. A certified break-apart fluorescence in situ hybridization assay was used to assess ALK rearrangements. The sequencing techniques used are detailed in Supplementary Table 1 of the germinal article.1 Additionally, each regional genetics center either did a concurrent analysis of all recommended molecular alterations in the six genes or used a sequential approach in which the EGFR and ALK assessments were done first and each of the other molecular alterations were then assessed until a mutation was found. Multiple alterations reported in this manuscript were found on the same tissue block.

Data Collection

The molecular genetic centers provided the IFCT with the results of molecular assessments of the six genes under investigation and histological typing measured by the referring pathologist. The following data were obtained: sex, ethnic origin, smoking history, family or personal cancer history, performance status, TNM stage, pathological diagnosis, and method of sample collection. Information on treatment types after biomarker analysis and outcomes (best response; first-line treatment and, when applicable, second-line treatment and date[s] of disease progression; and survival status) was reported.

Patients were treated on a routine basis. At the time the study was conducted, erlotinib and gefitinib were approved for first-line treatment of patients with EGFR mutations, whereas crizotinib was available only for second-line treatment of patients with ALK rearrangements. KRAS, BRAF, HER2, and PIK3CA mutations were targetable by drugs available either through clinical trials or off label. Other emerging targets such as ROS1, MET, and RET were not analyzed at the time of data collection.

Ethical Considerations

The study was approved by a national ethics committee for observational studies (National ethics committee for observational studies), by the French Advisory Committee on Information Processing in Material Research in the Health field, and by the National Commission of Informatics and Liberty according to French laws. All patients with NSCLC included in this study received information from their institution or referring clinician in line with French laws. Patients were not required to provide written informed consent to be included in the study.


The primary study objective was to determine the frequency of multiple molecular alterations among six genes that were routinely screened through a nationwide approach in consecutive patients with NSCLC. The secondary objectives were to describe the epidemiological particulars of this patient population and to compare their prognoses and responses to targeted or conventional therapies (overall survival [OS], progression-free survival [PFS], and response rate [RR]) with those observed in patients with single mutations.

Descriptive statistics, including median and range or quartiles for continuous variables and frequencies and percentages for categorical variables, were used. Median follow-up duration was defined as the time from the molecular analysis assessment date to the closing date of the analysis (December 31, 2015). First-line PFS was defined as the time from the molecular analysis assessment date to the date of first progression or death from any cause. Second-line PFS was defined as the time from initiation of second-line treatment to the date of second progression or death from any cause. OS was defined as the time from the molecular analysis assessment date to the date of death or final follow-up. Survival curves were estimated by the Kaplan-Meier method for the total population of patients with mutations and for groups of interest. We compared the groups of interest by using a two-sided log-rank test. Patient characteristics (with single or without double alterations and of each biomarker or biomarker association) were compared by using the chi-square test for qualitative variables or a nonparametric test for quantitative variables. Univariate Cox models were applied to select the most promising prognostic variables (threshold p = 0.20). A multivariate Cox model was then used to adjust for potential confounders (clinical or molecular characteristics associated with PFS or OS). Adjusted hazard ratios with 95% confidence intervals were calculated. All statistical tests were two sided, and a p value less than 0.05 was deemed statistically significant. All analyses were performed with SAS software, version 9.3 (SAS Institute Inc., Cary, NC).


Prevalence of Multiple Genetic Alterations in Nonsquamous NSCLC

Out of 17,826 patients, we identified 165 (0.93%) with multiple genetic alterations involving oncogenic drivers, including 162 double mutations (0.91%) and three triple mutations (0.02%). Double or triple mutations involving EGFR (including sensitizing and resistant mutations) represented 42.4% of all multiple alterations (n = 70). The most common co-mutations occurring de novo with EGFR were PIK3CA (n = 28), KRAS (n = 24), ALK (n = 10), and BRAF (n = 5) mutations. Concomitant mutations associated with KRAS represented 67.3% of all multiple alterations (n = 111), with the most frequent partners being PIK3CA (n = 53), EGFR (n = 24), ALK (n = 22), and BRAF (n = 9) mutations. The prevalence of each association has been detailed in Table 1.

Table 1

Prevalence of Different Double Molecular Alterations


EGFR 67 (40.6%) 24 (15%) 10 (6%) 5 (3%) 28 (17%) 0
KRAS 108 (65%) 24 (15%) 22 (13%) 9 (5%) 53 (32%) 0
ALK 37 (22.4%) 10 (6%) 22 (13%) 2 (1.2%) 2 (1.2%) 1 (0.6%)
BRAF 22 (13.3%) 5 (3%) 9 (5%) 2 (1.2%) 5 (3%) 1 (0.6%)
PIK3CA 88 (53%) 28 (17%) 53 (32%) 2 (1.2%) 5 (3%) 0
HER2 2 (1.2%) 0 0 1 (0.6%) 1 (0.6%) 0

ALK, anaplastic lymphoma receptor tyrosine kinase gene; PIK3CA, phosphatidylinositol-4,5-bisphosphate 3-kinase catalytic subunit alpha gene; HER2, erb-b2 receptor tyrosine kinase 2 gene.

