Document KRvMoeQk08M4yqMxKm6pNOZOo

Clinical and Cytogenetic Correlations in 63 Patients With Therapy-Related hlyelodysplastic Syndromes and Acute Nonlymphocytic Leukemia: Further Evidence for Characteristic Abnormalities of Chromosomes No. 5 and 7 By Michelle M. Le Beau, Kathy S. Albain, Richard A . Larson, James W. Vardiman, Elizabeth M. Davis, Richard R. Blough, Harvey M. Golomb, and Joi:et 0. Rowley Clinical, histologic, and cytogenetic features i n 63 pa- tients w i t h a therapy-related myelodysplastic syndrome (1-MDS) or acute nonlymphocytic leukemia (t-ANU) following cytotoxic chemotherapy or radiotherapy for a previous disease were analyred. Eleven patients had received only radiotherapy for the primary disorder. In most cases, high doses had been administered to treatment ports thot included the pelvic or spinal bone marrow. Twenty-one patients had received only chemotherapy for their primary disease, all for more than 1 year and a l l but one w i t h an alkylating agent, either alone or i n combination with other drugs. Thirty-one patients had received both radiotherapy and chemotherapy, either concurrently or sequentially. A clonal chromosomal abnormality was observed in marrow or blood cells from 61 of the 63 patients (97%). Fifty-five patients (87%) had a clonal abnormality of chromosomes no. 5 andlor 7 consisting of loss of a l l or part of the long arm of the chromosome. The critical chromosome region that was consistently deleted in all 17 patients with del(5q) comprised bands q23 to q32. In addition to nos. 5 and 7, five other chromosomes (no. 1, 4, 12, 14, and 18) were found to be nonrandomly involved. Both t-MDS and t-ANLL ore late complications of cytotoxic therapies thot have distinctive clinical and histologic features and are associated with characteristic aberrations of chromosomes no. 5 and 7. It seems likely that these t w o chromosomes contain genes involved in the pathogenesis of these hematopoietic neoplasms. J CIin Oncol4:325-345. C 1986b y American Society of Clinical Oncology. A DISTINCTIVE disorder of bone marrow morphology and function that terminates in a myelodysplastic syndrome (MDS) or in acute nonlymphocytic leukemia (ANLL! has been recognized as a late complication of cytotoxic therapy used in the treatment of both malipant and nonmalignant diseases. [-I5 Therapy-related MDS (t-MDS) and acute nonlymphocytic leukemia (t-ANLL) represent distinct clinical syndromes that exhibit imponant differences from ANLL that arises de In patients with no prior history of exposure to known mutagens, ANLL de novo usually appears suddenly. whereas t-ANLL is often preceded by a secondary MDS (or preleukemic phase) that may persist for several months. Frequently. all three hematopoietic cell lines. er>throid. myeloid, and megakaryocytic. are involved in this myelodysplastic process. and patients characteristically manifest an initial pancytopenia.b.!b [' Infection, bleeding. and anemia are common complications and ma> prove fatal even before overt leukemia evolves. Sunival times of patients with t-ANLL are often shon because this disorder has been less responsive to current foms of leukemia treatment than has ANLL de novo.IhIn the future, as a greater fraction of patients survive treatment with cytotoxic drugs or ionizing radiation for their primary concer. the incidence of t-MDS or t-ANLL arising secondary to such treatment may be expected to increase Characteristic nonrandom chromosome abnormalities are commonly observed in bone marrow cells of patients with t-MDS or tAAXLL.'.!.!O-" These abnormalities differ in their t>pe and frequency from those noted in ANLL From the Dcpartmenls of Medicine and Parholog!. and The Frun!Jin .McLrun .Memorial Research Instirrite. The Universrn qf Chi,-aee Sirhmitred Jill\ -7. 198.7: accepted No! 8 . 1985 Suppc1r:ed in p u n by LIS Department of Energy Conrracr DE.4CO2-.?0E1'1060.b\ h'arionul lnsririrres i f Health Grants .Yo. C.4-239.54 und C.4-25568, and h! the C'niversit? qf Chicago CJm er Rcscurch Foundation. M..M.L.B IS a Special Fellnn of rhr Lruhrmic Socien of America. .Address reprinr requests in Michelle M . Le Beau. PhD. B a t 120 l ' n n ersin qf Chicago Medical Center. SX4 I S M u d a n d d l r Chit-upo. IL 60627. C 1% h! .-\mcriran Socien of Clinical Oncolop. 0:32-183,s 96 (UV.~-00lRS3 oo!O Journal of Clinical Oncology, Vol 4, No 3 (March). 