|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
Journal of Clinical Oncology, Vol 23, No 10 (April 1), 2005: pp. 2248-2255 © 2005 American Society of Clinical Oncology. DOI: 10.1200/JCO.2005.03.119 Prognostic Significance of Allelic Imbalance at the c-kit Gene Locus and c-kit Overexpression by Immunohistochemistry in Pediatric OsteosarcomasFrom the Laboratoire de Biochimie et Biologie Moléculaire and Service de Pédiatrie Onco-Hématologie, Centre Hospitalier et Universitaire (CHRU) Hautepierre; Service d'Anatomie Pathologique, CHRU Strasbourg; Inserm U381, Strasbourg; Centre Anticancéreux Léon Bérard, Lyon; Institut Gustave Roussy, Villejuif; Service de Pédiatrie Oncologique, Institut Curie; Service de Pédiatrie Onco-Hématologie, Hôpital Trousseau, Paris; Service de Pédiatrie Onco-Hématologie, CHRU Nancy–Hôpital Brabois, Nancy; and Service d'Anatomopathologie, Centre Hospitalier Universitaire Côte de Nacre, Caen, France Address reprint requests to Natacha Entz-Werlé, MD, Pédiatrie Onco-Hématologie, CHRU Hautepierre, Avenue Molière, 67098 Strasbourg Cedex, France; e-mail: Natacha.entz-werle{at}chru-strasbourg.fr
PURPOSE: Since the recent development of biologic agents targeting oncogenes, increasing attention has been focused on determining the role of tyrosine kinase receptors in the pathogenesis of tumors. Our study was designed to investigate the status of region 4q12, which contains the candidate gene c-kit, and the expression of c-kit by immunohistochemistry (IHC). PATIENTS AND METHODS: Paired blood and biopsy specimens of 68 children treated for high-grade primary osteosarcomas were collected. Microsatellite analysis at two genomic sites containing c-kit gene was performed on paired DNA using a sensible fluorescent polymerase chain reaction technology. To confirm the DNA data, we studied c-kit protein expression by IHC in 56 available paraffin-embedded tumor tissues. RESULTS: The frequency of allelic imbalance (AI) at locus 4q12 was 39% in the overall population. In agreement with previous studies, we did not detect microsatellite instability, allowing us to hypothesize that this pathway is not implicated. Furthermore, the normal status at locus 4q12 was associated with a significantly better survival in the whole osteosarcoma population (P = .05). IHC overexpression of c-kit was concordant in all cases presenting an AI. However, normal status at locus 4q12 was correlated to an absence of c-kit protein expression in 19 (65.5%) of 29 informative cases. CONCLUSION: Allelotyping of locus 4q12, which contains the c-kit gene, could help pediatric osteosarcoma prognostic screening and showed a strong correlation with overexpression of c-kit protein. These results allowed us to hypothesize that, in some cases, a mutation of c-kit gene could lead to a protein overexpression.
High-grade osteosarcoma is the most common form of malignant bone cancers in children (5% of the neoplasia) and is observed mainly in teenagers and mostly localized in long bones.1,2 At present, there is no other worldly recognized prognostic factor than the presence of metastatic disease at diagnosis and the histologic response to preoperative chemotherapy established by Huvos et al,3 classifying children as good responders (GRs) or poor responders (PRs) to neoadjuvant chemotherapy. In the French group, with multimodality treatment, the 5-year survival rate was approximately 70% (OS873A and OS943B protocols). In recent years, the awareness of an earlier prognostic factor was the challenge of few studies. Increasing attention has been focused on determining the role of tyrosine kinase receptors in the pathogenesis of human tumors, including pediatric osteosarcomas, because of the recent development of biologic agents targeting these oncogenes.4 Furthermore, osteosarcomas seemed to be dependent on growth factors that act by autocrine or paracrine mechanisms.5 Indeed, some studies have shown that tyrosine kinase receptors and/or their ligands, as c-met or platelet-derived growth factor receptor, were involved in osteosarcomas oncogenesis.5–9 The oncogene c-kit is a tyrosine kinase receptor that binds the ligand stem-cell factor. The c-kit signal seems to play a central role in regulating normal cell differentiation, maturation, and proliferation. Concerning c-kit expression, a preliminary immunohistochemical (IHC) analysis was reported showing an extensive staining in pediatric solid tumors and, particularly, in a significant percentage of tested osteosarcomas,10 whereas c-kit overexpressions by