Characteristics of Patients Harboring Multiple Mutations

Multiple molecular alterations were observed more frequently in males (56.4% [p < 0.001]) with a median age of 67 years, in never-smokers or former smokers (34.7% versus 39.8%), and almost exclusively in patients with adenocarcinoma (83.6%). The epidemiology of patients harboring multiple molecular alterations is very similar to that of those harboring a single mutation, yet with a slightly higher percentage of never-smokers. These epidemiological data have been detailed in Table 2.

Table 2

Characteristics of Patients Harboring Multiple Molecular Alterations Compared With Those of Patients with Single or No Mutation


Characteristic Double Mutation (n = 165) Single Mutation (n = 7284) No Mutation (n = 2788) p Value
Sex, n (%) <0.001a
 Male 93 (56.4) 4119 (56.9) 1999 (72.3)
 Female 72 (43.6) 3125 (43.1) 765 (27.7)
 Missing 0 40 24
Age, y 0.06b
 Mean ± SD 66.77 ± 12.04 64.72 ± 11.10 64.76 ± 11.20
 Median 67.00 64.12 64.44
 Range 33.1–91.4 18.1–95.8 23.5–94.4
Smoking, n (%) <0.001a
 Smoker 30 (25.4) 1381 (34.7) 537 (41.1)
 Former smoker 47 (39.8) 1572 (39.5) 593 (45.4)
 Nonsmoker 41 (34.7) 1025 (25.8) 176 (13.5)
 Missing 47 3306 1482
PS, n (%) <0.001c
 0 30 (28.0) 1138 (31.5) 305 (25.3)
 1 48 (44.9) 1501 (41.5) 510 (42.4)
 2 21 (19.6) 643 (17.8) 243 (20.2)
 3 6 (5.6) 255 (7.1) 111 (9.2)
 4 2 (1.9) 79 (2.2) 35 (2.9)
 Missing 58 3668 1584
TNM stage, n (%) <0.001a
 Recurrence 17 (14.8) 403 (10.0) 115 (8.8)
 Stage I/II 15 (13.0) 645 (16.1) 215 (16.4)
 Stage III 8 (7.0) 491 (12.2) 219 (16.7)
 Stage IV 75 (65.2) 2476 (61.7) 764 (58.2)
 Missing 50 3269 1475
Histological type, n (%) <0.004a
 Squamous 2 (1.2) 122 (1.7) 225 (8.1)
 Adenocarcinoma 138 (83.6) 6068 (83.3) 2063 (74.0)
 Large cell 5 (3.0) 177 (2.4) 137 (4.9)
 Other 20 (12.1) 917 (12.6) 363 (13.0)

a Chi-square test.

b Variance analysis.

c Fisher's exact test.

PS, performance status.

Overall Outcome in Patients Harboring Multiple Mutations

No difference in OS was observed between patients harboring multiple versus single oncogene alterations (14.6 versus 16.3 months [p = 0.7]). After biomarker analysis, PFS was also shown to not be significantly influenced by a second molecular alteration (7.2 versus 9.8 months [p = 0.19]).

Objective RR (ORR) to first-line therapy (including targeted therapies and chemotherapies) did not differ between patients with double genetic alterations and those with a single mutation (44.8 versus 42.4%, respectively).

Prognostic Impact of Multiple Mutations

Patients harboring double mutations involving EGFR and another oncogene tended to experience worse prognosis than those with isolated EGFR activating mutation (OS = 17.7 versus 24.3 months [p = 0.23]). This negative impact on prognosis, although not statistically significant, seemed more marked when the co-alteration involved KRAS (OS = 13.3 versus 24.3 months [p = 0.17]) or PIK3CA (OS = 12.1 versus 24.3 months [p = 0.36]) (Fig. 1). A concomitant mutation with EGFR was significantly associated with poorer PFS under first-line treatment (7.5 versus 14.9 months [p < 0.0001]). This observation proved true when the co-alteration involved KRAS (7.1 versus 14.9 months [p = 0.03]) or PIK3CA (7.1 versus 14.9 months [p = 0.002]) but not when it involved ALK (10.3 versus 14.9 months [p = 0.23]) (see Fig. 2).