1986: PP 325-345 325 4 326 developing de novo. We reported previously that part or all of chromosomes no. 5 and.,, 7 was lost in cells from 13 of 26 (88%)t-MDS.t-.ANLL patients.' and this has been confirmed by others.'"" In contrast, only about 16% of patients with ANLL de novo have a similar abnormality of chromosomes no. 5 or 7 or both.': Moreover. the latter patients frequently have had significant occupational exposure to potential environmental carcinogens such as chemicals. solvents. or pesticides.".'5 Another major difference between the karyotypic pattern of t-ANLL and that of ANLL de novo involves the frequency of nonrandom structural chromosome rearrangements. In t-ANLL, one seldom finds the specific rearrangements that are closely associated with the distinct morphologic subsets of ANLL de novo such as the t( 15:17) in acute promyelocytic leukemia. the ((821) in acute myelogenous leukemia with maturation. or the inv(l6) in acute myelomonocytic leukemia with abnormal eosinophils.2 6 . ~ - This report details the clinical. histopathologic, and cytogenetic data on 63 patients with tMDS an&or t-XNLL. thus expanding the findings on our earlier series of 26 patients.'.' Based on these data. we have asked the following questions: (1) Was there a correlation between the type or dose of radiotherapy or chemotherapy and the development of t-.MDS/t-ASLL'? (2) What was the median time from priman. therapy to bone marrow dysfunction'? (3)Was the high frequency of abnormalities of chromosomes no. 5 and 7 confirmed in these cases and. if so. how were these abnormalities related to the developmenf 0.ft-ANLL'? (4)Was there a relationship between the chromosomal changes and either the primary diseasz or the type of primary therapy'? ( 5 ) Were abnormalities of chromosomes other than ai). 5 or 7 observed more frequently in these patient, with 1-XNLL than in patients with ANLL de novo in a concurrently analyzed series? (6) Did specific chromosomal abnormalities correlate with response to antileukemia therapy or with median survival after the diagnosis of t-ANLL? ( 7 ) By analysis of the chromosomal breakpoints in cells with a deletion of no. 5 or 7 , could the smallest region that is consistently deleted in every patient. ie. the critical region. be defined? The identification of this region is of crucial importance because it may contain genes LE BEAU ET AL whose altered function can result in hematopoietic neoplasia. MATERIALS A N D METHODS Clinicd and ,I.lorphologic Analwis The 63 patients described here include all those who were first seen at or were referred to the University of Chicago Medical Center with cl diagnosis of !-%IDS or t-ANLL that could be confirmed b) morphologic study of bone marrow specimens. Each patient had received cytoloxic therapy (chemotherapy. ra- diotherapy. or some combination of these) for an antecedent disease. Patients no. 2001 to 2026 and 2036 have been reported p r e v i o u ~ l y . ~T~h"ir~ty~-nine patients had been treated for their primary disease at the University of Chicago. whereas 24 had been treated at other institutions. Therefore. the total number of similarly treated patients at risk for developing t-ANLL in each primary-disease category is not available. Only one of the 24 patients referred from other institutlons was also treated at the Cniversity of Chicago for the subsequent leukemic process: the other 23 were examined at our institution on a consultative basis. or clinical information on these cases was fomarded to the Univenity ofChicago when a s p i m e n was sent for cytogenetic analysis. Clinical data were collected. and details concerning the pri- mary treatment were obtained from a review of each patient's medical history For the refeml cases. clinical data were gath- ered by direct communication uith the referring physician(s1. Radiation therapy pons. the dose of each treatment course in rad. and the doses and duration of each chemotherapy course were determined whenever possible for each patient. Patients were monitored until death or through Februa? 1985. The dura- tion of the secondary myelodysplastic phase. the type of treat- ment given for the leukemic'process and the response. and the survival from the time of initial bone marrou dysfunction were noted The diagnosis of t-MDS was made when the patient's periph- erd hlood and bone marrou showed features of dyspoiesis as defined by the French-.American-British ( F A B )criteria for MDS and described by us and others 3s characteristic of the changes seen in t-MDS.h.2'-3') Patients classified as having I-MDS had < 30% blasts in the marrow. whereas the diagnosis of overt tANLL was made =hen the percentage of blasts were > 30%. as determined from marrow aspirates or as judged from marrow biopsy sections when increased reticulin prevented aspiration. As noted in other a substantial number of cases could not readily be classified according to FAB criteria or were somewhat atypical: for the purpose of this report. unless other- wise noted. a11 of these cases are simply listed as t-AYLL Cyogeneric Andysis Cytogenetic analyses were performed with quinacrine fluorescence and tr).p.