Chromosome Laboratory, Section 5702, The Juliane Marie Center, Rigshospitalet, Henrik Harpestrengsvej 4, 2100 Copenhagen Ø, Denmark

Correspondence to: M K Andersen, Fax: 45 35 45 25 77

Dicentric chromosomes are observed in many malignant diseases including myelodysplasia (MDS) and acute myeloid leukemia (AML) and have often been observed in a subset of these diseases, namely therapy-related MDS (t-MDS) and AML (t-AML). Using fluorescence in situ hybridization (FISH) with centromere-specific probes, we investigated the frequency and type of dicentric chromosomes in 180 consecutive patients with t-MDS and t-AML and in 231 consecutive patients with de novo MDS and AML, whose karyotypes had been studied previously by conventional G-banding. Twenty-seven out of 180 patients with t-MDS or t-AML presented dicentric chromosomes compared to only seven out of 231 patients with de novo disease (P = 0.00003). A dic(1q;7p) was observed in 10 cases, a dic(5p;17q) was observed in six cases, whereas various isodicentric chromosomes were observed in six cases. Excluding these six cases with isodicentrics, all 25 patients with dicentric chromosomes had involvement of at least one of the chromosome arms 1q, 5p, or 7p resulting in monosomy for 5q or 7q, and/or trisomy for 1q. Patients with dicentric chromosomes presented significantly more often as t-MDS compared to patients without dicentrics (P = 0.046), and the presence of a dicentric chromosome was significantly related to previous therapy with alkylating agents (P = 0.026). Thus, only one out of 27 patients with a dicentric chromosome had not previously received an alkylating agent. A specific susceptibility to breakage at the centromere after exposure to alkylating agents is suggested and may explain the frequent loss of whole chromosomes, in particular chromosomes 5 and 7 in t-MDS and t-AML, if the breaks are not followed by rejoining. Leukemia (2000) 14, 105-111.

therapy-related myelodysplasia; therapy-related acute myeloid leukemia; dicentric chromosomes; alkylating agents

Introduction

Dicentric chromosomes are the result of translocations with a fusion of two chromosome arms and centromeric or pericentromeric sequences from both chromosomes involved. Two general types of dicentric chromosomes are observed: those derived from a fusion of non-homologous chromosome arms and the isodicentrics derived by a fusion of two short arms or two long arms of homologous chromosomes. Because the breakpoints in both participating chromosomes most often occur at, or very close to, the centromere, in a region presumably devoid of genes, these translocations do not directly result in gene rearrangement, as do the reciprocal balanced translocations. Most likely, the genetic consequences of these dicentric chromosomes merely relate to the characteristic loss of the arms reciprocal to those forming the dicentric or, to an apparent duplication of one of the chromosomes or chromosome arms involved.

So far, little is known about why chromosome aberrations arise in malignant diseases, including MDS and AML. However, in a subset of these disorders, namely t-MDS and t-AML, the etiology is well established as being previous therapy with an alkylating agent, treatment with a DNA topoisomerase II inhibitor or, in a small number of patients, previous radiotherapy.^1,2,3,4,5 As the two general types of cytostatic drugs or irradiation have all been shown experimentally to be mutagenic and to induce chromosome aberrations, there are good reasons to suggest that the leukemias as well as the chromosome aberrations are both a direct result of the previous therapy.

In previous studies of t-MDS and t-AML dicentric chromosomes have often been observed, and specific dicentric chromosomes such as the dic(1;7),^6,7 the dic(5;17),⁸ and the dic(7;12)⁹ have been discussed in separate reports. However, the occurrence of dicentric chromosomes has not been studied in general in a large series of patients with t-MDS and t-AML.

The present study was undertaken to analyze the frequency of dicentric chromosomes in our previously reported cases of t-MDS and t-AML compared to the frequency in de novo MDS and AML. Furthermore, we report all chromosomes involved in the formation of dicentrics and verify by FISH the results obtained by conventional karyotyping. We focused on dicentric whole-arm translocations, but also included cases presenting translocations with breakpoints close to the centromeres, because such cases in fact could represent dicentric chromosomes, as shown in a recent study.¹⁰ Finally, we have related the presence of dicentric chromosomes in patients with t-MDS and t-AML to the presence of other chromosome aberrations and to the most important clinical parameters, particularly the type of previous therapy, and whether the disease presented as either t-MDS or overt t-AML.

