|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
Originally published as JCO Early Release 10.1200/JCO.2005.04.4339 on March 20 2006 © 2006 American Society of Clinical Oncology. Association of the PDCD5 Locus With Lung Cancer Risk and Prognosis in Smokers
From the Department of Experimental Oncology and Laboratories, Thoracic Surgery, Istituto Nazionale Tumori, Milan, Italy; Asklepios Clinics, Pneumology Clinic, Munich-Gauting; Genefinder Technologies Ltd, Munich; Institute of Human Genetics, Molecular Oncogenetics Division, University Hospital, Tuebingen; Clinic Schillerhoehe, Pneumology Clinic, Stuttgart-Gerlingen, Germany; and Sequenom Inc, San Diego, CA. Address reprint requests to Tommaso A. Dragani, PhD, Department of Experimental Oncology, Istituto Nazionale Tumori, Via G. Venezian 1, 20133 Milan, Italy; e-mail: tommaso.dragani{at}istitutotumori.mi.it
PURPOSE: Whole-genome scan association analysis was carried out to identify genetic variants predictive of lung cancer risk in smokers and to confirm the identified variants in an independent sample. PATIENTS AND METHODS: A case-control study was performed using two pools consisting of DNA from 322 German smoking lung cancer patients and 273 healthy smoking controls, respectively. A replication study was carried out using 254 Italian lung adenocarcinoma (ADCA) patients and 235 healthy controls. RESULTS: Patients with genotypes GG or CG for the rs1862214 single nucleotide polymorphism, 5' upstream of the programmed cell death 5 (PDCD5) gene, compared with those with the common genotype CC showed an increased risk of lung cancer (odds ratio, 1.6; 95% CI, 1.2 to 2.1) and a higher incidence of poor clinical stage disease (hazard ratio [HR], 1.9; 95% CI, 1.1 to 3.4; P = .023), nodal involvement (HR, 1.9; 95% CI, 1.1 to 3.6; P = .033), and short-term survivorship (HR, 1.8; 95% CI, 1.2 to 2.6, P = .003). PDCD5 mRNA expression levels were approximately 2.4-fold lower in lung ADCA as compared to normal lung tissue. Human NCI-H520 cancer cells transfected with PDCD5 cDNA showed decreased colony-forming ability. CONCLUSION: These results suggest that the rs1862214 polymorphism in PDCD5 is predictive for lung cancer risk and prognosis, and that PDCD5 may represent a novel tumor suppressor gene influencing lung cancer.
Lung cancer is the leading cause of cancer mortality in the Western world. Indeed, this malignancy has a poor prognosis and, so far, therapeutic strategies have shown only a limited effect. Lung cancer risk is strongly associated with exposure to environmental carcinogens, in particular to cigarette smoke. Nevertheless, only approximately 10% of smokers develop lung cancer, and the disease also occurs in the absence of exposure to cigarette smoke. Therefore, the potential role of hereditary components in determining the risk of lung cancer in smokers should not be underestimated. Although this hypothesis has been questioned in some studies,1 epidemiologic data support the presence of genetic risk factors in lung cancer.2-7 Whole-genome analyses in mouse models have identified genetic loci that affect lung cancer risk. Indeed, genetic analysis of these models provided evidence of a complex inheritance of genetic predisposition to lung tumorigenesis8 and suggested that polygenic inheritance of predisposition to cancer might also occur in humans.9 In mice, whole-genome analyses require only a few hundred genetic markers because of the genetic homogeneity of inbred strains, the limited number of genetic recombinations in crosses, and the consequent tight linkage of genetic markers over wide chromosomal regions.8 Because of the genetic heterogeneity of the human population, complete analysis of the human genome by association studies may require a huge number of genetic markers and large sample collections. However, the presence of haplotype blocks in the human genome10 suggests that even a less dense marker coverage might be sufficient for a whole-genome analysis. Although the characterization of these haplotype blocks awaits completion,11 population-based association studies are complicated by complex linkage disequilibrium (LD) patterns, possible population admixture in the selection of cases or controls, differences in environmental factors, and other factors.12 In an effort to identify genetic variants that may predict individual risk of lung cancer in smokers, we conducted a large-scale case-control study using more than 80,000 single nucleotide polymorphisms (SNPs) located primarily in gene-based regions. We report the identification of an 80-kb genomic region on chromosome 19q12-q13.1 including the programmed cell death 5 gene (PDCD5, also called TFAR19 [TF1 cell apoptosis-related gene 19]) that conferred increased lung cancer risk in two independent sample collections and that may predict disease outcome.
