Background
In the last decade, great improvements have been made in the development of targeted therapies for lung cancer. To date, all approved targeted therapies have been aimed at adenocarcinomas of the lung, with adequate targets detected. Since patients with lung carcinoma usually present in an advanced stage, and the amount of available tissue is limited, it can be very difficult to reach a diagnosis and preserve sufficient tissue for molecular testing.
Molecular testing for epidermal growth factor receptor (EGFR) and anaplastic lymphoma kinase (ALK) mutations should be a current standard in pathology practice. The prevalence of EGFR mutations varies among different populations, and these mutations are present in higher frequencies in women, in light or never-smokers and in East Asian patients [
1‐
7]. Patients with specific mutations in the EGFR gene will respond differently to tyrosine kinase inhibitor (TKI) therapy. Patients with EGFR-mutated lung cancers have a better prognosis compared with patients without mutations, regardless of therapy type [
8]. ALK gene rearrangement, which is another targetable genetic change in lung adenocarcinomas, was discovered in 2007 [
9]. The frequency of ALK rearrangements is approximately 5 % in lung adenocarcinomas [
9,
10] and is higher in light or never-smokers and in younger individuals [
11‐
13]. Mutations in Kirsten rat sarcoma viral oncogene homolog (KRAS) are the most common mutations in lung adenocarcinomas, but currently, no targeted therapy is available. According to some studies, KRAS mutations are more prevalent in women, but other reports found no difference in regards to gender [
14‐
16]. Mutations in EGFR, ALK and KRAS are generally regarded as mutually exclusive [
8,
10,
11,
17].
The aim of this study was to present the EGFR, KRAS and ALK mutations in a representative cohort of patients with lung adenocarcinomas in Croatia and to correlate the mutational status with clinical data.
Discussion
We presented the results of the EGFR, KRAS and ALK mutational analyses in patients with lung adenocarcinomas in a prospective study, regardless of clinical stage. The analyses were performed only on patients who were newly diagnosed during a 6-month period at a single institution in Croatia. Importantly, the Department of Respiratory Diseases Jordanovac at the University Hospital Centre Zagreb is a referral center for lung diseases, and the majority of patients with lung carcinoma (almost 60 %) are diagnosed and treated here. According to available data from the Croatian National Cancer Registry for 2014, the number of patients with newly diagnosed carcinoma of the lung was 2915, and, of those, 1693 were diagnosed in our department (1693/2915, 58.1 %) [
19]. To date, this is the largest, most representative study of the mutational status of Croatian patients with lung adenocarcinomas. Our population was composed primarily of males (59.9 %) and a high proportion of smokers and former smokers (77.2 %); this population is practically identical to that of the recently published INSIGHT Study [
20] and is also similar to those in other studies [
21,
22].
In many published studies, a wide range in the prevalence of EGFR mutations has been reported in lung carcinomas, from 7.5 % in Norway [
23,
24], 8.8 % in a mixed ethnic population in the USA [
10], to 10 to 15 % in Europe [
20,
22], to more than 50 % in Asian countries [
25,
26]. In the study by Zaric et al., the authors detected the presence of EGFR mutations in 11.7 % of their patients [
22], while the INSIGHT Study showed a prevalence of 13.8 % of EGFR mutations in patients with lung carcinomas [
20]. Our prevalence of 15.7 % in EGFR-mutated lung adenocarcinomas seems to be among the higher rates in European countries and is closer to what was reported in the Russian study by Moiseyenko, where the prevalence was 19.8 % [
27].
What might be the reason for such a difference? There are some very important, and usually neglected differences, between published studies in regards to patient selection criteria (clinical and pathological), as well as the methods used for mutational analysis. In a French study by Vallee et al., one of the largest studies in Europe, [
28] 1403 tumor samples were analyzed, of which 1144 were adenocarcinomas and 101 were NSCLC-NOS; EGFR mutations were found in 14.7 % and 4.0 %, of adenocarcinomas and NSCLC-NOS, respectively. They used fragment analysis for exon 19 deletions and allele-specific PCR to detect the L858R mutation. Other mutations were not explored, and thus the true incidence of EGFR mutations in this population is almost certainly a bit higher than reported. Milella et al. presented the results of patients in clinical stages IIIB and IV, with histologically heterogeneous tumors (125 adenocarcinomas, 17 squamous cell lung carcinoma and 46 other lung carcinomas) [
21]. They also analyzed only exon 19 deletions and the exon 21 point mutation, and reported that 9.0 % of patients had EGFR mutations, but 40.4 % of their patient population was unable to be evaluated. Even when we analyzed what was apparently the lowest reported rate of EGFR mutations (7.5 %) in a study in Norway [
23], we see that, of 240 tumors, only 141 were actually lung adenocarcinomas, and only 11 % of patients had EGFR mutations. The patients included in that study all had operable lung cancers, and the investigators used a TheraScreen EGFR mutation kit (DxS, Manchester, UK) to detect 28 specific mutations; some samples were further analyzed by denaturing high-performance liquid chromatography and sequencing. In a more recent study in Norway, the prevalence of EGFR mutations in lung adenocarcinomas was found to be 9.4 % [
24]. However, the INSIGHT Study, which comprises results form 6 central European countries each with different inclusion criteria and testing methods, analyzed 1785 patients, of whom 1393 were diagnosed with adenocarcinomas [
20]. The prevalence of EGFR mutations in patients with adenocarcinomas was 15.4 %, which was practically identical to our findings. A Russian study analyzed only lung adenocarcinomas, which were assessed by PCR in a Cycler iQ Real Time Detection System (Bio Rad Laboratories, GmBH, Munich, Germany); the results revealed a prevalence of EGFR mutations that was higher compared with that in other European countries (19.8 %) [
27], but was still lower than that in Asian populations, in which the prevalence is above 40 %. In one study in China, the authors reported EGFR mutations in 66.3 % of consecutively collected lung adenocarcinomas, which were analyzed by sequencing [
25]. Another interesting point in the comparison of our results with those of the INSIGHT study was the frequency of specific mutations; while practically no differences were observed in the frequency of exon 19 deletions (the most common mutation type) and exon 20 insertions, the frequency of the L858R mutation was higher in our patients (representing 39.2 % of all mutations compared with 28.3 % in the INSIGHT Study) [
20]. To date, only one study with data on EGFR mutations in a Croatian population was published. This study, by Mohar et al., reported a prevalence of EGFR mutations of 19.8 % in tested patients [
29]. However, in this paper, the authors analyzed only cytological samples that were obtained from more than one hospital, and it is not clear from the paper whether the analysis was performed with or without patient pre-selection. Another possible reason for the difference between the reported prevalence rates and our results might be the higher number of female patients (53.9 % compared with 40.1 % in our study).