Figure 1

Overall survival (OS) in the presence of double mutation involving EGFR compared to EGFR mutated patients. (A) OS in patients harboring double mutation involving EGFR (EGFR+X) compared to EGFR mutated patients. (B) OS in the presence of EGFR and PI3K co-mutation compared to EGFR mutated patients. (C) OS in the presence of EGFR and ALK co-mutation compared to EGFR mutated patients. (D) OS in the presence of EGFR and KRAS co-mutation compared to EGFR mutated patients. IC95%, concentration that inhibits 95%; HR, hazard ratio; NR, not reached.



Figure 2

Progression-free survival (PFS) in the presence of double mutation involving EGFR compared to EGFR mutated patients. (A) PFS in the whole population of patients harboring double mutation involving EGFR (EGFR+X) compared to EGFR mutated patients. (B) PFS in the presence of EGFR and PI3K co-mutation compared to EGFR mutated patients. (C) PFS in the presence of EGFR and ALK co-mutation compared to EGFR mutated patients. (D) PFS in the presence of EGFR and KRAS co-mutation compared to EGFR mutated patients.


OS (17.7 versus 20.3 months [p = 0.57]) or PFS (10.3 versus 12.1 months [p = 0.93]) of patients harboring double genomic alterations involving ALK did not significantly differ from that of patients with ALK rearrangement patients, regardless of the associated mutation.

Patients harboring double mutations involving KRAS exhibited the same OS as those with tumors with KRAS mutation only (11.2 versus 13.4 months for the multiple alterations, [p = 0.28]). This observation proved valid regardless of the associated co-alteration, with median OS times of 13.6, 13.4, and 13.8 months for patients with KRAS/EGFR (p = 0.45), KRAS/ALK (p = 0.69), and KRAS/PIK3CA (p = 0.68) mutations, respectively. PFS was also similar between patients with double mutations including KRAS and those with single KRAS mutations (6.4 versus 7.2 months, respectively [p = 0.78]).

Predictive Impact of Multiple Mutations on Response to Targeted Therapy in Patients Harboring Concomitant Mutations

Response to first-line therapy with a EGFR TKI tended to be negatively influenced by a concomitant PIK3CA mutation (ORR = 60.9% for patients with EGFR mutation versus 45.5% for patients with a PIK3CA co-mutation [p = 0.30]). Among 24 patients with EGFR/KRAS co-mutation, data concerning first-line therapy were available for only seven patients. Only two received EGFR TKI as first-line therapy, resulting in one partial response and one case of disease progression. Three patients with EGFR/KRAS co-mutation received EGFR TKIs as second-line therapy, resulting in one partial response and two cases of progressive disease, whereas 172 patients with EGFR mutation who received this treatment in a second-line setting showed a 47.7% ORR and 23.8% rate of disease progression (Fig 3).


Figure 3

Response to targeted therapies in the presence of a double mutation compared to a single mutation. (A) Response to first-line treatment by EGFR-TKIs in patients harboring double mutation involving EGFR (EGFR+X) compared to EGFR mutated patients. (B) Response to second-line treatment by EGFR-TKIs in patients harboring double mutation involving EGFR (EGFR+X) compared to EGFR mutated patients. (C) Response to EGFR-TKIs in the presence of EGFR and PI3K co-mutation compared to EGFR mutated patients. (D) Response to crizotinib in patients harboring double molecular alterations involving ALK rearrangement (ALK+X) compared to ALK patients.


Patients treated with crizotinib in a first-line setting for NSCLC harboring ALK rearrangement concomitantly with other molecular alterations tended to experience a lower RR, although the number of patients was too small to achieve significant results (Fig. 3D). The only patient harboring ALK and EGFR mutations, along with two patients with KRAS/ALK mutations, exhibited stable disease as the best response to crizotinib, whereas patients with ALK rearrangement showed a 69.5% ORR to crizotinib. One patient with BRAF/ALK-rearranged adenocarcinoma showed a partial response.

No sufficient data were available concerning therapy targeting KRAS, BRAF, HER2. or PIK3CA in co-mutated patients.

Predictive Impact of Multiple Mutations on Response to Chemotherapy

No significant difference in RR to chemotherapy was observed between the populations of patients harboring single or multiple molecular aberrations. Response to first-line chemotherapy was not influenced by a concomitant mutation or rearrangement occurring with KRAS (ORR = 41.7% versus 37.2% and disease control rate = 58.3% versus 59.9% [p = 0.85] for multiple versus single mutation, respectively). No difference in terms of response to first-line chemotherapy has been identified for ALK-rearranged tumors, regardless of the partner (EGFR, KRAS, or BRAF). In a second-line setting, chemotherapy seemed more efficient when a second mutation was associated with KRAS, although the difference was not statistically significant (33.3% versus 15.5% ORR, respectively) (Fig 4).


