|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
Journal of Clinical Oncology, Vol 23, No 36 (December 20), 2005: pp. 9369-9376 © 2005 American Society of Clinical Oncology. DOI: 10.1200/JCO.2005.03.7127 MYC Amplification and Polysomy 8 in Chondrosarcoma: Array Comparative Genomic Hybridization, Fluorescent In Situ Hybridization, and Association With OutcomeFrom the Division of Orthopedic Oncology, Department of Pathology, and Center for Biostatistics, The Arthur James Cancer Hospital and Richard Solove Research Institute, The Ohio State University Medical School; and Columbus Children's Research Institute, Columbus, OH. Address reprint requests to Carl Morrison, MD, DVM, Director of Pathology Core Facility, 418-M SL, 320 W 10th Ave, Columbus, OH 43210; e-mail: morrison-4{at}medctr.osu.edu
PURPOSE: To identify recurrent regions of genomic gain or loss in chondrosarcoma in a clinically relevant and statistically valid fashion. MATERIALS AND METHODS: Array comparative genomic hybridization (CGH) results of 15 frozen tumor samples of high-grade chondrosarcoma for chromosome 8 are presented. A separate subset of 116 cartilaginous tumors with outcome data was used for validation. RESULTS: Array CGH identified gain at 8q24.12-q24.13, the region of the MYC (c-Myc) oncogene, as a frequent change in high-grade chondrosarcoma. In the validation arm of 116 cartilaginous tumors, MYC was frequently amplified in G2 (15%), G3 (20%), and dedifferentiated (21%) chondrosarcomas. No amplification was identified in samples of enchondroma and grade 1 chondrosarcoma. In samples without MYC amplification, polysomy 8 was a frequent finding in grade 1 (18%), grade 2 (31%), grade 3 (80%), and dedifferentiated (29%) chondrosarcomas, but was not found in any samples of enchondroma. MYC protein expression was identified in all samples with amplification, but was also frequent in the remaining samples without amplification or polysomy 8. Kaplan-Meier survival curves for overall survival showed a statistically significant difference for patients with MYC amplification or polysomy 8 (P = .034). Univariate analysis involving Cox proportional hazards models showed that grade (P = .003), polysomy 8 (P = .045), and MYC amplification (P = .053) correlated with shorter overall survival. By multivariate analysis, grade of chondrosarcoma (P = .026) was the only factor to reach statistical significance. CONCLUSION: MYC amplification and polysomy 8 can be used as markers of prognostic importance in chondrosarcoma. Molecular targeting of MYC expression may have therapeutic potential in the future for subsets of chondrosarcoma.
The majority of what is known about the genetic changes of conventional (central) type and dedifferentiated chondrosarcoma comes from a limited number of classical cytogenetic reports.1-4 The overall summary of these studies is that chondrosarcomas are karyotypically complex tumors, and any definitive conclusions about recurrent nonrandom genetic changes in this group of tumors are difficult. As molecular medicine and gene-targeted therapy progresses, an understanding of the genetic pathways involved in chondrosarcoma will be essential. One tool that has been widely used in the past to investigate genome-wide changes in a variety of tumors is metaphase comparative genomic hybridization (CGH). Although this type of analysis is useful, it is generally limited to resolution of genomic change of 10 to 15 million base pairs and often to just an entire chromosome or chromosome arm. Two different groups have previously used this type of analysis to investigate cartilaginous tumors.5,6 One group showed a limited resolution, with findings reported similar to classical cytogenetics. The second group initially reported similar findings6 but later, with an increased number of samples, reported gains of 20q12-qter, 8q24.1-qter, 20p, and 14q24-qter (24%) and losses of Xcen-q21, 6cen-q22, and 18cen-q11.2.7 Although this latter report provided important information, the resolution of specific genomic changes was still in the size of 10 to 15 million base pairs or greater. Array CGH is a relatively new microarray-based interrogation of regions of DNA loss or gain on a genome-wide basis. Compared with metaphase CGH, the resolution of genomic change with array CGH is potentially within 1 to 2 million base pairs of gain or loss.8 One such change noted from our array CGH data of 15 high-grade chondrosarcomas was amplification of 8q24.12-q24.13 over an approximate 2 to 3 million base pair region. The most notable gene of importance in this region related to tumor biology is MYC (v-myc myelocytomatosis viral oncogene homolog [avian]; alias c-Myc). In parallel to CGH and to validate our findings in a clinically relevant fashion, we used fluorescent in situ hybridization (FISH) on a different sample set consisting of more than 100 specimens of cartilaginous tumors with clinical follow-up. The findings presented here are an important first step in a genome-wide analysis of chondrosarcomas.