\in-Giemsa banding techniques on bone marrow cell, from aspirates or biopsy specimens or on peripheral b l d cells obtained ;It the time of diagnosis. The metaphase cells examined were obtained from direct preparations. from 74or 18-hour unstimulated cultures. or from methotrexate-synchronized cells cultured for 2 1 or 18 hours uith phytohemagglutinin-stimulated leukocyte-conditioned medium. Chromosome abnormalities are described according to the International Sys- CLINICAUCYTOGENETIC FEATURES OF t-ANLL 327 tem for Human Cytogenetic Nomenslature." The cnteria ~rdopted at the First International Workshop on Chromosomes in Leu- ''kemia were used for the identification of abnormal done> Statisticui MtTtliods The rrlarion\hip of chromosomdl abnormalities to variables describing cliniaal features were in\estigated. For evaludrions invol\ing unl? uncensored variablcs. ue used pnmaril) \tatter plots. comparisons of frequency di,tnhutions. and comparisons of means. Cencored variables were itudird ~ i t hthe techniques of sumival an;ll!xis Survival c u n e s and their associated median, were computed by the actuarial nierhod. with grouping intervals of 1 months. Descriptive statistics. correlations. chisquared statisrics. and actuanal survival c u n e s and their comparisons uere computed with programs of SPSS !Statiztical Package forthe Social Sciences). \Vhen cell sizes were too small for a valid approximation of the chi-squared distribution. Fisher's exact test or the analogous exact test developed h! Pagano and Hahursen ~ a u5sed." ,411 data processing and cornputatiom were done on a DEC-20 computer (Digital Equipment. Marlhoro. Masr,. accessed through the Biomedical Computation Facilities at the University of Chicago. RESULTS Clinical Characteristics The clinical characteristics of the 63 patients are listed in detail in Tables 1 to 3 and are summarized in Tables 4 and 5. There were 3 1 men and 32 women. ranging in age from 6 to 76 years at the time of diagnosis of treatment-associated bone marrow dysfunction. Based on morphologic findings. t-MDS without subsequent t-.4NLL developed in 19 patients (30%). 29 patients (46%) had t-MDS followed by t-ANLL, and 15 patients ( 2 4 4 )had t-ANLL only. It is not known whether any of the latter group had had an antecedent t-MDS because the population at risk had not been monitored prospectively. Table 4 lists the type of primary disease, the mode of primary therapy. and the time from primary therapy to initial bone marrow dysfunction. There were 1 3 patients (37%) with Hodgkin's disease, ten ( 16%) with non-Hodgkin's lymphoma, six ( 10%)with other hematologic malignant diseases. 21 (33%) with various solid tumors. and three ( 5 % )renal transplant recipients. Altogether. 21 patients ( 3 3 9 )had received chemotherapy alone for their primary disease. 1 1 (17%;)had received radiotherapy alone. and 31 (493;) had 'received both modalities. Of the patients who had received only chemotherapy, all had undergone more than I2 months of treatment. and all but one had received an alkylating agent (most often cyclophosphamide. nitrogen mustard. procarbazine. chlorambucil, or methyl CCNU) either alone or in conibination with other cytotoxic drugs. The single exception was a patient (no. 2052) who received oral doses of 5-fluorouracil (5-FL')daily for 5 years for squamous carcinoma of the rectum. In many cases. the primary chemotherapy doses had been sufficient to cause moderate to severe myelosuppression: but in ten casec. cyclophosphamide. melphalan. or chlwir-nbucil had k e n taken orally for a period of' many months ur years only at low daily doses that do not commonly cause significant myelosuppression. Three patients with colorectal carcinoma had received multiple courses of methyl CCNU and 5-FU for ;t period of 15 to 27 months. Three renal tranhplant recipients had received cyclophosphamide with or without azathioprine for 1.5 to 16 years. Nine of the 1 I patients who received only prima? radiotherapy had had the major portion or all of the pelvic bone marrow included within the initial treatment ports. Two patients with Hodgkin's disease (2.000 rad. 4,000 rad) and one with non-Hodgkin's lymphoma (4.000 rad) had received total nodal irradiation. Two patients with prostate carcinoma had received 5.000 rad to the entire pelvis plus an additional 2.OOO rad to the prostate. Two patients with endometrial carcinoma and one with