Materials and methods

One hundred and eighty consecutive cases of t-MDS and t-AML for which the cytogenetic and clinical characteristics have been published previously in detail,^{11,12,13,14,15,16,17} and a further 231 consecutive cases of de novo MDS and AML studied at our laboratory by conventional G-banding chromosome analysis of bone marrow cells, were re-analyzed. The cytogenetic nomenclature of ISCN (1995) was used. In 32 patients an apparently dicentric chromosome had been diagnosed by conventional G-banding and in an additional 25 patients translocations with breakpoints close to the centromeres had been observed.

In order to reassess the presence of dicentric chromosomes in the two groups of patients, two-color FISH was performed on methanol-acetic acid fixed bone marrow cells stored at -20°C from the time of diagnosis. Only in five patients out of 57 was material not available for FISH. Two of the following centromere-specific probes: CEP 1, 3, 6, 7, 10, 11, 12, 15, 17 (Vysis, Downers Grove, IL, USA) or the alpha satellite probes 1/5/19, 7, 8, 13/21, 14/22, 16, 20 (Oncor, Gaithersburg, MD, USA) were hybridized to metaphase cells using standard procedures as recommended by the manufacturers. FISH signals were visualized by a Zeiss Axioskop epifluorescence microscope (Oberkochen, Germany), and images were captured by the Quips Smart Capture FISH Imaging Software from Vysis. An average of 15 abnormal mitoses were examined, and only cases with two different and distinct signals from the centromeric region in at least five mitoses were classified as dicentrics. Whenever necessary whole chromosome painting probes for chromosomes 1, 5, 7 and 17 (Vysis) were applied to positively identify the dicentric chromosome. In six cases with a dic(5;17) and in one case with a dic(7;12) additional arm-specific painting probes for 5q, 17p, 7p, 7q, 12p, and 12q (AL Technologies, Arlington, VA, USA) were used.

Chromosome derivatives were classified as isodicentric in cases presenting a fusion of homologous whole chromosome arms and a FISH signal larger, most often twice as large as the centromeric signal from the normal chromosome.

Results

Dicentric chromosomes were identified by FISH analysis in 27 out of 180 patients with t-MDS or t-AML and in seven out of 231 patients with de novo MDS or AML (Table 1). This difference is highly significant: ² = 17.56, P = 0.00003. In 22 patients with t-MDS or t-AML the presence of a probably dicentric chromosome had previously been reported according to G-banding and in 17 of these patients the results were confirmed by FISH, whereas in five patients a centromeric signal from only one of the two chromosomes was present on the derivative. Similarly, the presence of a dicentric chromosome was verified in three out of six patients with de novo MDS or AML. In patients with translocations and breakpoints close to the centromere according to G-banding, 10 out of 15 patients with t-MDS or t-AML and four out of eight patients with de novo MDS or AML were verified by FISH as dicentric (Table 1). In four patients with t-MDS or t-AML and in one patient with de novo AML, bone marrow material was not available for FISH.

Table 2 shows the specific type of dicentric chromosome observed in 27 patients with t-MDS or t-AML and in seven patients with de novo MDS or AML, and the resulting gain or loss of whole, or almost whole chromosome arms. Most often a dic(1;7) (Figure 1), or a dic(5;17) was observed (Figure 2). All 25 patients with dicentrics derived by a fusion of the two arms from non-homologous chromosomes had involvement of at least one of 5p, 7p, or 1q. This resulted in monosomy for 5q or 7q and trisomy 1q, due to the loss of one whole chromosome 5 or 7, or in cases with involvement of 1q, the presence of two apparently normal chromosomes 1 in addition to the 1q participating in the formation of the dicentric. Conversely, in nine patients with isodicentrics all patients presented trisomy for the involved chromosome arm and eight out of nine patients simultaneously showed monosomy for the other arm of the chromosome (Table 2). None of the nine cases with isodicentrics involved chromosomes 5 or 7 and only one involved 1q. In 10 patients with t-MDS and t-AML for whom dicentrics were identified only by FISH, and in five patients in whom a suspected dicentric could not be confirmed by FISH, the previous karyotype was corrected according to our new findings as shown in Table 3.