Specimens Frozen specimens of 15 high-grade chondrosarcomas used for array CGH were acquired through the Midwestern Cooperative Human Tissue Network (Ohio State University, Columbus, OH). Specimens used for tissue microarray (TMA) validation (FISH) and clinical correlation consisted of 19 enchondromas, 44 grade 1 chondrosarcomas, 29 grade 2 chondrosarcomas, 10 grade 3 chondrosarcomas, and 14 dedifferentiated chondrosarcomas. The Ohio State University Institutional Review Board provided approval for this study. All samples of chondrosarcoma with less than complete negative surgical resection margins were excluded from this study.
Array CGH: Genomic DNA Isolation, Labeling, and Hybridization to Microarrays For DNA labeling, test and reference genomic DNA (10 µg) were digested using 10 µL of React3 10x buffer and 2.5 µL of EcoR1 enzyme (10 U/µL; Invitrogen, Carlsbad, CA). After digestion, 2 µg of the test and reference DNA was labeled via random priming using a BioPrime DNA Labeling Kit (Invitrogen Ltd, Paisely, United Kingdom) with Cy3-dUTP and Cy5-dUTP (Amersham, Piscataway, NJ). After removal of unincorporated nucleotides, purified labeled test and reference DNA were combined and mixed with 50 µg of human Cot1 DNA (Invitrogen), 100 µg of transfer RNA (Sigma, St Louis, MO), 20 µg of poly A, 20 µg of poly T, and 450 µL of TE buffer. Labeled DNA was then denatured and hybridized in a custom-made hybridization chamber at 65°C for 16 to 24 hours. The arrays were then washed and dried by centrifugation before scanning. All samples were labeled in the reverse direction and hybridized to two arrays, such that tumor DNA was evaluated with both Cy3 and Cy5 (dye-swap test). For controls, one normal male and female DNA were labeled and compared with a pool of normal male and female DNA, respectively.
Array CGH: Image Acquisition and Data Analysis Log2 expression ratios were computed for each spot on each array followed by a median normalization method, adding a scalar to each log ratio of a given array to shift the median log ratio of the array to zero. Normalized log ratios for duplicate spots within an array were averaged, and these values were then averaged for forward and reverse fluorophore (ie, dye-swap) experiments after reversing the sign of the log ratios for the dye-swap experiment. Clones were ordered by chromosomal position according to the University of California, Santa Cruz Golden Path draft (July 2003 update; Santa Cruz, CA) human genome sequence (http://genome.ucsc.edu). To reduce experimental noise, a smoothed local average [yi = 0.1x(i – 2) + 0.2x(i – 1) + 0.4xi + 0.2x(i + 1) + 0.1x(i + 2)] was computed for each spot on each array. The thresholds for the log2 ratio of gains and losses were set at the log2 ratio of +0.30 and –0.31, respectively, to include the upper and lower 5% of the median normalized fluorescent log2 ratio values for all spots (clones). To summarize the data for all samples and identify regions of recurrent gain or loss, changes for each chromosome were then evaluated using horizontal box plots. We identify potential recurrent gains or losses to be those regions with a gain or loss in at least 25% of the patient samples based on estimation that the probability of type I errors (false positives) in a given region is 0.0055 or less at this level. As a means of visually displaying the data for chromosome 8, each log2 ratio value of more than +0.30 or less than –0.31 was assigned a value of 1 or –1, respectively, whereas values between these two values were assigned a value of 0. This data was then imported into an Excel worksheet, and values of 1, –1, and 0 were displayed as a red, green, and black colorgram, respectively.
