Abstract
Background/Aim: Gefitinib and erlotinib were shown to be particularly effective in a clinically selected subpopulation of non-small cell lung cancer patients (NSCLC): adenocarcinoma histology, non-smoking status, Asian origin and female gender have been associated with improved clinical benefit compared to the unselected NSCLC population. The aim of the present study was to investigate the prognostic and predictive role of EGFR and KRAS analysis in advanced lung adenocarcinomas, selected according to clinical features associated to better response to EGFR tyrosine kinase inhibitors (TKIs), namely female gender and non-smoker or former light smoker status. Patients and Methods: EGFR and KRAS mutations and EGFR FISH status were assessed in 67 surgical samples. Results: EGFR and KRAS mutations were found in 16 (26.7%) and 12 (17.9%) patients, respectively. FISH analysis was positive in 34 (56.7%) patients. EGFR-mutated patients showed significantly longer survival when treated with EGFR TKIs (p=0.002, hazard ratio (HR)=0.036, 95% confidence interval (CI): 0.004 -0.303). KRAS mutations was found to be an independent negative prognostic factor in multivariate analysis (p=0.008, HR=3.52, 95%CI: 1.39-8.9). The prognostic value of EGFR FISH status was not confirmed in multivariate analysis (p=0.048, HR=0.47, 95%CI: 0.22-0.99). Conclusion: In a group of clinically selected patients, EGFR and KRAS analysis was able to define distinct molecular subsets of lung adenocarcinoma.
Lung cancer is a major cause of death from malignant diseases in the Western world. Despite therapeutic improvements, prognosis for advanced stage disease remains poor. Chemotherapy seems to have reached a plateau in terms of benefit and it is not likely to give further survival or disease-control advantages. The term ‘non small-cell lung cancer’ (NSCLC) refers to a heterogeneous group of diseases both from the clinical and the molecular point of view (1). Furthermore, the currently available clinical and pathological features are not sufficient for a satisfactory prediction of outcome. The study of molecular pathways involved in lung cancer development and progression has allowed the planning of new therapeutic strategies and may help in individuating NSCLC subgroups with different clinical behaviours.
Epidermal growth factor receptor (EGFR or ErbB-1) is a tyrosine kinase receptor belonging to the ErbB family; it is widely expressed in NSCLC (40-80% of NSCLC) and plays a significant role in carcinogenesis. It has emerged as one of the most relevant targets for cancer treatment.
Gefitinib and erlotinib are small-molecule reversible tyrosine kinase inhibitors (TKIs) that selectively target EGFR. They have a good toxicity profile and have obtained a 10-20% response rate in previously chemotherapy-treated Caucasian patients (2, 3). Erlotinib has been associated with survival benefit compared to best supportive care in second- or third-line therapy in a phase III trial (4, 5) and, more recently, gefitinib has demonstrated to be more effective than chemotherapy as a first-line treatment in a selected group of NSCLC patients (6, 7).
Some clinical and pathological features have been associated with a higher probability to benefit from TKIs: Asian origin, female gender, adenocarcinoma histology and non-smoking status. The presence of somatic mutations in the tyrosine kinase domain of EGFR (exon 18-21) involving the ATP-binding pocket of the receptor was shown to represent the most important predictive marker (6, 8-10). EGFR mutations, which in Caucasian populations are present in about 10-20% of NSCLCs (11), have been accepted as a criterion for selecting patients for gefitinib first-line treatment. Other markers have been suggested to characterize a subset of NSCLCs, but their role is still under evaluation. According to some authors, EGFR gene copy number may be associated with longer disease control in TKI-treated patients (12, 13). In contrast, somatic mutations in exon 2 (codon 12-13) of KRAS, encoding a GTPase downstream of EGFR, have been associated with primary resistance to EGFR inhibitors. EGFR and KRAS mutations seem to be mutually exclusive, which is consistent with the idea that different alterations of the same pathway are involved in lung carcinogenesis (14, 15).
The present study retrospectively investigated the prognostic and predictive role of a panel of molecular markers in a clinically selected population of advanced lung adenocarcinomas.
Patients and Methods
Patients. The study group was a cohort of 67 patients with advanced stage lung adenocarcinoma who underwent therapeutic or diagnostic surgery at the Department of Thoracic Surgery of the University of Padua from June 2005 to April 2009. Surgical samples were obtained after patients signed a written informed consent. All the patients were selected according to the presence of at least one of the following clinical features associated with better response to TKIs: non-smoker (having smoked fewer than 100 cigarettes in a lifetime) or former light smoker (having smoked fewer than 10 packs per year and having stopped smoking at least fifteen years before the sample collection) status and female gender. Staging was defined according to the 6th edition of the TNM classification (16). All patients had ECOG performance status 0 or 1.