In patients with t-MDS and t-AML the dicentric chromosomes were apparently true whole-arm translocations with fusion of the centromeres in all but three cases. In two out of six cases with a dic(5;17) the centromeres were separated by chromosomal material from 5q in one case and from 17p in another case, and the breakpoints were reassessed as dic(5;17) (q12 or q13; p11) and dic(5;17)(q11;p12), respectively. In the third case with a dic(7;12) the centromeric signals were separated by a small region of non-centromeric chromosomal material, by arm-specific painting probes demonstrated to derive from 12p as well as 7q (Figure 3).

In patients with t-MDS and t-AML an interesting and significant difference was observed between the two largest groups of dicentrics, the dic(1;7) and the dic(5;17), with regard to the presence of additional cytogenetic abnormalities (Table 4). Thus, a dic(1;7) was observed in eight out of 10 patients alone or together with one single (two patients) or two (one patient) other cytogenetic abnormalities, which was trisomy 8 in two cases. Conversely, all six patients with a dic(5;17) presented at least three other cytogenetic abnormalities (P = 0.007, Fisher's exact test, two-sided).

As far as presentation of the disease as t-MDS or as overt t-AML is concerned, 23 out of 27 patients with dicentric chromosomes presented as t-MDS, vs 97 out of 153 patients without dicentrics. This difference is significant (² = 3.94, P = 0.046). The result is not unexpected, since 17 out of 27 patients with a dicentric chromosome showed monosomy for 5q or 7q, an abnormality previously shown to be significantly related to t-MDS.^15,16,19 Patient age, latent period from primary tumor to development of t-MDS or t-AML, and response to intensive antileukemic chemotherapy, did not show any significant difference between patients with dicentric chromosomes and patients without dicentrics.

Finally, and most important, the presence of a dicentric chromosome was related to the general type of therapy for the primary tumor (Table 5). Twenty-six out of 27 patients with dicentric chromosomes had previously been treated with alkylating agents either alone (13 patients), combined with radiotherapy (five patients), or a topoisomerase II inhibitor (six patients), or combined with both (two patients). Only a single patient had previously received therapy not including alkylating agents, as she had had radiotherapy only. On the other hand, 37 out of 153 patients without dicentrics had received therapy not including an alkylating agent. Thus, the occurrence of dicentric chromosomes in general was significantly related to previous therapy with alkylating agents (² = 4.94, P = 0.026).

Discussion

The present study demonstrates that dicentric chromosomes in general are significantly more common in therapy-related leukemia as compared to de novo MDS and AML, and are easily overlooked or misinterpreted without FISH analysis. In patients with a dicentric chromosome diagnosed by G-banding this could be confirmed in only 71% by FISH, whereas 58% of patients with translocation breakpoints close to, but not at the centromere according to G-banding were shown by FISH in fact to represent dicentric chromosomes. This clearly demonstrates the superiority of the FISH technique for identification of this type of chromosome abnormality. However, as the centromere-specific FISH probes also cover the alpha satellite sequences flanking the centromeres, it can not be excluded that some breakpoints are not located in the centromere itself, but in the flanking repetitive sequences.

In t-MDS and t-AML the presence of dicentric chromosomes was significantly related to previous therapy with alkylating agents, indicating a specific susceptibility to chromosome breakage at the centromeres following alkylation. This observation may be of more general importance in the understanding of the cellular events leading to monosomy, particularly of chromosomes 5 and 7 in t-MDS and t-AML. Finally, the results may lead to a discussion of why only chromosome 1 or its long arm is duplicated, if 1q participates in the formation of a dicentric.

