FormalPara Key Points
Table 1
Full size table

1 Introduction

Lung cancer is the most common malignancy and the leading cause of cancer-related death both in China and worldwide [1,2,3]. Treatment of non-small-cell lung cancer (NSCLC) has changed dramatically in recent years with the advent of targeted therapies for different oncogenic drivers; hence, molecular analysis of NSCLC is routinely performed to screen for several major therapeutic targets, including activating mutations of the epidermal growth factor receptor (EGFR) and rearrangement involving the anaplastic lymphoma kinase (ALK) [4]. These two drivers have conventionally been considered to be mutually exclusive [5]. However, cumulative reports have revealed that concomitant occurrence of EGFR mutations and ALK rearrangement accounts for a small number of NSCLC cases, raising the issue of therapeutic strategies [6,7,8,9,10,11,12,13,14,15,16].

First-generation EGFR tyrosine kinase inhibitors (TKIs), including gefitinib, erlotinib, and icotinib, are effective in patients with EGFR mutations [17,18,19]. Crizotinib, a small-molecular TKI, also shows efficacy superior to that of conventional chemotherapy in ALK-positive patients [20]. Nevertheless, the efficacy of EGFR TKIs and crizotinib in EGFR/ALK double-positive patients has been controversial. Some reports have indicated that EGFR TKIs had a better response than ALK inhibitors in terms of objective response rate (ORR) and progression-free survival (PFS) [7, 9, 10], but others have come to the opposite conclusion [11, 14, 15, 21]. Moreover, the effectiveness of combination or sequential therapy remains to be elucidated because of the scant number of patients reporting with concomitant EGFR mutations and ALK rearrangements and receiving both TKIs. We performed a retrospective analysis of 22 patients harboring concomitant EGFR mutations and ALK rearrangements to analyze the therapeutic efficacy of first-generation EGFR TKIs and crizotinib in these patients compared with those with a single oncogenic driver and to highlight the effectiveness of sequential treatment with EGFR TKIs and crizotinib.

2 Patients and Methods

2.1 Patients

This study was a data-driven retrospective single institution clinical audit without a prospective protocol. We identified patients with NSCLC who underwent molecular analysis at the Shanghai Chest Hospital between January 2011 and March 2017. The aim of this study was to evaluate patients with advanced (stages IIIB–IV) disease with either EGFR mutations or ALK rearrangements and co-occurring EGFR/ALK aberrations who received first-generation EGFR TKIs, crizotinib, or both. Patients were excluded if they were diagnosed with early-stage IA–IIIA NSCLC, did not receive TKIs, or received next-generation TKIs first. Patients with ALK rearrangements that were weakly positive on immunohistochemistry (IHC) and not confirmed by fluorescent in-situ hybridization (FISH) were also excluded.

The primary outcome of this study was comparison of PFS in patients with EGFR mutations, ALK rearrangements, and EGFR/ALK co-alteration. Secondary outcomes included overall survival (OS) and ORR in these patients.

Positron emission tomography/computed tomography (PET/CT) or other radiological evaluation methods were used to confirm or exclude metastatic disease. Staging was performed according to the 7th edition of the tumor node metastasis (TNM) classification for NSCLC. The specimens used for molecular analyses were obtained from surgery, fine-needle small biopsies guided by CT or ultrasound, transbronchial biopsies, or malignant effusion cell blocks.

The study was approved by the institutional review board of the Shanghai Chest Hospital, and all patients signed written consent forms before invasive procedures and TKI treatments.

2.2 Molecular Analyses

EGFR mutations were evaluated with the direct sequence and amplification refractory mutation system (ARMS) according to the manufacturer’s protocol. DNA was extracted from various specimens, amplified with polymerase chain reaction (PCR), and analyzed with standard sequencing. Details of the two methods are previously described [22, 23].

ALK rearrangements were identified by IHC or FISH. Samples were sent as formalin-fixed and paraffin-embedded (FFPE) biopsies, and FISH tests were performed on these tissues using a Vysis ALK break-apart FISH probe kit (Abbott Molecular, Des Plaines, IL, USA). Patients were diagnosed as ALK FISH positive when ≥ 15% of the tumor cells showed split and/or isolated 3ʹ signals. ALK IHC was conducted with the anti-ALK (D5F3 Ventana) primary antibody combined with OptiView DAB IHC detection and OptiView amplification (Ventana Medical Systems, Inc., Tucson, AZ, USA). Patients were diagnosed as ALK IHC positive if strong granular cytoplasmic brown staining was present in tumor cells [24, 25].