Figure 4

Response to chemotherapy in the presence of a double mutation including KRAS compared to KRAS mutated patients. (A) Response to chemotherapy in patients harboring double mutation involving KRAS (KRAS+X) compared to KRAS mutated patients. (B) Response to chemotherapy in the presence of KRAS and EGFR co-mutation compared to KRAS mutated patients. (C) Response to chemotherapy in the presence of KRAS mutation and ALK rearrangement compared to KRAS mutated patients. (D) Response to chemotherapy in the presence of KRAS and PI3K co-mutation compared to KRAS mutated patients.


The impact of the most frequent combination of genomic alterations on the prognosis and response to conventional and targeted therapy has been summarized in Table 3.

Table 3

Impact of the Different Combinations of Genomic Alterations on Prognosis and Response to Conventional and Targeted therapy


Co-alteration Prevalence (n, % of Total Double Mutations) Prognostic Impact Response to Targeted Therapy Response to Chemotherapy
EGFR/KRAS n = 24 (15%) Detrimental impact of KRAS (PFS, trend for OS) Possible negative impact of KRAS on response to EGFR TKI None
EGFR/PIK3CA n = 28 (17%) Detrimental impact of PIK3CA (PFS, trend for OS) Possible negative impact of PIK3CA on response to EGFR TKI (ORR 61% vs. 45%) None
EGFR/ALK n =10 (6%) None None None
KRAS/PIK3CA n = 53 (32%) None NA None
KRAS/ALK n = 22 (13%) Detrimental impact of KRAS (trend for OS) Possible negative impact of KRAS on response to crizotinib None

PFS, progression-free survival; OS, overall survival; TKI, tyrosine kinase inhibitor; PIK3CA, phosphatidylinositol-4,5-bisphosphate 3-kinase catalytic subunit alpha gene; ORR, objective response rate; ALK, anaplastic lymphoma receptor tyrosine kinase gene; NA, not available.


Here we have presented the largest cohort of patients harboring multiple genomic alterations driving NSCLC (n = 165), with a prevalence of 0.93%. Little is known about this rare population of patients who constitute a real challenge for clinicians, especially from a therapeutic point of view. Deep sequencing will likely increase the incidence of multiple genetic alterations with undefined theranostic impact. Therefore, it seems crucial to decipher the effect of different combinations of basic oncogenes before facing complex molecular signatures.

A previous study reported 153 double mutations and seven triple mutations in a cohort of 5125 patients from China.6 In this study, the prevalence of double or more mutations was higher (3%) than in ours, yet the epidemiological characteristics of the two study populations were clearly different. Whereas the overall rate of mutations was similar (48.4% versus 49.5% in the IFCT Biomarkers France study1), the mutation profile was very different between the two study populations, with a predominance of EGFR mutations (36.2% versus 11%) and low rate of KRAS mutations (8.4% versus 29%) compared with in our white population. In another Chinese study involving stage IB resected lung adenocarcinomas, double mutations were found in 8.3% of cases (13 of 156). EGFR mutations were present in 44.2% of tumors and involved in 10 of 13 co-mutated tumors.7 Nevertheless, unlike our study, neither of these studies included ALK rearrangement and allowed analysis of either patient prognosis or response to targeted and conventional therapies. The Lung Cancer Mutation Consortium reported data on 1007 lung adenocarcinomas. Overall, 27 of 1007 patients (2.7%) harbored two oncogenic driver mutations. However, in this study, the spectrum of genes tested was larger (EGFR, KRAS, erb-b2 receptor tyrosine kinase 2 gene [ERBB2], AKT/serine threonine kinase 1 gene [AKT1], BRAF, mitogen-activated protein kinase kinase gene [MEK1], NRAS, and PIK3CA). The molecular epidemiology of this population was also slightly different, with (in particular) a higher prevalence of EGFR mutations (22%) and ALK rearrangements (8.5%). Furthermore, the double mutations observed in this study were different, including 14 with two small mutations and 13 with either a small mutation and ALK rearrangement (four); a small mutation and MET amplification (seven); concurrent ALK rearrangement and MET amplification (two) and one case had EGFR ex19del and AKT1. Of the 14 cases with two small mutations, 13 (92%) had a PIK3CA mutation in addition to another mutation, including nine with EGFR, two with BRAF, one with KRAS, and one with MEK1 mutation. In this study, however, neither prognosis nor outcome of these patients was reported.8 and 9