TMA
Immunohistochemistry and FISH
FISH for MYC amplification using a commercially available probe (LSI MYC; Vysis, Downers Grove, IL) was performed in accordance with the manufacturer's guidelines and performed manually. Briefly, formalin-fixed, paraffin-embedded TMA blocks were cut into 3–to 4–µm thick sections, incubated overnight at 56°C, deparaffinized, washed, digested with protease, formalin fixed, denatured, and hybridized at 37°C for 16 hours. The slides were then washed in a posthybridization wash, counterstained with 4'-6-diamidino-2-phenylindole, and covered with a coverslip. Specimens were evaluated with an Olympus BX51 microscope (Olympus Optical Company, LTD, Tokyo, Japan) under oil immersion at x100 magnification using the recommended filters. A ratio of the total number of MYC signals to the total number of CEP8 signals in at least 60 interphase nuclei with nonoverlapping nuclei in the tumor cells was determined. Cells with no signals or with signals of only one color were disregarded. Tumor cells displaying at least two centromeric chromosome 8 signals and multiple MYC signals, with an MYC:CEP8 ratio of
Statistical Analysis: FISH
Array CGH Chromosome 8 Chromosome 8 spans approximately 146 million base pairs and represents between 4.5% to 5% of the total DNA. The array used in this study contained 130 clones interrogating chromosome 8, giving coverage of one clone on average for every 1.1 mega base pairs. Copy number alterations on chromosome 8 representing gain for all samples were present in 4.0% of clones, whereas losses were present in 5.6% of clones (Fig 1A), compared with 4.3% and 5.8% for gain and loss, respectively, of all clones on other chromosomes (complete data not shown). Each column of the colorgram represents a single sample, with gains illustrated as red, losses illustrated as green, and no change illustrated as black for each clone in relation to the chromosome 8 ideogram. Overall, the copy number alterations for chromosome 8 varied from 1 to several million base pairs and seemed to be noncontiguous and random, except for one locus at 8q24.12-q24.13, the region of the MYC oncogene, involving clones RP11-89K10, RP11-79E8, and RP11-94M13. Results of the horizontal box plot (Fig 1B) show that gain at 8q24.12-q24.13 is the only change at the cutoff values of +0.3 and –0.3 present in greater than 25% of all samples. Each horizontal gray box represents one clone on the array. The line in the middle of each box represents the median value of 15 chondrosarcomas for that clone, the left and right ends of the box give the 25th and 75th percentiles, and the left and right whiskers give the 10th and 90th percentiles. The red-dashed lines indicate the threshold values for calling the filtered log ratio a gain or loss (±0.3). The region at 8q24.12-q24.13, the site of MYC, represented an amplicon of approximately 3.5 million base pairs at chr8:130,688,249 to 134,103,781. Besides the MYC oncogene, there are 11 other known genes either within or immediately adjacent to this region that include, from centromeric to telomeric, C8FW, NSE2, POU5FLC20, AK093424, AF336886, MGC27434, MLZE, DDEF1, ADCY8, D63477, and KCNQ3. No loci for which the 25th percentile approaches or exceeds the threshold value of –0.3, indicative of a recurrent region of genomic loss on chromosome 8, were identified.
FISH For the 116 samples represented in the TMA, there was no evidence of MYC amplification in any samples of enchondroma or grade 1 chondrosarcoma (Table 1). High-grade chondrosarcomas (grade 2 to 3 and dedifferentiated) all showed a similar rate of MYC amplification (15% to 21%). For all three samples of dedifferentiated chondrosarcoma with MYC amplification, the associated lower grade component did not show amplification. The pattern of MYC amplification for all samples consisted of tight clusters of signal generally restricted to one or two areas of the nucleus (Fig 2B). The presence of high-level amplification (> 20 to 25 copies per cell) was characteristic of all samples with MYC amplification and indicative of an intra- or interchromosomal tandem duplication (homogeneously staining regions) of the MYC gene. There was no difference in the occurrence of MYC amplification in axial versus appendicular tumors (P = .487).
In contrast to MYC amplification, polysomy 8 was identified in several grade 1 chondrosarcomas (18%). This was different from the results for enchondroma, in which no increased numbers of chromosome 8 were detected (P = .047). Although polysomy 8 was quite frequent in grade 1 chondrosarcomas, the rate was much higher (P = .003) in high-grade chondrosarcomas. Polysomy 8 was present in all grade 3 tumors (Fig 2D) and in approximately one third of all grade 2 and dedifferentiated tumors, without MYC amplification. Of the nine total samples in the validation study with MYC amplification, two showed polysomy 8, and the others showed no increased CEP8 signals.
Immunohistochemistry
Survival Analysis Of the 116 patients in this study, 53 met the requirement for PFS, and of this group, 47 met the requirement for OS. The six patients in the PFS group not meeting the criteria for OS analysis consisted of patients with locally recurrent resectable disease but with a limited follow-up after this event. For patients with an adverse event, the median follow-up to the time of event was 12 months (average, 40 months; range, 3 to 156 months). For patients with no adverse event, the median follow-up was 84 months (average, 108 months; range, 60 to 276 months). Of this group of 53 patients, there were 36 adverse events, which resulted in the death of the patient in 25 instances. There was a near even distribution of axial (45%) versus appendicular tumors (55%) for this group of patients, with almost three fourths of the patients (72%) having grade 1 or 2 chondrosarcoma versus grade 3 or dedifferentiated chondrosarcoma. In this group of 53 patients, there were five patients with MYC amplification and 19 patients with polysomy 8.