For many years the alkylating agents have been recognized as the most important etiological factor for t-MDS and t-AML, characterized by monosomy for chromosomes 5 or 7 or deletions of various parts of the long arms of the two chromosomes.^{11,12,13,14,15,16,17,18,19,20,21,22,23,24} More recently, the DNA topoisomerase II inhibitors have also been shown to be leukemogenic, in most instances administered in combination with platinum derivatives or an alkylating agent.^25,26,27,28 These leukemias primarily present balanced translocations to chromosome bands 11q23 or 21q22²⁹ with rearrangement of the MLL and the AML1 genes, but also less frequently other balanced rearrangements such as the inv(16)(p13q22), and the t(15;17)(q22;q11) known from de novo MDS and AML.^30,31

The whole-arm translocations leading to dicentric chromosomes as observed in t-MDS and t-AML in the present study differ generally from the balanced chromosome aberrations. Thus, the dicentric chromosomes are significantly related to previous therapy with alkylating agents, whereas the balanced translocations are related to therapy with DNA topoisomerase II inhibitors. Furthermore, the breakpoints of the dicentric chromosomes possibly do not directly result in gene rearrangement as the recurrent balanced aberrations, as they occur at, or very close to the non-coding centromeric region of the chromosome. Finally, the dicentrics are always unbalanced, because the reciprocal arms of the two chromosomes are lost, or if the long arm of chromosome 1 is involved, this arm is duplicated. Thus, in the future as has been done in the past, we suggest that the dicentric chromosomes, including 5p and 7p and leading to monosomy 5q and 7q, should be classified together with interstitial deletions of the long arms and the complete loss of the two chromosomes.

Previously, it has been demonstrated experimentally that breakage with loss of the telomeric region of a chromosome results in cellular arrest followed by a loss of the whole chromosome at one of the following cell divisions.³² As alkylating agents induce double strand breaks of DNA, this could lead to loss of the telomeric sequences and subsequently loss of the whole chromosome by non-disjunction or by anaphase lagging. The present study demonstrates that another non-coding part of the chromosome, the centromere, may be particularly sensitive to breakage following alkylation, because breakage is a prerequisite of the development of a dicentric chromosome. If breaks occur at the centromere without homologous religation, only chromosome arms undergoing rearrangement and fusion with other chromosome fragments, for instance by the formation of a dicentric chromosome, remain as functional and consistent elements. Isolated chromosome arms, on the contrary, are eliminated, as are chromosomes devoid of the telomere. Centromeric breakage could represent an important mechanism for the development of monosomy 5 and monosomy 7 in t-MDS and t-AML as well as for the development of hypodiploidy in malignancy in general.

The results of the present study support the existence of one genetic pathway in t-MDS and t-AML characterized by deletions of various parts of the long arms of chromosomes 5 or 7 or loss of the whole chromosome 7 as the primary cytogenetic event.¹⁶ In this model deletions or loss of the short arm of chromosome 17 or loss of a whole chromosome 17 were secondary cytogenetic events. A dic(5p;17q) has previously been closely related to mutations of the p53 gene located at 17p13, and to a complex karyotype,⁸ and our results support the latter association.

As demonstrated, the presence of a dicentric chromosome in most cases resulted in loss of the arms reciprocal to those forming the dicentric, except for cases including 1q. Such cases were always associated with the presence of two apparently normal chromosomes 1, resulting in trisomy for 1q. A primary duplication of a whole chromosome 1 could precede and perhaps predispose to subsequent formation of a dicentric chromosome. Alternatively, the formation of a dicentric chromosome containing 1q could precede and perhaps predispose to a subsequent duplication of the whole chromosome. Trisomy for chromosome 1, however, is an uncommon cytogenetic finding in t-MDS and t-AML,^{14,15,16,17,18,19,20,21,22,23,24} and the presence of dicentrics apparently does not result in duplication of other chromosomes or parts of chromosomes, except for 1q and the involved arms of the isodicentric chromosomes.

An interesting new model for the concurrent development of a dicentric chromosome or another unbalanced translocation containing a supernumerary 1q, has recently been proposed in advanced multiple myeloma.³³ In this disease decondensation of pericentromeric heterochromatin was observed and suggested to predispose to chromosome recombination at this region. A simultaneous formation of triradial configurations due to duplication of 1q was seen in some cases. This mechanism by which a translocation and a whole arm duplication of 1q develop simultaneously, should also be considered in t-MDS and t-AML with dicentric chromosomes containing 1q. If in a myeloma cell 16p is the partner and if in a myeloid cell 7p is the partner, the two types of cell could perhaps achieve a specific proliferative advantage, due to a combination of monosomy for one chromosome arm and trisomy for 1q. As in the study of advanced stage myeloma, demethylation of the centromeric region should be considered as the primary injury also in t-MDS and t-AML. Our results indicating that alkylation of the centromeric region could be the eliciting event in t-MDS and t-AML, could also be the mechanism in advanced stage myeloma, because almost all these patients have previously received alkylating agents, being the cytostatic drug most extensively used in this disease.

