|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
Originally published as JCO Early Release 10.1200/JCO.2005.03.8224 on March 20 2006 © 2006 American Society of Clinical Oncology. Expression Profile–Defined Classification of Lung Adenocarcinoma Shows Close Relationship With Underlying Major Genetic Changes and Clinicopathologic Behaviors
From the Division of Molecular Carcinogenesis, Center for Neurological Diseases and Cancer, Nagoya University Graduate School of Medicine; Department of Pathology and Molecular Diagnostics, Aichi Cancer Center Hospital; Department of Thoracic Surgery, Aichi Cancer Center Hospital; Division of Molecular Oncology, Aichi Cancer Center Research Institute, Nagoya, Japan; and the Japan Biological Informatics Consortium, Tokyo, Japan. Address reprint requests to Takashi Takahashi, MD, PhD, Division of Molecular Carcinogenesis, Center for Neurological Diseases and Cancer, Nagoya University Graduate School of Medicine, 65 Tsurumai-cho, Showa-ku, Nagoya 466-8550, Japan; e-mail: tak{at}med.nagoya-u.ac.jp
PURPOSE: This study was conducted to gain insight into the relationship between expression profiles and underlying genetic changes, which are known to be important for the pathogenesis of lung cancers. METHODS: Expression profiles of 18,175 unique genes and three major targets for genetic changes, p53, epidermal growth factor receptor (EGFR), and K-ras, were investigated in 149 patients with non–small-cell lung cancer, including 90 patients with adenocarcinoma to determine their relationships with various clinicopathologic features and Gene Ontology (GO) terms. RESULTS: This study successfully established a basis for expression profile-defined classification, which can classify adenocarcinomas into two major types, terminal respiratory unit (TRU) type and non–TRU type. Our GO term–based identifier of particular biologic processes, molecular functions, and cellular compartments clearly showed characteristic retention of normal peripheral lung features in TRU type, in sharp contrast to the significant association of non–TRU type with cell cycling and proliferation-related features. While significantly higher frequency of EGFR mutation was observed in TRU type, we found that the presence of EGFR mutations was a significant predictor of shorter postoperative survival for TRU type, independent of disease stage. We were also able to identify a set of genes in vivo with significant upregulation in the presence of EGFR mutations. CONCLUSION: This study has shed light on heterogeneity in lung cancers, especially in adenocarcinomas, by establishing a molecularly, genetically, and clinically relevant, expression profile-defined classification. Future studies using independent patient cohorts are warranted to confirm the prognostic significance of EGFR mutations in TRU-type adenocarcinoma.
Lung cancer is the leading cause of cancer-related death in developed countries.1,2 While the current classification of neoplasia is largely based on the histologic features observed under the microscope, marked variations in clinical behavior are sometimes evident even within a particular histologic type, and adenocarcinomas are known to exhibit the highest degree of morphologic and clinical diversities.3 Thus, more detailed, accurate, and objective means to classify non–small-cell lung cancer (NSCLC) tumors, especially adenocarcinomas, are greatly anticipated not only for a better understanding of the pathogenesis, but also for improving our currently inadequate diagnostic capabilities to help develop more effective treatment modalities. The recent development of microarray technologies has made it possible to correlate gene expression profiles in individual cases with various clinical parameters.4-9 To date, various groups including our own have reported that expression profiling can recapitulate morphologic classification of NSCLCs, and some studies also showed that adenocarcinomas can be subclassified additionally.10-13 However, these previously reported subclassifications vary considerably from study to study, making it difficult to reconcile their findings or reach any definite conclusions. It should also be noted that previous expression profiling studies provided little information with regard to the relationship of expression profiles with underlying genetic changes, which are known to be important for the pathogenesis of lung cancers.14 The identification of activating mutations of epidermal growth factor receptor (EGFR) is one of the most intriguing recent discoveries in the field of lung cancer research.15,16 EGFR mutations are present in a subset of pulmonary adenocarcinomas,15-20 and tumors with this mutation have been shown to be highly sensitive to gefitinib.15-17,19 Good clinical response to gefitinib has been observed most frequently in female, Japanese and other Asian ethnicities, nonsmoking patients with adenocarcinomas.21,22 We have also reported that EGFR mutations are more prevalent in the terminal respiratory unit (TRU) -type adenocarcinomas,23 which we have proposed as a characteristic subset, representing adenocarcinomas with the peripheral airway epithelium as their putative origin.23,24 In this study, we concurrently analyzed global expression profiles and EGFR, p53, and K-ras mutation status in 149 patients with NSCLC, including 90 patients with adenocarcinoma, in order to establish a both genetically and clinicopathologically relevant expression profile-defined classification. We used this classification to identify the significantly higher prevalence and clear prognostic impact of EGFR mutations in one of the two major subtypes identified in adenocarcinoma.