The median OS time was 5.0 years. A Kaplan-Meier survival curve (Fig 3) for OS showed a statistically significant difference (log-rank test survivor function, P = .034) for patients with MYC amplification (median, 1.1 year; range, 0.5 to 12.0 years; 95% CI, 0.5 years to
Univariate analysis involving Cox proportional hazards models showed that grade (P = .003), polysomy 8 (P = .045), and MYC amplification (P = .053) correlated with shorter OS (Table 3). For univariate analysis of PFS, only grade (P = .035) was important. For multivariate analysis of OS, grade of chondrosarcoma (P = .026) was the only factor to reach statistical significance. For multivariate analysis of PFS, none of the factors analyzed reached a statistically significant P value. Expression of the MYC protein was too frequent for any meaningful statistical analysis using these models.
At the current time, the only proven treatment for chondrosarcoma is complete resection. As surgery has improved over the last 30 to 40 years, the OS rate for chondrosarcoma has remained flat.10 This emphasizes the need to understand the genetic pathways involved in chondrosarcoma as gene-targeted therapy progresses. In the context of markers of diagnostic or prognostic importance, unique site-specific genome copy number alterations do not have to identify a specific gene but merely a segment of DNA from several kilo base to mega base pairs in size that contain one or more potential genes of interest.11 There are several limitations to this study in the context of biologic interpretation of the results that relate to this issue. The presence of polysomy 8 in a high proportion of patients with a poor outcome, with or without MYC amplification, possibly implies other candidate oncogenes on chromosome 8 not discovered by our analysis. Other important oncogenes on chromosome 8 could be amplified at lower levels that we did not detect because we set our threshold values of gain or loss to detect those genomic regions with the greatest change. To further complicate this issue, some of these potential oncogenes, such as tribbles homolog 1 (Drosophila) (C8FW, or TRIB1) and development and differentiation enhancing factor 1 (DDEF1), are within 2 to 3 mega base pairs of MYC. Another distinct possibility is that polysomy 8 acts as a marker of chromosomal instability in chondrosarcoma. As strictly a prognostic marker, neither of these arguments has little value because the mere identification of the specific genomic change can serve as a clinical end point. In regard to development of potential molecular-targeted therapies, these issues could be critically important. Data analysis of array CGH data does influence the final results of any such study, and a number of somewhat complicated models have been put forth for this purpose12,13 but at the risk of introducing systematic errors that are difficult to identify. We were particularly concerned with this problem given the known karyotypic complexity of high-grade chondrosarcomas. We did combine data from two measurements with dye reversal as one step to reduce errors. Our threshold of gain or loss was a statistically simple model based on the distribution of overall changes and designed to emphasize those loci with the highest level of change. In addition, we smoothed the average ratio on neighboring array elements to improve the thresholding process at the cost of potentially blurring the boundaries of any such change. Undoubtedly, this type of analysis will miss some loci of gain or loss occurring at lower threshold values. We have initially accepted this latter fault in an attempt to initially identify the most obvious areas of genomic change in this group of tumors. The results of our immunohistochemical studies would support that MYC is at least one important oncogene in the 8q24.12-q24.13 region. All of our samples with MYC amplification showed high levels of MYC protein expression. That MYC was expressed in a number of samples without MYC amplification, some of which showed polysomy 8, is not surprising. Both MYC amplification and protein overexpression have been demonstrated in a wide variety of tumor types.14-16 Amplification is only one mechanism of protein overexpression of a gene, and other mechanisms, such as increased transcription or translation and/or decreased mRNA degradation, could play a role in samples without such genomic changes. Our survival data support the assertion that polysomy 8 and/or MYC amplification is a much better marker of prognostic importance than MYC protein expression. This is not surprising given similar findings for other