Cytogenetic Aberrations in Ewing Sarcoma: Are Secondary Changes AssociatedWith Clinical Outcome?

Carl-Magnus Kullendorff, MD,1* Fredrik Mertens, MD,2 Mikael Donner, MD,3

Thomas Wiebe, MD,3 Måus Åkerman, MD,4 and Nils Mandahl, PhD2

Background. Ewing sarcoma is associatedwith a nonrandom pattern of primary and sec-ondary chromosomal aberrations. Whereas thefinding of rearrangements of chromosome 22,usually in the form of a balanced translocationt(11;22)(q24;q12), is important diagnostically,nothing is known about the potential prognos-tic impact of the secondary chromosomal ab-errations. Procedure. During a 13-year-period,short-term cultured tumor samples from 21children and young adults with Ewing sarcomawere cytogenetically analyzed successfully. Re-sults. Clonal chromosome aberrations were de-tected in 18 patients, 17 of whom had the char-acteristic t(11;22)(q24;q12) or variants thereof.The most frequent secondary change was +8,

followed by +12, +2, +5, +9, +15, and gain ofmaterial from the long and short arms of chro-mosome 1. The only recurrent secondarychange that was restricted to tumors from theten patients that were dead at latest follow-upwas gain of 1q material. Furthermore, all threepatients with tumors with chromosome num-bers over 50 had died, and the only patient witha tumor karyotype lacking chromosome 22 re-arrangement was alive without evidence of dis-ease. Conclusions. These data and previouslypublished results indicate that the karyotypicpattern not only may be of diagnostic signifi-cance but also may be important prognosti-cally. Med. Pediatr. Oncol. 32:79–83, 1999.© 1999 Wiley-Liss, Inc.

Key words: Ewing sarcoma; chromosome aberration; cytogenetics; clinical outcome

INTRODUCTION

Ewing sarcoma (ES) is the second most common ma-lignant bone tumor in children and young adults [1]. Itoccurs most frequently in the second decade of life and israre in children under 5 years of age and in adults over 30[2]. Current treatment protocols, which combine chemo-therapy, surgery, and sometimes radiation, have in-creased the survival rates to more than 50%. Some un-favorable prognostic factors have been identified, e.g.pelvic and sacral sites for primary tumors, increased se-rum lactate dehydrogenase levels, metastatic disease atpresentation, and poor response to chemotherapy. How-ever, it is often difficult to predict the clinical outcome inthe individual case [3].

The cytogenetic hallmark of ES is rearrangements ofchromosome 22, band q12 [4]. In about 90% of the cases,these are detected as a translocation t(11;22)(q24;q12)and rarely in the form of the variant translocationst(21;22) or t(7;22) [5]. The molecular consequences ofthe various translocations have been clarified recently.The gene involved on chromosome 22 is alwaysEWS,which is fused toFLI1, ERG, or ETV1 on chromosomes11, 21, and 7, respectively [6]. The consistency andspecificity of these translocations at the chromosome andDNA levels have made cytogenetic and molecular ge-netic investigations important diagnostically. Becausepolymerase chain reaction (PCR) analysis provides a

more rapid answer to the question of whether any of theES-associated rearrangements are present and, further-more, does not require cell culturing, PCR often is pre-ferred to cytogenetic analysis in the diagnostic setting.

Rearrangements of theEWSgene could also be visu-alized rapidly by fluorescent in situ hybridization (FISH)analysis of interphase nuclei [6]. However, moleculargenetic analysis will not provide any information aboutthe chromosomal rearrangements that may occur duringclonal evolution. More than 80% of the cases witht(11;22) have secondary changes, many of which clearlyare nonrandom [7]. The most common additional aber-ration is gain of chromosome 8, which is detected inalmost 40% of the cases, followed in decreasing order of

1Department of Pediatric Surgery, University Hospital, Lund, Sweden2Department of Clinical Genetics, University Hospital, Lund, Sweden3Department of Pediatrics, University Hospital, Lund, Sweden4Department of Pathology, University Hospital, Lund, Sweden

Grant sponsor: Children’s Cancer Fund of Sweden; Grant sponsor:Swedish Cancer Society; Grant sponsor: Swedish Association for Can-cer and Traffic Victims.