Patients A series of 149 patients with NSCLC, comprising 90 patients with adenocarcinomas, 35 patients with squamous cell carcinomas, 18 patients with large-cell carcinomas, four patients with adenosquamous carcinomas, and two patients with large cell neuroendocrine carcinomas, who successfully underwent potential curative resections between December 1995 and December 1999, were obtained from a file at the Department of Pathology and Molecular Diagnostics, Aichi Cancer Center, Nagoya, Japan. For the postoperative survival analysis, 82 of the patients with adenocarcinoma met the criteria for inclusion, while the remaining eight patients were excluded because of postoperative treatment with gefitinib. The median follow-up period was 77 months (range, 6 to 108 months) when all the eligible patients were included and 92 months (range, 64 to 108 months) when deceased patients were excluded. All the tumor specimens were embedded in OCT compound (Sakura Finetechnical Co Ltd, Tokyo, Japan) and stored at –80°C after the requisite approval from the institutional review board and patients' written informed consent had been obtained.
Acquisition of Expression Profiles
Bioinformatic Analysis
Mutation Analysis of EGFR, p53, and K-ras Genes
Gene Ontology Term–Based Identifier and Other Statistical Analysis
The
Expression Profiled-Defined Classification of NSCLCs We first used unsupervised hierarchical clustering to classify all 149 samples using the 4,834 most variably expressed transcripts in order to attain a molecular classification based on the similarity of genome-wide expression patterns in individual tumors. The resultant clusters accurately recapitulated the well-established, widely used histologic classification of NSCLC, while the large cell cluster was considerably mixed with other histologic types (Fig 1). SAM analysis disclosed the presence of distinct sets of genes with up- or downregulation for each of the clusters, and the gene sets were found to be similar to those previously reported by us13 and others10,11 (data not shown). Expression profiling data were confirmed by both real-time reverse transcriptase polymerase chain reaction and immunohistochemical analyses.
In this study, we concurrently performed extensive searches for mutations in the EGFR, p53, and K-ras genes, which are strongly believed to contribute to the development of lung cancers, in order to investigate whether the expression profile-defined subtypes have any relationship to the presence or absence of these genetic changes. As was expected from the clearly accurate recapitulation of the conventional histologic classification of NSCLC and previous reports on these three gene alterations, highly significant associations were observed between EGFR mutations and the corresponding adenocarcinoma branches (36% v 0%; P < .001) as well as between p53 mutations and the branch consisting mostly of squamous cell carcinomas and large cell carcinomas (67% vs 33%; P < .001).
Expression Profile-Defined Two Major Types of Adenocarcinomas
To gain additional insight into the molecular and biologic nature of these two major expression profile-defined adenocarcinoma subtypes (ie, non-TRU and TRU types), SAM analysis was performed first to select differentially expressed genes. A total of 1,657 genes passed prefiltering at a significance level of < 0.1% false-discovery rates in the SAM analysis, and 293 of these genes showed differences by a factor of more than 2 between their expression levels in TRU- and non-TRU types. These 293 genes consisted of 201 with higher expression in the TRU type and 92 with higher expression in the non-TRU type. In order to better understand the underlying functional distinctions between TRU and non–TRU types, the SAM-identified gene sets were subjected to our GO term identifier of differentially utilized functions and other characteristics, which was developed in our laboratory for this study. With this identifier, GO terms related to nine biologic processes, six molecular functions, and two cellular components were extracted as occurring significantly more frequently in TRU-type adenocarcinomas, exhibiting clear relation to normal lung functions (Table 1). In contrast, GO terms identified as those specific to non-TRU type distinctively included those related to cell cycling and proliferation, suggesting an inherently aggressive nature of non–TRU type adenocarcinomas.