2.3 Treatment and Follow-Up

Patients in three cohorts were treated with EGFR TKIs (gefitinib [Iressa, AstraZeneca Pharmaceuticals], erlotinib [Tarceva, Roche], or icotinib [Conmana, Betta]), crizotinib, or both. Clinical evaluation was performed every 4–6 weeks according to RECIST 1.1 (Response Evaluation Criteria in Solid Tumors) [26]. PFS was defined from the initiation of TKIs to radiographic or clinical progression or the last follow-up time. OS was calculated from pathological diagnosis of stage IIIB/IV NSCLC to death or last follow-up time. The median follow-up time was 33 months (range 4–100), and the last follow-up time was 23 January 2019.

2.4 Statistical Analyses

We eliminated the large difference in the numbers of patients among the EGFR-mutant or ALK-rearranged cohort, and the ALK/EGFR co-altered cohort. EGFR-mutant or ALK-rearranged patients included for survival analysis were randomly selected from those with a single oncogenic driver who received first-generation EGFR TKIs or crizotinib and had complete medical records. The ratio of selected patients to double-positive patients who received EGFR TKIs or crizotinib was 1:5. The randomization process using Excel 2013 (Microsoft, Redmond, WA, USA) was as follows: (1) every patient was given a random number using the RANDBETWEEN function, ranging from 0 to 1; (2) patients were ranked according to the assigned number; and (3) the top patients needed for analysis were selected. To test the representativeness of the selected patients, their baseline characteristics were compared with those of the whole population, including patients with missing data.

To compare baseline characteristics, we used the chi-squared test for categorical variables, Fisher’s exact test for small samples, and the Wilcoxon rank sum test for continuous variables. Survival curves were generated to compare PFS and OS using Kaplan–Meier methods and further compared using the log-rank test. A p value  < 0.05 was considered statistically significant. All analyses were performed using the Statistical Package for Social Science (IBM; Armonk, NY, USA) version 22.0 for Windows. Figures were created using GraphPad Prism 7 (San Diego, CA, USA).

3 Results

3.1 Patient Characteristics

Between January 2011 and March 2017, a total of 5816 patients underwent both EGFR-mutation and ALK-rearrangement analysis. In total, 2392 patients had EGFR mutations, 503 had ALK alterations, and 26 had concomitant EGFR mutations and ALK alterations. The frequency of concomitant EGFR mutations and ALK rearrangement was about 0.45% (26/5816) of the entire NSCLC patient cohort, 1.1% (26/2418) of EGFR-mutation patients, and 4.91% (26/529) of ALK-positive patients.

Excluding patients diagnosed with stage IA–IIIA, who had not received TKIs, or who had received next-generation TKIs first left 1328, 259, and 23 patients with advanced (stage IIIB–IV) NSCLC in the EGFR-mutant, ALK-rearranged, and EGFR/ALK co-altered cohorts, respectively. Table 1 presents the demographic profiles and clinico-pathological characteristics of the three groups. The median age at diagnosis of stage IIIB/IV NSCLC for EGFR/ALK co-altered, EGFR-mutant, and ALK-rearranged groups was 61, 60, and 56 years, respectively. Patients from the ALK-rearranged cohort were younger than those from the EGFR-mutant or EGFR/ALK cohorts (p < 0.01). The specimens for diagnosis and molecular analysis were obtained in various ways (p < 0.01). There was no difference between the three groups with respect to sex, smoking status, pathologic type, performance status, or stage at initiation of TKIs. The EGFR-mutation types also showed no difference between the EGFR/ALK co-altered and EGFR-mutant groups.

Table 1 Baseline demographic profiles and clinico-pathological characteristics of patients with EGFR (+), ALK (+), and EGFR/ALK non-small-cell lung cancer
Full size table

3.2 Response and Survival Analyses

In the EGFR-mutant, ALK-rearranged, and EGFR/ALK co-altered cohorts, respectively, 367, 56, and 1 patient with stage IIIB/IV NSCLC and missing data received TKIs. Using the abovementioned procedure, 95 patients with EGFR mutations and 60 with ALK rearrangements were randomly selected from those with complete medical records. The patient selection process is outlined in Fig. 1.