*Correspondence to: C.-M. Kullendorff, Department of Pediatric Sur-gery, University Hospital, S-221 85 Lund, Sweden. E-mail: carl-magnuskullendorff@pedlund.postnet.se

Received 18 June 1997; Accepted 12 August 1998

Medical and Pediatric Oncology 32:79–83 (1999)

© 1999 Wiley-Liss, Inc.

frequency by +12, +2, +20, der(16)t(1;16)(q11–21;q11–13), +14, +7, +5, +9, and +21 [5]. The clinical impor-tance of these secondary changes remains unknown.Still, less than 130 cytogenetically abnormal ES caseshave been reported, and the largest series have includedinformation on 10–20 patients [8–11], making clinical-cytogenetic correlations difficult. The main aim of thepresent study, which includes data on 21 children andyoung adults with ES that have been analyzed success-fully cytogenetically during a 13-year period, was tosearch for possible associations between cytogenetic fea-tures and clinical outcome.

PATIENTS AND METHODS

During the period from 1984 to 1996, 22 patients withES were analyzed cytogenetically. In the present study,only 21 cases in which short-term cultured tumorsamples had yielded analyzable mitoses were included.All patients (nine girls and 12 boys) were under 25 yearsof age (range, 8–22 years; mean, 14 years), and they werefollowed for at least 5 months or until death. The ana-lyzed samples were obtained at diagnosis in 17 cases, atrelapse only in three cases, and at both diagnosis andrelapse in one case (Table I).

Cell cultures from tumor biopsies, fine-needle aspira-tion (FNA) biopsies, and bone marrow were processedfor cytogenetic analysis as described by Mandahl et al.[12] and Åkerman et al. [13]. Bone marrow samples werecultured for 24 hours and 48 hours, and cell cultures fromtumor biopsies and FNA biopsies were harvested after3–10 days. The karyotypes, but not the clinical details, ofcases 2, 6, 15, 16, 18, and 19 have been reported previ-ously [13,14].

All of the material was used for initiating short-termcultures and obtaining G-banded metaphase chromo-somes of good quality. Thus, none of the tumors could beanalyzed with reverse transcriptase (RT)-PCR forEWSrearrangements, nor were there any cells in fixative orslides with unstained cell spreads, precluding the possi-bility to complement the cytogenetic analyses with inter-phase or metaphase FISH techniques. During the periodfrom 1984 until 1989, the patients were treated accordingto a protocol [Scandinavian Sarcoma Group (SSG)-IV]that was designed by the SSG [15]. This protocol is basedon the administration of six different drugs (actinomycin,adriamycin, bleomycin, cyclofosfamide, methotrexate,and vincristine) and local therapy with surgery and/orlocal irradiation. From 1990 on, the treatment was ac-cording to the SSG-IX protocol [16], which is a com-bined modality protocol consisting of alternating coursesof vincristine-adriamycin-ifosfamide (VAI) and cisplati-num-adriamycin-ifosfamide (PAI) at 3-week intervals incombination with surgery and/or local irradiation.

The diagnosis of ES was made on biopsy specimens

and FNA with cytological, immunohistochemical, andultrastructural analyses [14]. In eight patients (cases 1, 4,5, 6, 7, 11, 13, and 20), the diagnosis was based onhistopathologic examination of surgical biopsies supple-mented with ancillary diagnostics [17]. Immunohisto-chemical analysis was also carried out in all eight pa-tients, and electron microscopic examination was carriedout in three patients [18]. Antibody to CD 99 (MIC2antigen) was not available during the first years of thestudy. Therefore, supplementary immunohistochemicalstaining for CD 99 has now been carried out for all casesthat were not stained primarily. All tumors showed astrong homogenous positive staining.

