Comparative Genomic Hybridization of Postirradiation Sarcomas Maija Tarkkanen, M.D., Ph.D. 1,2 Tom A. Wiklund, M.D., Ph.D. 3 Martti J. Virolainen, M.D., Ph.D. 4 Marcelo L. Larramendy, Ph.D. 1,2,5 Nils Mandahl, Ph.D. 6 Fredrik Mertens, M.D., Ph.D. 6 Carl P. Blomqvist, M.D., Ph.D. 3 Erkki J. Tukiainen, M.D., Ph.D. 7 Markku M. A. Miettinen, M.D., Ph.D. 4 A. Inkeri Elomaa, M.D., Ph.D. 3 Y. Sakari Knuutila, Ph.D. 1,2 1 Department of Medical Genetics, Haartman Insti- tute, University of Helsinki, Helsinki, Finland. 2 Laboratory of Medical Genetics, Helsinki Univer- sity Central Hospital, Helsinki, Finland. 3 Department of Oncology, Helsinki University Cen- tral Hospital, Helsinki, Finland. 4 Department of Pathology, Haartman Institute, University of Helsinki, Finland. 5 Laboratorio de Citogenética y Cátedra de Citolo- gia, Facultad de Ciencias Naturales y Museo, Na- tional University of La Plata, La Plata, Argentina. 6 Department of Clinical Genetics, Lund University Hospital, Lund, Sweden. 7 Department of Plastic Surgery, Helsinki Univer- sity Central Hospital, Helsinki, Finland. Supported by the Clinical Research Institute of Helsinki University Central Hospital, Helsinki Uni- versity Central Hospital, Leiras Research Founda- tion, Swedish Cancer Society, and Nordic Cancer Union. Address for reprints: Maija Tarkkanen, M.D., Ph.D., Department of Medical Genetics, Haartman Insti- tute, University of Helsinki and Laboratory of Med- ical Genetics, Helsinki University Central Hospital, Haartmaninkatu 3, 4th Flr. (Box 21), FIN-00014 University of Helsinki, Finland; Fax: 358-(0)9-191 26788; E-mail: maija.tarkkanen@helsinki.fi Received December 7, 2000; revision received May 23, 2001; accepted June 19, 2001. BACKGROUND. Radiotherapy is a known risk factor for sarcoma development. Postirradiation sarcomas arise within the radiation field after a latency period of several years and usually are highly malignant. Very little is yet known about their genetic changes. METHODS. Twenty-seven postirradiation sarcomas were analyzed by comparative genomic hybridization, which allows genome-wide screening of DNA sequence copy number changes. RESULTS. Copy-number aberrations were detected in 20 (74%) tumors. The mean number of aberrations per tumor was 5.3 with gains outnumbering losses. The most frequent gains affected the minimal common regions of 7q11.2-q21 and 7q22 in 30% and 7p15-pter in 26%. Gain of 8q23-qter was detected in 22%. The most frequent losses affected 11q23-qter and 13q22-q32 in 22%. In osteosarcomas, the most frequent aberration was loss of 1p21-p31, in malignant fibrous histiocytomas (MFH) gain of 7cen-q22, and in fibrosarcomas gain of 7q22. The findings in postirradiation osteosarcomas and MFHs were compared with findings in sporadic osteosarcomas and MFHs, reported previously by the authors. In sporadic osteo- sarcomas, gains outnumbered losses, but, in postirradiation osteosarcomas, losses were more frequent than gains. Loss at 1p was rare in sporadic osteosarcoma (3%) but frequent (57%) in postirradiation osteosarcomas. Gains at 7q were frequent both in postirradiation and sporadic MFH. CONCLUSIONS. According to previous studies on different types of sporadic sarco- mas, gains at 7q or 8q are associated with poor prognosis or large tumor size. Thus, the frequent gains at 7q and 8q might have been responsible in part for the poor prognosis of postirradiation sarcomas. Also, however, some of their clinical fea- tures, i.e., high malignancy grade, late diagnosis, and central location, are associ- ated with a poor prognosis. Cancer 2001;92:1992– 8. © 2001 American Cancer Society. KEYWORDS: postirradiation sarcoma, comparative genomic hybridization, angiosar- coma, osteosarcoma, malignant fibrous histiocytoma (MFH), fibrosarcoma. The majority of soft tissue and bone sarcomas arises spontaneously, but, in some patients, ionizing radiation can be identified as a predisposing factor. Criteria given by Cahan et al.1 define postirradi- ation sarcomas as follows: evidence of initial nonmalignancy, devel- opment of the second malignant tumor in the irradiated tissue, a relatively long latency period, and histologic confirmation of sarcoma. Later, the first criterion was modified to include also malignant tu- mors of a different kind than the subsequent sarcoma.2,3 The most common primary malignancies are breast carcinoma and cancers of the female reproductive organs according to the population-based study by Wiklund et al.4 Median time interval between radiotherapy and the development of sarcoma in that study was 13.2 years, and the 1992 © 2001 American Cancer Society most common histologic types were osteosarcoma, malignant fibrous histiocytoma (MFH), and fibrosar- coma.4 The prognosis is poor, with 5-year survival rates varying from 11–29%.4,5 The prognosis is, thus, clearly worse than in sporadic bone and soft tissue sarcomas with 5-year overall survival rates of 65% and 64%.6,7 Very little is known about the genetic changes involved in the tumorigenesis of postirradiation sar- comas. According to a recent study and review by Mertens et al.8 regarding cytogenetic changes in post- irradiation sarcoma, these tumors have complex karyotypes with loss of 3p21-pter being more frequent than in sporadic sarcomas. Also, polyclonal tumors with near-diploid chromosome numbers were ob- served.8 Another recent cytogenetic study and review pointed out two distinct patterns: 1) polyclonal karyo- types, often with simple and balanced translocations, preferentially observed after long-term culturing and 2) monoclonal chromosomal alterations observed in highly aneuploid and complex karyotypes, usually de- tected in short-term cultures or in xenografts.9 Alter- ations of the RB1 and TP53 tumor suppressor genes are frequent10,11 and according to the study by Nakan- ishi et al.10, the frequency of TP53 mutations is higher in postirradiation sarcomas than in sporadic sarco- mas. Comparative genomic hybridization (CGH) allows for a genome-wide screening of losses, gains, and am- plifications of DNA sequence copy numbers12,13 (for reviews of amplifications and losses see Knuutila et al.,14,15 which also are available at URL http://www. helsinki.fi/;lgl_www/CMG.html). In the current study, CGH was used to study 27 postirradiation sar- comas for the purpose of analyzing changes relevant to the tumorigenesis of these tumors. In addition, the findings in postirradiation osteosarcoma and MFH were compared with findings in sporadic osteosar- coma and MFH, reported previously by us.16,17 To the best of our knowledge, this is the first report on po- stirradiation sarcoma studied by CGH (search covered Medline and pre-Medline, 1992–2000, online literature databases). MATERIALS AND METHODS Materials Twenty-seven postirradiation sarcomas of 26 patients were analyzed by CGH. The tumors were selected to represent the most common histologic subtypes of postirradiation sarcoma, i.e., osteosarcoma, MFH, and fibrosarcoma,4 on the basis of availability of fresh- frozen or paraffin-embedded tumor tissue. As angio- sarcoma is very rare as a sporadic tumor but is a typical postirradiation sarcoma, this subtype was in- cluded also. Twenty-four tumors were identified from the Finnish Cancer Registry4 and from patients treated by the Soft Tissue Sarcoma group at the Helsinki Uni- versity Central Hospital.6 Three tumors (Cases 4, 11, and 12) were from Sweden, and their CGH results have been published previously.8 The material consisted of three angiosarcomas, seven