Patients and Tissue Samples The initial study included 322 male smokers with lung cancer who were recruited at the Pneumology and Thoracic Surgery Clinics in Stuttgart-Gerlingen and the Pneumology Clinic in Munich-Gauting (both in Germany), and 273 healthy smokers recruited at the same sites and at the Dermatology Department of the University Hospital in Tuebingen, Germany (Table 1). The Italian replication sample included 254 cases of pathologically documented lung adenocarcinoma (ADCA) patients who underwent surgery at Istituto Nazionale Tumori (Milan, Italy), and 235 healthy controls (blood donors) enrolled at the same institute. Personal files were recorded to obtain clinical data (Tables 1 and 2). All subjects involved in our studies signed a written informed consent, and the institutional ethics committees of participating institutions approved the study protocol.
For the German discovery samples, DNA was obtained from blood samples using a desalting method (Gentra Systems, Minneapolis, MN) and quantitated using Pico green reagents and a Fluorometer (Fluoroskan Ascent CF; Labsystems, Chicago, IL). Genomic DNAs of the Italian patients and controls were extracted from a small piece of nontumor tissue excised during surgery or from peripheral blood samples using the automatic DNA extractor Extragen 8C (Talent, Trieste, Italy). Total RNA was extracted from 20 paired nontumor lung parenchyma and lung tumor tissue and additional 26 nontumor lung parenchyma of Italian lung ADCA patients using RNeasy Midi kit (Qiagen, Valencia, CA). cDNA was synthesized from RNA of each sample using the SuperScript First-Strand Synthesis System (Invitrogen, Carlsbad, CA).
SNP Markers and Genotyping DNA pools were formed by combining equimolar amounts of each sample as described.14 For pooled DNA assays, 25 ng of case and control DNA pools was used for amplification at each site. All polymerase chain reaction (PCR) and homogeneous MassEXTEND15 reactions were conducted using standard conditions. Relative allele frequency was estimated on the basis of the area under the peak calculated from mass spectrometry measurements of four aliquots, as described.15 For individual genotyping, the same procedure was applied with the following differences: only 2.5 ng DNA was used and only one mass spectrometry measurement was taken. The following gene-specific primers were used to genotype the marker SNP rs1862214: PCR primer 1—5'-ACGTTGGATGCCTGTTTTCCTGATCACACC-3'; PCR primer 2—5'-ACGTTGGATGTGCAGTCCTGAAAGGAAACC-3'; extension primer—5'-CACACCACTCACTGCAC-3'.
Cloning of PDCD5 From Human Lung Tissue
Real-Time Quantitative PCR
Transfection and In Vitro Clonogenic Assay
Statistical Analysis
Identification of a Lung Cancer Susceptibility Locus on 19q12-q13.1 Using a screening strategy based on DNA pooling,20,21 we examined DNA samples from 322 German smokers clinically diagnosed with lung cancer and 273 age-matched healthy smokers (Tables 1 and 2) in a genome-wide study consisting of 83,715 SNPs. Individual genotyping identified 160 SNPs that differed significantly (P < .05) between cases and controls. To verify true genetic effects by independent replication, we also studied a group of 254 Italian lung cancer patients and 235 unaffected controls. One of the loci identified was located on chromosome 19q12-q13.1. A C/G polymorphism (rs1862214) located 35 kb upstream of the PDCD5 gene was associated with lung cancer susceptibility. Analysis of the German discovery sample revealed an increased frequency of the rare G allele in lung cancer patients (0.29) as compared with controls (0.22; Table 3). Frequencies of the CG and GG genotypes were higher among patients (41% and 9%, respectively) than controls (37% and 3%, respectively). Hardy-Weinberg genotypic proportions were taken into account for in both groups (Table 3). A significant association with lung cancer risk was found in the discovery set of cases, with an OR for GG homozygous versus CC homozygous patients of 3.5 (95% CI, 1.3 to 9.5; Table 3). Whereas comparison of the common CC genotype with CG heterozygous or with the G allele carrier status revealed no statistically significant associations, the relative ORs were more than 1 (Table 3). In this sample, the presence of 32% female controls and the lack of matched female cases might have decreased the statistical association. Indeed, logistic analysis without sex adjustment indicated a significant association for the G allele carrier status (OR, 1.4; 95% CI, 1.03 to 2.01), and the homozygous comparison (GG v CC) showed a stronger association in the sex-unadjusted (data not shown) than in the sex-adjusted analysis.
