Lung cancer is a lethal but common disease with a 5-year survival rate of approximately 5%–15% [1]. Molecular targeted therapy is emerging as an effective therapeutic strategy for many cancers [2–4]. The following molecularly targeted genes are potentially relevant to oncogenic events and have demonstrated remarkable significance: the v-raf murine sarcoma viral oncogene homolog B1 (BRAF), human epidermal growth factor receptor-2 (HER2), phosphoinositide 3′-kinase (PI3K), v-Ki-ras 2 Kirsten rat sarcoma viral oncogene homolog (KRAS), discoidin domain receptor 2 (DDR2), epidermal growth factor receptor (EGFR), anaplastic lymphoma kinase (ALK), c-ros oncogene 1 (ROS1), and ret proto-oncogene (RET). ALK rearrangement represents a new oncogene driver [5, 6].

The echinoderm microtubule-associated protein like 4 (EML4)-ALK in non-small cell lung cancer (NSCLC) was first discovered as an oncogenic driver gene in 2007; the EML4-ALK fusion gene was generated by an inversion in the short arm of chromosome 2 [6]. According to the breakpoint on EML4 (from exon 2 to exon 20), 13 variants of the EML4-ALK fusion gene have been found [7]. Recently, the TRK-fused gene (TFG), the kinesin family member 5B (KIF5B), and the kinesin light chain 1 (KLC1) were reported to invert and fuse to ALK in NSCLC [5, 8, 9].

Compared with other genetic abnormalities in NSCLC, the frequency of the ALK rearrangement is approximately 1%–6% in unselected NSCLC [10–12]. Patients with ALK rearrangement are highly sensitive to crizotinib, an oral tyrosine kinase inhibitor (TKI) of the c-Met proto-oncogene (c-Met), ALK, and ROS1 genes. The powerful and specific therapeutic efficacy of this drug on ALK-rearranged NSCLC led to the approval by the Food and Drug Administration (FDA) in the United States [13]. Several studies have shown that most of the patients who harbor this chromosomal abnormality have benefited from targeted therapy. A randomized phase III study reported that, in the subgroup of ALK-rearranged NSCLC, patients treated with crizotinib showed higher response rate and longer progression-free survival (PFS) compared with those treated with the standard second-line chemotherapies (docetaxel or pemetrexed) [14]. Other studies showed dramatic clinical benefits associated with crizotinib in patients with ALK-rearranged NSCLC [15, 16].

The tumors from patients with ALK rearrangements are characterized by distinct histologic features including a solid or acinar growth pattern, a cribriform structure, the presence of mucous cells and abundant extracellular mucus, a lack of lepidic growth, and nuclear pleomorphism [17]. Nevertheless, these histologic parameters are of insufficient sensitivity and specificity to detect ALK rearrangements, and therefore, histomorphology should not replace confirmatory molecular or immunohistochemical studies [17]. The histomorphology of ALK-rearranged lung cancer cannot be used as a screening method. Several methods including reverse transcription-polymerase chain reaction (RT-PCR), fluorescence in situ hybridization (FISH), and immunohistochemistry (IHC) are currently used to identify ALK translocations in patients with NSCLC. In the present study, we performed ALK FISH in NSCLC cases at the Sun Yat-sen University Cancer Center (SYSUCC) using the gold standard method, Abbott ALK break-apart probe. We analyzed the clinicopathologic features of the patients, their survival status, and the relationship between the clinicopathologic features and ALK rearrangement. In previous studies, ALK rearrangement represented a unique molecular subset of NSCLC with no overlap with cancers that featured alterations in the EGFR or KRAS genes [18, 19]. According to EGFR and ALK status, we retrospectively studied the responses of patients to traditional therapies compared with targeted therapies.

The identification of ALK rearrangements in NSCLC is pivotal to guide the appropriate treatment with ALK-TKI. In this study, we used an ALK break-apart probe assay to test 487 NSCLC cases and identified 44 (9.0%) with ALK rearrangements, which was in accordance with the results of previous investigations [25–28]. The frequency of ALK rearrangements was 20.3% (25/123) in young patients with the adenocarcinoma subtype who were non-smokers, which was also consistent with the previously reported results [24, 27, 29, 30]. Our study also revealed other characteristics of NSCLC with ALK rearrangements, such as advanced stage (70.5%, 31/44) and wild-type EGFR (97.7%, 43/44). Additionally, patients with advanced-stage tumors and poorly differentiated tumors had a shorter time to cancer progression in comparison with those with early-stage tumors and well differentiated tumors.