Fig. 1
figure 1

Flow diagram outlining the patient selection process. ALK anaplastic lymphoma kinase, EGFR epidermal growth factor receptor, NSCLC non-small-cell lung cancer, TKI tyrosine kinase inhibitor

Full size image

Finally, 95, 60, and 22 patients from the EGFR-mutant, ALK-rearranged, and EGFR/ALK co-altered groups were included for survival analysis. Among the EGFR/ALK co-altered cohort, ten patients received EGFR TKIs (gefitinib, erlotinib, icotinib) alone, three patients received crizotinib alone, and nine patients received both EGFR TKIs and crizotinib. Therefore, a total of 19 patients previously received EGFR TKIs, and 12 patients were treated with crizotinib. All patients in the EGFR/ALK co-altered cohort experienced disease progression as assessed by RECIST while receiving EGFR TKIs and crizotinib.

The baseline characteristics of the entire EGFR-mutant cohort (all patients with EGFR-mutant stage IIIB/IV NSCLC treated with EGFR TKIs), selected EGFR-mutant patients, and EGFR/ALK co-altered patients who received EGFR TKIs were well-balanced (Table 2). The clinico-pathological characteristics of the entire ALK-rearranged cohort (all stage IIIB/IV NSCLC ALK patients treated with crizotinib), selected ALK-rearranged patients, and EGFR/ALK co-altered patients who received crizotinib were also well-balanced (Table 3). As a result, the randomly selected patients were considered to satisfactorily represent the entire EGFR-mutant and ALK-rearranged cohorts.

Table 2 Baseline characteristics of the entire EGFR-mutant cohort, selected EGFR-mutant patients, and EGFR/ALK co-altered patients who received EGFR TKIs
Full size table
Table 3 Baseline characteristics of the entire ALK-rearranged cohort, selected ALK-rearranged patients, and EGFR/ALK co-altered patients who received crizotinib
Full size table

The ORR for patients treated with EGFR TKIs was 62.1% (59/95) for EGFR-mutant and 63.2% (12/19) for double-positive patients. The difference was not statistically significant (p = 0.93). The ORR for patients treated with crizotinib was 65% (39/60) for ALK-rearranged and 66.7% (8/12) for EGFR/ALK double-positive patients. The difference was not statistically significant (p = 1.00).

PFS was defined from the initiation of TKIs to radiographic or clinical progression or the last follow-up time. The median PFS of EGFR/ALK co-altered patients treated with EGFR TKIs was 10.3 months, which was not statistically different from that of EGFR-mutant patients (11.4 months; hazard ratio [HR] 0.96; 95% confidence interval [CI] 0.59–1.57; p = 0.87) (Fig. 2a). Additionally, the median PFS of crizotinib was 11.1 months and 12.5 months for EGFR/ALK double-positive and ALK-rearranged patients, respectively. No  statistically significant difference was found between these two groups (HR 1.39; 95% CI 0.69–2.80; p = 0.28) (Fig. 2b).

Fig. 2
figure 2

a Kaplan–Meier curves for progression-free survival of selected EGFR-mutant patients and EGFR/ALK double-positive patients treated with EGFR TKIs. b Kaplan–Meier curves for progression-free survival of selected ALK-rearranged patients and EGFR/ALK double-positive patients treated with crizotinib. ALK anaplastic lymphoma kinase, EGFR epidermal growth factor receptor

Full size image

OS was calculated from pathological diagnosis of stage IIIB/IV NSCLC to death or last follow-up date, with 68% (15/22), 72% (68/95), and 63.3% (38/60) of patients known to have died at the data cut-off date in the EGFR/ALK co-alteration, EGFR-mutation, and ALK-rearrangement groups, respectively. Median OS was 36.8, 27.1, and 36.2 months (p = 0.034) (Fig. 3). Although no statistically significant difference was found between the EGFR/ALK co-existing group and the selected EGFR-mutant group (HR 0.66; 95% CI 0.40–1.08), there was a trend toward increased long-term survival for double-positive patients.