Frequencies of the susceptibility G allele were similar in the control groups of both discovery and replication samples (0.22; Table 3). The frequency of the G allele was also increased in Italian lung cancer patients (0.27), with a detected OR of 2.1 (95% CI, 1.3 to 3.3) for patients carrying the G allele (GG or CG) as compared with CC homozygous patients (Table 3). A combined analysis of the two samples resulted in an OR of 1.6 (95% CI, 1.2 to 2.1; P = .0004) for individuals carrying the G allele (Table 3). A trend toward an increased risk was observed from heterozygous to homozygous patients carrying the rare allele, with ORs of 1.5 (95% CI, 1.1 to 2.0) and 2.9 (95% CI, 1.5 to 5.7), respectively (Table 3).
LD in the PDCD5 Gene Region
Association of rs1862214 SNP With Lung Cancer Prognosis Analysis of the rs1862214 polymorphism with respect to prognostic factors in Italian lung cancer patients revealed a significant positive association between the rare G allele and advanced clinical stage, nodal status, or overall survival. Indeed, the allele G carrier status showed a significant association with higher clinical stages (hazard ratio [HR], 1.9; 95% CI, 1.1 to 3.4; P = .023). The association was statistically significant also for the heterozygous genotype group (HR, 1.9; 95% CI, 1.1 to 3.4; P = .028) but not for the GG homozygous group (HR, 3.3; 95% CI, 0.7 to 15.3). Although the HR was higher in the latter group, the lack of association might be due to the low number of patients with the GG genotype and the small number of events detected in this group (Table 4).
Presence of lymph node metastases was more frequent in patients carrying the G allele than in CC homozygous patients (HR, 1.9; 95% CI, 1.1 to 3.6; P = .033), and the CG and GG groups maintained an excess value over CC homozygous patients, although the statistical significance was borderline (P = .054 and P = .051, respectively; Table 4). Median follow-up of patients alive at the end of the follow-up period was 56.5 months (Table 2). Kaplan-Meier survival curves analysis showed a significant difference among the three genotype groups (log-rank P = .0013), with median survival times of 79.8, 40.2, and 19.0 months associated with the CC, CG, and GG genotypes, respectively (Fig 2). Cox proportional hazard analysis of survival, adjusted for sex, smoking habits, and age at diagnosis (in decades) showed that patients carrying the CG and GG genotype had a poorer survival than patients with the CC genotype, with a trend in excess HR associated with copy number of the G allele (HR, 1.7; 95% CI, 1.2 to 2.5; P = .007 for the CG genotype and HR, 2.4; 95% CI, 1.1 to 5.6; P = .034 for the GG genotype; Table 4). Overall comparison of patients carrying the G allele (at homozygosity or heterozygosity) versus CC homozygous patients revealed a significant excess of deaths associated with the G allele carrier status (HR, 1.8; 95% CI, 1.2 to 2.6; P = .003; Table 4), with a median follow-up at death of 40.2 months for the G allele carriers and of 79.8 months for CC homozygous ADCA patients.
PDCD5 mRNA Expression Is Downregulated in Lung Cancer Because genomic DNA and RNA were available from nontumor and tumor lung tissue of 46 Italian lung adenocarcinoma patients, samples with different rs1862214 genotypes (24 CC, 17 CG, and five GG) were analyzed for PDCD5 expression to identify a possible correlation between SNP genotypes and mRNA levels. However, the mRNA levels in healthy lung tissue were similar in the common CC genotype as compared with G-allele carriers (data not shown). Comparison of PDCD5 mRNA levels in nontumor and tumor lung tissue from the same patient revealed approximately 2.4-fold lower expression in tumor tissue (median –  Ct values 4.4 [SE, 0.2] and 3.5 [SE, 0.2] in nontumor and tumor tissue, respectively; Fig 3; P = .006). Differences between nontumor and tumor tissue in PDCD5 mRNA levels did not correlate with the genotype at the rs1862214 polymorphism (data not shown).