Tumors with ALK rearrangement were found to exhibit papillary and acinar growth patterns [18], mucinous cribriform patterns [17], and a solid signet-ring cell pattern [24]. In addition, the coexistence of glandular and squamous morphologies was also reported in previous studies [18, 24]. In our study, acinar, papillary, and mucinous adenocarcinomas were found in ALK-rearranged cases. Interestingly, in our study, we also detected ALK rearrangements in a case of squamous cell carcinoma. Furthermore, 4 ALK-rearranged adenocarcinomas were observed in patients with focal squamous differentiation. This suggests that a test for ALK rearrangement should be performed in patients with classical morphologic patterns, mixed squamous and adenocarcinoma patterns as well as in patients with squamous carcinoma in case of a wrong therapeutic decision.

ALK rearrangement was generally reported to be found in tumors with wild-type EGFR and KRAS [7, 19, 27, 31, 32]. However, some investigators also observed the coexistence of ALK rearrangement and EGFR mutations [33–35]. It was reported that patients who harbor a concurrent EGFR mutation and an ALK rearrangement may partially respond to EGFR inhibitors [36–38]. Interestingly, we identified one patient with a concomitant ALK rearrangement and EGFR mutation who had undergone therapy with erlotinib. Nevertheless, this patient experienced disease progression less than 1 month later as confirmed by CT scan. The mechanism of resistance might be affected by different signal transduction pathways. Sasaki et al. [38] reported that in in vitro studies the co-expression of an EGFR mutation and the EML4-ALK fusion gene might lead to resistance to targeted therapies. It is therefore important to determine the optimal combination of a given EGFR-TKI and ALK inhibitor or other new therapeutic regimens for patients with a concomitant EGFR mutation and an ALK rearrangement.

In an analysis of OS and PFS between ALK rearrangement-positive and ALK rearrangement-negative groups, no statistical significance was observed. The reason might be that the patients with ALK rearrangements, in our study, were not treated with targeted therapy or they were treated for only a short time. Li et al. [39] reported that EGFR/KRAS mutation status appeared to be significantly associated with neither PFS nor OS if these patients did not receive targeted therapies. This hypothesis is consistent with the results of our study. All evidence supports the hypothesis that gene subtypes could only predict the therapeutic response but not the survival benefit until patients received a given molecular targeted therapy.

Patients with EGFR mutations demonstrated a superior PFS after treatment with molecular targeted therapy compared with those treated with traditional platinum-doublet chemotherapy [40]. However, in our study, disease progression was observed in two patients with ALK rearrangement after they were treated with crizotinib, and the two patients were considered resistant to crizotinib. Presently, the mechanism of resistance to ALK inhibitors is thought to be a compensatory mechanism or the occurrence of drug-resistant gene mutation, such as a mutation in ALK, EGFR, KRAS, or BRAF [41, 42]. Rapid disease progression can deprive the patients of a second chance for survival. Spaans et al. [42] suggested that a combination targeted therapy that simultaneously inhibits multiple resistance pathways would elicit a better clinical response. However, the toxicity of combination drugs requires further investigation.

There are some limitations in our study. First, this was a retrospective study and all cases were selected within a single hospital. Second, many patients with ALK rearrangement or EGFR mutations could not receive targeted therapy or they had a short medication duration due to budget limitations. Therefore, we were not able to exclude geographic and demographic variations that might have influenced the outcome of the study.

In conclusion, the patients who harbor ALK rearrangements tend to be relatively young, non-smokers or light smokers with the adenocarcinoma subtype. Prior to treatment, it is necessary to assess ALK rearrangement status to determine the appropriate therapeutic regimen. Because the underlying mechanism of the partial response to targeted therapy is unknown, we will continue to investigate this mechanism in patients with NSCLC in the future.