oncogenes, such as HER-2, in other types of cancer.17 Amplification of MYC in chondrosarcoma was first reported approximately 15 years ago by the same group in two separate publications.18,19 In these studies, amplification of MYC was noted in two of nine chondrosarcomas and one osteosarcoma. These studies were limited by the use of dot blot and Southern blot procedures, which do not distinguish between true gene amplification and polysomy 8, and lacked a significant number of patients with clinical follow-up. The only report to date of array CGH in chondrosarcoma involves a single chondrosarcoma cell line.20 In this study, gene expression profiling was combined with array CGH in the FSCP-1 cell line to identify up- or downregulated expression of genes corresponding to regions of DNA loss or gain. The results of this study in the FSCP-1 cell line revealed an overall heterogeneous pattern of DNA gains and losses, with MYC being the only gene with a marked upregulation of expression located in a region of DNA copy number gain. In regard to therapeutic potential of MYC amplification and protein overexpression, many facets of cancer-related biologic activities, including transformation, immortalization, blockage of cell differentiation, and induction of apoptosis,21 have been identified. Recent studies have shown that inactivation of the MYC oncogene alone can reverse tumorigenesis in various cancer cell lines.22,23 Of therapeutic importance, short interfering RNA-mediated gene silencing of MYC has also recently been demonstrated.24 Of added importance, this downregulation of MYC expression resulted in an increased sensitivity of HeLa cells to radiation used in this latter study. To summarize, we have shown that MYC is amplified and overexpressed in a frequent percentage (15% to 20%) of high-grade chondrosarcomas. This change has little utility in the diagnostic assessment of chondrosarcoma because enchondroma and grade 1 chondrosarcoma showed no evidence of MYC amplification. Polysomy 8 may have some utility in this regard because it was identified in almost 20% of grade 1 chondrosarcomas but none of the enchondromas. For grade 2 and higher chondrosarcomas, MYC amplification and polysomy 8 are prognostic markers of poor outcome. The therapeutic implications of this finding are less clear, but potential methodologies of MYC inactivation have been demonstrated.
The authors indicated no potential conflicts of interest.
We thank Scott Jewel, PhD, and Laurie Johnson of the Midwestern Human Cooperative Tissue Network who helped in the procurement of samples for this study.
Supported by the developmental funds of C.M. as a joint venture of the Department of Pathology and Department of Orthopedic Oncology at The Ohio State University Medical School. Authors' disclosures of potential conflicts of interest are found at the end of this article.
1. Mandahl N, Gustafson P, Mertens F, et al: Cytogenetic aberrations and their prognostic impact in chondrosarcoma. Genes Chromosomes Cancer 33:188-200, 2002[Medline] 2. Tallini G, Dorfman H, Brys P, et al: Correlation between clinicopathological features and karyotype in 100 cartilaginous and chordoid tumours: A report from the Chromosomes and Morphology (CHAMP) Collaborative Study Group. J Pathol 196:194-203, 2002[CrossRef][Medline] 3. Bridge JA, Bhatia PS, Anderson JR, et al: Biologic and clinical significance of cytogenetic and molecular cytogenetic abnormalities in benign and malignant cartilaginous lesions. Cancer Genet Cytogenet 69:79-90, 1993[CrossRef][Medline] 4. Sjogren H, Orndal C, Tingby O, et al: Cytogenetic and spectral karyotype analyses of benign and malignant cartilage tumours. Int J Oncol 24:1385-1391, 2004[Medline] 5. Ozaki T, Wai D, Schafer KL, et al: Comparative genomic hybridization in cartilaginous tumors. Anticancer Res 24:1721-1725, 2004[Medline] 6. Larramendy ML, Tarkkanen M, Valle J, et al: Gains, losses, and amplifications of DNA sequences evaluated by comparative genomic hybridization in chondrosarcomas. Am J Pathol 150:685-691, 1997[Abstract] 7. Larramendy ML, Mandahl N, Mertens F, et al: Clinical significance of genetic imbalances revealed by comparative genomic hybridization in chondrosarcomas. Hum Pathol 30:1247-1253, 1999[CrossRef][Medline] 8. Inazawa J, Inoue J, Imoto I: Comparative genomic hybridization (CGH)-arrays pave the way for identification of novel cancer-related genes. Cancer Sci 95:559-563, 2004[CrossRef][Medline] 9. Morrison C, Marsh W Jr, Frankel WL: A comparison of CD10 to pCEA, MOC-31, and hepatocyte for the distinction of malignant tumors in the liver. Mod Pathol 15:1279-1287, 