PDCD5 Overexpression Inhibits Colony Formation In Vitro Amplification of the PDCD5 gene from human lung cDNA produced two different transcripts: the annotated variant NM_004708, consisting of six exons and encoding a protein of 125 amino acids, and an additional alternatively spliced variant that lacks exons 3 to 5 and is translated into a truncated protein product of 40 residues (GenBank Access No. DQ208400; http://www.ncbi.nlm.nih.gov/Genbank/). Mammalian expression vectors containing PDCD5 isoforms were prepared and transfected into A549 and NCI-H520 human lung cancer cell lines, whose endogenous levels of PDCD5 mRNA were approximately 0.8- and 0.3-fold lower, respectively, as compared with a pool of healthy lung tissue from 20 patients (data not shown). Overexpression of either PDCD5 transcript in the human lung cancer cell line A549 did not lead to significant variation in number or size of colonies grown in both high- and low-serum conditions (data not shown). By contrast, a significant inhibitory effect on colony formation was observed in NCI-H520 cells, with the number of colonies more than 0.5 mm in diameter approximately 6-fold lower (P = .006) in cells transfected with the full-length PDCD5 transcript than in control cells (Fig 4A and B). Cells overexpressing the short PDCD5 variant showed an approximately 1.5-fold reduction in clonogenicity, but the difference from the vector-transfected cells was not statistically significant (Fig 4A and B).
Genetic loci affecting lung cancer risk have been identified in mouse models, but the relationship with the risk of human lung cancer has not yet been established. One reason for the gap between the characterization of genetic factors involved in this disease in humans and experimental models is the difficulty in carrying out family-based studies for lung cancer because this disease rarely clusters in families and has a poor prognosis, making a sib-pair study design difficult. The strong association of lung cancer risk with smoking habits further complicates family-based linkage studies. Despite these difficulties, a recent genetic linkage study in lung cancer families reported a significant linkage of a single locus on chromosome 6q23-25 to lung cancer risk.23 This result supports the idea that genetic factors influence lung cancer risk in humans, but does not rule out a potential polygenic mode of inheritance of such risk, in analogy with the mouse model. Indeed, the observation of a single locus linked to lung cancer risk may simply reflect the relatively low power of that study to detect additional loci. Because the study in pedigrees included rare families with multiple lung cancer cases, whereas the large majority of lung cancer cases occur without familial clustering, the role of this locus in nonfamilial lung cancer by population-based association studies remains to be established. In our present genome-wide association study, we identified several chromosomal regions putatively associated with lung cancer risk in smokers. Replication of these findings in an independent sample for which follow-up data were available verified the validity of the observed associations and their importance in different populations. We identified a significant association of a SNP on chromosome 19, 35 kb 5' of the PDCD5 gene, with lung cancer risk. The association was first observed in a collection of German lung cancer cases consisting only of male smokers with a range of histologic subtypes and later confirmed in an independent population of Italian lung ADCA patients of both sexes, with mixed smoking habits. Despite the heterogeneity of the discovery and replication samples, the significant association between the rs1862214 SNP and lung cancer risk in both populations points to an important role for this chromosomal region in lung cancer predisposition. Moreover, in the replication sample, the same allele (G) associated with an increased lung cancer risk (Table 3) was also associated with poor survival, higher clinical stage, and nodal involvement (Table 4). This finding further supports the involvement of the