The identification of these highly robust expression profile-defined subtypes of adenocarcinomas led us to conduct a similar unsupervised hierarchical clustering analysis using an independent data set. The Stanford data set consisting of 34 patients with lung adenocarcinoma10 was analyzed by means of unsupervised hierarchical clustering based on the expression profiles of the 30 top-ranked genes in terms of distinctive expression in the non-TRU, TRU-a, or TRU-b types. This resulted in the clear visualization of two major branches with a subcluster, which appeared to correspond well to the three expression profile-defined adenocarcinoma subtypes identified in this study (Fig 3).
Significant Association of Expression Profile-Defined Adenocarcinoma Subtypes With Clinicopathologic Characteristics We examined the relationship between various clinicopathologic features and the two expression profile-defined adenocarcinoma subtypes (ie, TRU and non-TRU adenocarcinomas; Table 2 ). TRU-type adenocarcinomas were seen significantly more frequently than non-TRU types in females (P = .005) and never-smokers (P < .001). In contrast, a detailed microscopic examination showed that invasive growth and the presence of necrosis, indicative of higher malignant potential/appearance and a rapidly proliferating nature, were characteristically more prevalent in non-TRU types (P < .001 for both invasive growth and necrosis), in accordance with the predominance of proliferation and cell cycling-related GO terms distinctive for non–TRU type tumors. Multivariate logistic regression analysis with age, sex, smoking status, and pathologic stage as variables identified never-smoking status as the only significantly associated variable (P = .001). As for postoperative prognosis, we noted that patients belonging to the two apparent clusters under the TRU-type adenocarcinoma branch seen in Figure 2A, namely TRU-a and TRU-b, were distinct in terms of their postoperative prognoses. While TRU-a type had a prognosis similar to that of non–TRU type, the prognosis for TRU-b type adenocarcinomas was significantly better than for non–TRU type (P = .021; Fig 4). This seemed to be consistent with the fact that microscopic examination showed that apparent invasive growth occurred much less frequently in TRU-b type adenocarcinomas than in the other types of adenocarcinomas (P < .001).
Significant Association of Expression Profile-Defined Adenocarcinoma Subtypes With EGFR Mutation Status We examined whether any of the three major genetic alterations in NSCLCs were significantly associated with the expression profile-defined adenocarcinoma subtypes (Fig 2B). The presence of EGFR mutations was found to be significantly more prevalent in TRU-type adenocarcinomas than in non–TRU type adenocarcinomas (45.3% v 21.6%; P = .026). We also noted that the two apparent clusters (ie, TRU-a and TRU-b) under the TRU-type adenocarcinomas branch modestly differed in the prevalence of EGFR mutations, with a higher EGFR mutation frequency for TRU-b type (52.6%) than for TRU-a type (41.2%) adenocarcinomas. In contrast, K-ras and p53 did not correlate with the subtypes of adenocarcinomas (P = .33 for p53; P = .17 for K-ras), although interesting inverse correlations of mutation frequencies were observed between these genetic alterations and EGFR mutations. Non–TRU type adenocarcinomas contained the highest percentage of tumors carrying p53 and/or K-ras mutations (41% for p53; 16% for K-ras), followed by TRU-a type (29% and 12%, respectively) and TRU-b type (21% and 0%, respectively).
Prognostic Significance of EGFR Mutations in TRU-Type Adenocarcinoma
Search for Genes With Significant Differential Expression in the Presence or Absence of EGFR Mutations We used a significance level of 5% false-discovery rate for selecting genes associated with the presence of EGFR mutations in all adenocarcinoma cases, since none were selected when a 0.1% false discovery rate was used (Table 4). Five genes were identified with more than two fold upregulation in tumors with EGFR mutations, while additional 11 genes showed upregulation of 1.5 to two times in association with the presence of EGFR mutations. We also searched for the genes differentially expressed specifically within TRU-type adenocarcinomas with significantly higher prevalence of EGFR mutations. Although it was necessary to use a quite high false-positive discovery rate (10%) for selecting genes by SAM, a single gene, GGTLA4, was identified with more than two fold upregulation in the presence of EGFR mutations. In addition, five genes showed upregulation of 1.5 to two times at the same level of significance (Table 4).