More basic research, is required to confirm our conclusions, and may lead to a better understanding of the complicated mechanisms by which chromosome aberrations develop in malignant diseases.

Acknowledgements

This work was supported by grants from HS forskningspulje 1997, the Danish Cancer Society, and Anders Hasselbalchs fond til leukaemiers bekaempelse. We are indebted to lnge-lise Frost Andersen for excellent and careful technical assistance in performing the FISH analyses, and to Harry Cowan for help in the preparation of the manuscript.

1 Levine EG, Bloomfield CD. Leukemias and myelodysplastic syndromes secondary to drug, radiation, and environmental exposure. Semin Oncol 1992; 19: 47-84, MEDLINE

2 Thirman MJ, Larson RA. Therapy-related myeloid leukemia. Hematologic complications of cancer. Hematol Oncol Clin North Am 1996; 10: 293-320, MEDLINE

3 Park DJ, Koeffler HP. Therapy-related myelodysplastic syndromes. Semin Hematol 1996; 33: 256-273, MEDLINE

4 Van Leeuwen FE. Risk of acute myelogenous leukemia and myelodysplasia following cancer treatment. Baillières Clin Haematol 1996; 9: 57-85, MEDLINE

5 Pratt CB, Pui C-H. Second tumors after treatment with anticancer drugs. In: Powis G, Hacker MP (eds). The Toxicity of Anticancer Drugs. Pergamon Press: New York, 1991, 28-43.

6 Scheres JMJC, Hustinx TWJ, Geraedts JPM, Leeksma CHW, Meltzer PS. Translocation 1;7 in hematologic disorders: a brief review of 22 cases. Cancer Genet Cytogenet 1985; 18: 207-213, MEDLINE

7 Kibbelaar RE, Mulder J-WR, van Kamp H, Dreef EJ, Wessels HW, Beverstock GC, Haak HL, Raap AK, Kluin PM. Nonradioactive in situ hybridisation of the translocation t(1;7) in myeloid malignancies. Genes Chromos Cancer 1992; 4: 128-134, MEDLINE

8 Wang P, Spielberger RT, Thangavelu M, Zhao N, Davis EM, Iannantuoni K, Larson RA, Le Beau MM. dic(5;17): a recurring abnormality in malignant disorders associated with mutations of TP53. Genes Chromos Cancer 1997; 20: 282-291, MEDLINE

9 Johansson B, Arheden K, Hoglund M, Othzén A, Békássy AN, Turesson I, Heim S, Mitelman F. Fluorescence in situ hybridization analysis of whole-arm 7;12 translocations in hematologic malignancies. Genes Chromos Cancer 1995; 14: 56-62, MEDLINE

10 Berger R, Coniat MBL. Centric and pericentric rearrangements in hematopoietic malignancies. Leukemia 1999; 13: 671-678, MEDLINE

11 Pedersen-Bjergaard J, Philip P, Mortensen BT, Ersbøll J, Jensen G, Panduro J, Thomsen M. Acute nonlymphocytic leukemia, preleukemia, and acute myeloproliferative syndrome secondary to treatment of other malignant diseases. Clinical and cytogenetic characteristics and results of in vitro culture of bone marrow and HLA typing. Blood 1981; 57: 712-723, MEDLINE

12 Pedersen-Bjergaard J, Philip P, Pedersen NT, Hou-Jensen K, Svejgaard A, Jensen G, Nissen N1. Acute nonlymphocytic leukemia, preleukemia, and acute myeloproliferative syndrome secondary to treatment of other malignant diseases. II. Bone marrow cytology, cytogenetics, results of HLA typing, response to antileukemic chemotherapy, and survival in a total series of 55 patients. Cancer 1984; 54: 452-462, MEDLINE