In 11 patients (cases 2, 3, 8, 10, 14, 15, 16, 17, 18, 19,and 21), the definitive diagnosis was based on FNA cy-tology supplemented with electron microscopic exami-nation of aspirated material for all patients and on stain-ing for CD 99 in three patients who were all positive[19]. During the first years of the study period, two pa-tients (cases 9 and12) were analyzed with routine stain-ing only. All 21 patients have been reexamined recently,and the diagnoses of ES were established.

RESULTS

Clinical and cytogenetic data are summarized in TableI. Clonal chromosome aberrations were detected in 18 ofthe 21 tumors. The patients in cases 7, 9, and 11 werestudied only at relapse. Prior to the cytogenetic analysis,these patients had been treated with chemotherapy andlocal irradiation. A t(11;22)(q24;q12) or a variant thereofwas found in 17 patients and was the sole change foundin two patients. In one patient (case 15), no rearrange-ment of chromosomes 11 or 22 could be recognizedamong the four cytogenetically abnormal metaphasecells, which were of relatively poor technical quality, thatwere obtained at diagnosis. At relapse 10 years later, arelated but more complex clone was seen. At this time,the banding quality was better, and a del(22)(q12) wasclearly identified. The most frequent secondary rear-rangements were trisomies for chromosome 8 (eight pa-tients); chromosome 12 (five patients); and chromo-somes 2, 5, 9, and 15 (three patients each). Three patientseach showed loss of 1p material, with 1p36 as the com-monly deleted band, and gain of 1q material. The unbal-anced t(1;16) in case 18, which led to partial loss of 16qand gain of 1q24-qter, resulted in the smallest sharedgain of 1q material. One ES patient (case 6) had a clonalrearrangement that did not involve chromosome 22. Allabnormal cells in cultures from both FNA and tumorbiopsy at diagnosis had a t(15;19)(q14;p13) as the solestructural aberration. The lack of clonal chromosome ab-errations in cases 5, 8, and 14 most probably was due toovergrowth in vitro of nonneoplastic cells. The culturing

80 Kullendorff et al.

times for these three patients, however, were not differentfrom those for the patients with clonal aberrations.

At the time of latest follow-up, ten patients had died:nine from treatment failure and one from treatment com-plications. One patient (case 14) had a normal tumorkaryotype, one patient (case 13) had a t(11;22) as thesole change, and eight patients, including all three pa-tients with tumors with more than 50 chromosomes, hada t(11;22) together with additional changes. The onlyrecurrent secondary change that was restricted to thegroup with poor clinical outcome was gain of 1q mate-rial. The three patients (cases 1, 16, and 18) with this

aberration died 12, 88, and 14 months, respectively, afterdiagnosis.

DISCUSSION

In some hematologic neoplasms, the paradigmatic ex-ample of which is chronic myeloid leukemia, there is agood correspondence between presence of secondarychromosomal rearrangements and poor clinical outcome[4]. For sarcomas, so far, there is only indirect supportfor a phenotypic impact of secondary chromosome aber-rations. Preliminary data indicate that trisomy 8 may be

TABLE I. Clinical and Cytogenetic Data on 21 Patients With Ewing Sarcoma

CaseSex/age(years) Site

Clinicalcoursea

Survival(months) Sampleb Statusc Karyotype

1 M/8 Os ileum DOD 12 TB D 52–59,XY,+del(1)(p13),+2,+6,+7,+8,t(11;22)(q24;q12),+12,+13,+13,+17,+20[cp4]/46,XY[31]

BM D 52–56,XY,+del(1)(p13),+2,+6,+7,+8+9,t(11;22)(q24;q12),+12,+13,+13,+15,+17[cp5]

BM R 57,XY,+del(1)(p13),+2,+6,+7,+8,+9,t(11;22)(q24;q12),+12,+13,+18,+18,+20[4]