osteosarcomas, eight MFHs, and nine fibrosarcomas (Table 1). The median time interval between the first, primary tumor and subsequent postirradiation sarcoma was 11 years (range, 4 –33 yrs). Fifteen samples were from paraffin- embedded material, and 12 were fresh-frozen tumor samples. In addition, 3 skin samples from irradiated but tumor-free areas from Patients 5, 20, and 26 were analyzed by CGH. Comparative Genomic Hybridization CGH was performed according to standard proce- dures13 with the modification of using a mixture of fluorochromes conjugated to dCTP and dUTP nucle- otides for nick translation.18 Hybridizations, washings, and ISIS digital image analysis (Metasystems GmbH, Altlussheim, Germany) were performed as described elsewhere.19 In each CGH experiment, a negative con- trol (peripheral blood DNA from a healthy donor) and a positive control (DNA from a tumor with known DNA copy number aberrations) were included. Based on our earlier reports and the control results, 1.17 and 0.85 were used as cutoff levels for gains and losses, respectively. Gains exceeding the 1.5 threshold were termed high-level amplifications. All CGH results were confirmed by a 99% confidence interval.20 RESULTS DNA sequence copy number aberrations were de- tected in 20 of 27 (74%) postirradiation sarcomas (Ta- ble 2). The mean number of changes per tumor was 5.3, with means of 3.0 gains, 2.3 losses, and 0.3 high- level amplifications per tumor. Gains were more fre- quent than losses by a ratio of 1.3:1. No copy number aberrations were detected in seven tumors: five were paraffin-embedded samples, and two were fresh-fro- zen tissue samples. The most frequent DNA sequence copy number gains affected the minimal common regions of 7q11.2- q21 and 7q22 in 30%, 7p15-pter in 26%, and 8q23-qter and 15q24-qter in 22% of the 27 samples (Table 3). The most frequent losses affected 11q23-qter and 13q22- q32, seen in 22% of the samples. Recurrent high-level amplifications (green to red ratio . 1.5) were not detected. All aberrations in all tumor types are shown in Figure 1. CGH of Postirradiation Sarcoma/Tarkkanen 1993 Postirradiation Angiosarcomas DNA sequence copy number aberrations were de- tected in two out of three postirradiation angiosarco- mas, both showing a single gain, one in 11q14-q23 and the other in 8q23-qter (Table 2). Postirradiation Osteosarcomas DNA sequence copy number aberrations were de- tected in six out of seven postirradiation osteosarco- mas. The mean number of aberrations per sample was 8.4, with means of 3.7 gains, 4.7 losses, and 0.4 high- level amplifications per sample. Losses were more fre- quent than gains with the ratio of 1.3:1. The most frequent aberration was loss of 1p21-p31 in 4 samples (57%). Postirradiation MFHs DNA sequence copy number aberrations were de- tected in five out of eight postirradiation MFHs. The mean number of aberrations per sample was 2.9 (1.8 gains, 1.1 losses, and no high-level amplifications). Gains were more frequent than losses (1.6:1). The most frequent aberration was gain of 7cen-q22, which was seen in 3 samples (38%). Postirradiation Fibrosarcomas DNA sequence copy number aberrations were de- tected in seven out of nine postirradiation fibrosarco- mas. The mean number of aberrations per sample was 6.6 (4.2 gains, 2.3 losses, and 0.6 high-level amplifica- tions with the ratio of 1.8:1 of gains to losses). The TABLE 1 Clinical and Histopathologic Characteristics of 27 Postirradiation Sarcomas Casea Gender Ageb Primary tumor Postirradiation sarcomac Type of sarcoma sampled Location Prior chemotherapy Time interval to PIS (yrs) 1 F 75 Breast carcinoma Angiosarcoma G4 P/F Skin of breast No 4 2 F 78 Bladder carcinoma Angiosarcoma G4 LR/P Abdominal wall No 5 3 F 43 Breast