Lung Cancer with Epidermal Growth Factor Receptor Exon 20 Mutations Is Associated with Poor Gefitinib Treatment Response

Epidermal growth factor receptor (EGFR), a transmembrane glycoprotein, is related to the proliferation and resistance to apoptosis of cells. A loss of constraints on the EGFR due to deregulation, amplification, or mutations may result in a malignant change of cells (1, 2). A series of EGFR kinase domain mutations have been discovered in non-small cell lung cancer (NSCLC), primarily adenocarcinoma, in recent years (3–5). The mutations exist in exons 18 to 21, which encompass most of the tyrosine kinase binding domain of EGFR (6). The clinical courses and survival of lung cancer patients with EGFR mutations can be changed with the treatment of gefitinib (7–9), a small molecule that is a selective EGFR tyrosine kinase inhibitor (10).

Patients with EGFR mutations and effective responsiveness to gefitinib in previous studies were found to have mutations primarily in exons 18, 19, and 21 (7, 9, 11–19). These are substitutions for G719 in the nucleotide-binding loop of exon 18, in-frame deletions within exon 19, and substitutions for L858 or L861 in the activation loop in exon 21. Mutations in exon 20 could be in-frame duplication and/or insertion, such as A767_V769dupASV or H773_V774insH (20). They could also be point mutations. Examples of reported exon 20 mutations in literature are V765M, S768I, H773R, and T790M (9, 11, 21–24). Only a handful of reports have described lung cancer with EGFR mutations in exon 20, and the population was small (8, 20, 21, 24–28). This is probably because of the innate smaller percentage of exon 20 mutations in the EGFR mutations in lung cancer. Experience is even limited about gefitinib treatment on lung cancer with exon 20 mutations (8, 21, 25). Cell models in vitro have shown that transformation by D770_N771insNPG of exon 20 brings about resistance to gefitinib (29). In the patient group of the study conducted by Chou et al. (8), the five patients with exon 20 mutations had a shorter duration of gefitinib response than those with other mutations. However, more clinical data are required. To increase our understanding of the influence of exon 20 mutations on gefitinib treatment response, we investigated the clinical features of lung cancer with exon 20 mutations and reviewed relevant literature reports about exon 20 mutations.

Patients. The study group included lung cancer patients diagnosed at the National Taiwan University Hospital between January 2000 and June 2007. All patients underwent complete lung cancer staging, including bronchoscopy, computed tomography of the head, chest, and abdomen, and whole-body bone scintigraphy. The clinical data from the patients, including demographic information, clinical staging, cell histology, and treatments received and responsiveness were recorded. Smoking status was recorded with direct statements of patients at the time of admission for management of lung cancer. Patients who had smoked <100 cigarettes in their lifetime were categorized as never smokers. Those who smoked cigarettes within 1 year of the diagnosis were categorized as current smokers. The others were categorized as former smokers. Lung cancer histology was defined according to the WHO pathology classification (30). Tumor specimens obtained by either surgical or needle biopsy/aspiration procedures, including primary lung lesions, malignant effusion cell blocks, and other distant metastases, were sequenced for mutational analysis. Patients screened for EGFR mutations were (a) those who were in the studies of clinical trial (31), (b) those who underwent fine-needle biopsies or thoracentesis for pleural effusions after July 2004 when consecutive recruitment for EGFR mutations was started in the hospital, (c) those whose resected tumors were retrospectively sequenced, or (d) those who were recruited for retrospective NSCLC studies (32, 33). Written informed consent for use of tissue in molecular analysis was acquired from patients at the time of procurement of tumor specimens. This study was approved by the Institutional Review Board of the National Taiwan University Hospital.