cervical carcinoma received radiotherapy to the entire pelvis by extemalbeam techniques in addition to radium implants. A second patient with carcinoma of the cervix was treated with a radium implant alone. Of the two radiotherapy patients who did not receive pelvic irradiation, one with laryngeal carcinoma was treated with a laryngeal pon only; the other, with metastatic carcinoma of the lung, was treated with mediastinal, hepatic, and femoral ports. The 31 patients who had received both radiotherapy and chemotherapy were a heterogeneous group. In 16 cases, the radiotherapy and chemotherap! had been administered approximately concurrently. In another nine cases, the radiotherapy had been administered as initial treatment. and chemotherapy was administered only later. at the time of relapse of the primary disease. In most of these cases. high doses (3.000 rad to 6.000 rad) had been administered to bone marrow-containing regions over total nodal field3 or pelvic or spinal ports. In the remaining six patients, radiotherapy followed initial che- 328 fe l f 0 f x U -.0-) -g Y0 .E- c 2 3 Y E. P 3e `1E m f isYI e .E- c 0 & c e4 Y 0 .H.c- c0, e 0 -53 U .-.sc-- U c ai -0 c0 ' LE BEAU ET AL 0. 0 N I 0t L 6 t 2 + et a U "l % 2 'A z c v) - 3 Ue 0. I. L 00 x+ 22 N X -a3 ry Ua c c rl 2 8 CLINICAUCYTOGENETIC FEATURES OF t-ANLL 329 -az + xx + Lp* 22 26X 6 4 s+ + au ua 5+ + 4 o o ua ua u0 c c t a U tC ua ua + a U fS 2.2. " oz ee s 2. G 2. s3 2".-" o2.w2. ee 2 20h. 0 (Y 9 v) I I 1 I" %-8 0 Ue z I- 6.*- .- A CLINICAVCYTOGENETIC FEATURES OF t-ANLL 33 1 . 332 LE BEAU ET AL Table 4. Primary Disease, Type of Therapy, a n d Time to Bone Marrow Dysfunction Primary Therapy Mean Time 'to Bone Marrow Primarv Disease No. of ' ChemaPatients theraov Radiotheraw Both Dysfunction (mo) Hodgkin's disease 23 4 4 17 63 Nan-Hodgkin's lymphoma 10 3 1 6 53 Other hematologic malignant 6 4 0 2 64 diseases' Solid tumorsT 21 8 8 5 65 Renal transplants 3 2 0 1 101 Total 63 21 11 31 64 Mean time to bone marrow dysfunction (mo) Median time to bone marrow dysfunction (mo) Range (mo) 64 50 22-192 59 65 34 60 10-183 20-132 56 *Hairy-cell leukemia, one; multiple myeloma, five. +Lung carcinoma, two;breast, four, colorectal, four; cervicoliendometrial, four; ovarian, two; prostate, two; anus, one; glioblastoma, one; larynx, one. motherapy. and more commonly. only local fields were irradiated. Table 4 shows the mean times from initial treatment to subsequent bone marrow dysfunction for all patients (median. 56 months) and for the subgroups based on primary disease and primary therapy. Neither the comparison according to primary disease nor that according to primarytherapy modality was statistically significant despite the unexpected trend of a shorter time to bone marrow dysfunction for the radiothsrapyalone group than for either the chemotherapyalone p o u p or the group receiving both modali- ties. These latency intervals. however. may not be accurate because of the variable observation of these patients after the completion of primary therapy. The median age at the time of bone rixirrw dysfunctinn \.vas 55 \cars overall. but was lower for the patients hith Hodgkin's disease (37 years) and for the renal transplant recipients (40 years):this observation probahlv reflects the low& ages of these patienis at th; time of primary treatment. In Table 5. the three secondaq bone marrow disorders (t-XIDS. t->IDS-+t-ASLL. t-ANLL) are categorized according to the primary disease and the primary therapy. There were no statistically significant differences in these distribu- tions. The intervals between primary treatment and subsequent bone marrow dysfunction were 43, 5 I , and 67 months for patients with t-MDS alone. t-MDS evolving into t-ANLL. and tANLL alone. respectively. Although interesting. this trend did not reach statistical significance. The median duration of the m>elodqsplastic phase in the 29 patients in whom t-ANLL subsequently developed was 5 months. Table 5. Primary Disease, Primary Therapy, and Secondary Disorder Number of Patients With Secondary Disorder t-MDS t-MDS 1 t-ANLL only t-ANLL only Primary disease lymphoma Other hematologic malignant disease Solid tumor Renal transplant Primary therapy Chemotherapy only Radiation only Both modalities 7 19 31 98 01 88 44 7 17 7 2 4 2 5 3 7 Total 19 29 15 CLINICAL'CYTOGENETIC FEATURES OF t-ANLL Rrspotise io Trenimctit ntid S u n Y d A j k r Boric Marro,c. D!,sfiiric.iioti Of the 19 patients with [-\IDS only. most died of infection. hemorrhage. or some other coniplication of pancjmpenia. Treatment u'as supportive only: none of the patients were given intensive antileukemia chemotherapy. Two