PDCD5 region in the etiology of lung cancer. The rs1862214 variation is part of a haplotype block that contains the PDCD5 gene. This gene is a strong candidate for the modulation of lung cancer risk because of its described biologic functions and its involvement in the regulation of cell apoptosis. Indeed, PDCD5 levels are significantly increased in cells undergoing apoptosis, and the rapid translocation of the protein from the cytoplasm to the nucleus of apoptotic cells precedes the externalization of phosphatidylserine and the fragmentation of chromosome DNA.24 Moreover, the introduction of anti-PDCD5 antibody suppresses apoptosis,25 supporting the critical role of PDCD5 in this pathway. Since transcription may also be affected by long-range regulatory sequences, we investigated whether the rs1862214 SNP, which is 35 kb upstream of the PDCD5 gene, has a functional role in PDCD5 mRNA expression in normal and tumor lung tissue. However, in healthy lung parenchyma PDCD5 mRNA levels were similar in individuals with the common CC genotype as compared with G-allele carriers. Preliminary functional studies in vitro using human lung cancer cell lines transfected with recombinant mammalian expression vectors for the PDCD5 gene indicated that overexpression of this gene significantly inhibits colony growth in NCI-H520 cells but does not modulate colony formation in A549 cells. The reasons for such cell-specific effect are not known. Perhaps the ability of PDCD5 in modulating clonogenicity may be related to specific somatic alterations of cancer cells. Analysis of a shorter mRNA variant encoding only the 40 N-terminal amino acid residues of PDCD5 in NCI-H520 cells showed that the inhibitory activity of PDCD5 is almost completely lost, indicating that important functional domains are present in the truncated portion. However, mRNA levels of this shorter transcript were approximately 10-fold lower than those of the full-length variant in nontumor lung tissue, and the relative quantity of the two isoforms did not differ in lung tumor tissue (data not shown). Additional studies, including silencing of the PDCD5 gene and studies in animal models, may provide further insight into the role of PDCD5 in lung cancer risk and development.
The authors indicated no potential conflicts of interest.
Linkage disequilibrium: Nonrandom association of linked genes. This is the tendency of the alleles of two separate but already linked loci to be found together more frequently than would be expected by chance alone. PDCD5 (programmed cell death 5): This gene encodes a protein expressed in tumor cells during apoptosis independent of the apoptosis-inducing stimuli. Before apoptosis induction, this gene product is distributed in both the nucleus and cytoplasm. Once apoptosis is induced, the level of this protein increases and by relocation from the cytoplasm, it accumulates in the nucleus. Although its exact function is not defined, this protein is thought to play an early and universal role in apoptosis. PDCD5 is also called TFAR19 (TF1 cell apoptosis-related gene 19). SNP (single nucleotide polymorphism): Genetic polymorphisms are natural variations in the genomic DNA sequence present in greater than 1% of the population, with SNP representing DNA variations in a single nucleotide. SNPs are being widely used to better understand disease processes, thereby paving the way for genetic-based diagnostics and therapeutics. TFAR19 (TF1 cell apoptosis-related gene 19): This gene encodes a protein expressed in tumor cells during apoptosis independent of the apoptosis-inducing stimuli. Before apoptosis induction, this gene product is distributed in both the nucleus and cytoplasm. Once apoptosis is induced, the level of this protein increases and by relocation from the cytoplasm, it accumulates in the nucleus. Although its exact function is not defined, this protein is thought to play an early and universal role in apoptosis. TFAR19 is also called PDCD5 (programmed cell death 5).
Some of the nonsmoking patients were recruited with the help of Associazione Marta Nurizzo, Brugherio, Italy.