In patients with EGFR mutation, we observed a detrimental effect of a concomitant KRAS or PIK3CA mutation on prognosis. PFS was logically lower when one of these two oncogenes, which are implicated in downstream signaling pathways, was mutated. With a prevalence of 10% to 15% for EGFR and 30% for KRAS reported in white patients, the probability of observing a concomitant mutation of these two oncogenes can be assumed to be about 3%.10 However, this association appears much more rarely (n = three of 73111 and n = 29 of 51256), with descriptions in the literature limited to isolated clinical cases or short series,6, 11, 12, 13, 14, and 15 so that these mutations have usually been considered to be mutually exclusive.4 In our study, we have reported 27 cases of EGFR/KRAS–co-mutant tumors. PFS was twice as low for patients with EGFR/KRAS mutation as for patients with EGFR mutation. KRAS mutations were shown to cause the loss of the guanosine triphosphatase activity, along with its subsequent feedback regulation. KRAS then continuously activates BRAF and PIK3CA, independently of EGFR. This can explain why KRAS mutations are usually associated with a poor RR to an EGFR TKI.16, 17, and 18 However, responses in the short term have been reported in small series.12 and 15 In our study, first-line treatment was available for only seven patients, with only one treated with an EGFR TKI, resulting in a partial response. KRAS and EGFR mutations were likely carried by two different clones, with initial response but shorter PFS accounted for by the emergence of a KRAS-resistant clone, as observed in the series of Benesova et al., in which patients had partial responses but short PFS (3, 5, and 7 months12). Taken together, all these data explain, as reported in Li’s large Chinese cohort in which only 13.1% of patients received EGFR TKIs in the first-line setting, why clinicians are reluctant to prescribe EGFR TKIs in the presence of KRAS mutation, regardless of EGFR status. Nevertheless, our study, which is in line with other small series, seems to suggest that EGFR TKIs could be considered for these patients as first-line therapy. On the other hand, when these agents were given in a second-line setting, poor results (one partial and two progressive responses) were found, probably because previous chemotherapy proved effective on EGFR but not on a KRAS clone, this latter mutation being associated with chemoresistance.18 Obviously, the very limited number of patients harboring double mutations involving EGFR does not allow drawing definitive conclusions.

Up to now, PIK3CA has been shown to be mutated in 5% to 10% of squamous cell carcinomas and 2% to 3% of adenocarcinomas, which is associated with disappointing results when PIK3CA or mammalian target of rapamycin inhibitors are used.19PIK3CA mutations also seem to be associated with poor RR to EGFR TKIs.20 Moreover, introducing PIK3CA mutation into EGFR mutant cell lines was reported to abrogate sensitivity to EGFR TKIs.21 PFS in our study also tended to be lower for patients with PIK3CA/EGFR mutation than for patients with EGFR mutation (p = 0.02). However, although responses to EGFR TKI tended to be influenced by a concomitant PIK3CA mutation, the 45.5% RR indicates that these agents should nonetheless be considered as first-line treatment of EGFR/PIK3CA co-mutated tumors. Chaft et al. reported objective and durable responses to EGFR TKI for three EGFR/PIK3CA co-mutated tumors.22 Moreover, although Eng et al. corroborated a detrimental impact of an additional PIK3CA mutation in EGFR-mutant adenocarcinomas on prognosis, sensitivity to EGFR TKIs was not significantly affected and median duration response was the same in both groups.23

Because of a similar epidemiology, the association of EGFR mutations and ALK rearrangement has been studied extensively. These two alterations are often considered mutually exclusive.24 and 25 Among 1683 NSCLC tumors analyzed by Gainor et al. (including 301 EGFR-mutated and 75 ALK-rearranged tumors), no co-mutation was detected.25 These data are concordant with the results of a meta-analysis conducted by Wang et al. in which only 0.75% of patients (three of 399) with EGFR mutations also harbored ALK rearrangement.26 However, numerous clinical cases (n = 43) have been reported and collected in a recent review of case series, leading to the conclusion of a 1.5% prevalence for ALK rearrangement in patients with EGFR mutation.27 In addition, Rosell et al. reported on a population of 95 EGFR-mutated adenocarcinomas (EURTAC trial), with a 15.8% prevalence of concomitant ALK rearrangement, although this result was probably accounted for by the low specificity of the polymerase chain reaction assay.28 Because of the co-expression in IHC of ALK fusion protein and abnormal EGFR (specific antibodies of different mutations), these two molecular abnormalities can co-exist in the same cell.27, 29, and 30 In our study, a concomitant mutation of EGFR, KRAS, or PIK3CA had no detrimental impact on PFS or OS. Response to first-line treatment by chemotherapy seemed very similar, regardless of the second mutation. Nevertheless, although the number of patients was too small, a co-alteration could likely change sensitivity to crizotinib: two of two patients with ALK/KRAS mutation and one patient with ALK/EGFR mutation showed only stable disease as the best response. Obviously, these findings involve too limited a number of patients not to consider crizotinib as a first-line choice in patients with ALK/EGFR or ALK/KRAS mutations. In the particular context of patients with EGFR/ALK mutations, it is very difficult to define the best treatment sequence. The allelic fraction of each alteration in the next-generation sequencing report should help in choosing the clone to target first. To the best of our knowledge, there are no data on combined treatment with ALK and EGFR TKI in this setting.