IHC have been reported in the progression of several adult cancers.4,11–13 Mostly, mutations or deletions in exons 9, 11, 13, or 17 have been described in gastrointestinal stromal tumors, mastocytosis, germ cell tumors, or myelodysplastic syndrome.11–13 These mutations resulted in a constitutive activation of the c-kit protein that could be inhibited by the oral agent STI571.4 However, some of the overexpressions observed by IHC could be a result of chromosome rearrangement. At present, to our knowledge, there is no report concerning c-kit status at gene level in pediatric osteosarcoma. Thus, the purpose of our study was to determine the status of region 4q12, which contains the c-kit gene, at diagnosis in 68 high-grade pediatric osteosarcoma patients treated with the same therapeutic management in a multicenter protocol. Then, a microsatellite study using a sensible fluorescent polymerase chain reaction (PCR) was performed to characterize the presence of chromosomal alterations in the region containing the c-kit gene. Furthermore, to confirm the DNA data, an IHC analysis allowed us to study the c-kit expression in 56 patients of our cohort because analysis at RNA level was not possible as a result of the small size of tumor samples obtained at diagnosis by biopsy. Both types of results were correlated with histologic responses to preoperative chemotherapy and survival rates.
Patients Sixty-eight pediatric patients from 20 different French centers (Table 1) were included in our molecular study from November 1994 to December 2002. These primary high-grade osteosarcoma patients were treated according to OS94 French protocol guidelines.3B
After establishing the diagnosis, osteosarcomas were treated with a preoperative chemotherapy lasting 4 months, using either the combination of an anthracycline (doxorubicin) and high-dose methotrexate or the combination of etoposide, ifosfamide, and high-dose methotrexate. The postoperative treatment was determined by the histologic system established by Huvos et al3; GRs were treated with the same initial regimen, and PRs received a second-line chemotherapy, comprising either the combination of etoposide and ifosfamide in case of preoperative therapy with an anthracycline and high-dose methotrexate or the combination of cisplatin and doxorubicin. At present, the statistical evaluation is under progress.3B In this histopathologic grading, the GR subgroup was defined by more than 95% tumor necrosis.
Case Material
DNA Extraction: Microsatellite Analysis D4S3019 and D4S428 microsatellites, surrounding the c-kit gene (4q12), were analyzed on paired normal and biopsy DNA (for primer descriptions, see http://www.gdb.org and http://www.ncbi.nlm.nih.gov/genemap99). DNA from both paired samples (10 ng) were amplified by PCR in a total volume of 25 µL using 0.125 µL of Taq polymerase and 4 pmol of both forward and Cy5-labeled reverse primers. PCR was carried out in an Omnigen Hybaid Thermocycler (Hybaid Ltd, Ashford, United Kingdom) using the following protocol: 5 minutes at 95°C, 35 cycles of 1 minute at 95°C, 1 minute at 55°C, and 1 minute at 72°C, followed by 5 minutes at 72°C. After separation on denaturing urea gel on a sequencer (Amersham Pharmacia Biotech, Uppsala, Sweden), amplified fragments were directly detected and quantified using an Alfwin Fragment Analyzer software package (Amersham Pharmacia Biotech). This technique allowed highly reproducible and sensitive quantification. The presence of additional peaks in heterozygous samples will be described as microsatellite instability (MSI). A modification of the allele ratio in tumor DNA compared with normal DNA will be described as an allelic imbalance (AI). The intensity of the AI was directly proportional to the percentage of tumor cells, and the AI was calculated as follows: AI percentage = absolute value [(Bb/Ba) – (Tb/Ta) x 100/(Bb/Ba)], in which Ba and Bb represent the height of the two alleles in the control blood and Ta and Tb represent the height of the two alleles in the tumor. An allelic variation greater than a cutoff of 20% corresponded to the presence of a significant AI, as previously described.14 In our study, each alteration was confirmed, at least, by a duplicate PCR. The homozygous status addressed patients whose analyzed blood and tumor DNAs had identical allele on both chromosome 4q12 regions. In these patients, who corresponded to a noninformative population, losses of heterozygosity or amplifications could not be identified with these homozygous microsatellites.