Expression profiles in a given tumor can be regarded as the outcome of complex influences resulting from the accumulated genetic changes important for the pathogenesis as well as from differentiation commitment of the progenitor cells. In this comprehensive study, we were able to establish an expression profile-defined, highly robust classification of adenocarcinomas, the relevance of which is clearly supported by the inclusion of various molecular genetic and clinicopathologic distinctions. The two major subtypes (ie, TRU and non-TRU types) feature differential expressions of a large number of genes, thus indicating their significant difference in terms of gene usage.23 The results obtained by using our GO term–based identifier support their different nature additionally. TRU-type tumors are characterized by biologic processes, which are important for the maintenance of peripheral lung functions.29 Similarly, molecular functions related to TRU type also appear to reflect retention of their progenitor cells' characteristics. In contrast, most non–TRU type associated GO terms are related to cell cycling and cellular proliferation, which appear to be consistent with their microscopic appearance of high-grade characteristics. The marked distinctions in clinical features between TRU and non-TRU subtypes also support the robustness of the present expression profile-defined classification. Notable clinical features of TRU type are significantly higher proportions of females and nonsmokers, whereas the result of multivariate analysis indicates nonsmoker status, but not sex as an independently associated factor. Thus, this tumor type appears to arise from peripheral lung airway cells under much less influence of smoking and to retain its progenitor's characteristics as indicated by the GO term-based analysis. TRU-b type had the most favorable prognosis with less frequent invasive growth and expressed various differentiation markers at higher levels even when compared with TRU-a type, suggesting that TRU-b type retains features of normal peripheral lung airway cells better than the other types, but may progress to the TRU-a type. Our study clearly shows that the presence of EGFR mutations is significantly associated with TRU type adenocarcinoma. In fact, 45.3% of TRU-type adenocarcinomas carried EGFR mutations in contrast to a 21.6% occurrence in non-TRU type. In contrast, p53 and K-ras did not show significant differences in terms of their mutation frequencies, but there was an interesting inverse propensity with the highest occurrence in non-TRU type. It should also be noted that the presence of EGFR mutations was significantly associated with shortened postoperative survival, specifically for TRU-type adenocarcinoma, independently of disease stage. Indeed, 5-year survival rates of patients with stage II/III disease of TRU type were found to be 25% in the presence of EGFR mutations and 78% in their absence. These findings suggest the potential clinical usefulness of concurrent analysis of expression profile-defined subtypes and EGFR mutation status to select candidates for intensive adjuvant therapy, possibly with gefitinib. In conclusion, this comprehensive study has shed light on the existence of heterogeneity in lung cancers, especially adenocarcinomas, by establishing a genetically and clinicopathologically relevant, expression profile-defined molecular classification. Additional studies using independent patient cohorts will be necessary before any definitive conclusion can be reached with regard to the prognostic significance of the presence of EGFR mutations in TRU-type adenocarcinoma.
Microarray data preprocessing. Fluorescence intensities of scanned images were quantified and normalized using Feature Extraction 7.5 software (Agilent Technologies, Palo Alto, CA). After the LOWESS dye normalization was calculated for each feature, the transcript abundance of a gene was defined as the ratio of its intensity compared with that of the signal of the reference RNA obtained from a pooled mixture of 20 lung cancer cell lines. Genes that were flagged in more than 10 samples were excluded from further analysis. The following three kinds of flags provided by the Feature Extraction software were used: IsSaturated, UnifOL and WellabobeBG. If 50% of inlier pixels show intensity values above the saturation threshold of 65,502, the feature is flagged as "IsSaturated." The nonuniformity outlier algorithm flags anomalous features and local backgrounds based on statistical deviations from the Agilent noise model. The feature background-subtracted signal was then compared with the noise of its background, and defined as "well above the background" when the signal was stronger than 2.6 background standard deviation. In addition, genes with expression levels varying by a factor of less than three across the sample set of interest were eliminated because they were unlikely to be informative. Hierarchical clustering. Cluster (Eisen MB, Spellman PT, Brown PO, et al. Proc Natl Acad Sci U S A 95:14863-14868, 1998) was used for average linkage clustering of both genes and array by means of median centering and normalization, and TreeView (Eisen MB, Spellman PT, Brown PO, et al. Proc Natl Acad Sci U S A 95:14863-14868, 1998) was used to display the results. Statistical analysis for microarray data. For the ranking and selection of genes differently expressed in the groups that were compared, we used significance analysis of microarray (Tusher VG, Tibshirani R, Chu G. Proc Natl Acad Sci U S A 98:5116-5121, 2001). In this analysis, each of the patient subtypes was treated as unpaired, two-class data (ie, squamous cell