Fig. 3
figure 3

Kaplan–Meier curves for overall survival of selected EGFR-mutant, selected ALK-rearranged, and EGFR/ALK double-positive patients. ALK anaplastic lymphoma kinase, EGFR epidermal growth factor receptor

Full size image

Among a group of patients who were treated with both EGFR TKIs and crizotinib, only one patient received crizotinib before first-generation EGFR TKIs, and eight patients received crizotinib after failure of EGFR TKIs. In this subgroup, the ORR was 55.6% (5/9) and 66.7% (6/9), respectively, for EGFR TKIs and crizotinib. The median PFS of crizotinib as a sequential therapy after failure in EGFR TKIs was 15.0 months, which exhibited no statistically significant difference from that of ALK-altered patients who received crizotinib (12.5 months; HR 1.10; 95% CI 0.51–2.37; p = 0.80). Four patients received crizotinib as first-line therapy (three only received crizotinib; one received both crizotinib and an EGFR TKI), and the ORR for crizotinib in this subgroup was 75.0% (3/4).

4 Discussion

Our study suggests that the coexistence of EGFR mutations and ALK rearrangement accounts for about 0.45% (95% CI 0.28–0.62) of NSCLC, 1.1% (26/2418) of EGFR-mutant disease, and 4.91% (26/529) of ALK-positive disease. First-generation EGFR TKIs and crizotinib in patients with concomitant EGFR mutations and ALK rearrangement were equally efficacious as in patients with single-gene alterations. EGFR/ALK co-altered patients also appeared to have longer OS than patients with EGFR-mutant disease.

Remarkable progress has been made in the treatment of NSCLC due to the discovery of several critical activating pathways. Among these pathways, EGFR-activating mutations and ALK rearrangements are routinely screened for and clinically relevant [27]. Although these two driver alterations were previously considered mutually exclusive [5], recent reports have shown that EGFR/ALK double-positive patients account for a small proportion of NSCLC [6,7,8,9,10,11,12,13,14,15,16]; however, previous data has been inconsistent with regard to the frequency of double-positive patients in NSCLC, ranging from 0.1 to 1.6% [9, 15, 21, 28]. The prevalence of double-positive patients in our study is within this range when using ARMS and IHC/FISH techniques to detect EGFR mutations and ALK rearrangement. Won et al. [15] reported the frequency of EGFR and ALK coexistence in ALK-positive patients to be 4.4% when using Sanger sequencing, but this increased to 15% with high-sensitivity next-generation sequencing (NGS). Therefore, as more advanced techniques emerge, the proportion of double-positive patients could be much higher than previously expected, which will have implications for the treatment of these patients.

First-generation EGFR TKIs such as gefitinib, erlotinib, and icotinib provide survival benefits over conventional chemotherapy and have revolutionized the therapy of patients with NSCLC with EGFR-activating mutations [17,18,19], as has crizotinib, a TKI targeting ALK in activating rearrangements, for ALK-positive patients [20]. These TKIs play irreplaceable roles in managing NSCLC with a single oncogenic driver, but there is no consensus on their effects in double-positive patients because of the limited number of reported cases.  Lou et al. showed that the median PFS was 11.2 months for EGFR/ALK co-altered patients treated with EGFR TKIs and 13.2 months for EGFR-mutant patients. A less favorable result for crizotinib was found in double-positive patients, with a median PFS of 1.9 months compared with 6.9 months in ALK-rearranged patients, although no statistical significance was demonstrated because of the small number of patients included in this analysis [9]. Sweis et al. [12] illustrated a poor response to both EGFR and ALK TKIs with a disease control rate (DCR) of 33.3% (1/3) and 25% (1/4), respectively, among four patients with coexisting EGFR mutations and ALK rearrangement. However, some reports showed encouraging responses when EGFR and ALK TKIs were given to these patients. Baldi et al. [6] reported a double-positive long-term survivor achieving 3 years of stable disease using erlotinib and a partial response to crizotinib afterwards. Lo Russo et al. [8] conducted a literature review of 100 cases, among which 51 cases received EGFR TKIs and 37 received ALK TKIs. Disease control or disease response were reported as best response in 69.8% and 43.4% compared with 79.5% and 51.3% of reviewed cases treated with EGFR and ALK TKIs, respectively. It is hard to draw firm conclusions from these results, partly because most of these patients were presented in single case reports, and no large-scale study was conducted. In our study, the ORR for EGFR TKIs and crizotinib for dual-positive patients was 63.2% (12/19) and 66.7% (8/12), and the median PFS was 10.3 months and 11.1 months, respectively, which showed no statistically significant difference from patients with single gene drivers. Thus, we may conclude that both EGFR and ALK TKIs were effective in the treatment of EGFR/ALK double-positive patients.