patients received daily oral doses of cis-retinoic acid. and one recei\.eddaily subcutaneous injections of cytarabine at low doses: none of these patients showed an) evidence of response. The 43 patients in whom t-ANLL de\.eloped were treated in various ways. depending in part on the year of diagnosis and the treating institution. Thirteen patients received supporti\ e care only. either because the! refused treatment or because of a moribund state at the time of diagnosis. Twenty-eight patients received conventional AXLL remission induction chemotherrtp!. with cytarabine and either 6-thioguanine or an anthracycline drug. and five patients received high-dose cytarabine therapy ( 3 gm/m' e v e n 11 hours). Five patients had a complete remission: four of these had not had an antecedent t-MDS. Three of the complete responders subsequently relapsed and died. one died in remission after bone marrow transplantation. and one is alive with no evidence of disease after 57 months. None of the patients with partial or no response to treatment are alive. Most patients died of infection or bleeding during prolonged periods of bone marrow hypoplasia following intensive chemotherapy. The median survival time for all 63 patients from the time of initial bone marrow dysfunction was 8 months. No significant differences in survival were noted based on age. sex. primary disease. or primary therapy. There was. hotvever, a statistically significant difference ( P < .OS) in median survival among patients wjth t-MDS only ( 6 months). those with t-MDS that evolved into t-ANLL ( 9 months). and those with t-.ANLL only ( 5 months). This result may be influenced. of course. bj- the possibility that an antecedent myelodysplastic phase had been undetected in the last group of patients. who had frank t-.ANLL at diagnohs. Cytogetietic ,haIyses The results of the cytogenetic anal!ses are shown in Table 6. None of the 62 patlent5 had 333 received antileukemia therapy before chromosomal anal~sisA. nalyses were performed on bone marron cells in 34 cases and on cultured unstimulated peripheral blood cells in the remaining nine cases. Clonal chromosomal abnornialities Here observed in 61 of the 63 patients (97%). Consistent clonal abnormalities leadin: to i i x y of the u hole chromosome I*T !O.iS of par? of [hi. j m s arm of chromosomes no. 5 or 7. or both of these chrorncicomes. uere observed in 55 of the 61 patients (90%)u ith abnormal karyotypes (87% of all patients). Abnormalities of chromosome no. 5 occurred in 1 1patients; six had a loss of one chromosome no. 5 . and eight others had a deletion of the long arm of this chromosome (delccqr].Loss of all or part of the long arm of one chromosome no. 7 was noted in 24 patients; '71 had nionosom! 7. Patient 2.039 had an unbalanced translocation resulting in monosomy of the long arm of no. 7 . and patients 2027 and 2029 had deletions of the long ann. An additional 17 patients had abnormalities of both no. 5 and no. 7 . In nine of these 17 patients. the aberration of no. 5 was a del(5q): in four patients. loss of 7q resulted either from a deletion (two patients) or from an unbalanced translocation (two patients). Thus. among the 55 patients. 14 patients had a - 5. 17 patients had a del(5q), 34 patients had a - 7 . and sei'en patients had a loss of the long ami of no. 7 that resulted from an unbalanced translocation in three patients. A del(5q) was the most coninion structural aberration in our series. Although abnormalities of nos. 5 and 7 were frequently obsened in the same clone. these abnormalities may in fact occur independently. That is. the number of cases in which aberrations of both nos. 5 and 7 were observed was similar to that expected to occur by chance. given the high frequency with which abnormalities of either one or the other of these chromosomes were present in these patients. The most common single abnormality was loss of one no. 7 : this karyotype was observed in eight patients. On the other hand. of the 3 I patients with abnormalities of no. 5 . only one (patient no. 2022) had an alteration of this chromosome [del(Sq)las the sole karyotypic change. In eight patients it was possible to determine whether an abnormality of no. 5 or no. 7 occurred first because we were able to study chromosome et-olution in sequential samples (three L I c ! - 334 LE BEAU ET AL Table 6. Results of Cytogenetic Analysis of 63 Patients With Therapy-Related MDS or AN11 Patient No. Specimen Method' No. of Metophase Cells Percentage Abnormal Koryotype f 200 1 2002 2003 2004 2005 2006 2007 2008 BM BM PB BM PB BM BM BM PB 2009 2010 201 1 2012 2013 BM BM PB BM EM 2014 2015 2016 BM BM PB BC 2017 2018 2019 2020 BM BM BM BM 202 1 2022 2023 2024 2925 2026 3.327 BM BM PB BM SM BM EM 2028 BM 2029 2030 203 1 BM BC PB direct 24 h 24 h, 48 h direct 24 h, 48 h direct 24 h 24 h 24 h, 48 h direct direct, 24 h 24 h 24 h 24 h 24 h 24 h 24 h, 48 h 24 h 24 h 24 h 24 h 24 h direct, 24 h direct 24 h direct 24 h 24 h 24 h direct, 24 h 24 h 24 h 24 h, 48 h 18 90 4 6 , W 4 4 , X X . - 5 , - 7 . - 1 6 , - 1 8 , + 2 m o r 11 100 48,XX, - 5 , - 16, -2l,t(8;?)