A total of 20 patients underwent surgery with diagnostic intent. In 47 cases, surgery had both a diagnostic and a therapeutic purpose; for 10 of them, it resulted in an exploratory thoracotomy only. After treatment, all the patients were followed according to a regular follow-up visit schedule at the Department of Thoracic Surgery of the University of Padua or the Department of Medical Oncology of the Istituto Oncologico Veneto of Padua.
Molecular analysis. All paraffin-embedded samples were evaluated by a pathologist in order to assess the tumour tissue quality and quantity. Molecular analysis was performed when the tumour tissue comprised more than 50% of the entire sample. DNA was extracted using a QIAmp DNA kit (Qiagen, Milan, Italy), according to the manufacturer's instructions. Agarose gel electrophoresis was used to control its quality and quantification was performed by spectrophotometry.
A total of 100 ng of genomic DNA was used for polymerase chain reaction (PCR) to amplify exons 18, 19, 20, 21 of EGFR gene and exon 2 of KRAS gene by using primers, sequences and amplification conditions according to previous studies (8, 17). PCR products were purified using a pre-sequencing kit (Amersham Biosciences, Little Chalfont Buckinghamshire, UK). Subsequently, they were sequenced with both forward and reverse primers using BigDye Terminator v1.1. Cycle Sequencing Kit (Applied Biosystems, Foster City, CA, USA). Automated sequencing was carried out using an ABI 3130xl DNA sequence (Applied Biosystems).
Fluorescence in situ hybridisation (FISH) was performed on 4 μm paraffin-embedded sections after deparaffinisation in xylene and dehydration on 100% ethanol. Sections were pretreated on microwaves with 0.01M citrate buffer (pH 6.0) and digested with proteinase K (25 μg/ml). Each section was hybridised with LSI EGFR/CEP7 (Vysis/Abbott, Downers Grove, IL, USA), labelled with SpectrumOrange and SpectrumGreen respectively after simultaneous denaturation using Hybrite system (Vysis). Post-hybridisation washings were performed according to standard protocols for LSI probes. Analyses were carried out with a fluorescence Zeiss microscope Axioplan 2 plus (Zeiss, Göttingen, Germany) equipped with a single-band filter, specific for the fluorochromes used. Two observers independently scored 50 tumour cells, each in at least four tumour areas. The interpretation of the EGFR FISH results was made according to the Colorado scoring criteria (18).
All molecular analyses were performed in a blinded fashion without access to the patient clinical characteristics or treatment outcome.
Statistical analysis. Overall survival (OS) was calculated from diagnosis until death from any cause or last follow-up visit. Progression-free survival (PFS) was calculated from the beginning of the treatment to documented disease progression or death. Categorical data were analysed by the chi-square test. The Kaplan-Meier method was used to estimate survival rates and comparative analyses were carried out by log-rank test. All p-values were two-sided and a p-value<0.05 was considered statistically significant. The statistical analyses were performed using GraphPad Prism 5 Software (GraphPad, La Jolla, CA, USA). The Cox proportional hazard model was used for the multivariate analysis performed in SPSS version 14.0 (SPSS, Inc., Chicago, IL, USA).
Results
Study population. Demographic and clinical data are shown in Table I. Almost half of the patients were women; 57% of patients had never smoked and 22% had stopped smoking at least fifteen years before the collection of the sample. The median age was 64 years. Seventeen patients had received chemotherapy before the sample collection, for six patients as preoperative and for eleven patients as palliative therapy. Bronchioloalveolar carcinoma features (BAC) were present in 16 cases (23.9%); in 11 of them this component was more than 80%.
Chemotherapy was the first-line therapy in 45 cases; six patients had been enrolled in clinical trials and received erlotinib as first-line therapy, three of them faced second-line chemotherapy. In the whole group, second-line therapy was chemotherapy for 16 patients and erlotinib for 11 of them. As third-line treatment, eight patients received chemotherapy and three were treated with erlotinib.
Molecular analysis. EGFR mutation analysis was performed in 60 cases out of 67 (89.6%). In seven cases, poor sample quality precluded further analysis. Tyrosine kinase domain mutations were found in 16 (26.7%) cases. Patients with positive tests were either never smokers (81.3%) or former light smokers (18.7%). Never smoker status was the only feature able to predict the presence of EGFR mutations (p=0.04). Twelve of these genomic alterations involved exon 19 (75%) and consisted of deletions or deletions/insertions resulting in the loss of four amino acids (LREA), located next to K745 residue, thought to be essential for ATP-binding. Two cases showed a missense mutation in exon 21 affecting the activation loop of the receptor (L858R). An exon 20 insertion was found in two samples. This genetic alteration is thought to be associated with a worse response to TKIs, compared to alterations of exons 18, 19 and 21. No exon 20 T790M mutation was found, but no patients received TKI treatment before molecular analysis (Table II).