carcinoma Angiosarcoma G4 P/F Skin of breast No 6 4 F 17 Ewing sarcoma Skeletal osteosarcoma OB G4 P/F Humerus Yes 6 5 F 73 Breast carcinoma Extraskeletal osteosarcoma CB G4 P/F Armpit No 23 6 F 57 Giant cell tumor of bone Extraskeletal osteosarcoma CB G4 P/F Lower lege No 33 7 (7) F 39 Ovarian carcinoma Extraskeletal osteosarcoma G3 P/P Abdominal wall Yes 11 8 (27) F 66 Uterus carcinoma Skeletal osteosarcoma G3 P/P Pelvis No 7 9 (15) F 52 Ovarian carcinoma Extraskeletal osteosarcoma FS G3 P/P Pelvis, pararectal space Yes 11 10 (6) F 27 Breast carcinoma Skeletal osteosarcoma CB G4 P/P Scapula Yes 7 11 F 51 Breast carcinoma MFH NOS G4 P/F Shoulder, superficial Yes 4 12 M 37 Astrocytoma G2–3 MFH NOS G4 P/F Scalp No 5 13 F 51 Breast carcinoma MFH MYX G3 LR/P Thoracic wall No 11 14 F 60 Breast carcinoma MFH ST-PL G4 P/P Shoulder No 10 15 (19) F 63 Uterine cervix carcinoma MFH GCT G3 P/P Abdominal wall No 12 16 (30) F 82 Breast carcinoma MFH G4 P/P Subcutis of proximal humerus No 16 17 (12) F 59 Papillary thyroid carcinoma MFH MYX G2 P/P Neck No 21 18 F 34 Chondrosarcoma MFH ST-PL G4 P/F Amputation stump Yes 12 19 (28) F 73 Larynx carcinoma Fibrosarcoma G3 P/P Neck No 14 20a F 50 Thyroid carcinoma Fibrosarcoma G2 P/F Neck No 18 20b F 50 Thyroid carcinoma Fibrosarcoma G3 LR/F Neck No 18 21 F 36 Breast carcinoma Fibrosarcoma G4 P/P Thoracic wall No 6 22 F Breast carcinoma Fibrosarcoma G4 P/F Thoracic wall No 11 23 (33) M 84 Lymphoma Fibrosarcoma G3 P/P Back, scapular region No 7 24 (18) F 66 Breast carcinoma Fibrosarcoma G4 P/P Thoracic wall, axilla No 15 25 (5) F 20 Hodgkin disease Fibrosarcoma G4 P/P Thoracic wall Yes 11 26 M 52 Lymphoma Fibrosarcoma G4 P/F Thoracic wall Yes 8 G: grade. a Corresponding case numbers in Wiklund et al., 19914 in parentheses. b Age at diagnosis of postirradiation sarcoma. c Subtypes of postirradiation sarcoma (PIS), OB: osteoblastic, CB: chondroblastic, FS: fibrosarcomatous, MYX: myxoid, GCT: giant cell type, ST-PL: storiform-pleomorphic. d Type of sarcoma sample, P: primary tumor, LR: local recurrence, P: paraffin sample, F: fresh-frozen sample. e Soft tissue tumor attached to the tibia. 1994 CANCER October 1, 2001 / Volume 92 / Number 7 most frequent aberration was gain of 7q22, which was seen in 5 samples (56%). Skin Samples from Irradiated Tumor-Free Areas No copy number aberrations were seen in the skin samples taken from irradiated but tumor-free areas from Patients 5, 20, and 26. DISCUSSION The current study describes genetic changes in postir- radiation sarcomas focusing on the most common histologic subtypes. Overall, the most frequent aber- rations were copy number gains affecting chromo- some 7, with the minimal common regions of 7q11.2- q21 and 7q22 in 30% and 7p15-pter in 26% of the tumors. Of these, 7q11.2-q21 and 7q22 were affected most frequently in MFH and fibrosarcoma, and 7p15- pter was affected more often in osteosarcoma and fibrosarcoma. As reported previously by us, gains af- fecting chromosome 7 are frequent also in sporadic MFH17 and sporadic osteosarcoma.16 To the best of our knowledge, there is not yet any data on copy number changes, studied by CGH, in sporadic fibro- sarcoma. Expression of the hepatocyte growth factor HGF (gene located in 7q21) is common in sarcomas,21 and, thus, this gene may be a target for copy number increases. No proof for such an association has been presented yet, and the possible other targets for gains on chromosome 7 are not yet known. Gains were also common at the distal part of 8q (8q23-qter), most frequently in osteosarcoma. Gains at distal 8q have been also found to be frequent in spo- radic osteosarcoma,16 chondrosarcoma,20 and leiomy- osarcoma.22 Furthermore, gain of the entire chromo- some 8 is frequent in Ewing