Molecular Epidemiology of EGFR Mutations in Asian Patients with Advanced Non-Small-Cell Lung Cancer of Adenocarcinoma Histology – Mainland China Subset Analysis of the PIONEER study

Epidermal growth factor receptor (EGFR) mutations are the strongest response predictors to EGFR tyrosine kinase inhibitors (TKI) therapy, but knowledge of the EGFR mutation frequency on lung adenocarcinoma is still limited to retrospective studies. The PIONEER study (NCT01185314) is a prospective molecular epidemiology study in Asian patients with newly diagnosed advanced lung adenocarcinoma, aiming to prospectively analyze EGFR mutation status in IIIB/IV treatment-naïve lung adenocarcinomas in Asia. We report the mainland China subset results. Eligible patients (≥20 yrs old, IIIB/IV adenocarcinoma and treatment-naïve) were registered in 17 hospitals in mainland China. EGFR was tested for mutations by amplification refractory mutation system using biopsy samples. Demographic and clinical characteristics were collected for subgroup analyses. A total of 747 patients were registered. Successful EGFR mutation analysis was performed in 741, with an overall mutation rate of 50.2%. The EGFR active mutation rate is 48.0% (with 1.3% of combined active and resistance mutations). Tobacco use (>30 pack-year vs. 0–10 pack-year, OR 0.27, 95%CI: 0.17–0.42) and regional lymph nodes involvement (N3 vs. N0, OR 0.47, 95%CI: 0.29–0.76) were independent predictors of EGFR mutation in multivariate analysis. However, even in regular smokers, the EGFR mutation frequency was 35.3%. The EGFR mutation frequency was similar between diverse biopsy sites and techniques. The overall EGFR mutation frequency of the mainland China subset was 50.2%, independently associated with the intensity of tobacco use and regional lymph nodes involvement. The relatively high frequency of EGFR mutations in the mainland China subset suggest that any effort to obtain tissue sample for EGFR mutation testing should be encouraged.

Therefore, it is of great importance to determine the prevalence of EGFR mutations frequency. Studies in different populations identified some subgroups (adenocarcinoma histology, women, never-smokers and East Asian ethnic origin) with higher EGFR active mutation rate [8,9]. However, several retrospective studies in resectable and advanced lung adenocarcinoma have suggested that gender was not an independent factor for EGFR mutation [10,11]. Therefore, it is significant to identify demographic and clinical characteristics associated with EGFR mutations, to allow identifying subpopulations of patients at high risk of harboring EGFR mutation, thus allowing the oncologists to decide which first-line treatment may offer the most benefits.
Most studies assessed EGFR mutation status in a clinical trial setting, or retrospectively based on archival tissue samples which might not sufficiently comply with real-life practice on an epidemiological point of view. Therefore, prospective studies of EGFR mutation status in clinical settings are needed.
A prospective epidemiological study has been conducted in Spain to screen for EGFR mutations in lung cancer [12], but no prospective studies were carried out for Asian patients. The PIO-NEER study (NCT01185314) is a large prospective molecular epidemiology study in Asian patients with newly diagnosed advanced lung adenocarcinoma, aiming to assess their EGFR mutation status [13]. This study enrolled a total of 1510 patients in 51 investigational sites in 7 Asian countries/regions (mainland China, Hong Kong, India, Philippines, Taiwan, Thailand and Vietnam) to investigate EGFR mutation frequency and to correlate these mutations with patients' demographic and clinical characteristics, as well as with various tissue sampling techniques. In the present report, we present the results for the subset of patients from mainland China.

Study Design and Patients
The PIONEER study (NCT01185314) is an epidemiological, multicenter, open-label and noncomparative study of EGFR mutation status in patients with newly diagnosed advanced (stage IIIb/IV) NSCLC. Eligibility criteria were: 1) 20 years old; 2) histologically or cytologically confirmed advanced lung adenocarcinoma (stage IIIb/IV); and 3) treatment-naïve. Patients were registered in 17 hospitals in mainland China.
The study was performed according to the Declaration of Helsinki and good clinical practice guidelines. The study was approved by the Ethics Committees of all study centers. All patients provided a written informed consent for participation in the study and the use of tumor samples (tissue or cytology).