Gefitinib treatment and evaluation of effectiveness. Patient information regarding treatment with gefitinib was obtained from a search of records of the Department of Pharmacy, National Taiwan University Hospital. When administered, gefitinib was taken orally at a dosage of 250 mg/d. Baseline assessments were usually done within 2 weeks before treatment. For patients who were in the clinical trial (31), computed tomography of all the measurable tumor sites was done every 8 weeks. Response to treatment, time to progression, and survival were retrieved from the data set of the clinical trial. For patients who did not participate in clinical trials, chest radiography was routinely done and assessed every 2 to 3 weeks to evaluate responses after the start of treatment. A chest computed tomography scan (including liver and adrenal glands) was done every 2 to 3 months as routine clinical practice and as needed to confirm the response and progression of disease. With measurable tumor sizes, treatment responses were defined as progressive disease, stable disease, partial response, and complete response according to the criteria of the Response Evaluation Criteria in Solid Tumors group (34). Patients with complete remission or partial remission were regarded as responders to gefitinib, whereas the others were regarded as nonresponders. Progression-free survival was calculated from the date of initiation of gefitinib treatment to the date of disease progression. For patients receiving gefitinib, survival duration was calculated from the start of gefitinib to the death of the patients. For patients who did not receive gefitinib, survival duration was calculated from the diagnosis of lung cancer to death. Patients who were not deceased were censored at the date of last contact in this institution.

Mutational analysis of EGFR. Tumor specimens, including paraffin blocks or frozen tissues of surgical specimens, fine-needle biopsies, and pleural effusions, were procured. These formalin-fixed and paraffin-embedded blocks were retrieved from the Department of Pathology, National Taiwan University Hospital. The mutational analysis of EGFR genes has been described previously (32). Briefly, DNA was derived from tumors embedded in paraffin blocks using a QIAmp DNA Mini kit (Qiagen). The tyrosine kinase domain of the EGFR coding sequence exons 18 to 21 were amplified, and independent PCR amplifications were purified and sequenced in an automatic ABI Prism 3700 DNA Analyzer.

The frozen lung cancer tissues were obtained at surgery, immediately snap frozen in liquid nitrogen, and stored until use. Total mRNA was extracted from resected cancer tissue using a RNA extraction kit (RNeasy Mini kit; Qiagen). The four exons (exon 18-21) that code for the tyrosine kinase domain of the EGFR gene were amplified with primers and PCR conditions as reported previously (35). Reverse transcription-PCR amplicons were purified and sequenced.

All sequencing reactions were done in both forward and reverse directions using tracings from at least two PCR.

Statistical analyses. All categorical variables were analyzed with χ² tests. Significance was determined if P < 0.05. All analyses were done using the SPSS software (version 13.0 for Windows; SPSS).

Characteristics of lung cancer patients. There were 4,737 NSCLC patients seen in the National Taiwan University Hospital between January 2000 and June 2007. Specimens from 515 (11%) patients were examined for mutations of the EGFR tyrosine kinase domain. They included 195 surgical specimens, 101 fine-needle biopsies (either echo guided, computer tomography guided, or bronchoscopic), and 219 cell block preparations of pleural effusion. There were 441 adenocarcinomas, 69 squamous carcinomas, and 5 small cell carcinomas. Of the total 515 patients, 201 were smokers, 50 former smokers, and 151 current smokers. Two hundred and fifty-three (49%) patients had EGFR mutations. The mutations were more frequent in never smokers than former or current smokers (74% versus 26%; P < 0.001), in adenocarcinomas than nonadenocarcinomas (96% versus 4%; P < 0.001), and in females than in males (61% versus 39%; P < 0.001). Two hundred and twenty-five (89%) patients had exon 19 and/or 21 mutations, and 13 patients had exon 18 mutations (5%).

We identified 23 lung cancer patients with EGFR exon 20 mutations, solely or with other mutations (Table 1). Nine (39%) had coexisting mutations in EGFR exons other than exon 20. Six patients (cases 2, 5, 7, and 11-13) were in a clinical trial of first-line gefitinib treatment (31), whereas the others were not in clinical trials. Clinical results on four patients had been reported previously (33, 36). The demographic information was as follows: 19 (83%) were female, the median age at which lung cancer was diagnosed was 65 years (ranging between 40 and 87 years), and 20 (87%) were never smokers with the other 3 being current smokers. Regarding the cell type of cancer, 22 patients had adenocarcinomas. The other one, who was a current smoker, had squamous carcinoma. Therefore, the exon 20 mutations predominantly occurred in females, never smokers, and adenocarcinoma, as with the other types of EGFR mutations. Half (52%) of the patients were stage IV at the diagnosis of lung cancer.