sarcoma and related tumors.23 Taken together, overrepresentation of se- quences at 8q seems to have a significant role in different sarcoma types. Chromosomal band 8q24 contains the MYC locus, but amplification of MYC is an uncommon event in sporadic osteosarcoma ac- cording to Ladanyi et al.24 However, recently an in- creased MYC copy number was seen 7 of 16 osteosar- comas25 and overexpression of MYC was frequent particularly in osteosarcomas which later relapsed.26 MYC amplification has been reported only in a small TABLE 2 DNA sequence Copy Number Changes in 27 Postirradiation Sarcomas by Comparative Genomic Hybridization (CGH) Case CGH findings 1 111q14-q23 2 no changes 3 18q23-qter 4 11pter-q23, 12p22-q21, 22q33-qter, 13q24-q25, 15p, 25cen-q31.1, 19pter-q31, 211q13-qter, 115cen-q15, 119p, 120q 5 21p13-p31, 22p23-pter, 23cen-p14, 26q12-qter, 17pter-q21, 210p11.2-q22, 112p, 112cen-q21/1112q13-q21, 213q22-qter, 114q/1114q31-qter, 115q22-qter, 218cen-q21 6 11q21-q31, 18q/118q23-qter, 114q24-qter, 115q22-qter, 117cen- p13, 2Xq 7 21p, 12p11.2-p12, 26q15-q22, 17p, 29p21-pter, 210, 211, 212q22-q23, 213q14-q32, 114q, 218q 8 21p21-p32, 24q27-qter, 18q21.3-qter, 211p13-pter, 2X 9 21p21-p31, 23cen-p21, 14q, 26p21.3-pter, 17p15-pter, 18q, 29p21-pter, 210p, 211, 212p11.2-pter, 213q, 214q24-qter, 115q24-qter, 220p 10 no changes 11 26p12-q16, 17cen-q22 12 15p, 17pter-q32, 110q24-q26, 213q, 115q 13 22pter-p23, 22q33-qter, 14q31.1-q33, 17cen-q31, 211q23-qter, 218q12-qter 14 11pter-q41, 12p, 13p12-pter, 15p12-p14, 28p21-pter, 19q21-q32, 112p11.2-p12, 120q11.2-qter 15 26p22-p23, 211q23-qter 16 no changes 17 no changes 18 no changes 19 111q22 20a 22q22-q31, 16p21.1-p21.3, 16q25-qter 20b 16p11.2-pter, 16q24-qter, 17q22-qter, 114q22-qter, 116p, 117 21 21q32-qter, 22p11.2-pter, 22q21-q24, 13p21-pter, 23q12-qter, 26p21.3-q23, 27p11.2-pter, 17q11.2-q31, 28p11.2-pter, 29p, 210pter-q21, 213q, 216, 117cen-p12, 218q12-qter, 221q, 2Xp11.4-qter 22 no changes 23 17, 18, 112 24 no changes 25 11q21-qter, 12q21-qter, 13pter-q22, 15, 17pter-q22/117p14-pter, 18, 114q, 115q, 2Xq13-qter 26 11q31-qter, 12/112q33-qter, 23, 14pter-q13, 15p/115p14-p15.2, 25q13-q15, 16q, 17, 29p, 110p11.2-pter, 110q24-qter, 111pter-q13, 211q22-qter, 112p11.2-pter, 213q21-qter, 115q21- qter, 120q, 121q/1121q22, 122q/1122q13, 1Xq13-qter CGH results of Cases 4, 11, and 12 have been published previously.8 TABLE 3 Most Frequent DNA-sequence Copy Number Aberrations in 27 Postirradiation Sarcomas by Comparative Genomic Hybridization (CGH) Gains n (%) Losses n (%) 7q11.2-q21 8 (30) 11q23-qter 6 (22) 7q22 8 (30) 13q22-q32 6 (22) 7p15-pter 7 (26) 1p21-p31 4 (15) 8q23-qter 6 (22) 6q15-q16 4 (15) 15q24-qter 6 (22) 9p21-pter 4 (15) 5p12-p14 5 (19) 10p11.2 4 (15) 14q24-qter 5 (19) 18q12-q21 4 (15) 1q21-q23 4 (15) Xq13-qter 4 (15) 1q31 4 (15) 2p11.2-p12 4 (15) 12p11.2-p12 4 (15) CGH of Postirradiation Sarcoma/Tarkkanen 1995 fraction of chondrosarcomas27 and soft tissue sarco- mas.28 The gains at 8q affect large chromosomal re- gions and the minimal common regions are slightly different in different tumor subtypes. Thus there prob- ably are several target genes yet to be identified. The most common losses affected 11q23-qter and 13q22-q32, both most frequently in osteosarcoma. Loss of distal 11q is common in several types of can- cer15 and also in sporadic MFH.17 Loss of distal 11q was not detected by us in sporadic osteosarcoma, but these changes could have been masked by the fre- quent gains in 11q in this tumor type.16 Candidate genes for losses in 11q include ATM (11q22-q23) and PPP2R1B (11q22-q24), but no data about their possi- ble role in tumorigenesis of sarcomas exists yet. Inter- estingly, ATM is involved in cellular responses to ion- izing radiation. Also, loss of distal 13q is frequent in several cancer types15 and is seen also in both spo- radic osteosarcoma16 and MFH,17 but no target genes in this region have been identified yet. FIGURE 1. Summary of gains, high-level amplifications, and losses of DNA sequence copy number in 27 postirradiation sarcomas analyzed by CGH. Losses are shown on the left and gains on the right side of the chromosomes. Each line represents a genetic aberration seen in one sample. High-level amplifications are shown with thick lines. Notations a and b refer to Case 20. 