IHC Slides were examined blindly by two independent investigators. Immunostaining was scored according to the percentage of positive-stained tumor cells and the intensity (weakly, moderately, or strongly positive; Table 2). We codified the immunostaining intensity as follows: 0 for no expression, + for a weak overexpression of c-kit, ++ for a moderate expression, and +++ for a strongly positive staining. We arbitrary decided to use a cutoff of 10%. When we considered this individualized 10% positive-cell subgroup of patients, the allelotyping results showed either a normal status or AI and then allowed us to confirm this percentage as a possible limit between positive and negative IHC samples. The survival statistical analysis was based on this cutoff.
Statistics Data were computed using SPSS 11.0 for Windows (SPSS Inc, Chicago, IL). Survival function was estimated using the Kaplan-Meier test to evaluate the prognostic value of our results. A multivariate analysis using the Cox model was performed to correlate allelotyping results with survival adjusted to the quality of response to preoperative chemotherapy. The log-rank test was used to compare survival, and the  2 test was used to analyze correlation between AI or IHC results and subgroups of patients. To correlate the AI and IHC results, an Armitage tendency test were used. The Mann-Whitney U test, which is especially suitable for small subgroups of patients, was also performed (statXact 3.0; Cytel Software, Cambridge, MA).
The purpose of our study was to determine, by allelotyping, the possible involvement of locus 4q12, which contains the c-kit gene, in 68 high-grade osteosarcoma patients at diagnosis to define new target markers for prognosis. In this group of patients, the overall survival (OS) rate at 5 years was 80%, and the statistical analysis of the GR and PR subgroups revealed representative patient groups of the whole protocol population. The prognostic significance of response to chemotherapy persisted in this cohort (P = .002). The 5-year OS rate for the GR and PR patients was 97% and 76%, respectively (data not shown).
Locus 4q12 Seems to Be Frequently Altered in Pediatric High-Grade Osteosarcomas
The two selected microsatellites defined a region containing the c-kit gene in 4q12, allowing us to improve the sensitivity of our technique because homozygous alleles are not informative in this technique. In a single patient, when both microsatellites were heterozygous and, thus, informative for the allelotyping, no discordance was observed at the two sites, allowing us to consider patients as informative for these loci if they were heterozygous for both markers or if one microsatellite was altered or normal and the second one was homozygous. The analysis of locus 4q12 showed an overall AI frequency of 39% (22 of 56 informative samples; Table 3). The percentage of AI in the different subgroups was 29% in the GRs and 52% in the PRs. However, no correlation was observed between the presence of AI and response to chemotherapy. Interestingly, the unmodified status at this locus was correlated in the whole population and in the PR subgroup with a significantly better OS (P = .05 and P = .02, respectively; Figs 1A and 1B). The single event in the GR subgroup could not allow us to obtain such statistical information. In this analysis, the OS rates in the normal subgroup and in the rearranged subgroup were 94% and 70%, respectively, at 5 years. A multivariate analysis restricted to the relevant risk factor, corresponding at least to the quality of response to chemotherapy, seemed to reveal a significant correlation between AI and a worse survival adjusted for response to chemotherapy (rate ratio = 3.1; P = .05).
IHC Results of c-kit Protein Expression Because most of the studies focused on c-kit were performed only by IHC and because the allelotyping that we performed at locus 4q12 lacked an accurate specificity for c-kit gene analysis, we decided to correlate our microsatellite results at 4q12 with the expression of c-kit by IHC in the 56 patients whose paraffin-embedded samples were available. These analyses showed a detectable expression of c-kit protein in 32 (57%) of 56 analyzed slides using a cutoff of 10% (Table 2). In all cases of positivity, c-kit protein localization was cytoplasmic (Fig 2), with no evidence of nuclear localization. We only obtained 18 paraffin-embedded specimens out of the 22 samples with AI. Interestingly, all of these 18 cases showed c-kit protein overexpression in IHC. Only one sample (patient 30) showed approximately 5% positive cells with moderate staining and a focal distribution, which allowed us to consider this patient as normal because no patient with AI showed 5% of overexpressing cells. The homozygous patients in allelotyping analysis could be also included in the population analyzed by IHC, increasing the number of informative patients for IHC study. A heterogeneous distribution (focal) through the tumor sample was observed in seven patients (patients 10, 13, 26, 30, 40, 54, and 66; Fig 3).