13 Pedersen-Bjergaard J, Philip P. Cytogenetic characteristics of therapy-related acute nonlymphocytic leukemia, preleukemia and acute myeloproliferative syndrome: correlation with clinical data for 61 consecutive cases. Br J Haematol 1987; 66: 199-207, MEDLINE

14 Pedersen-Bjergaard J, Philip P, Larsen SO, Jensen G, Byrsting K. Chromosome aberrations and prognostic factors in therapy-related myelodysplasia and acute nonlymphocytic leukemia. Blood 1990; 76: 1083-1091, MEDLINE

15 Pedersen-Bjergaard J, Philip P, Larsen SO, Andersson M, Daugaard G, Ersbøl J, Hansen SW, Hou-Jensen K, Nielsen D, Sigsgaard TC, Specht L, Østerlind K. Therapy-related myelodysplasia and acute myeloid leukemia. Cytogenetic characteristics of 115 consecutive cases and risk in seven cohorts of patients treated intensively for malignant diseases in the Copenhagen series. Leukemia 1993; 7: 1975-1986, MEDLINE

16 Pedersen-Bjergaard J, Pedersen M, Roulston D, Philip P. Different genetic pathways in leukemogenesis for patients presenting with therapy-related myelodysplasia and therapy-related acute myeloid leukemia. Blood 1995; 86: 3542-3552, MEDLINE

17 Pedersen-Bjergaard J, Timshel S, Andersen MK, Andersen A-ST, Philip P. Cytogenetically unrelated clones in therapy-related myelodysplasia and acute myeloid leukemia: experience from the Copenhagen series updated to 180 consecutive cases. Genes Chromos Cancer 1998; 23: 337-349, MEDLINE

18 Le Beau MM, Albain KS, Larson RA, Vardiman JW, Davis EM, Blough RR, Golomb HM, Rowley JD. Clinical and cytogenetic correlations in 63 patients with therapy-related myelodysplastic syndromes and acute nonlymphocytic leukemia: further evidence for characteristic abnormalities of chromosomes Nos 5 and 7. J Clin Oncol 1986; 4: 325-345, MEDLINE

19 Johansson B, Mertens F, Heim S, Kristofferson U, Mitelman F. Cytogenetics of secondary myelodysplasia (sMDS) and acute nonlymphocytic leukemia (sANLL). Eur J Haematol 1991; 47: 17-27, MEDLINE

20 Kantarjian HM, Keating MJ, Walters RS, Smith TL, Cork A, McCredie KB, Freireich EJ. Therapy-related leukemia and myelodysplastic syndrome: clinical, cytogenetic, and prognostic features. J Clin Oncol 1986; 4: 1748-1757, MEDLINE

21 Rowley JD, Golomb HM, Vardiman JW. Nonrandom chromosomal abnormalities in acute nonlymphocytic leukemia in patients treated for Hodgkin disease and non-Hodgkin lymphomas. Blood 1977; 50: 759-770,

22 Iurlo A, Mecucci C, Orshoven AV, Michaux J-L, Boogaerts M, Noens L, Bosly A, Louwagie A, Van den Berghe H. Cytogenetic and clinical investigations in 76 cases with therapy-related leukemia and myelodysplastic syndrome. Cancer Genet Cytogenet 1989; 43: 227-241, MEDLINE

23 Zaccaria A, Alimena G, Baccarani M, Bilstrom R, Carbonell F, Castoldi GL, Fuscaldo K, Hecht F, Hossfeld DK, Mitelman F, Rosti G, Sandberg AA, Tassinari A, Testoni N, Tura S. Cytogenetic analysis in 89 patients with secondary hematologic disorders. Results of a cooperative study. Cancer Genet Cytogenet 1987; 26: 65-74, MEDLINE

24 Whang-Peng J, Young RC, Lee EC, Longo DL, Schechter GP, DeVita VT. Cytogenetic studies in patients with secondary leukemia/dysmyelopoietic syndrome after different treatment modalities. Blood 1988; 71: 403-414, MEDLINE

25 Ratain MJ, Kaminer LS, Bitran JD, Larson RD, Le Beau MM, Skosey C, Purl S, Hoffman PC, Wade J, Vardiman JW, Daly K, Rowley JD, Golomb HM. Acute nonlymphocytic leukemia following etoposide and cisplatin combination chemotherapy for advanced non-small-cell carcinoma of the lung. Blood 1987; 70: 1412-1417, MEDLINE