1996 CANCER October 1, 2001 / Volume 92 / Number 7 When the findings in postirradiation osteosar- coma and MFH were further compared with findings in sporadic osteosarcoma and sporadic MFH, as re- ported previously by us,16,17 some differences were found: Although the mean number of aberrations per tumor was similar in postirradiation and sporadic os- teosarcomas, in sporadic osteosarcomas gains were more frequent than losses (1.9:1), but, in postirradia- tion osteosarcomas, losses outnumbered gains (1.3:1). Loss at 1p was seen in only one sporadic osteosar- coma (3%), but loss at 1p was frequent (57%) in post- irradiation osteosarcomas. Loss at 1p is a frequent event in different types of cancer15 and is seen also in malignant mesothelioma,29 a cancer strongly associ- ated with exposure to asbestos. Conversely, gains at Xp and 6p were frequent in sporadic osteosarcomas but were not seen in postirradiation osteosarcomas. At present, we believe that there is no data on CGH findings in extraskeletal osteosarcomas and, therefore, the CGH findings in postirradiation osteosarcomas (three skeletal and four extraskeletal tumors) were compared with skeletal osteosarcomas.16 In postirradiation MFH, the mean number of ab- errations per tumor was somewhat lower than in spo- radic MFH, which could be because of the slightly larger proportion of tumors without any aberrations in the current study. In postirradiation MFH, the most frequent aberration was gain of 7cen-q22, seen in 38%. In sporadic MFH, gains affecting chromosome 7 also were frequent: 7cen-q22 in 17% and 7q32 in 24%. Because of a lack of CGH data on sporadic tumors, similar comparisons can not be carried out at present in angiosarcoma or fibrosarcoma. A great variability characterizes the findings of the present study, from cases with complex aberrations to cases with a few aberrations to cases with a single or no aberration. Several factors can explain why no copy number changes were detected in some cases: CGH does not detect balanced aberrations, and, thus, tu- mors containing only such changes will appear nor- mal by CGH. Furthermore, CGH detects only clonal aberrations occurring in a relatively large subset of the tumor cell population, and changes present in a sub- clone can remain undetected by CGH.13 Thus, intra- tumoral heterogeneity30 and polyclonality8 can pre- vent detection of existing changes. The analysis of the primary tumor and its local recurrence in Case 20 is illustrative of such heterogeneity. These samples share two imbalances but show also unique copy number changes, the local recurrence showing a more com- plex aberration pattern than the primary tumor. The local recurrence was detected more than 2 years after the primary tumor was operated. Also, the recurrence was a high-grade tumor, whereas the primary tumor was of Grade II. Thus clonal evolution during tumor progression is a likely explanation for the different aberration patterns in this tumor pair. According to the recent study and review by Mertens et al.,8 loss at 3p (3p21-pter) is frequent in postirradiation sarcomas by cytogenetic analysis (12 out of 18 cases). In the current study, loss at 3p was detected in only three tumors with the