2 M/8 Femur Alive NED 123 TB D 50,XY,+2,+8,+9,t(11;22)(q24;q12),+12[2]/50,idem,−8,add(16)(q24),+mar[3]

3 F/8 Femur DOCT 6 FNA D 50,XX,+4,+5,+8,t(11;22)(q24;q12),+16[6]/46,XX[5]4 F/8 Vertebra Alive NED 5 TB D 47,XX,+8,t(11;22)(q24;q12)[7]/46,XX[4]5 M/9 Rib Alive NED 142 TB D 46,XY[20]6 M/10 Os ileum Alive NED 65 FNA D 46,XY,t(15;19)(q14;p13)[15]/45,idem,−Y[7]

TB D 46,XY,t(15;19)(q14;p13)[2]/46,XY[2]7 F/12 Pelvis DOD 35 TB R 45–46,X,−X,der(1)add(1)(p?)add(1)(q?),del(3)(q21),−4,

del(6)(q16),−9,add(10)(q?),?t(11;22)(q24;q12),−12,−15,add(18)(q?),+1−4mar[cp6]

8 F/12 Tibia Alive NED 120 FNA D 46,XX[22]9 F/13 Metacarpal DOD 98 TB R 46,XX,der(1)t(1;12)(p36;q22),t(3;17)(q21;p13),del(11)

(q13),der(12)t(11;12)(q13;q22)t(12;22)(p11;q12),der(22)t(12;22)(p11;q12)[4]/46,XX[4]

10 F/14 Fibula Alive NED 52 FNA D 46,XX,?t(11;22)(q24;q12),inc[2]11 F/14 Tibia DOD 114 TB R 46,XX,t(1;12)(p21;p12),t(11;22)(q24;q12)[5]/46,XX[7]12 M/15 Femur Alive 19 TB D 46,XY,t(11;22)(q24;q12)[7]/46,XY[2]13 M/16 Tibia DOD 21 TB D 46,XY,t(11;22)(q24;q12)[2]/46,XY[2]14 M/16 Pelvis DOD 41 FNA D 46,XY[24]15 M/16 Vertebra Alive WD 119 TB D 49–50,XY,add(1)(p36),−2,+5,+15,+15[cp4]

FNA R 44–47,Y,add(X)(p11),add(1)(p22),?ins(1;?)(p13;?),−2,der(4)t(1;4)(p22;p1?5)ins(4;?)(p1?5;?),+5,+8,del(10)(q23),−14,+15,−17,−18,ins(18;?)(q21;?),−19,add(21)(q21),del(22)(q12),inc[cp6]

16 M/17 Femur DOD 88 FNA D 55,X,der(Y)t(Y;1)(q12;q21),+der(Y)t(Y;1),+5,+6,+8,+8,+9,t(11;22)(q24;q12),+12,der(15)t(2;15)(p11;p13),+17,+20[14]

17 M/19 Metatarsal Alive NED 18 FNA D 46,XY,der(11)(11pter-p15::11q23-11q24::22q12-22q13::11p15-cen-11q23::22q13-22qter),der(22)t(11;22)(q24;q12)[4]

18 F/21 Scapula DOD 14 FNA D 47–50,XX,t(11;22)(q24;q12),+12,+12,der(16)t(1;16)(q24;q23),+18,+21[cp3]

19 F/21 Knee Alive NED 33 FNA D 47,XX,+8,t(11;22)(q24;q12),del(13)(q14)[14]20 M/22 Os calcaneum DOD 45 TB D 53,XY,+2,+8,+10,t(11;22)(q24;q12),+12,+14,+15[2]/46,

XY[15]21 M/22 Femur Alive NED 18 FNA D 47–49,XY,add(22)(q12),inc[3]

aDOD, dead of disease; NED, no evidence of disease; WD, with disease; DOCT, dead of complications to treatment.bTB, tumor biopsy; BM, bone marrow; FNA, fine-needle aspiration biopsy.cD, at diagnosis; R, at relapse.