plus large-cell carcinoma [SQ-LA] subtype v adenocarcinoma [AD] subtype, or TRU subtype v non-TRU subtype). All data are shown as log scale, the number of permutations was set at 1,000, and "Random Number Seed" was generated each time. The significance levels (ie, false-discovery rate), used for each of the analyses is available in Results. Gene Ontology terms analysis. Gene Ontology (GO; Ashburner M, Ball CA, Blake JA, et al. Nat Genet 25:25-29, 2000) analysis was employed to highlight functionally distinct biologic features of the gene set specific for each of the patient subtypes. Three compressed database files, such as Hs.seq.all.gz, Hs.data.gz, and LL_tmpl.gz, were downloaded from the UniGene FTP site (ftp://ftp.ncbi.nih.gov/repository/UniGene/Homo_sapiens/ and ftp://ftp.ncbi.nih.gov/refseq/LocusLink/) on January 20, 2005. First, the Hs.seq.all database file was used to link the GenBank accession number corresponding to each of the spots on the microarray to its corresponding UniGene ID. Next, the Hs.data database file was used to link the UniGene IDs to their corresponding Locus IDs. Finally, the GO terms corresponding to the Locus IDs were parsed by means of the LL_tmpl database file. All these procedures were automatically processed with the aid of our newly developed program written with Perl. Eventually, 12,745 known genes among the 18,175 unique genes on the microarray chip could be assigned to GO terms, and linked to about 67,000 of these terms. In addition, 139 of 201 TRU-defining genes, and 70 of 92 non–TRU-defining genes could be assigned to at least one GO term. The frequency of a GO term appearing in either a TRU or a non-TRU gene set was compared with that in the entire set of 12,745 genes and analyzed by means of Fisher’s exact test to identify which GO terms were significantly over- or under-represented in a gene set of interest.
Although all authors completed the disclosure declaration, the following author or immediate family members indicated a financial interest. No conflict exists for drugs or devices used in a study if they are not being evaluated as part of the investigation. For a detailed description of the disclosure categories, or for more information about ASCO’s conflict of interest policy, please refer to the Author Disclosure Declaration and the Disclosures of Potential Conflicts of Interest section in Information for Contributors.
Dollar Amount Codes (A) < $10,000 (B) $10,000-99,999 (C)
EGFR (epidermal growth factor receptor): Also known as HER-1, EGFR belongs to a family of receptors (HER-2, HER-3, HER-4 are other members of the family) and binds to the EGF, TGF-  , and other related proteins, leading to the generation of proliferative and survival signals within the cell. It also belongs to the larger family of tyrosine kinase receptors and is generally overexpressed in several solid tumors of epithelial origin. TRU (terminal respiratory unit): A physioanatomical unit of the terminal airway in the lung that is conducted by single respiratory bronchioles to a few alveoli. This unit is a primary site of gas exchange. The surface is covered by a characteristic epithelium (ie, pneumocytes). Expression profile: The expression pattern of genes, selected from a particular cell or tissue type, generally obtained by a vari- ety of high-throughput methods, such as microarray and serial analysis gene expression (SAGE). Hierarchical clustering: An analytical tool used to find the closest associations among gene profiles and specimens under evaluation. Gene ontology: Allows for annotating genes and their products with a limited set of attributes, with the three organizing principles being molecular function, biological process, and cellular component. The development of structured, controlled vocabularies (ontologies) that describe gene products in terms of these organizing principles in a species-independent manner is a constantly evolving process.Hs.seq.all, Hs.data, and LL_tmpl: UniGene (www.ncbi.nlm.nih.gov/UniGene) is an experimental system maintained at the National Cen- ter for Biotechnology Information for automatically partitioning GenBank sequences into a nonredundant set of gene-oriented clusters. Each UniGene cluster contains sequences that represent a unique gene, as well as related information such as the tissue types in which the gene has been expressed and map location. Hs.seq.all contains all Homo Sapience sequences and the information included in clusters, Hs.data contains various information related to each UniGene ID, and LL_tmpl contains various information related to each Locus ID. Perl (practical extraction and report language): A programming language especially designed for processing text. Because of its strong text processing abilities, Perl is used extensively in areas such as bioinformatics and Web programming. SAM (significance analysis of microarrays): A statistical technique using established software that determines the significance in changes of gene expression seen in microarray analysis (eg, cDNA and oligonucleotide microarrays), which measures the expression of thousands of genes and identifies changes in expression between different biologic states. On the basis of changes in gene expression relative to the standard deviation of repeated measurements, SAM assigns a score to each gene. When scores are greater than an adjustable threshold, permutations of repeated measurements are used by SAM to estimate the percentage of such genes identified by chance, the false discovery rate (FDR). In addition, SAM correlates gene expression data to a wide range of clinical parameters, including treatment, diagnosis categories, and survival time.