Gefitinib treatment on patients with EGFR exon 20 mutations. Medical records from all 23 patients were reviewed in detail. Sixteen of the 23 patients had received gefitinib treatment: 13 as first-line treatment, 1 as second-line treatment, and 2 as third-line treatment (Table 1). All of the gefitinib treatment courses were given as the sole therapy, without other concurrent chemotherapy agents or radiotherapy for the primary lung tumor. At the start of gefitinib treatment, 2 were stage IIIb and the other 14 were stage IV. Nine had point mutations in exon 20 and seven had insertion/duplication mutations in exon 20 as listed in Table 2. Eight (50%) had more than one mutation. Among the coexisting mutations other than exon 20, seven were in exon 21 (six L858R and one L861Q) and one in exon 18 (G719A).

Four (25%) patients (cases 2, 4, 6, and 12 in Table 2) reached partial response after gefitinib treatment, and 12 (75%) were nonresponders to gefitinib. For the 4 responders to gefitinib, the median progression-free survival was 79 weeks (range, 61-95) and the median survival duration after gefitinib was 23 months (range, 20-32). Three of these four patients had a coexisting mutation, L858R in exon 21, in addition to exon 20 mutations (Table 2).

In the 12 nonresponders, the median progression-free survival was 8 weeks (range, 1-17) and the median survival duration after gefitinib treatment was 7 months (range, 1-30 months). For the nonresponders, gefitinib was the only treatment in 2 patients. The other 10 patients had undergone a median of 3 (range, 1-5) regimens of chemotherapy other than gefitinib. However, 9 of these 10 patients were not responsive to any chemotherapy (stable disease or progressive disease as maximal response), with the exception of 1 patient (case 8) who reached a partial response to docetaxel as the third-line regimen.

Clinical courses of patients without gefitinib treatment. Seven patients with exon 20 mutations had never received gefitinib treatment (Table 2, cases 17-23). Three patients (cases 19, 20, and 23) received tumor resection in stage Ib of lung cancer. No recurrence was noted after the operations. Case 21 underwent operation for his stage IIIa tumor after neoadjuvant chemotherapy, but he expired 1 month later due to unrelated complications without evidence of cancer recurrence. The other three patients (cases 17, 18, and 22) received chemotherapy for cancer control. Case 17 achieved stable disease after first-line gemcitabine. The tumor progressed 6 months later, and he continued with further chemotherapy treatments. Cases 18 and 22 failed to respond to chemotherapy (gemcitabine plus cisplatin for case 18 and paclitaxel with carboplatin for case 22), and both of them expired due to cancer-related respiratory complications.

In the large-scale study conducted by Shigematsu et al. (20), 130 of the total 617 lung cancer patients had EGFR mutations distributed within exons 18 to 21. These mutations were associated with several clinical features: a higher incidence in never smokers compared with smokers, a higher incidence in adenocarcinoma compared with cancers of other cell types, and a higher percentage in female patients. The data in our study supported these findings. The 23 patients with exon 20 mutations in our study exhibited a similar predominance of females, never smokers, and adenocarcinomas. The similarity of these clinical features did not include the response to gefitinib, however.

As shown by cell lines and molecular studies, the mutant EGFR can transform fibroblasts and pulmonary epithelial cells without exogenous epidermal growth factors (3, 24). Gefitinib limits the growth of cells transformed via the mutant EGFR, which may explain the positive clinical responsiveness to gefitinib of lung cancer accompanied by mutant EGFR. For example, in the study of recurrent lung cancer after surgery (37), gefitinib was effective in 24 of 29 patients with EGFR mutations but only effective in 2 of 21 patients without mutations. Mutations were localized primarily to exons 19 and 21, with the exception of 3 patients who had mutations in exon 18. Furthermore, in another study of chemotherapy-naive lung cancer patients in stage IIIb and IV with EGFR mutations, 12 of 16 patients had objective response to gefitinib (16). Similarly, all the sensitive mutations were in exons 19 and 21.