Cytogenetics in Ewing Sarcoma 81

associated with more advanced disease in liposarcomaswith t(12;16)(q13;p11) as the primary abnormality [20]and that trisomy 2 may be restricted to local recurrencesand distant metastases in synovial sarcomas with thecharacteristic t(X;18)(p11;q11) [21].

Few of the previous reports on cytogenetic findingsin ES have commented on the possible clinical impor-tance of secondary chromosomal changes. Hattinger etal. [11], combining their own results with previouslypublished information, suggested that the presence ofder(16)t(1;16) might be associated with advanced stagesof Ewing tumors. The t(1;16), however, is far from spe-cific for Ewing tumors, and the exact breakpoints mayvary [5,11]. Thus, it is likely that it is the resulting im-balances, of which gain of 1q is the most consistent, thatare of pathogenetic importance. In our series of ES, theonly patient (case 18) with a tumor karyotype containinga der(16)t(1;16) died 14 months after diagnosis. Further-more, the two patients (cases 1 and 16) who showed gainof 1q material through other mechanisms also died oftheir disease 12 months and 88 months, respectively,after diagnosis. Thus, our findings add support to theconclusions by Hattinger and coworkers [11].

In the present series, all patients with tumor karyo-types containing more than 50 chromosomes (cases 1,16, and 20) died from metastatic disease. Chromosomenumbers over 50 have been detected in 17 previouslypublished ES [5]. In nine of these cases, the prognosticsignificance of hyperdiploidy could be evaluated. Onepatient was alive without any evidence of disease, threepatients were dead of disease, three patients presentedwith metastases at diagnosis, and two patients were alivewith disease. Additional indirect support for the interpre-tation that complex karyotypes in ES may be associatedwith poor clinical outcome may be derived from the re-sults on other solid tumor types, in which there usuallyhas been a good correspondence between increasing cy-togenetic complexity and poor clinical outcome [22]. Fi-nally, in a retrospective analysis of the DNA content in37 ES, Dierick and coworkers [23] found that all patientswith nondiploid tumors had died within 5 years afterdiagnosis.

Due to the facts that ES are exceedingly rare tumorsand that many of them now are investigated only byusing molecular genetic techniques, it may well be dif-ficult to reach any firm conclusions with regard to theprognostic importance of secondary chromosome aber-rations unless multiple pediatric oncology centers com-bine their results. Without continuing to analyze the fullkaryotypes in these cases, we will never be able to an-swer questions regarding additional changes and progno-sis. The observations in the present study and in previousstudies that chromosome numbers over 50 and gain of 1qmaterial, through der(16)t(1;16) or other mechanisms,could be associated with particularly poor prognosis and

that tumors without cytogenetic evidence of chromosome22 rearrangements may be less aggressive need to besubstantiated in larger series.

REFERENCES

1. Young JL, Miller RW. Incidence of malignant tumors in U.S.children. J Pediatr 1975;86:245–258.

2. Miller RW. Contrasting epidemiology of childhood osteosarcoma,Ewing tumor and rhabdomyosarcoma. Natl Cancer Inst Monogr1981;56:9–14.

3. Pizzo PA, Poplack DG, Horowitz ME, et al. Solid tumors ofchildhood. In: De Vita VT, Hellman S Jr., Rosenberg SA, editors.Cancer: Principles and Practice, 4th ed. Philadelphia: Lippincott;1993. p 1738–1791.

4. Heim S, Mitelman F. Cancer Cytogenetics, 2nd ed. New York:Wiley-Liss; 1995.

5. Mitelman F. Catalog of Chromosome Aberrations in Cancer, 5thed. New York: Wiley-Liss; 1994.

6. Desmaze C, Brizard F, Turc-Carel C, et al. Multiple chromosomalmechanisms generate an EWS/FLI1 or an EWS/ERG fusion genein Ewing tumors. Cancer Genet Cytogenet 1997;97:12–19.