Interestingly, although the poor prognosis associated with KRAS mutations has been well established,18 these data remained valid in the presence of a concomitant molecular alteration, even if the latter was known to be associated with good prognosis. Moreover, response to chemotherapy was not improved by the concomitant presence of ALK rearrangement or EGFR mutation (Fig. 4). We can thus conclude that the detrimental impact of KRAS mutations on prognosis and response to platinum-based chemotherapy proved independent of other oncogenes.

In conclusion, we have herein reported on the largest cohort of patients with tumors harboring multiple molecular alterations. The tenet of mutual exclusivity of genomic aberrations toward each other is likely to be questioned. Our results suggest a detrimental effect of concomitant PIK3CA or KRAS mutations on both prognosis and sensitivity to EGFR TKI in patients harboring EGFR mutations. Even if the number of patients is low, responses of these tumors to TKIs seem to remain high enough to consider these agents as a first- or second-line regimen. The determination of the respective allelic fraction of each mutation would be of interest to assess the relative importance of each mutation and help in the choice of the first-line therapy.


Dr. Sabourin and Dr. Lemoine received grants from the French National Cancer Institute during the conduct of this study. This research work was supported by grants from the French Cooperative Thoracic Intergroup.

Supplementary Data

Download file

Supplementary Data


  • 1 F. Barlesi, J. Mazieres, J.P. Merlio, et al. Routine molecular profiling of patients with advanced non-small-cell lung cancer: results of a 1-year nationwide programme of the French Cooperative Thoracic Intergroup (IFCT). Lancet. 2016;387:1415-1426 Crossref
  • 2 A.T. Shaw, D.W. Kim, K. Nakagawa, et al. Crizotinib versus chemotherapy in advanced ALK-positive lung cancer. N Engl J Med. 2013;368:2385-2394 Crossref
  • 3 I.B. Weinstein. Cancer. Addiction to oncogenes—the Achilles heal of cancer. Science. 2002;297:63-64 Crossref
  • 4 H. Shigematsu, L. Lin, T. Takahashi, et al. Clinical and biological features associated with epidermal growth factor receptor gene mutations in lung cancers. J Natl Cancer Inst. 2005;97:339-346 Crossref
  • 5 S. Chevrier, L. Arnould, F. Ghiringhelli, B. Coudert, P. Fumoleau, R. Boidot. Next-generation sequencing analysis of lung and colon carcinomas reveals a variety of genetic alterations. Int J Oncol. 2014;45:1167-1174
  • 6 S. Li, L. Li, Y. Zhu, et al. Coexistence of EGFR with KRAS, or BRAF, or PIK3CA somatic mutations in lung cancer: a comprehensive mutation profiling from 5125 Chinese cohorts. Br J Cancer. 2014;110:2812-2820 Crossref
  • 7 Y.S. Wen, L. Cai, X. Zhang, et al. Concurrent oncogene mutation profile in Chinese patients with stage Ib lung adenocarcinoma. Medicine (Baltimore). 2014;93:e296 Crossref
  • 8 M.G. Kris, B.E. Johnson, L.D. Berry, et al. Using multiplexed assays of oncogenic drivers in lung cancers to select targeted drugs. JAMA. 2014;311:1998-2006 Crossref
  • 9 L.M. Sholl, D.L. Aisner, M. Varella-Garcia, et al. Multi-institutional oncogenic driver mutation analysis in lung adenocarcinoma: the Lung Cancer Mutation Consortium experience. J Thorac Oncol. 2015;10:768-777 Crossref
  • 10 C. Greenman, R. Wooster, P.A. Futreal, M.R. Stratton, D.F. Easton. Statistical analysis of pathogenicity of somatic mutations in cancer. Genetics. 2006;173:2187-2198 Crossref
  • 11 C.Q. Zhu, G. da Cunha Santos, K. Ding, et al. Role of KRAS and EGFR as biomarkers of response to erlotinib in National Cancer Institute of Canada Clinical Trials Group Study BR.21. J Clin Oncol. 2008;26:4268-4275 Crossref
  • 12 L. Benesova, M. Minarik, D. Jancarikova, B. Belsanova, M. Pesek. Multiplicity of EGFR and KRAS mutations in non-small cell lung cancer (NSCLC) patients treated with tyrosine kinase inhibitors. Anticancer Res. 2010;30:1667-1671
  • 13 S.W. Han, T.Y. Kim, Y.K. Jeon, et al. Optimization of patient selection for gefitinib in non-small cell lung cancer by combined analysis of epidermal growth factor receptor mutation, K-ras mutation, and Akt phosphorylation. Clin Cancer Res. 2006;12:2538-2544 Crossref