Data collection
Demographic data of patients were collected, including gender, age, and smoking status. Disease data included date of first NSCLC diagnosis, histological type, AJCC stage, nodal status, and distant metastases. Standardized case report forms were used to record the data in accordance with the protocol's instructions. Smoking was assessed using two ways. First, patients were classified according to their actual smoking status (never-smoked means that the subject smoked no cigarettes during his entire lifetime; ex-smoker means that the subject no longer smokes; occasional smoker means that the subject smokes, but not every day; and regular smoker means that the subject smokes every day). Second, smoking patients were classified according to their tobacco consumption, in pack-year.

EGFR Mutation Analysis
Tumor samples were obtained from primary or metastatic lesions, and were handled and stored following laboratories' quality control requirements. Biopsy site and technique were recorded. Cytological samples were accepted only when histological material was unavailable. After tumor DNA extraction, EGFR mutation was analyzed at Laboratory of Medical Oncology, Cancer Hospital, Chinese Academy of Medical Sciences & Peking Union Medical College; Beijing Key Laboratory of Clinical Study on Anticancer Molecular Targeted Drugs by an amplification refractory mutation system (ARMS)-based EGFR mutation detection kit (EGFR RGQ PCR kit, Qiagen, Crawley, UK). This kit allows the detection of 29 mutations in the EGFR gene.

Statistical Analyses
Statistical analysis was performed by the Statistics and Programming department of Astra Zeneca Japan using SAS 8.2 (SAS Institute, Cary, NY, USA). The per protocol analysis (PPS) set was used for all statistical analyses. Mutations prevalence and corresponding 95% confidence intervals (95%CI) were calculated using the Wilson score method. Associations between mutations and demographic and clinical characteristics were analyzed by χ2 tests or Fisher's exact tests, as appropriate. Characteristics associated with mutations with a P-value <0.05 were included in a multivariate logistic model. All analyses were two-sided and a P-value <0.05 was considered significant. In the multivariate analysis, a P-value <0.01 was considered significant.
Age, time from original diagnosis, malignant pleural effusion, and primary tumor were not associated with M+ status. Tumor grade seemed to be significantly associated with EGFR mutation frequency; however, this variable was documented for a little number of patients, mainly due to the limited amount of tissue available for diagnosis, and was therefore not included in the multivariate analysis. The distribution of EGFR M+ according to biopsy site, type and technique was relatively similar (S1 Table). Patients with lymph nodes biopsy didn't show lower EGFR mutation frequency compared with other biopsy sites (47.2% vs 50.8%, P = 0.4591).
Multivariate analysis identified pack-year and regional lymph node involvement as independent predictors of EGFR M+. Patients smoking 10-30 (OR 0.54, 95%CI: 0.37-0.78) or  Never-smoked means that the subject smoked no cigarettes during his entire lifetime; ex-smoker means that the subject no longer smokes; occasional smoker means that the subject smokes, but not every day; and regular smoker means that the subject smokes every day. more than 30 (OR 0.27, 95%CI: 0.17-0.42) pack-year had less chance to harbor an EGFR mutation than those smoking less than 10 pack-year. Tumors with regional lymph nodes involvement (stage N1-2, OR 0.71, 95%CI: 0.44-1.14; and N3, OR 0.47, 95%CI: 0.29-0.76) were also less likely to harbor EGFR mutations than patients without lymph nodes involvement (N0) ( Table 2).

Types of EGFR mutation
The most common mutations were exon 19 (Table 3). Packyear and regional lymph nodes involvement were also identified as independent factors significantly associated with EGFR active mutations by univariate and multivariate analysis (S2 Table  and S3 Table). Additionally, nine patients provided both histology and cytology samples, among which one patient had mutation results that did not match.