Although we found no difference between ALK-rearranged and EGFR/ALK co-altered patients with respect to OS, there was a tendency towards longer OS for double-positive patients compared with those with EGFR mutations alone. This was probably related to their effective response to both EGFR and ALK TKIs. In contrast to our results, Lou et al. reported that double-positive patients had relatively shorter OS than those with either an EGFR mutation or an ALK rearrangement [9], possibly because of the smaller sample size.

Among the 22 patients included for analysis in our study, nine patients received both EGFR TKIs and crizotinib. EGFR TKIs were administered before crizotinib in the majority of cases (eight patients). In this subgroup, we found that the ORRs were 55.6% (5/9) and 66.7% (6/9) for EGFR TKIs and crizotinib, respectively. The median PFS for crizotinib as sequential therapy after failure in EGFR TKIs was 15.0 months, which exhibited no statistical significance compared with that of patients with an ALK rearrangement who only received crizotinib (p = 0.80). This all suggests that sequential treatment with EGFR and ALK TKIs is effective in treating double-positive patients. Furthermore, EGFR TKIs did not influence the efficacy of crizotinib as a subsequent therapy after progression on EGFR TKIs. Four patients received crizotinib as first-line treatment, with an ORR of 75% (3/4). Given the limited number of patients who received crizotinib before EGFR TKIs in our study, it was infeasible to determine the optimal sequence of administration.

Finding biomarkers to predict the efficacy of EGFR and ALK TKIs is another challenge. The burden of EGFR mutations in the double-positive patients may be a predictor of response to EGFR TKIs. Meanwhile, Yang et al. [16] proposed that relative phospho-ALK and phospho-EGFR levels, as well as the level of phosphorylation of downstream proteins in the signaling pathway, could predict the efficacy of EGFR TKIs and crizotinib.

Different theories exist concerning the coexistence of EGFR mutations and ALK rearrangement in NSCLC. Cai et al. [29] demonstrated that the two gene alterations were not from the same tumor cell and that coexistence was due to intratumor heterogeneity. However, Yang et al. [16] found mutant EGFR proteins co-expressed with an ALK fusion protein in the same cell population using IHC, and Sasaki et al. [30] showed that co-activation of EGFR signaling occurred in an ALK-rearranged cell line.

As mentioned, EGFR/ALK co-altered patients can benefit from sequential treatment with both TKIs and have long-term survival, so a more effective way to find these co-altered patients is needed. NGS is now widely available, and more EGFR/ALK patients can be identified using highly sensitive NGS [15]. Therefore, we recommend that NGS would be better performed in all patients with advanced NSCLC to find all targetable gene alterations; additionally, in areas with limited access to NGS, patients should also be analyzed for both EGFR mutations and ALK rearrangements using conventional methods.

This was a retrospective single-center study, which cannot accurately reflect the entire double-positive patient population. Although the study comprised the highest number of patients with concomitant EGFR mutations and ALK rearrangements in one institution, to our knowledge, the sample size was still relatively small. Additionally, chemotherapy and radiation therapy could have been used in different lines of treatment, which might interfere with the result.

5 Conclusions

EGFR/ALK double-positive patients accounted for 0.45% (26/5816) of all patients with NSCLC, 1.1% (26/2418) of EGFR-mutant patients, and 4.91% (26/529) of ALK-positive patients. Both first-generation EGFR TKIs, including gefitinib, erlotinib, and icotinib, as well as crizotinib, were effective in these patients, which showed no statistically significant difference from patients with a single gene driver in terms of PFS and ORR. The sequential treatment with EGFR TKIs and crizotinib appears promising, and EGFR TKIs did not influence the efficacy of crizotinib as a subsequent treatment in these patients.