  • 14 A.G. Pallis, A. Voutsina, A. Kalikaki, et al. “Classical” but not “other” mutations of EGFR kinase domain are associated with clinical outcome in gefitinib-treated patients with non-small cell lung cancer. Br J Cancer. 2007;97:1560-1566 Crossref
  • 15 A. Choughule, R. Sharma, V. Trivedi, et al. Coexistence of KRAS mutation with mutant but not wild-type EGFR predicts response to tyrosine-kinase inhibitors in human lung cancer. Br J Cancer. 2014;111:2203-2204 Crossref
  • 16 E. Massarelli, M. Varella-Garcia, X. Tang, et al. KRAS mutation is an important predictor of resistance to therapy with epidermal growth factor receptor tyrosine kinase inhibitors in non-small-cell lung cancer. Clin Cancer Res. 2007;13:2890-2896 Crossref
  • 17 H. Linardou, I.J. Dahabreh, D. Kanaloupiti, et al. Assessment of somatic k-RAS mutations as a mechanism associated with resistance to EGFR-targeted agents: a systematic review and meta-analysis of studies in advanced non-small-cell lung cancer and metastatic colorectal cancer. Lancet Oncol. 2008;9:962-972 Crossref
  • 18 N. Guibert, M. Ilie, E. Long, et al. KRAS mutations in lung adenocarcinoma: molecular and epidemiological characteristics, methods for detection, and therapeutic strategy perspectives. Curr Mol Med. 2015;15:418-432 Crossref
  • 19 C. Fumarola, M.A. Bonelli, P.G. Petronini, R.R. Alfieri. Targeting PI3K/AKT/mTOR pathway in non small cell lung cancer. Biochem Pharmacol. 2014;90:197-207 Crossref
  • 20 J.Y. Chen, Y.N. Cheng, L. Han, et al. Predictive value of K-ras and PIK3CA in non-small cell lung cancer patients treated with EGFR-TKIs: a systemic review and meta-analysis. Cancer Biol Med. 2015;12:126-139
  • 21 J.A. Engelman, T. Mukohara, K. Zejnullahu, et al. Allelic dilution obscures detection of a biologically significant resistance mutation in EGFR-amplified lung cancer. J Clin Invest. 2006;116:2695-2706 Crossref
  • 22 J.E. Chaft, M.E. Arcila, P.K. Paik, et al. Coexistence of PIK3CA and other oncogene mutations in lung adenocarcinoma-rationale for comprehensive mutation profiling. Mol Cancer Ther. 2012;11:485-491 Crossref
  • 23 J. Eng, K.M. Woo, C.S. Sima, et al. Impact of concurrent PIK3CA mutations on response to EGFR tyrosine kinase inhibition in EGFR-mutant lung cancers and on prognosis in oncogene-driven lung adenocarcinomas. J Thorac Oncol. 2015;10:1713-1719 Crossref
  • 24 A.T. Shaw, B.Y. Yeap, M. Mino-Kenudson, et al. Clinical features and outcome of patients with non-small-cell lung cancer who harbor EML4-ALK. J Clin Oncol. 2009;27:4247-4253 Crossref
  • 25 J.F. Gainor, A.M. Varghese, S.H.I. Ou, et al. ALK rearrangements are mutually exclusive with mutations in EGFR or KRAS: an analysis of 1,683 patients with non-small cell lung cancer. Clin Cancer Res. 2013;19:4273-4281 Crossref
  • 26 Y. Wang, S. Wang, S. Xu, J. Qu, B. Liu. Clinicopathologic features of patients with non-small cell lung cancer harboring the EML4-ALK fusion gene: a meta-analysis. PLoS One. 2014;9:e110617 Crossref
  • 27 L. Baldi, M.C. Mengoli, A. Bisagni, M.C. Banzi, C. Boni, G. Rossi. Concomitant EGFR mutation and ALK rearrangement in lung adenocarcinoma is more frequent than expected: report of a case and review of the literature with demonstration of genes alteration into the same tumor cells. Lung Cancer. 2014;86:291-295 Crossref
  • 28 Rosell R, Massuti Sureda B, Costa C, et al. LBA31- Concomitant actionable mutations and overall survival (OS) in EGFR-mutant non-small-cell lung cancer (NSCLC) patients (p) included in the EURTAC tri…[abstract]. Accessed June 27, 2016.
  • 29 T. Sasaki, J. Koivunen, A. Ogino, et al. A novel ALK secondary mutation and EGFR signaling cause resistance to ALK kinase inhibitors. Cancer Res. 2011;71:6051-6060 Crossref
  • 30 J.J. Yang, X.C. Zhang, J. Su, et al. Lung cancers with concomitant EGFR mutations and ALK rearrangements: diverse responses to EGFR-TKI and crizotinib in relation to diverse receptors phosphorylation. Clin Cancer Res. 2014;20:1383-1392 Crossref