2002 10. Fiorenza F, Abudu A, Grimer RJ, et al: Risk factors for survival and local control in chondrosarcoma of bone. J Bone Joint Surg Br 84:93-99, 2002[CrossRef][Medline] 11. Hill C, Hunter SB, Brat DJ: Genetic markers in glioblastoma: Prognostic significance and future therapeutic implications. Adv Anat Pathol 10:212-217, 2003[CrossRef][Medline] 12. Price TS, Regan R, Mott R, et al: SW-ARRAY: A dynamic programming solution for the identification of copy-number changes in genomic DNA using array comparative genome hybridization data. Nucleic Acids Res 33:3455-3464, 2005 13. Myers CL, Dunham MJ, Kung SY, et al: Accurate detection of aneuploidies in array CGH and gene expression microarray data. Bioinformatics 20:3533-3543, 2004 14. Jenkins R, Qian J, Lieber M, et al: Detection of c-myc oncogene amplification and chromosomal anomalies in metastatic prostatic carcinoma by fluorescence in situ hybridization. Cancer Res 57:524-531, 1997 15. Melhem M, Meisler A, Finley G, et al: Distribution of cells expressing Myc proteins in human colorectal epithelium, polyps, and malignant tumors. Cancer Res 52:5853-5864, 1992 16. Liao D, Dickson R: C-Myc in breast cancer. Endocr Relat Cancer 7:143-164, 2000[Abstract] 17. Borg A, Tandon AK, Sigurdsson H, et al: HER-2/neu amplification predicts poor survival in node-positive breast cancer. Cancer Res 50:4332-4337, 1990 18. Castresana JS, Barrios C, Gomez L, et al: Amplification of the c-myc proto-oncogene in human chondrosarcoma. Diagn Mol Pathol 1:235-238, 1992[Medline] 19. Barrios C, Castresana JS, Ruiz J, et al: Amplification of c-myc oncogene and absence of c-Ha-ras point mutation in human bone sarcoma. J Orthop Res 11:556-563, 1993[CrossRef][Medline] 20. Schorle CM, Verdorfer I, Finger F, et al: Comparative analysis of imbalances in genomic DNA and mRNA expression levels in chondrosarcoma-derived cell line FSCP-1. Int J Oncol 25:1651-1660, 2004[Medline] 21. Nesbit C, Tersak J, Prochownik E: MYC oncogenes and human neoplastic disease. Oncogene 18:3004-3016, 1999[CrossRef][Medline] 22. Jain M, Arvanitis C, Chu K, et al: Sustained loss of a neoplastic phenotype by brief inactivation of MYC. Science 297:102-104, 2002 23. Shachaf CM, Felsher DW: Tumor dormancy and MYC inactivation: Pushing cancer to the brink of normalcy. Cancer Res 65:4471-4474, 2005 24. An J, Xu QZ, Sui JL, et al: Downregulation of c-myc protein by siRNA-mediated silencing of DNA-PKcs in HeLa cells. Int J Cancer 117:531-537, 2005[CrossRef][Medline] Submitted August 4, 2005; accepted September 23, 2005.
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
2, were considered consistent with amplification of the MYC gene. Tumor cells displaying at least two centromeric chromosome 8 signals and an equal number of MYC signals, with an MYC:CEP8 ratio of less than 2, were considered consistent with no amplification of the MYC gene. Tumor cells displaying multiple CEP8 signals and an approximately equal number of MYC signals, with a somewhat random distribution of both, were considered polysomy 8. 
2 analysis was used for comparison of individual surgical and pathologic variables with MYC amplification and polysomy 8. Both univariate and multivariate logistic regression analyses were performed to investigate relationships between surgical and pathologic variables and outcome. For statistical purposes, site and grade as variables were reduced to two categories (appendicular v axial and grades 1 and 2 v grade 3 and dedifferentiated, respectively). For progression-free survival (PFS), patients were required to have time of first adverse event recorded or a minimum of 5 years of follow-up with no adverse event. In addition to the requirements for PFS, for overall survival (OS), patients were required to have a minimum of 5 years of follow-up after the first adverse event if this was a resectable recurrence that did not result in patient death. Only patients with initial complete tumor resection with adequate margins were used for PFS or OS analysis. Only patients with chondrosarcoma were included in survival analysis, with exclusion of all patients with enchondroma. Specifically, there were 23 grade 1, 15 grade 2, seven grade 3, and eight dedifferentiated chondrosarcomas in the group for survival analysis. Patients not included in the survival analysis and present in the univariate or multivariate analysis were excluded as a result of less than 5 years of follow-up. Kaplan-Meier curves for PFS and OS were compared using the log-rank test. 
) or polysomy 8 (median, 2.0 years; range, 0.5 to 15.0 years; 95% CI, 0.5 to 12.0 years). The results for PFS (