Nine of 56 patients were not informative by allelotyping. A significant correlation between allelotyping and IHC was observed in 36 of 47 informative patients (Table 2). However, in the 29 samples with a normal allelotyping status (Fig 4), only 19 samples showed no or very weak (5%) c-kit protein expression. Patients 6, 8, 10, 21, 31, 32, 50, 55, 59, and 66 showed normal allelotyping with a positive IHC analysis.
Interestingly, the statistical survival analysis of c-kit protein overexpression tends to have the same prognostic value as the presence of AI in statistical allelotyping analysis (Fig 1A). Indeed, the presence of c-kit overexpression tended to be correlated with a worse outcome (Fig 1C; P = .1). Considering our cutoff, the 5-year OS rate in the subgroup of patients with no staining was 100%, and the 5-year OS rate in the subgroup with positive staining was 85%. The proportion of 4q12 locus AI within each class of increasing IHC positivity was 10 of 18 AI in the +++ subgroup, three of 18 AI in the ++ subgroup, and six of 18 AI in the + subgroup (Table 2). The Armitage tendency test showed a statistical correlation between the percentages of AI within each class of increasing IHC positivity (P < .001). Furthermore, the Mann-Whitney U test, which is especially suitable for small subgroups, confirmed this correlation between the 4q21-positive AI and overexpression of c-kit protein (P < .001).
Allelotyping at 4q12 is a new and informative screening method in pediatric high-grade osteosarcomas. Our study analyzed the status of the region 4q12 containing the c-kit gene at diagnosis in a homogeneous population of pediatric osteosarcoma patients who were treated with the same therapeutic management. To our knowledge, even if, in many adult cancers, few mutations have been described, no molecular study of IHC investigating the locus 4q12 and c-kit protein expression has been published in such a large group of osteosarcomas. Therefore, microsatellite analysis at DNA level was achieved in 68 paired blood samples and biopsies to obtain rapid and sensitive information concerning this locus. Because our diagnostic samples, obtained from 20 different centers, were not immediately frozen at biopsy, study at the RNA level, like transcriptome analysis, was not available. Furthermore, the mRNA extraction required a large amount of tumor cells that we do not have in our small diagnostic specimens. Our microsatellite analysis was achieved in 68 paired blood and fresh-frozen biopsy specimens. AI only were detected, and no MSI were detected, allowing us to speculate that the repair error (RER) phenotype is not a main mechanism involved in the oncogenesis of osteosarcoma, which is in contrast to colorectal cancers,18,19 as previously published for 58 patients of this cohort.8 One of the Bethesda consensus microsatellites (D5S346) was used in the previous published data of this cohort8,16,17 and did not show any MSI. Indeed, the MSI was explored, at least, by a set of 14 informative markers, and therefore, we could hypothesize that this MSI pathway is not implicated in resistance to alkylating drugs, which are drugs known to be frequently administered in this malignant bone tumor. In this population, the observation of an average AI frequency of 39% using two microsatellites surrounding the c-kit gene confirmed the interest of this region in osteosarcomas. In addition, we showed a significant correlation between AI at locus 4q12 and a worse prognosis in the overall population, which is in agreement with the higher AI frequency observed in the PR subgroup and the fact that two out of the three dead patients had persistent alterations at the locus 4q12 in paired biopsy and surgical resection samples (data not shown). This specific region was not previously described as frequently altered in osteosarcomas either in cytogenetic or in comparative genomic hybridization studies.20–22 AI at the c-kit locus seems to be a new prognostic factor. Because we focus on a region containing the c-kit gene, in which most of