Lung cancer is the leading cause of cancer-related death in worldwide, and accounts for one third of all deaths (1). Current treatment of lung cancer is still based on traditional surgery. Unfortunately, most lung cancer patients, especially non-small cell lung cancer (NSCLC) patients present advanced disease that is surgically unresectable. It is current practice to treat these patients with a combination of chemotherapy and external beam irradiation. But treatment outcomes for NSCLC remain unsatisfactory, with low long-term survival rates (<15%). Emerging TARGETED therapy with epidermal growth factor receptor tyrosine kinase inhibitors (EGFR-TKIs) are currently being evaluated in NSCLC.

EGFR is a receptor tyrosine kinase (TK) which is widely distributed in mammalian epithelial cells, glial cells, fibroblasts, keratinocytes and other cell surface. EGFR family includes four structurally-related tyrosine kinase receptors: EGFR (ErbB-1), HER2/c-neu (ErbB-2), Her3 (ErbB-4). These receptors are related to 70% of cancer (2). Abnormal activation of EGFR can promote tumor cell proliferation, differentiation, migration. EGFR has been found to be overexpressed in a variety of human malignancies (2,3). Activation of EGFR results in the initiation of a diverse range of cellular signaling pathways, including cell proliferation and protection of the cell from apoptosis (4). In 2004, Lynch et al (5) and Paez et al (6) proposed the EGFR gene mutations in lung cancer can be predicted, which is considered a milestone for NSCLC individualized targeted molecular therapies. These mutations are located in EGFR exons 18–21 (7) and ~90% of the sensitizing mutations are in-frame deletions in exon19, and the point mutation L858R in exon 21 (8). These mutations cluster around the ATP binding pocket of the tyrosine kinase domain of EGFR, leading to ligand-independent activation of the receptor and prolonged activation time compared to wild-type EGFR. Although exon 20 mutations were less, it is associated with resistance to anti-EGFR therapies (9–11).

The treatment of NSCLC using EGFR-TKIs, such as gefitinib and erlotinib was confirmed, showing promising outcomes for some patients with NSCLC, especially those with mutated EGFR (5,6,12). EGFR mutations increase sensitivity to TK inhibitors, most likely through induction of critical structural modifications of the ATP-binding site in the TK domain (13). EGFR mutation testing may be prognostically important to guide identify potential responders (or non-responders) to therapy. Therefore, a rapid, sensitive and reliable method for EGFR mutations testing is required.

The classic method for detecting EGFR mutations is direct sequencing. However, sequencing for routine testing is impeded by relatively high cost, requirement of sophisticated equipment and technical expertise, which are often only available in specialized molecular pathology laboratories (14). HRM is a recently developed technique that shows great potential for somatic mutations (15), but rarely uses serum samples. In 1989, studies have shown that the DNA in the plasma of cancer patients had tumor characteristics (16). In 1994, Sorenson et al and Vasioukhin et al detected the same gene mutation in tumor and plasma, this discovery confirmed that a part of free DNA from tumor existed in plasma (17,18). However, the clinical development of EGFR mutation detection was impeded by the low sensitivity of conventional methods, because the tumor DNA levels in peripheral blood is extremely low. In this study, we aimed to establish a sensitive and reliable HRM method for routine EGFR mutation screening using serum, which will benefit many NSCLC patients, especially those surgically unresectable.

Personalized medicine for cancer has raised new hope of cancer treatment, and become widely used in clinical settings over the last decade. Based on specific genetic information of various cancers, it had become possible to determine the genetic type of cancer, select the appropriate treatment, then enhance drug efficacy and reduce toxicity. However, the effectiveness of these new introduced molecular targeted drugs was closely dependent on the presence of specific genetic mutations in the tumor context (20–22). Gene mutation testing is the premise to using molecular-targeting drugs. EGFR gene mutations have recently been identified as a predictor for first-line EGFR-TKIs sensitivity (23). It was shown that 70% of EGFR mutation-positive NSCLC patients were effective to EGFR-TKIs, while only 10% of EGFR mutation- negative NSCLC patients (24). In addition, the responses of patients with different EGFR mutation types to EGFR-TKIs treatment were also different (25–27). Thus, it is of great importance to detect EGFR mutation status and type for NSCLC patients, which can effectively predict the benefit of taking an EGFR-TKI, and allow the physician to prescribe the most suitable therapy.