FISH analysis was performed successfully in 60 cases (89.6%). Thirty-four patients showed increased gene copy number (56.7%): seven tumour samples showed EGFR gene amplification (11.7%), 27 carried high polysomy (≥4 copies of EGFR gene in ≥40% of the cells) (45%) (Table II). Only five out of 27 patients showed high polysomy in less than 60% of the cells. Ten patients carried both EGFR mutations and gain of gene copy number. Neither clinical features nor histological differentiation grading were found to be correlated with FISH-positive status.
KRAS mutations in codon 12 or 13 were found in 12 patients out of 62 analysed (19.4% of the screened population). In the study group, KRAS mutations were not significantly associated with male gender or smoking status. Six KRAS-mutated patients were also FISH positive, whereas only one EGFR-mutated patient carried KRAS mutation.
Survival. The median follow-up period was 97 weeks and the median OS was 99 weeks (Figure 1A). Such a result would have been impressive if it referred to an unselected advanced NSCLC population. In fact, it may have been influenced by clinical selection, special staging subgroups or metastatic sites (e.g. single ipsi- or contralateral lung metastasis, single brain metastasis radically resected or treated with radiosurgery). The percentage of study patients with these features was far greater than that encountered in daily clinical practice because the study group was represented by patients undergoing a surgical procedure (data not shown).
The OS of 67 advanced NSCLC patients was studied in reference to EGFR and KRAS status. Prognosis of patients with EGFR mutation was not significantly different (Figure 1B; median OS: 99 weeks vs. 92 weeks, p=0.87). On the contrary, better survival was noticed in the FISH-positive group, regardless of treatment (Figure 1C; median OS: 177 weeks vs. 57 weeks, p=0.048, hazard ratio (HR)=0.47, 95% confidence interval (CI): 0.22-0.99).
The presence of KRAS mutation in this group of patients appeared to be correlated with worsening outcome. Patients with KRAS mutation had significantly poorer PFS (data not shown; median PFS: 11 weeks vs. 28 weeks, p=0.001, HR=4.42, 95% CI: 1.31-14.90) and OS was also significantly different (Figure 1D; median OS: 44 weeks vs. 125, p=0.03, HR= 3.1, 95% CI: 1.15-8.4).
Multivariate analysis confirmed that only KRAS status was an independent prognostic marker (p=0.008, HR=3.52, 95% CI: 1.39-8.9). Tumour stage, gender, age and smoking status were also tested for their influence on the outcome, but none of them was shown to affect survival significantly in the study population.
Predictive role of EGFR genomic alterations. To assess the predictive role of mutational and FISH status in the study population, OS was analysed in relation to the presence of genomic alterations and treatment with erlotinib. The group of patients with mutated samples had a better outcome when treated with the TKI in comparison to patients with mutated samples who never received this kind of treatment (Figure 2A; median survival: 177 weeks vs. 51 weeks, p=0.002, HR=0.03, 95% CI: 0.004-0.3066). In wild-type EGFR patients, no significant difference in survival was found according to erlotinib treatment (Figure 2B). On the contrary, EGFR FISH status did not appear to affect survival of the erlotinib-treated patients. No clinical feature was significantly associated with better outcome in the erlotinib-treated group. The predictive role of KRAS mutation was not verified, as positive patients were never treated with TKIs in the study population.
Since patients received more than one line of treatment, PFS after first-line treatment was also investigated. Patients with EGFR mutations or increased gene copy number receiving TKI treatment showed longer PFS (median PFS: 57 weeks) in comparison to unselected patients treated with chemotherapy (median PFS: 28 weeks) (Figure 2C; p=0.045, HR=0.4, 95% CI: 0.16-0.98).
Discussion
This study examined both EGFR and KRAS mutations, and EGFR gene copy number in a clinically selected population of patients with advanced stage lung adenocarcinoma to evaluate their prognostic and predictive implications. The role of those molecular markers was investigated in a population with clinical features suggesting better response to TKIs. The whole population showed better clinical outcome than a usual advanced stage NSCLC population, probably because only patients eligible for a surgical manoeuvre were considered in the study. Nevertheless, molecular markers were able to identify patient subgroups with different clinical behaviour.
In the study population, both EGFR mutations and gain of gene copy number were found in a higher percentage of patients, compared to an unselected NSCLC population, which indirectly confirms previously published data about clinical-pathological features predicting EGFR alterations (2, 5, 19-20).