26 Pui C-H, Behm FG, Raimondi SC, Dodge RK, George SL, Rivera GK, Mirro J, Kalvinsky DK, Dahl GV, Murphy SB, Crist WM, Williams DL. Secondary acute myeloid leukemia in children treated for acute lymphoid leukemia. New Engl J Med 1989; 321: 136-142, MEDLINE

27 Pedersen-Bjergaard J, Daugaard G, Hansen SW, Philip P, Larsen SO, Rørth M. Increased risk of myelodysplasia and leukemia after etoposide, cisplatin, and bleomycin for germ-cell tumours. Lancet 1991; 338: 359-363, MEDLINE

28 Pedersen-Bjergaard J, Sigsgaard TC, Nielsen D, Gjedde SB, Philip P, Hansen M, Larsen SO, Rørth M, Mouridsen H, Dombernowsky P. Acute monocytic or myelomonocytic leukemia with balanced chromosome translocations to band 11q23 after therapy with 4-epi-doxorubicin and cisplatin or cyclophosphamide for breast cancer. J Clin Oncol 1992; 10: 1444-1451, MEDLINE

29 Pedersen-Bjergaard J, Rowley JD. The balanced and the unbalanced chromosome aberrations of acute myeloid leukemia may develop in different ways and may contribute differently to malignant transformation. Blood 1994; 83: 2780-2786, MEDLINE

30 Dissing M, Le Beau MM, Pedersen-Bjergaard J. Inversion of chromosome 16 and uncommon rearrangements of the CBFB and MYH11 genes in therapy-related acute myeloid leukemia: rare events related to DNA-topoisomerase II inhibitors? J Clin Oncol 1998; 16: 1890-1896, MEDLINE

31 Andersen MK, Johansson B, Larsen SO, Pedersen-Bjergaard J. Chromosomal abnormalities in secondary MDS and AML. Relationship to drugs and radiation with specific emphasis on the balanced rearrangements. Haematologica 1998; 83: 483-488, MEDLINE

32 Sandell LL, Zakian VA. Loss of a yeast telomere: arrest, recovery, and chromosome loss. Cell 1993; 75: 729-739, MEDLINE

33 Sawyer JR, Tricot G, Mattox S, Jagannath S, Barlogie B. Jumping translocations of chromosome 1q in multiple myeloma: evidence for a mechanism involving decondensation of pericentromeric heterochromatin. Blood 1998; 91: 1732-1741, MEDLINE

Figure 1  A case of dic(1;7) examined by FISH. (a) Centromere specific signals for chromosome 1 (red) and chromosome 7 (green). The dic(1;7) contains two closely related centromeric signals (arrow). (b) The same metaphase spread with whole chromosome painting probes for chromosome 1 (green) and chromosome 7 (red).

Figure 2  A case of dic(5;17) examined by FISH. (a) Centromere specific signals for chromosomes 1 and 5 (red) and chromosome 17 (green). The dic(5;17) contains two closely related centromeric signals (arrow). (b) The same metaphase spread with whole chromosome painting probes for chromosome 5 (green) and chromosome 17 (red).

Figure 3  The case of dic(7;12) examined by FISH. (a) Centromere specific signals for chromosome 7 (green) and chromosome 12 (red). The centromeric signals on the dicentric chromosome are clearly separated (arrow). (b) Another metaphase spread with arm specific probes for 7q (red) and 12p (green). The fragment between the two centromeres contains chromosomal material from 12p as well as 7q (arrow).

Table 1  Results of conventional G-banding compared with the results of FISH analysis for dicentric chromosomes

Table 2  Number of patients and types of dicentric chromosomes observed in 180 therapy-related and in 231 de novo cases of MDS and AML

Table 3  Revision of the previously reported karyotype in 10 patients with and five patients without dicentric chromosomes and t-MDS or t-AML based on the results of FISH analysis

Table 4  Number of additional cytogenetic abnormalities in patients with t-MDS and t-AML and dicentric chromosomes

Table 5  Type of previous therapy in t-MDS and t-AML with and without dicentric chromosomes