7. Mugneret F, Lizard S, Aurias A, et al. Chromosomes in Ewing’ssarcoma. II. Nonrandom additional changes, trisomy 8 andder(16)t(1;16). Cancer Genet Cytogenet 1988;32:239–245.

8. Turc-Carel C, Aurias A, Mugneret F, et al. Chromosomes in Ew-ing’s sarcoma. I. An evaluation of 85 cases and remarkable con-sistency of t(11;22)(q24;q12). Cancer Genet Cytogenet1988;32:229–238.

9. Douglass EC, Rowe ST, Valentine M, et al. A second nonrandomtranslocation, der(16)t(1;16)(q21;q13), in Ewing sarcoma and pe-ripheral neuroectodermal tumor. Cytogenet Cell Genet1990;53:87–90.

10. Ladanyi M, Heinemann FS, Huvos AG, et al. Neural differentia-tion in small round cell tumors of bone and soft tissue with thetranslocation t(11;22)(q24;q12). Hum Pathol 1990;21:1245–1251.

11. Hattinger CM, Rumpler S, Ambros IM, et al. Demonstration of theder(16)t(1;16)(q12;q11.2) in interphase nuclei of Ewing tumors.Genes Chromosomes Cancer 1996;17:141–150.

12. Mandahl N, Heim S, Arheden K, et al. Three major cytogeneticsubgroups can be identified among chromosomally abnormal soli-tary lipomas. Hum Genet 1988;79:203–208.

13. Åkerman M, Dreinho¨fer K, Rydholm A, et al. Cytogenetic studieson fine-needle aspiration samples from osteosarcoma and Ewing’ssarcoma. Diagn Cytopathol 1996;15:17–22.

14. Åkerman M, Alvegård T, Eliasson J, et al. A case of Ewing’ssarcoma diagnosed by fine needle aspiration. Light microscopy,electron microscopy and chromosomal analysis. Acta OrthopScand 1988;59:589–592.

15. Alvegård TA, Solheim O¨ , Åkerman M, et al. Ewing’s sarcoma. Apreliminary report from the Scandinavian Sarcoma Group. ActaOncol 1989;28:59–64.

16. Elomaa I, Blomqvist C. Preliminary results of the SSG IX proto-col in Ewing’s sarcoma. Acta Orthop Scand 1994;65(Suppl256):61.

17. Dahl I, Åkerman M, Angervall L. Ewing’s sarcoma of bone. Acorrelative cytological and histological study of 14 cases. ActaPathol Microbiol Scand 1986;94:353–369.

18. Akhtar M, Ashraf A, Sabbah R. Aspiration cytology of Ewing’s

82 Kullendorff et al.

sarcoma. Light and electron microscopic correlations. Cancer1985;56:2051–2060.

19. Fanning CV, Katz R, Ayala A. Fine needle aspiration of Ewing’ssarcoma of bone. An analysis of 57 specimens. Acta Cytol1985;29:932.

20. Mandahl N, Mertens F, Åman P, et al. Nonrandom secondarychromosome aberrations in liposarcomas with t(12;16). Int J On-col 1994;4:307–310.

21. Mandahl N, Limon J, Mertens F, et al. Nonrandom secondary

chromosome aberrations in synovial sarcomas with t(X;18). Int JOncol 1995;7:495–499.

22. Mitelman F, Johansson B, Mandahl N, et al. Clinical significanceof cytogenetic findings in solid tumors. Cancer Genet Cytogenet1997;95:1–8.

23. Dierick AM, Langlois M, Van Oostveldt P, et al. The prognosticsignificance of the DNA content in Ewing’s sarcoma: a retrospec-tive cytophotometric and flow cytometric study. Histopathology1993;23:333–339.

Cytogenetics in Ewing Sarcoma 83