Several studies on the gefitinib treatment response of tumors with EGFR exon 20 mutations have been published, albeit with relatively small patient numbers (Table 3). The effective treatment percentage in these studies ranged from 0% to 75% (8, 21, 22, 25). In our study, 4 of 16 patients reached a partial response after gefitinib treatment, with a significantly longer progression-free duration. Taken together, the investigations listed in Table 3 give an average 34% (11 of 32) gefitinib response rate, which is much lower than that for mutations in exons 19 and 21.

In the in vitro study conducted by Greulich et al., mutations in exon 20 (D770_N771insNPG) confer EGFR mutant cell resistance to gefitinib in contrast to mutations in other exons (29). These transformed cells with exon 20 mutations were sensitive to the irreversible inhibitor CL-387785, and development of alternative kinase inhibition strategies was suggested (6, 29, 38). This accounts for the lower gefitinib responsive rate in treating patients with exon 20 mutations. However, as mentioned earlier, there are still some patients with exon 20 mutations who respond to gefitinib treatment clinically. Clearly, the cellar and molecular studies do not completely explain the clinical outcomes observed.

What contributes to this disparity is not currently understood. One possibility is the diverse distribution of mutations in exon 20 (Table 4). Theoretically, mutations might be in-frame duplications, in-frame insertions, in-frame duplications plus insertions, or point mutations (11, 20, 21). Patients with point mutations may have different amino acid substitutions and may thereby have different responsiveness to gefitinib treatment. For example, both cases 3 and 4 in our study had two point mutations at identical amino acid locations. The mutation in exon 21 is the same, L858R, but mutations in exon 20 were replaced by different amino acids. Case 4, the responder to gefitinib, had mutation R776H, whereas case 3 had mutation R776G and did not respond to gefitinib. Another possible cause of different treatment outcomes is the coexistence of mutations in other exons of the EGFR tyrosine kinase domain. Combining our study and the reports in literature, there were 5 patients having mutation S768I who were treated with gefitinib. They had different coexisting EGFR mutations, however, and the responses to gefitinib were also different (responders had G719C or L858R, whereas nonresponders expressed G719A or V769L; refs. 9, 22). Whether these mutational differences explain gefitinib responsiveness is not clear, and further studies are still required for clarification.

Another interesting finding is the high prevalence of exon 20 mutations coexisting with other mutations (Table 4). In published reports, of the 32 patients treated with gefitinib, only 14 (44%) had single exon 20 mutations. In these, patients with a single exon 20 mutation had a lower gefitinib response rate (21%) than those with coexisting mutations (44%). Besides, exon 20 mutations presenting as insertions, duplications, or deletions are all single-site mutations and appeared in 10 patients. Only one of these was responsiveness to gefitinib, a much lower response rate than exon 20 point mutations. Another particular exon 20 mutation T790M led to unresponsiveness in all three patients who possessed it, although this was a point mutation coexisting with L858R (23, 31, 32). Overall, patients who had T790M or the insertion/duplication/deletion mutations in exon 20 seemed to have unfavorable response to gefitinib treatment.

The major limitation of this report is its retrospective nature. EGFR mutations in exon 20 are rare, and we extensively recruited reports in the literature to investigate the role of exon 20 mutations in the treatment of NSCLC patients. Some prospective trials are currently ongoing (46, 47), and we look forward to future clinical results on exon 20 mutations.

In conclusion, we reported a larger group of lung cancer patients with mutations in EGFR exon 20 than those previously published. Together with cases reported in literature, mutations in exon 20 were conducive to lower response rate to gefitinib treatment generally, but variability existed between cases. Further clinical and laboratory studies are required to improve the treatment of lung cancer with EGFR exon 20 mutations.

C.-H. Yang, C.-J. Yu, and J.-Y. Shih have received honoraria from AstraZeneca; C.-H. Yang has a research funding and is a consultant to AstraZeneca.

We thank Dr. Ron Freund for editing this article.