minimal com- mon region of 3cen-p14. This discrepancy might have been due to complex marker chromosomes that con- tain material from 3p, not detectable by conventional cytogenetic analysis, or due to intratumoral heteroge- neity with subclonal variation.30 In conclusion, according to our findings, postirra- diation osteosarcomas differ genetically from sporadic osteosarcomas,16 most notably in more frequent losses of DNA sequences in postirradiation osteosar- coma, whereas gains are more frequent in sporadic osteosarcoma. Furthermore, loss at 1p was rare in sporadic osteosarcoma but frequent in postirradiation osteosarcoma. Copy number gains at 7q were frequent in postirradiation sarcomas, especially in MFH and fibrosarcoma. In a previous study of sporadic MFH, we demonstrated that gain of 7q32 is associated with shorter metastasis-free and overall survival.17 Also, gain of distal 8q was frequent in postirradiation sar- comas, especially in osteosarcoma. This aberration is associated with poor prognosis in sporadic osteosar- coma,16 chondrosarcoma,20 and Ewing tumors.23 Gain of distal 8q also has been associated with large tumor size in leiomyosarcoma22 and synovial sarcoma.31 Thus, the frequent gains at 7q and 8q may explain, in part, the poor prognosis of patients with postirradia- tion sarcomas. However, the poor outcome of these patients also may be due to some of the typical clinical features of postirradiation sarcomas: high malignancy grade, late diagnosis, and central location. REFERENCES 1. Cahan WG, Woodard HQ, Higinbotham NL, Stewart FW, Coley BL. Sarcoma arising in irradiated bone: report of 11 cases. Cancer 1998;82(1):3–25. 2. Laskin WB, Silverman TA, Enzinger FM. Postradiation soft tissue sarcomas: An analysis of 53 cases. Cancer 1988;62: 2330 – 40. 3. Arlen M, Higinbotham NL, Huvos AG, Marcove RC, Miller T, Shah IC. Radiation-induced sarcoma of bone. Cancer 1971; 28:1087–99. 4. Wiklund TA, Blomqvist C, Raty J, Elomaa I, Rissanen P, Miettinen M. Postirradiation sarcoma. Analysis of a nation- wide cancer registry material. Cancer 1991;68:524 –31. 5. Robinson E, Neugut AI, Wylie P. Clinical aspects of postir- radiation sarcomas. J Natl Cancer Inst 1988;80:233– 40. 6. Wiklund T, Huuhtanen R, Blomqvist C, Tukiainen E, Viro- lainen M, Virkkunen P, et al. The importance of a multidis- ciplinary group in the treatment of soft tissue sarcomas. Eur J Cancer 1996;32A:269 –73. CGH of Postirradiation Sarcoma/Tarkkanen 1997 7. Saeter G, Alvegard TA, Elomaa I, Stenwig AE, Holmstrom T, Solheim OP. Treatment of osteosarcoma of the extremities with the T-10 protocol, with emphasis on the effects of preoperative chemotherapy with single-agent high-dose methotrexate: a Scandinavian Sarcoma Group study. J Clin Oncol 1991;9:1766 –75. 8. Mertens F, Larramendy M, Gustavsson A, Gisselsson D, Ry- dholm A, Brosjo O, et al. Radiation-associated sarcomas are characterized by complex karyotypes with frequent rear- rangements of chromosome arm 3p. Cancer Genet Cytogenet 2000;116:89 –96. 9. Chauveinc L, Dutrillaux AM, Validire P, Padoy E, Sabatier L, Couturier J, et al. Cytogenetic study of eight new cases of radiation-induced solid tumors. Cancer Genet Cytogenet 1999;114:1– 8. 10. Nakanishi H, Tomita Y, Myoui A, Yoshikawa H, Sakai K, Kato Y, et al. Mutation of the p53 gene in postradiation sarcoma. Lab Invest 1998;78:727–33. 11. Brachman DG, Hallahan DE, Beckett MA, Yandell DW, Weichselbaum RR. p53 gene mutations and abnormal reti- noblastoma protein in radiation-induced human sarcomas. Cancer Res 1991;51:6393– 6. 12. Kallioniemi A, Kallioniemi OP, Sudar D, Rutovitz D, Gray JW, Waldman F, et al. Comparative genomic hybridization for molecular cytogenetic analysis of solid tumors. Science 1992; 258:818 –21. 