Discussion
EGFR mutations are currently the strongest response predictor to EGFR-TKI therapy [14]. In addition, with the establishment of first-line treatment using EGFR-TKI [6,7], the knowledge of EGFR mutation status in treatment-naïve advanced NSCLC, especially lung adenocarcinoma, is of great importance for clinical decision. Previous data were generally derived from retrospective studies and clinical trials, which might not reflect real-life. PIONEER is a large epidemiological study prospectively investigating the EGFR mutation status in Asian treatment-naïve advanced NSCLC patients [13], and we report the results from the mainland China subset.
In our study (PPS population), the overall EGFR mutation frequency was 50.2%, similar to previous studies in East Asian patients [6,15], but much higher than in Caucasian lung adenocarcinoma patients [12,16]. Besides a different ethnicity distribution, EGFR mutations are also The significance level for variable selection was 0.01. NX in regional lymph nodes were excluded from the analysis. suggested to be more common in adenocarcinoma, women, and never-smokers [8,9]. Rekhtman et al. reported that "pure" squamous cell lung cancer, verified by immunohistochemistry as Np63 (+)/TTF-1(-), lacked EGFR mutations [17], thus leaving lung adenocarcinoma as the major candidate for EGFR screening in NSCLC. In the present prospective study, we also found that the impact of gender was not significant after adjustment for smoking status, and did not have any interaction with smoking (data not shown).
Smoking status was identified as an independent predictor of EGFR mutation in the present study. We observed that the EGFR mutation incidence decreased with the increasing tobacco consumption, which is similar to a retrospective by Li et al. in resectable NSCLC [18]. When smoking was classified into never-smoked, ex-smoker, occasional smoker, and regular smoker according to their smoking frequency, there was a wide gap between never-smoked and eversmoked patients. However, even in regular smokers, EGFR mutation was still found in 35.3% patients, indicating the necessity of EGFR mutation testing in this group of patients.
Another interesting finding of the present study was that EGFR mutation frequency independently decreased with the involvement of regional lymph node, and patients with N3 disease had the least chance of harboring an EGFR mutation. This finding is supported by a study by Na et al. in a Korean population (OR 0.2, 95%CI: 0.0-0.7, P = 0.02) [19]. There is little chance that this observation might be due to the tumor sample volume because the ARMS approach has a well-known detection sensitivity, and the EGFR M+ frequency according to biopsy sites was comparable. The plausible explanation still warrants further investigations. EGFR T790M mutation is associated with shorter progression-free survival with EGFR-TKI therapy in patients with NSCLC [20,21]. However, the prevalence of the EGFR T790M mutation varied by population and technique. Su [21]. In this prospective cohort of treatment-naïve advanced lung adenocarcinoma, the T790M mutation was found in only 0.4% of our Chinese patients using the ARMS assay, accounting for  [20]. This might be due to the smaller sample size in their study, and by the different detection method. The EGFR exon 20 insertion mutation was also suggested to induce a poor responses to EGFR-TKI [22], and was reported in about 2% of NSCLC or lung adenocarcinoma [23][24][25], which is similar to our report (2.0%).
The present study has some limitations. Indeed, despite all our efforts, we were unable to recruit all lung cancer patients, mainly because some of them did not consent, which could introduce a bias. Furthermore, tobacco consumption was self-reported, and might suffer from some under-reporting.

Conclusion
In conclusion, from the PIONEER prospective study, the overall EGFR mutation frequency of the mainland China subset was 50.2%. The EGFR active mutation rate is 48.0% (with 1.3% of combined active and resistance mutations). The relatively high frequency of EGFR mutation in this population and the comparable EGFR M+ rate among diverse biopsy sites and techniques suggest that any effort to gain tissue samples for EGFR mutation testing should be encouraged.
Supporting Information S1