We thank Tetsuo Noda, MD, PhD, and Yoshio Miki, MD, PhD, (Cancer Institute, Tokyo, Japan) for their encouragement throughout the study, and Kaori Hayashi (Aichi Cancer Center, Nagoya, Japan) for her excellent technical assistance with the molecular genetic experiments.
Supported by in part by a grant-in-aid for scientific research on priority areas and a grant-in-aid for scientific research from the Ministry of Education, Culture, Sports, Science and Technology of Japan, and a grant from the Japan New Energy and Industrial Technology Development Organization. T. Takeuchi, S.T., and Y.Y. contributed equally to this work. Terms in blue are defined in the glossary, found at the end of this article and online at www.jco.org. Authors' disclosures of potential conflicts of interest and author contributions are found at the end of this article.
1. Minna JD: Neoplasms of the lung, in Kasper DL, Braunwald E, Fauci AS, et al (eds): Harrison's Principals of Internal Medicine (Ed 16). New York, NY, McGraw-Hill, 2001, pp 506-516 2. Satistics and Information Department, Minister's Secretariat: Vital Statistics of Japan 2001. Tokyo, Japan, Ministry of Health, Labor and Welfare, 2003; pp 384-411, 3. Travis WD, Brambilla E, Mueller-Hermelink HK, et al: Pathology and genetics of tumours of the lung, pleura, thymus and heart, World Health Organization classification of tumors. Lyon, France, IARC Press, 2004 4. Perou CM, Sorlie T, Eisen MB, et al: Molecular portraits of human breast tumours. Nature 406:747-752, 2000[CrossRef][Medline] 5. Alizadeh AA, Eisen MB, Davis RE, et al: Distinct types of diffuse large B-cell lymphoma identified by gene expression profiling. Nature 403:503-511, 2000[CrossRef][Medline] 6. Shipp MA, Ross KN, Tamayo P, et al: Diffuse large B-cell lymphoma outcome prediction by gene-expression profiling and supervised machine learning. Nat Med 8:68-74, 2002[CrossRef][Medline] 7. Singh D, Febbo PG, Ross K, et al: Gene expression correlates of clinical prostate cancer behavior. Cancer Cell 1:203-209, 2002[CrossRef][Medline] 8. van 't Veer LJ, Dai H, van de Vijver MJ, et al: Gene expression profiling predicts clinical outcome of breast cancer. Nature 415:530-536, 2002[CrossRef][Medline] 9. Hedenfalk I, Ringner M, Ben-Dor A, et al: Molecular classification of familial non-BRCA1/BRCA2 breast cancer. Proc Natl Acad Sci U S A 100:2532-2537, 2003 10. Garber ME, Troyanskaya OG, Schluens K, et al: Diversity of gene expression in adenocarcinoma of the lung. Proc Natl Acad Sci U S A 98:13784-13789, 2001 11. Bhattacharjee A, Richards WG, Staunton J, et al: Classification of human lung carcinomas by mRNA expression profiling reveals distinct adenocarcinoma subclasses. Proc Natl Acad Sci U S A 98:13790-13795, 2001 12. Beer DG, Kardia SL, Huang CC, et al: Gene-expression profiles predict survival of patients with lung adenocarcinoma. Nat Med 8:816-824, 2002[Medline] 13. Tomida S, Koshikawa K, Yatabe Y, et al: Gene expression-based, individualized outcome prediction for surgically treated lung cancer patients. Oncogene 23:5360-5370, 2004[CrossRef][Medline] 14. Meyerson M, Carbone D: Genomic and proteomic profiling of lung cancers: Lung cancer classification in the age of targeted therapy. J Clin Oncol 23:3219-3226, 2005 15. Paez JG, Janne PA, Lee JC, et al: EGFR mutations in lung cancer: Correlation with clinical response to gefitinib therapy. Science 304:1497-1500, 2004 16. Lynch TJ, Bell DW, Sordella R, et al: Activating mutations in the epidermal growth factor receptor underlying