There are several methods for EGFR mutation testing, such as PCR-SSCP, gene chips and gene sequencing, with the gene sequencing being the current gold standard (28). HRM is a technique recently developed on the basis of real-time PCR amplification, which shows great potential for somatic mutations. A double stranded DNA binding dye is utilized in melting analysis in order to characterize primer-related non-specific amplification (or primer dimer) for detection of a specific target (29). Compared with traditional real-time PCR, HRM which adopt saturated dye have higher resolution, even single-base change can make the melting temperature of PCR products different (30). All the processes of PCR amplification followed by HRM take place in the same tube during a real-time run in <2 h (31). Therefore, HRM is the method of choice for rapid EGFR mutation screening.

In our study, we confirmed that both FFPE and fresh frozen tissue can be used for the HRM method. Mutation rates of FFPE and fresh frozen tissue are compatible; fresh frozen tissues showed no better than FFPE as others have reported (29). In addition, we found that HRM method could be used to differentiate 2310–2311insGGT and Q787Q mutation in exon 20 if known positive mutations control were properly set. However, HRM failed to identify the two mutation types in exon 21; sequencing had to be done to discriminate them.

Owing to the fact that tissue samples cannot be obtained from most of the lung cancer patients, the use of EGFR mutation screening in tissue samples is limited in practice. Plasma or serum which has free DNA from tumor provides as good alternative. However, the tumor DNA levels in peripheral blood are extremely low, which needs a more sensitive method for detection. The HRM assays described herein successfully identified EGFR mutations not only in tissue specimens but also in serum samples from NSCLC patients. The concordance rate between serum and fresh frozen tissues was better than in a previous report (32). The HRM assays acted perfectly in serum from NSCLC patients in stage II-IV, and no false negative was found. In addition, for NSCLC patients in stage I, most mutations can also be detected despite the 18.2% false negative rate. The sensitivity and specificity for serum EGFR screening by HRM was 91.67 and 100%, respectively. Therefore, HRM assays using serum samples for EGFR mutation screening can act as a good alternative for tissue screening, provide convincing and valuable results for the physician, and benefit surgically unresectable NSCLC patients.

The reported EGFR mutation rate in NSCLC patients differs and varies from 5 to 48.3% (30,32,33). It was shown that EGFR mutations were found in ~10% of cases of NSCLC in North America and 30–50% of NSCLC patients of East Asian descent (7). The frequency of the activating mutations is higher in Asian populations (34). In our research, the total EGFR mutation rate was 55.56%, which is the highest so far. This result may suggest the EGFR mutation rate of people in eastern China must be higher than other regions. In addition, many reports suggest major activating EGFR mutations are primarily in exons 19 and 21, with the inframe deletion in exon 19 and L858R accounting for 90% of reported mutations, additional mutations do occur in exons 18 and 20 that account for 8% of mutants (35). However, we found in this region of china, EGFR mutation occurs in exon 20 more often, especially Q787Q, and then in exons 19 and 21. Therefore, we should not only detect EGFR mutation in exons 19 and 21, but must attach great importance in EGFR exon 20 mutation detection in this region. We recommend that EGFR mutations in exons 18–21 should be detected before prescribing an EGFR-TKI for NSCLC patients from eastern China.

There have been debates on the relationships between EGFR mutation state and clinical characteristics (14,33,36). Our data showed EGFR mutations were more frequently observed in never-smokers, females, and those with adenocarcinoma histology, and there was no correlation between different clinical stages. Our results are comparable to those of Sriram et al (14) and Takano et al (33).

In summary, we have confirmed the feasibility of HRM for screening EGFR mutations in serum. Serum EGFR gene mutation screening by HRM assays are fast, relatively cheap, and robust, and can benefit surgically unresectable NSCLC patients; provide rapid scientific reference to the physician, and guide them to prescribe the most suitable therapy.