the mutations were described in various cancers, and because no valuable RNA were available at diagnosis in this OS94 protocol, we compared our DNA results to IHC data. The concordance of allelotyping results to IHC in 36 of 47 informative patients and the statistical significance of the Armitage tendency and the Mann-Whitney U tests confirmed the role of c-kit in osteosarcomas and the significant correlation between c-kit overexpression and AI. The discordance in patients without any microsatellite abnormalities (patients 6, 8, 21, 31, 32, 50, 55, and 59) could probably be explained in these cases by the presence of mutations, according to previously published results in other cancers.11,12,13 Thus, it seems that a frequent event in osteosarcomas is a cytoplasmic overexpression, which could be linked to 4q12 chromosomal rearrangement. Such overexpressions correlated with DNA rearrangement in allelotyping could be a result of the amplification of the gene, but we could not rule out the possibility of the loss of a normal allele and the presence of mutation on the other allele. A real-time quantitative PCR analysis of the c-kit gene and the search for c-kit mutations would definitely confirm such a model. In addition, a clonal heterogeneity of the tumor would explain, for example, the discordance in patients 10, 30, and 66 because the overexpression was weakly and heterogeneously distributed. In these three latter patients, this low number of tumor cells with an amplification of c-kit would not be detected by allelotyping. Then, the use of laser capture to ensure DNA from different localizations in the primary tumor would overpass the problem of tumor heterogeneity. Interestingly, the normal status in allelotyping at this locus was significantly correlated in the whole population and in the PR subgroup with a better OS, allowing us to use c-kit allelotyping as a prognostic marker in pediatric osteosarcoma. The multivariate analysis confirmed this prognostic significance. Furthermore, this prognostic consequence was also found as a statistical tendency in IHC results. The low number of events could be involved in the failure of statistical correlation. Nevertheless, the DNA microsatellite analysis seemed to have an accurate prognostic impact that could be used as a new diagnostic molecular tool. Thus, microsatellite analysis and IHC could serve as tools to complete the prognosis of response to preoperative chemotherapy in pediatric osteosarcomas at diagnosis and probably for designing specific immunotherapies to complement existing treatments for long-term control of patients who had a worse prognosis. In conclusion, our large and homogenous cohort of 68 patients was investigated at a region containing a putative therapeutic target gene that, in the future, could have important consequences for osteosarcomas treatment with tyrosine kinase inhibitors. In addition, at diagnosis, the locus 4q12 containing the c-kit gene might have a prognostic role at DNA level in this bone cancer. Considering the allelotyping results and the significant concordance with IHC, c-kit gene seemed to play an important role in osteosarcomas oncogenesis as an oncogene. Future studies should focus our search for c-kit oncogenic mutations and determination of quantitative status of this gene.
The authors indicated no potential conflicts of interest.
We thank all the anatomopathologists, especially Anne De Muret, MD, Jeanne Champigneulle, MD, Corinne Bouvier, MD, and Jean-Marc Guinebretière, MD, and the clinicians of the French Society of Pediatric Oncology involved in the bone tumor group. We also thank Evelyne Neuville, Eliane Mengus, Véronique Kussaibi, Sylvie Delacourt, and Odile Regine for expert technical assistance, Nicolas Meyer for helpful discussions and statistical analysis, and Noëlle Dupouys for the registration of patient clinical data.
Supported by funds from a national Programme Hospitaliere de Recherche Clinique. Authors' disclosures of potential conflicts of interest are found at the end of this article.