a Toulouse University Hospital, Paul Sabatier Toulouse University, Toulouse, France

b Aix-Marseille University, Assistance Publique Hôpitaux de Marseille, Marseille, France

c Brest University Hospital, Brest, France

d Pontchaillou Rennes University Hospital, Rennes, France

e Gustave Roussy Cancer Campus, Villejuif, France

f Strasbourg University Hospital, Strasbourg, France

g Rennes University Hospital, Rennes, France

h Tours University Hospital, Tours, France

i Bordeaux University Hospital, Pessac, France

j Aix Marseille University, Marseille, France

k Bordeaux Nord Aquitaine Polyclinic, Bordeaux, France

l Louis Pradel University Hospital, Bron, France

m Louis Pasteur Hospital, Colmar, France

n Bretagne Atlantique Vannes Hospital, Vannes, France

o Rouen University Hospital, Rouen, France

p Paris-Sud Hospital, Villejuif, France

q French Cooperative Thoracic Intergroup, Paris, France

r Grenoble University Hospital, Grenoble, France

Corresponding author. Address for correspondence: Julien Mazières, MD, PhD, Thoracic Oncology Unit, Respiratory Disease Department, Hôpital Larrey, CHU Toulouse, Chemin de Pouvourville, 31059 Toulouse Cedex, France.

Readers of this article may receive CME credit. Further information can be found at

Disclosure: Dr. Barlesi reports grants from AstraZeneca, Eli Lilly Oncology, F. Hoffmann–La Roche Ltd., Novartis, Pfizer, and Pierre Fabre and personal fees from Boehringer Ingelheim, Dai-Ichii Seiko, GlaxoSmithKline, AstraZeneca, Eli Lilly Oncology, F. Hoffmann–La Roche Ltd., Novartis, Pfizer, and Pierre Fabre. Dr. Léna has received grants from Roche; personal fees from AstraZeneca, Boehringer Ingelheim, Bristol-Myers Squibb, Merck, Pierre Fabre Oncology, Pfizer, and Lilly; and nonfinancial support from Pfizer, Lilly, and Roche. Dr. Besse has received grants from Novartis. Dr. Moreau has received nonfinancial support from Pfizer. Dr. Sabourin has received personal fees from Boehringer Ingelheim, AstraZeneca, and Roche and grants from Roche outside the submitted work. Dr. Lemoine has received personal fees from Roche, AstraZeneca, Pfizer, Boehringer Mannheim and nonfinancial support from Boehringer Mannheim and Roche outside the submitted work. In addition, Dr. Lemoine has a patent (13306589.6) issued. Dr. Moro-Sibilot reports grants from Pfizer during conduct of this study and personal fees from Pfizer, Roche, and Novartis outside the submitted work. Dr. Mazieres has received grants from Roche and Bristol-Myers Squibb and personal fees from Roche, Novartis, Pfizer, Bristol-Myers Squibb, Merck Sharp and Dohme, and AstraZeneca and has also served as consultant for Roche, Novartis, Pfizer, Bristol-Myers Squibb, Merck Sharp and Dohme, and AstraZeneca. The remaining authors declare no conflict of interest.

Commentary by Stefan Zimmermann

You believe that oncogenic molecular alterations are mutually exclusive? There are exceptions. This largest-ever cohort of NSCLC patients harboring multiple molecular alterations represents 0.9% of the French Biomarkers France database, accounting for 162 out of 17’664 patients. Interestingly, while OS was not decreased by the presence of double mutations, PFS under first-line therapy was decreased, especially so for EGFR mutated tumors harboring a concomitant KRAS mutation. Concomitant mutations associated with KRAS account for two thirds of all multiple alterations, with a detrimental effect on prognosis and response to platinum-based chemotherapy independent of other oncogenes. Interestingly, about 1.5% of EGFR mutated tumors contained a coexisting ALK rearrangement, with numbers too small to allow any conclusion on the treatment efficacy. Allelic frequencies were not reported, precluding any conclusion in this dataset on the clonal or subclonal nature of these alterations; the recent publication of the TRACERx study do show that driver mutations in EGFR, MET, BRAF are almost always clonal, with heterogeneous driver alterations such as PIK3CA occurring later in evolution. This highlights the challenged posed by tumor heterogeneity, most obvious in the context of targeted therapy.