The present study confirmed the predictive role of EGFR tyrosine kinase domain mutations. For the overall survival analysis, erlotinib-treated patients were considered as a unique group, since most patients received two or more lines of therapy and Rosell et al. demonstrated no difference in survival between patients with mutations receiving gefitinib in first- or second-line treatment (11). Despite the relatively small size of the sample, OS was affected by TKI treatment in the EGFR-mutated group with statistical significance.
In the study population, EGFR mutations were not correlated with better outcome or better response to chemotherapy. Previous retrospective studies demonstrated prolonged survival for patients with EGFR mutation irrespective of treatment (21-23). The results of the present study suggest that in a clinical homogeneous population, EGFR mutation is not an independent prognostic factor, as shown in recent studies in surgically resected lung adenocarcinomas (24, 25).
Gain of EGFR gene copy number was found in approximately half of the screened patients (56.7%). OS of erlotinib-treated patients was not affected by FISH status. A recent meta-analysis study showed notable heterogeneity both in positivity range (8.1%-61.1%) and in sensitivity and specificity for predicting response to TKIs (26). Despite the possible inadequacy of response rate as a clinical outcome, the study concluded that FISH analysis clearly has less predictive accuracy compared to EGFR mutation analysis. Moreover, in the present study 17.6% of FISH-positive patients also carried KRAS mutations. Their negative impact on clinical outcome may have influenced the overall survival results, overcoming the predictive role of FISH analysis. Evidently, new larger studies are required to investigate the clinical behaviour of EGFR FISH-positive, KRAS-negative patients treated with EGFR TKIs.
Available data about the prognostic role of EGFR FISH status are still controversial and they do not support the hypothesis that EGFR copy number represents a prognostic factor (27-29). However, in the study population, FISH-positive patients showed better outcome irrespectively of treatment.
The impact of KRAS alterations on prognosis has already been studied in lung cancer. KRAS alterations are more frequent in current or former smokers than in those who have never smoked (25% vs. 6%) and in adenocarcinomas compared to other histologies (26% vs. 16%) (29). Two different meta-analyses suggested an association of KRAS with poorer survival in lung cancer (30, 31). The prognostic role, was limited to adenocarcinoma patients (31), supporting the idea of different biological signalling pathways according to histological classification. No sufficient confirmation has yet been found in well-designed multivariate analyses. More recent studies have investigated the prognostic role of KRAS in surgically treated NSCLC. In a Japanese study, KRAS did not have a statistically significant prognostic role (25). Marks et al. studied the prognostic value of EGFR and KRAS mutations and outlined two distinct lung adenocarcinoma groups with different clinical behaviour: patients with EGFR mutation patients showed better outcome compared to those with KRAS mutation (24). The present study confirmed the impact of KRAS mutations on prognosis in a clinical selected group of advanced adenocarcinomas, highlighting the role of KRAS as an independent prognostic marker.
In the present study, 19.4% of the screened population carried KRAS mutations and this molecular marker was significantly associated with worse prognosis, in terms of PFS and OS. It represents a marker able to find patients with worse prognosis out of a clinically homogeneous population.
A previous retrospective study analysed the role of the three molecular markers in advanced NSCLC (32). In that study, KRAS mutations were able to overcome the predictive value of FISH. Moreover, the percentage of KRAS-positive patients carrying also gain of EGFR gene copy number (8/14, 57%) was similar to the present results (6/12, 50%). KRAS mutations are usually mutually exclusive to EGFR mutations (33, 34) with very rare tumours containing both genes mutated (35). Consequently, they have a limited role as predictive markers, but the confirmation of the independent prognostic role for KRAS supports the idea of distinct biological features and should be further investigated prospectively in relationship to different treatment regimens.
The results of the present study should be confirmed in a larger group of patients and in early stage disease. In the perspective of personalised treatment for lung cancer, further studies should focus their attention on with KRAS mutation patients demonstrating the limits of the currently available treatments, in order to develop newly designed target therapies (36).
Conclusion
In a clinically selected population of lung adenocarcinomas, analysis of molecular markers is able to further characterize tumours, providing information about prognosis and response to EGFR TKIs. The predictive role of EGFR mutations was confirmed, whereas EGFR copy number did not affect OS of EGFR TKI-treated patients. KRAS exon 2 mutations were demonstrated to be an independent prognostic marker in advanced lung adenocarcinoma, whereas the prognostic role of EGFR gene copy number was not confirmed in the multivariate analysis. EGFR tyrosine kinase domain mutations are not associated with different outcome irrespective of treatment.
Acknowledgements
We acknowledge Martina Boscaro for assistance in writing and reviewing the manuscript.
Footnotes
-
↵* Both authors contributed equally to this work.
- Received September 1, 2010.
- Revision received October 29, 2010.
- Accepted November 1, 2010.
- Copyright© 2010 International Institute of Anticancer Research (Dr. John G. Delinassios), All rights reserved