Supported in part by grants from Associazione and Fondazione Italiana Ricerca Cancro (AIRC and FIRC) and Fondo Investimenti Ricerca de Base (FIRB) to T.A.D. M.S., P.M., and S.K. contributed equally to this manuscript. R.D., T.A.D., and A.B. contributed equally to this manuscript. 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. Braun MM, Caporaso NE, Page WF, et al: Genetic component of lung cancer: Cohort study of twins. Lancet 344:440-443, 1994[CrossRef][Medline] 2. Ahlbom A, Lichtenstein P, Malmstrom H, et al: Cancer in twins: Genetic and nongenetic familial risk factors. J Natl Cancer Inst 89:287-293, 1997 3. Sellers TA, Potter JD, Bailey-Wilson JE, et al: Lung cancer detection and prevention: Evidence for an interaction between smoking and genetic predisposition. Cancer Res 52:2694s-2697s, 1992 4. Sellers TA, Bailey-Wilson JE, Elston RC, et al: Evidence for Mendelian inheritance in the pathogenesis of lung cancer. J Natl Cancer Inst 82:1272-1279, 1990 5. Yang P, Schwartz AG, McAllister AE, et al: Lung cancer risk in families of nonsmoking probands: Heterogeneity by age at diagnosis. Genet Epidemiol 17:253-273, 1999[CrossRef][Medline] 6. Gauderman WJ, Morrison JL: Evidence for age-specific genetic relative risks in lung cancer. Am J Epidemiol 151:41-49, 2000 7. Wu PF, Lee CH, Wang MJ, et al: Cancer aggregation and complex segregation analysis of families with female non-smoking lung cancer probands in Taiwan. Eur J Cancer 40:260-266, 2004[CrossRef][Medline] 8. Dragani TA: 10 Years of mouse cancer modifier loci: Human relevance. Cancer Res 63:3011-3018, 2003 9. Dragani TA, Canzian F, Pierotti MA: A polygenic model of inherited predisposition to cancer. FASEB J 10:865-870, 1996[Abstract] 10. Wall JD, Pritchard JK: Haplotype blocks and linkage disequilibrium in the human genome. Nat Rev Genet 4:587-597, 2003[CrossRef][Medline] 11. The International HapMap Consortium: The International HapMap Project. Nature 426:789-796, 2003 12. Vieland VJ: The replication requirement. Nat Genet 29:244-245, 2001[CrossRef][Medline] 13. Nelson MR, Marnellos G, Kammerer S, et al: Large-scale validation of single nucleotide polymorphisms in gene regions. Genome Res 14:1664-1668, 2004 14. Buetow KH, Edmonson M, MacDonald R, et al: High-throughput development and characterization of a genomewide collection of gene-based single nucleotide polymorphism markers by chip-based matrix-assisted laser desorption/ionization time-of-flight mass spectrometry. Proc Natl Acad Sci U S A 98:581-584, 2001 15. Bansal A, van den Boom D, Kammerer S, et al: Association testing by DNA pooling: An effective initial screen. Proc Natl Acad Sci U S A 99:16871-16874, 2002 16. Barratt BJ, Payne F, Rance HE, et al: Identification of the sources of error in allele frequency estimations from pooled DNA indicates an optimal experimental design. Ann Hum Genet 66:393-405, 2002[CrossRef][Medline] 17. Weir BS: Genetic Data Analysis 2: Methods for Discrete Population Genetic Data. Sunderland, MA, Sinauer Associates Inc, 1996 18. Kaplan EL, Meier P: Non-parametric estimation from incomplete observations. J Am Stat Assoc 3:457-481, 1958[CrossRef] 19. Peto R, Peto J: Asymptotically efficient rank invariant test procedures. J R Stat Soc 135:185-206, 1972 20. Kammerer S, Roth RB, Hoyal CR, et al: Association of the NuMA region on chromosome 11q13 with breast cancer susceptibility. Proc Natl Acad Sci U S A 102:2004-2009, 2005 21. Kammerer S, Roth RB, Reneland R, et al: Large-scale association study identifies ICAM gene region as breast and prostate cancer susceptibility locus. Cancer Res 64:8906-8910, 2004 22. Liu H, Wang Y, Zhang Y, et al: TFAR19, a novel apoptosis-related gene cloned from human leukemia cell line TF-1, could enhance apoptosis of some tumor cells induced by growth factor withdrawal. Biochem Biophys Res Commun 254:203-210, 1999[CrossRef][Medline] 23. Bailey-Wilson JE, Amos CI, Pinney SM, et al: A major lung cancer susceptibility locus maps to chromosome 6q23-25. Am J Hum Genet 75:460-474, 2004[CrossRef][Medline] 24. Chen Y, Sun R, Han W, et al: Nuclear translocation of PDCD5 (TFAR19): An early signal for apoptosis? FEBS Lett 509:191-196, 2001[CrossRef][Medline] 25. Rui M, Chen Y, Zhang Y, et al: Transfer of anti-TFAR19 monoclonal antibody into HeLa cells by in situ electroporation can inhibit the apoptosis. Life Sci 71:1771-1778, 2002[CrossRef][Medline] Submitted September 30, 2005; accepted December 16, 2005.
Related Editorial
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
2 method.