responsiveness of non-small-cell lung cancer to gefitinib. N Engl J Med 350:2129-2139, 2004 17. Pao W, Miller V, Zakowski M, et al: EGF receptor gene mutations are common in lung cancers from "never smokers" and are associated with sensitivity of tumors to gefitinib and erlotinib. Proc Natl Acad Sci U S A 101:13306-13311, 2004 18. Kosaka T, Yatabe Y, Endoh H, et al: Mutations of the epidermal growth factor receptor gene in lung cancer: Biological and clinical implications. Cancer Res 64:8919-8923, 2004 19. Mitsudomi T, Kosaka T, Endoh H, et al: Mutations of the epidermal growth factor receptor gene predict prolonged survival after gefitinib treatment in patients with non-small-cell lung cancer with postoperative recurrence. J Clin Oncol 23:2513-2520, 2005 20. Shigematsu H, Lin L, Takahashi T, et al: Clinical and biological features associated with epidermal growth factor receptor gene mutations in lung cancers. J Natl Cancer Inst 97:339-346, 2005 21. Fukuoka M, Yano S, Giaccone G, et al: Multi-institutional randomized phase II trial of gefitinib for previously treated patients with advanced non-small-cell lung cancer. J Clin Oncol 21:2237-2246, 2003 22. Kris MG, Natale RB, Herbst RS, et al: Efficacy of gefitinib, an inhibitor of the epidermal growth factor receptor tyrosine kinase, in symptomatic patients with non-small cell lung cancer: A randomized trial. JAMA 290:2149-2158, 2003 23. Yatabe Y, Kosaka T, Takahashi T, et al: EGFR mutation is specific for terminal respiratory unit type adenocarcinoma. Am J Surg Pathol 29:633-639, 2005[CrossRef][Medline] 24. Yatabe Y, Mitsudomi T, Takahashi T: TTF-1 expression in pulmonary adenocarcinomas. Am J Surg Pathol 26:767-773, 2002[CrossRef][Medline] 25. Eisen MB, Spellman PT, Brown PO, et al: Cluster analysis and display of genome-wide expression patterns. Proc Natl Acad Sci U S A 95:14863-14868, 1998 26. Tusher VG, Tibshirani R, Chu G: Significance analysis of microarrays applied to the ionizing radiation response. Proc Natl Acad Sci U S A 98:5116-5121, 2001 27. Ashburner M, Ball CA, Blake JA, et al: Gene Ontology: Tool for the unification of biology. The Gene Ontology Consortium. Nat Genet 25:25-29, 2000[CrossRef][Medline] 28. Yatabe Y: Role of expression of thyroid transcription factor-1 in pulmonary adenocarcinoma, in Hayat MA (ed): Immunohistochemistry and In Situ Hybridization of Human Carcinomas. New York, NY, Elsevier Science/Academic Press, 2004, pp 169-179 29. Veldhuizen EJ, Haagsman HP: Role of pulmonary surfactant components in surface film formation and dynamics. Biochim Biophys Acta 1467:255-270, 2000[Medline] Submitted August 12, 2005; accepted October 27, 2005.
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2006 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
TOP
ABSTRACT
INTRODUCTION
tibs/SAM/index.html
2 test or Fisher's exact test were used for comparisons of proportions, while multivariate logistic regression analysis was performed to examine associations of the expression of expression profile-defined subtypes with various clinical parameters. The Kaplan-Meier method was used to estimate survival as a function of time, and survival differences were analyzed with the log-rank test. Cox proportional hazards modeling was performed to identify which independent factors might jointly have a significant effect on survival. All the analyses were performed with Stata software (version 7; Stata Corp, College Station, TX), and the two-sided significance level was set at P < .05. 
$100,000 (N/R) Not Required