13. Kallioniemi OP, Kallioniemi A, Piper J, Isola J, Waldman FM, Gray JW, et al. Optimizing comparative genomic hybridiza- tion for analysis of DNA sequence copy number changes in solid tumors. Genes Chromosomes Cancer 1994;10:231– 43. 14. Knuutila S, Bjorkqvist AM, Autio K, Tarkkanen M, Wolf M, Monni O, et al. DNA copy number amplifications in human neoplasms. Review of comparative genomic hybridization studies. Am J Pathol 1998;152:1107–23. 15. Knuutila S, Autio K, Bjorkqvist AM, Aalto Y, El-Rifai W, Hemmer S, et al. DNA copy number losses in human neo- plasms. Am J Pathol 1999;155:683–94. 16. Tarkkanen M, Elomaa I, Blomqvist C, Kivioja A, Kello- kumpu-Lehtinen P, Bohling T, et al. DNA sequence copy number increase at 8q - a potential new prognostic marker in high-grade osteosarcoma. Int J Cancer 1999;84:114 –21. 17. Larramendy ML, Tarkkanen M, Blomqvist C, Virolainen M, Wiklund T, Asko-Seljavaara S, et al. Comparative genomic hybridization of malignant fibrous histiocytoma reveals a novel prognostic marker. Am J Pathol 1997;151:1153– 61. 18. El-Rifai W, Larramendy M, Bjorkqvist AM, Hemmer S, Kn- uutila S. Optimization of comparative genomic hybridiza- tion using fluorochrome conjugated to dCTP and dUTP nucleotides. Lab Invest 1997;77:699 –700. 19. Tarkkanen M, Huuhtanen R, Virolainen M, Wiklund T, Asko- Seljavaara S, Tukiainen E, et al. Comparison of genetic changes in primary sarcomas and their pulmonary metas- tases. Genes Chromosomes Cancer 1999;25:323–31. 20. Larramendy ML, Mandahl N, Mertens F, Blomqvist C, Kiv- ioja AH, Karaharju E, et al. Clinical significance of genetic imbalances revealed by comparative genomic hybridization in chondrosarcomas. Hum Pathol 1999;30:1247–53. 21. Fukuda T, Ichimura E, Shinozaki T, Sano T, Kashiwabara K, Oyama T, et al. Coexpression of HGF and c-Met/HGF re- ceptor in human bone and soft tissue tumors. Pathol Int 1998;48:757– 62. 22. El-Rifai W, Sarlomo-Rikala M, Knuutila S, Miettinen M. DNA copy number changes in development and progression in leiomyosarcomas of soft tissues. Am J Pathol 1998;153:985– 90. 23. Tarkkanen M, Kiuru-Kuhlefelt S, Blomqvist C, Armengol G, Bohling T, Ekfors T, et al. Clinical correlations of genetic changes by comparative genomic hybridization in Ewing tumors. Cancer Genet Cytogenet 1999;114:35– 41. 24. Ladanyi M, Park CK, Lewis R, Jhanwar SC, Healey JH, Huvos AG. Sporadic amplification of the MYC gene in human os- teosarcomas. Diagn Mol Pathol 1993;2:163–7. 25. Stock C, Kager L, Fink FM, Gadner H, Ambros PF. Chromo- somal regions involved in the pathogenesis of osteosarco- mas. Genes Chromosomes Cancer 2000;28:329 –36. 26. Gamberi G, Benassi M, Bohling T, Ragazzini P, Molendini L, Sollazzo M, et al. C-myc and c-fos in human osteosarcoma: prognostic value of mRNA and protein expression. Oncology 1998;55:556 – 63. 27. Castresana JS, Barrios C, Gomez L, Kreicbergs A. Amplifica- tion of the c-myc proto-oncogene in human chondrosar- coma. Diagn Mol Pathol 1992;1:235– 8. 28. Barrios C, Castresana JS, Ruiz J, Kreicbergs A. Amplification of the c-myc proto-oncogene in soft tissue sarcomas. Oncol- ogy 1994;51:13–7. 29. Lee WC, Balsara B, Liu Z, Jhanwar SC, Testa JR. Loss of heterozygosity analysis defines a critical region in chromo- some 1p22 commonly deleted in human malignant me- sothelioma. Cancer Res 1996;56:4297–301. 30. Orndal C, Rydholm A, Willen H, Mitelman F, Mandahl N. Cytogenetic intratumor heterogeneity in soft tissue tumors. Cancer Genet Cytogenet 1994;78:127–37. 31. Skytting BT, Szymanska J, Aalto Y, Lushnikova T, Blomqvist C, Elomaa I, et al. Clinical importance of genomic imbalances in synovial sarcoma evaluated by comparative genomic hybrid- ization. Cancer Genet Cytogenet 1999;115:39–46. 1998 CANCER October 1, 2001 / Volume 92 / Number 7