(~23;?).1(15;?) ( p l 1;?), t m a r l , + mar2, t 2mar3, + mar4 17 100 46,XY,t(1;17)(~36;q21) 10 30 46.W45,XX. - 7 12 92 46.W45,XX. - 4, -5,-14,t(Sl;?)(q22;?), + 2mar 4 100 43,XX. - 5,- 7, - 12,del(6)(q15orq21), +min 6 75 46,XYI45,XY, - 5, - 7, - 20, + Zmar 14 93 46,W'45,XX,-5, - 7, - 12,- 17, -22,t(l;?) (~3?6;?),1(13;?)(q34;?),1(14;?1);?()~,d1el(15) (q2?4), + 4mar 9 100 46,XX,-5,-7,-17,+8,del(6)(q13),+2mar 12 77 46,W43,XX, - 5, - 13, - 18 (15%)/44, M ,- 5, - 18,t(13;?)(pl l;?) (62%) 13 30 46.Xy1'46,XY. - 5, + mar 13 93 46,XY/47,XY, + 8 21 90 46,W46,XX, - 7, - 16, - 21, + t(2;7)(2q7q),del (2)(q33),de1(1l)(q22), + t(l l;?)(p15;?),t(14;?) +( p l 1;?), der(2l )t(21;2 1;21;16)(21 pter- q22::21qll-q22::21ql 1--tq22:: 16~13-qter) 23 80 46.W45,XX. -7,inv(l)(p36q13) 13 100 45,XY,-7 10 60 46,XY/44,XY, - 7, - 16,t(l7;?)(pl l;?), del(2O)(qllql3)(10%)/43,XY, - 5, - 7, - 12, - 16,t(4;?)(pl6;?),t(l7;?),de1(20+)m,ar (2006)/43,X, t(.Y;?)(q 12;?), same (30%) 12 100 45,XY. - 5, -7,de1(3)(p13), Lder(5)t(5;17) (ql?l;qll) 16 94 46,XY,'45,XY, - 7,de1(5)(q13q33) (38%)1 45,XY, same, t( 12; 17)(q1?3;p1?2) (4406) 13 100 45,XY. - 7,del(3)(pl?3),de1(8)(q22),del(l6)(pl2) 8 88 46,XY.'44,XY, - 2, - 3, - 6, - 12, - 15, - 21, - 22,dei(2)(pll ),del( 1 l)(pl?4),t( 14;?)(q32;?), +t(l7;?)(pl3;?),deI(22)(ql?l), 5mar 6 0 46,XX 6 100 45,XX. - 13, - 14, t t(13;14)(p13;ql l),de1(5) (413q33)+ 22 100 45,XX, - 7,t(3;9)(q29;~21) (73%)/ 44,X, -X,some (27%) +25 100 5 1 , X Y , - 7 , - 1 4 , + 1 , + 2 . + 6 , + 8 , + 1 5 . 21, +mar 5 100 45,XX, - 7,t(3;3)(q21;q26) 30 100 45,XY. -7,t(2;3)(~21;q27) 7 100 44,XX. - 15, - 16, - 19,de1(7)(qll), + der( 19)t(19; 15;8)( 19pter- ql3:: 15ql l-q26::8q21+qter) +19 100 59,XY, iX, - 7, - 12, + 1, 2, +4, + 11, +T 15, 18, i19, +20, +21, ~ 2 1t,del(2) +(q2?3q3?1),t t(B;?)(q2?4;?), 2mor 15 13 46,XY:46,XYI del(7)(q34) 20 100 45,XX,-7 8 100 43,XX, - 7,- 8, - 10, -22,+mar (50%)/43, XX,same,t(l7;?)(p13;?) (37?/0)/43,XX,same, 1(17;?),t(4;12)(qZZorq23,p13) (13Yo) CLINICAVCYTOGENETIC FEATURES OF t-ANLL 335 Table 6. Results of Cytogenetic Analysis of 63 Patients With Therapy-Related MDS or AN11 (Cont'd) Patient No. Specimen Method* No. of Metaphase Cells Percentage Abnormal KarvotvDe 2032 2033 2034 2035 2036 2037 2039 2040 204 1 2042 2043 2044 2045 2046 2047 2048 2049 2050 205 1 2052 BM BM BM BM BC PB BM PB BM BC BM PB BC BM BM, PB BM BM BM BM PB BM BM 48-h MTX direct, 24 h 24-h MTX 24-h MTX 24 h 24 h 20 17 20 26 2 2 48-h MTX 26 24 h. 24-h MTX 13 24 h, 48-h MTX 24 h 24 h, 72 h 24 h 24 h 19 30 56 23 17 24 h 48-h MTX 24 h, 48-h MTX 18 22 25 direct 48-h MTX 24 h 34 14 18 24 h, 48-h MTX 24 h, 48-h MTX 24 h 20 22 23 100 45,X. - X, - 7, -der(X)t(X;7)(p22;p1?5),?(1;41 (p36;pl2),deI(5)(ql lq34).del(l2)(~1?2) (80?6)/4 related single cell obnormalities 24 46,XX,45,X, - X, - 12, - lb,del(3)(q21),del (5)(q13q33),i(21q), i2mor 100 47,U, + 8 77 46,XY:46,XY, - 7, t del( l)(p32) (12%)/47,XY, +del(l) (31%)/49,XY, +8,+9,+del(l) +(19%)/5O,XY,same, 15 (8%) 100 46,XY/47,XY, - 5, - 12, - 15,del(2)(q33),del 50 (7)(q22q35),del(l9)(q12)-, del( 19),t(l2;15) -(12q15q), 2r (25%)149,XY, - 5,de1(2), -+de1(7), +mor, 3r. ldmin (25O/O)i 45,XY, - 5, - 12, - 15,de1(2),deI(7), del(l9),t(12;15), r (25?b) 73 46,XY 46,XY,del(5)(q13q32wq33) (6O0/0)/ 45,XY. - 1, - 13, - 18, 4del(l)(ql?l), +dic (1;4)(cen or pll;pl6),de1(2)(q23q36) 100 45,XX, - 5, - 13, - 15, +t(13q15q), - marl -(3196)!45,XX,-5, -13,-15,+t(13q15q), mar2 (15%)/7 single cell abnormalities, same, -1i- 3mor +15 46,XY846,XY, - 7, der(1)t(1;7)(p 1 1;p 11) 100 46,XX. - 7,t(21;?)(q22;?), +mar 93 46,xX46,XX,t(l5; 17)(q22;q2 1) 96 46,W45,XX, - 7 76 46,XY,43,XY, - 7, - 17, - 18, - 19, t(6;1O)(p25;qZZ),de1(9)(q12), -inv(11)(p15q13). t( 16:?)(p 1?1;?), der(lP)t(17;19)(ql3;q21)(65%) 44,XY, - 7, - 17, - 19,inv(l 1), -der(19),t(20;?)(q13;?),del(lO)(pl3)(12%) 68 46,xx'45,XX, - 7 (52%)/45,XX, - 7,del (5)(q15q31) (l6Y0) 91 46,XYi45,XY, - 7 (82%)/45,XY, - 7,de1(5)(q23q32) (9%) 68 46,XYl48,XY. - 4, -7. - 14, - 16, + der(4)t(4;14)(q3 1;ql l).del(5)(q13q33), t(9;?)