1. Huvos AG: Bone Tumors, Diagnosis, Treatment, and Prognosis (ed 2). Philadelphia, PA, WB Saunders, 1991, pp 85-155, pp 343-393 2. Dahlin DC, Umi KK: Bone Tumors, General Aspects and Data on 8542 Cases. Springfield, IL, Charles C Thomas, 1986, pp 227-307 3. Huvos AG, Rosen G, Marcove RC: Primary osteogenic sarcoma: Pathologic aspects in twenty patients after treatment with chemotherapy, in bloc resection and prosthetic bone replacement. Arch Pathol Lab Med 101:14-18, 1977[Medline] 3. Violas P, Kohler R, Mascard E, et al: Conservative surgical treatment of osteogenic sarcoma of the limb in children and adolescents [in French]. Rev Chir Orthop Reparatrice Appar Mot 86:675-683, 2000[Medline] 3. Kalifa C, Guinebretière JM, Gentet JC, et al: Comparison of doxorubicin versus etoposide-ifosfamide in addition to HDMTX as preoperative chemotherapy in osteosarcoma: A randomized trial by the French Society of Pediatric Oncology. Proc Am Soc Clin Oncol 21:409a, 2002 (abstr 1634) 4. Scappaticci FA, Marina N: New molecular targets and biological therapies in sarcomas. Cancer Treat Rev 27:317-326, 2001[CrossRef][Medline] 5. Sulzbacher I, Traxler M, Mosberger I, et al: Platelet-derived growth factor-AA and -alpha receptor expression suggest an autocrine and/or paracrine loop in osteosarcoma. Mod Pathol 13:632-637, 2000[CrossRef][Medline] 6. Benini S, Baldini N, Manara MC, et al: Redundancy of autocrine loops in human osteosarcoma cells. Int J Cancer 80:581-588, 1999[CrossRef][Medline]
7. McGary EC, Weber K, Mills L, et al: Inhibition of platelet-derived growth factor-mediated proliferation of osteosarcoma cells by the novel tyrosine kinase inhibitor STI571. Clin Cancer Res 8:3584-3591, 2002 8. Entz-Werle N, Schneider A, Kalifa C, et al: Genetic alterations in primary osteosarcoma from 54 children and adolescents by targeted allelotyping. Br J Cancer 88:1925-1931, 2003[CrossRef][Medline] 9. Oda Y, Naka T, Takeshita M, et al: Comparison of histological changes and changes in nm23 and c-met expression between primary and metastatic sites in osteosarcoma: A clinicopathologic and immunohistochemical study. Hum Pathol 31:709-716, 2000[CrossRef][Medline] 10. Smithey BE, Pappo AS, Hill DA: C-kit expression in pediatric solid tumors: A comparative immunohistochemical study. Am J Surg Pathol 26:486-492, 2002[CrossRef][Medline]
11. Rubin BP, Singer S, Tsao C, et al: Kit activation is a ubiquitous feature of gastrointestinal stromal tumors. Cancer Res 61:8118-8121, 2001 12. Fritsche-Polanz R, Jordan JH, Feix A, et al: Mutation analysis of c-kit in patients with myelodysplastic syndromes without mastocytosis and cases of systemic mastocytosis. Br J Haematol 113:357-364, 2001[CrossRef][Medline]
13. Tian Q, Frierson HF, Krystal GW, et al: Activating c-kit gene mutations in human germ cell tumors. Am J Pathol 154:1643-1647, 1999
14. Schneider A, Borgnat S, Lang H, et al: Evaluation of microsatellite analysis in urine sediment for diagnosis of bladder cancer. Cancer Res 60:4617-4622, 2000
15. Miller SA, Dykes DD, Polesky HF: A simple salting out procedure for extracting DNA from human nucleated cells. Nucleic Acid Res 16:1215, 1988
16. Boland CR, Thibodeau SN, Hamilton SN, et al: A National Cancer Institute workshop on microsatellite instability for cancer detection and familial predisposition: Development of international criteria for the determination of microsatellite instability in colorectal cancer. Cancer Res 58:5248-5257, 1998
17. Umar A, Boland CR, Terdiman JP, et al: Revised Bethesda Guidelines for hereditary nonpolyposis colorectal cancer (Lynch syndrome) and microsatellite instability. J Natl Cancer Inst 96:261-268, 2004 18. Karran P: Microsatellite instability and DNA mismatch repair in human cancer. Semin Cancer Biol 7:15-24, 1996[CrossRef][Medline] 19. Leach FS, Nicolaides NC, Papadopoulos N, et al: Mutations of a mutS homolog in hereditary nonpolyposis colorectal cancer. Cell 75:1215-1225, 1993[CrossRef][Medline] 20. Ozisik YY, Meloni AM, Peier A, et al: Cytogenetic findings in 19 malignant bone tumors. Cancer 74:2268-2275, 1994[CrossRef][Medline] 21. Ozaki T, Schaefer KL, Wai D, et al: Genetic imbalances revealed by comparative genomic hybridization in osteosarcomas. Int J Cancer 102:355-365, 2002[CrossRef][Medline] 22. Squire JA, Pei J, Marrano P, et al: High resolution mapping of amplifications and deletions in pediatric osteosarcoma by use of CGH analysis of cDNA microarrays. Genes Chromosomes Cancer 38:215-225, 2003[CrossRef][Medline] Submitted March 19, 2003; accepted December 24, 2004.
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2005 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
TOP
ABSTRACT
INTRODUCTION