However, in all the studies presented here, and in practically all others, a statistically significant correlation was found between gender and EGFR status (mutations are more prevalent in females) and between smoking and EGFR mutation status (mutations are more prevalent in never smokers and former smokers). This is in accordance with our results. It is known that higher EGFR mutation frequencies are observed in female, non-smoking patients of Asian origin [
25,
26], reaching more than 60 % in this population. A similar, higher EGFR mutation frequency (52 %) was also found by our study group when only females who are non-smokers were analyzed. The limitation of this finding is that the number of these patients was relatively low (50 patients), but nevertheless, the similarity is obvious.
KRAS mutations are present in many different tumors, including carcinomas of the lung, as KRAS is one of the most commonly mutated genes in human cancer [
30]. In lung carcinomas, KRAS mutations are more common in smokers [
31‐
33]. KRAS mutations occur in approximately 30 % of lung adenocarcinomas in Caucasians [
30,
34‐
36] and in approximately 10 % of lung adenocarcinomas in Asians [
34,
37]. To date, no effective therapy that targets KRAS has been released, although some therapeutic molecules are currently under investigation [
38]. In lung adenocarcinomas, mutations usually occur in codon 12 of exon 2, followed by codon 13 of exon 2 (3-5 %) and rarely (<1 %) in codon 61 of exon 3 [
39]. Interestingly, in a Chinese population composed of patients with lung adenocarcinomas who are also smokers, the incidence of KRAS mutations was found to be 14.0 %, while that of non-smokers was 3.4 % [
40]. Two studies of Asian patients with NSCLC [
41,
42], as well as meta- and pooled analyses [
43,
44], demonstrated that the presence of KRAS mutations is a poor prognostic factor in Asians with NSCLC, while its relationship to prognosis in cases of NSCLC in non-Asian patients is still debatable. Additionally, the predictive role of KRAS mutations is not yet completely clear (for review see [
45]). Our results demonstrated a slightly higher incidence of KRAS mutations in lung adenocarcinomas than was previously reported for a Caucasian population, which might be due to the smoking habits of this Croatian population. Furthermore, in our population, the correlation of KRAS mutations with smoking proved to be statistically significant even after stratification for gender, which was not the case for the correlation of KRAS mutations with gender after stratification for smoking. The distribution of specific types of mutations in our sample was also similar to previously published data.
At the other end of the mutational landscape of lung adenocarcinomas are ALK gene rearrangements, which occur in 1-6 % of NSCLCs worldwide [
46‐
48]. The meta-analysis conducted by Fengzhi Zhao et al., which included a total of 6950 patients, showed that the prevalence of ALK translocations was 6.8 % in NSCLC patients; this study did not select for ethnicity, but once again demonstrated that ALK translocations are practically mutually exclusive with EGFR and KRAS mutations [
49]. They did find that EGFR and ALK alterations occurred simultaneously in 15/6950 patients (0.2 %). Their meta-analysis showed, for the first time, a higher frequency of ALK gene rearrangement in a non-Asian population (8.5 %, 173/2044) without selection for NSCLC, compared with 6.1 % (299/4906) in an Asian population. In a study by Wong et al., the prevalence of ALK gene rearrangement in an East Asian population was 5 % [
48]. By contrast, Fu et al. found 44 patients with ALK gene rearrangement, (43/382, 11.3 % of patients with lung adenocarcinoma, and 1/73, 1.3 % of patients with squamous cell lung carcinoma) [
50]. The prevalence of ALK gene rearrangement in the Chinese population also differs based on smoking status, as the prevalence is lower in smokers (2.9 %) than in non-smokers (7.2 %) (for review see 42). Based on the fact that IHC and FISH are highly concordant [
51,
52], and that IHC was economically feasible, we decided to test the tumors of our patients by IHC only. Our IHC results are in agreement with published results, as are their correlation with gender (higher frequency in females) and smoking status (lower frequency in former smokers and smokers). Smoking and the higher percentage of males in our study group are also the likely causes of the lower incidence of ALK translocations in our population.
Although only 2.4 % of the initial number of patients who were eligible for molecular analysis did not have adequate tissue/tumor cells for any tests, no additional adequate material was available for ALK immunohistochemistry in 8.6 % of the tested samples, which stresses the importance of proper handling of tumor samples in order to have enough for all necessary analyses. This is an important issue that will probably be solved by the introduction of panel tests, in which not only can mutations be analyzed simultaneously, but a smaller amount of material is needed for molecular testing.