(p24;?),del(l2)(pl2pl3),t(21;?) (q22;?), +marl - 5 (36%)/47,XY, - 3,some +(16%)/47,XY,same, mar2 - 5 +100 48,XX, t 1, 11,del(5)(ql3q33) 100 46,XY, t(3;4)(q28orq29;q2 1orq22),del(5) (q23q32),del(9)(q2?2) 100 47.U. - 7, - 18,del(4)(q22q31), del(s)(q22932),de1(9)(q13),de1(21)(q22) (83%)/43,XX,sarne, - 21 (17%) 95 46,XYi45,XY, - 7 100 45,XY, - 7 100 4 5 , X X , - 7 , - 1 0 , - 1 7 , - 2 0 , - 2 0 , 4 8 . del(5)(ql4q34), t der( lO)t(10;17)(q22; p1?3),t(12;?)(pl l;?), +der(20)t(7;20)(p13; p1?3), +marl (57%);46,XX, - 7, - 10, - 17, -- 20, - 20, + 8,de1(5). +der(lO),t(l2;?), der(20). mor2, + mor3 (22Y0)!5 related single cell abnormalities t j ji i I - nnn LE BEAU ET AL Table 6. Results of Cytogenetic Analysis of 63 Patients With Therapy-Related MDS or AN11 (Cont'd) Patient No. Specimen Method* No. of Metaphase Cells Percentage Abnormal Karyotype I 2053 BM 24 h, 48-h MTX 20 80 46,%'44,XX, - 7, - 18 (25%)/44,XX. - 7, - 20, del(5)(q12q31) (55%) 2054 BM 24 h, 48-h MTX 22 91 46,XY/46,XY,del(5)(q13q34) (9?/0)/45, XY, - 6 ,- 15, - 17, i 8 , +der(17)t(6;?;17) (q13;?;pl3),de1(5) (23%)/46,XY,same, imor 2055 BM 48-h MTX 20 80 46,W44,XX,-3,-17,-18,-22,del(5) (912q31), +der( 12)t(l2;12)(pl 1; q21), der(lb)t(l6;22)(q24;q13), +der(17)t(3;17)(ql l;p13) (35%)/45,m, same, +mar (45%) 2057 PB 48-h MTX 22 95 46,XY/47,XY, +8,t(l5;17)(q22;q21) (55%)/46,XY, - 21, -21, +8,t(15; 17), +der(21)t(21;21)(pll;qll or cen;cen) 2058 BC 24 h (45%) 18 100 45,XX, - 5, - 12, + 8,t(& ;8)(q3?4;q2?2),dic(& ; 12)(p14;q13),t(18;?)(pl 1;?) (39%)/45,XX, some, - 6 ,+der(6)t(6;1l)(p22;q13) (61%) 2059 2060 BM BM 24 h, 48-h MTX 24 h, 48-h MTX 21 21 90 46,XY/45,XY, - 7 +100 47,XY,t(l9;?)(q13;?), mor1 (19%)/46,XY, -7,t(l9;?),t(u;?) (q25;?),del(lJ)(pl3), + marf(43%)/44, XY, - 7, -2~,t(l9;?),t(lJ:?),del (11) (19?6)/46,XY, - 7,t(19;?),t(u;?), del(lJ), + mar3 (10?/0)/2 relotedringle cell ibnormalities 2061 BM 24 h 20 90 46,XYt'46,XY, - 16,de1(4)(q21q31),de1(5) (q15q31),t( 18;19)(p23;pl3),t(9;17; 13) (q22;pl3;ql?4),t( 14;?)(pl3;?), t(21;21) ( p l lorpl2;q21), +marl (40h)/45,XY, - 17, some (20?/0)/45,XY,- 17,same, + f(lo%)/ 45,XY, - 10. - 17,some. mar2'2reloted 2062 BC 24 h single cell abnormalities 19 100 45,XY, - 21, - 21, t( 1;6)(q25:ql3), de1(7)(q22q32 orql lq22),deI(P)(ql3q22), +marl (47%)/45,XY, - 21, -2l,t(l;6), de1(7),de1(9), ider(21)t(21;2 1)(pl2;q 11), del(S)(q13q33) (37%)/3 reloted single cell abnormalities 2063 BC 24 h 28 0 46,XX PB 24 h, 48-h MTX 2044 BM 24 h 21 90 46,XYi45,XY, - 7,(859/0)/45, -- -___ ~ XY, -7,de1(6)(qZlq24) (5%) Abbreviations: BM, bone wrrow; PB, peripheral blood; BC, bone morrow biopsy. 'Dwect refersto direct preporations; 24 h and 48 h refer to cells cultured for 24 or 48 hours; 48-h MTX refers to cells cultured for 4 8 hours with phytohemogglutinideukocyte-conditioned medium and synchronized with methotrexate. ?Patient no. 2022 had a constitutional chromosomal abnormality consisting of o Robertsonion tronslocafion between chromosomes no. 13 and 14. patients) or because multiple clones showing evolution were identified in the initial sample (five patients). In six of the eight patients, the abnormality of chromosome no. 7 occurred first. Clonal abnormalities not involving chromo- somes no. 5 or 7 were noted in only six patients. Two of these (patients no. 9041 and 2057) had a t( 15;17)(q22:q21) and the characteristic rnorphologic features of acute promyelocytic leukemia." Patients no. 20 12 and 2034 each had only a gain CLINICAL'CYTOGENETIC FEATURES OF t-ANLL Marker chromosomes Dresent u:v) 14c ." I234~670910111213141516!1I019202122KY CHROMOSOME NUMBER Fig 1. Distribution of the gain and loss of whole chromosomes from the 61 patients with t-MDS or t- ANLL who had clonal abnormalities. of no. 8. Patient no. 2057 had both a t( 15:17)and a gain of no. 8. Various structural rearrangements were noted in the remaining two cases (patients no. 2003 and 3013). The frequencies of gain or loss of a whole chromosome and of structural rearrangements are illustrated in F i p 1 and 2 . With the esception of a gain of no. 8 that was noted in ten patients. a 120 TRANSLOCATION DELETION CHROMOSOME NUMBER Fig 2. Distribution of the involvement of individual chromosomes in structural rearrangements from the 61 patients with t-MDS or t-AN11 who had clonal abnormalities. 337 pain of other chromosomes was infrequent. A gain of chromosomes no. 5 or 7 was never seen: however. among patients whose karyotypes shon ed chromosome loss. monosomy for these two chromosomes was observed most frequently. Losses of nos. 12. 16, and 18 were the next most frequent. being observed in five patients each. Chromosomes no. 3. 5. 7. IO. 12. 13. 13. 16. 17. 22. and Y were never gained. and no. I , 2. 9. 1 I . 19. and Y were never noted art~ungthe lossec. Of the sxucturnl rearrangrments obs e n d (Fig 2 ) . a deletion of 5q was noted in I7 patients. and translocations involving no. 17 were seen in 13 patients. Of the 13 translocations of no. 17. six involved a breakpoint at band p13. and three. a breakpoint at band q 2 l . Rearrangements of chromosomes no. IO. 1.7. 16. 18. 20. 22. S . and Y were seldom observed. Comparison of the Freqrtetic:\. of Chrornosotnnl AbtiortnnliriPs in t-A+'LL atid AVLL De NCHYI To determine whether the frequency and types of abnormalities observed in patients with tMDS and t-.4NLL differed from those observed in our 0u.n concurrent series of 130patients with ANLL de novo (79 of whom [56%] had an abnormal karyotype). we compared the number of anomalies of each chromosome in the two groups of patients by using Fisher's exact test.?' The frequency of alterations of seven chromosomes. namelj-.no. 1.3.5.7. 13, 14, and 18. was found to differ significantly ( P < . O S ) between these groups of patients. In each case, abnormalities of these chromosomes occurred more frequently in patients with therapy-related disease (Fig 3). As expected, the greatest differences were found for chromosomes no. 5 and 7 . Twenty-eight of the 140 patients ( 2 0 9 )with ANLL de novo had abnormalities of nos. 5 and/or 7 , as compared with S5 of 63 (87%)patients with therapy-related disease. Moreover. only 22 of the patients with ANLL de novo ( 16%) had losses of the long ami or of the entire chromosome no. 5 or 7 , whereas these aberrations occurred in all of the 55 patients with therapy-related disease. The frequency of patients with clonal abnormalities involving both nos. 5 and 7 floss or deletion of the long arm) was much hipher in t-MDS!t-ANLL ( I 7 of 63. 77%) than in ANLL de novo (five of 140. 4%)).The frequehcy of abnormalities of chromosomes no. 5 and 7 in MDS arising de novo is even lon,er 338 60- v+) 50- zw 5 40- 0 140 PtS PNLL de nova 63 Pts Therapy-Related I-HDS or 1-ANLL CHROMOSOME Fig 3. Comparison of the frequency of involvement of specific chromosomes in our patients with ANLL de novo and in those with t-MDS or t-ANLL. than that seen in patients with ASLL de novo. and thus it is markedly lower than that observed in patients with t-.LIDS. Of 49 patients with MDS de novo who were evaluated at our institution. 20 ( 41%) had clonal chromosomal abnormalities, and only six of these patients 1 2 4 ) had aberrations involving no. 5 or 7.'! Moreover. two of the six patients had a history of occupational esposure to potential mutagens. The presence of an abnormality of chromosomes no. 1,4. 12. 14, or 18 in t-.\NLL was not invariably independent of abnormalities of nos. 5 and/or 7. Specifically. most patients with therapy-related disease who had a rearrangement of chromosome no. 4 also had an abnormality of nos. 5 and/or 7 ( P = .W). and all 14 patients with an abnormality of no. 13also had resrrange- ments of TW. Sandor 7 ( P = .OI4).In contrast, the crccumnce o f m 3bnOrmahty of chromoE C ~ ~ CnoS . 1. 11. or IS not associated with the presence of aberrations of nos. 5 and!or 7 . The nature of the abnormalities of nos. 1 , 4. 13, 14. and 18 was similar for the patients with therapy-related disease and those with ANLL de novo. with chromosome loss and translocations accounting for most of the aberrations. Delerions of Chromosomes tio. 5 orid 7 utiti the Critical Regioris We have used the kaqotypes showing Ioss of only part of 5q or 7q to define more precisely the .location of genes involved in the pathogenesis of LE BEAU ET AL t-MDSlt-ANLL. Seventeen patients were found to have a del(5q). hhich in every case was interstitial rather than terminal. In interstitial deletions, two breaks occur in the long arm, and the intermediate segment is lost. By comparing the chromosomal breakpoints in these patients, one can identify the smallest region that is consistently deleted, ie. [he critical region. The breakpoints and the segments that were deleted are illustrated in Fig 4. In most of the patients, the proximal breakpoint was in bands q l 1 to q l 3 , and the distal breakpoint was in band q33 or q34; the critical region consisted of bands q23 to q32. A similar analysis for the critical region of 7q was not possible because only four patients in our series had a del(7q). Relationship of Clinical Characteristics to Chromosomal Abnormalities Analjsis of the involvement of specific chromosomes among male and female patients revealed no significant associations with sex, nor 5" [:I -I----- - --l- - --- 5 Fig 4. Diogram of the banding pattern of chromosome no. 5 illustrating the chromosomal breakpoints and deletions in 17 patients, Each vertical bar represents the region that was deleted; the numbers above the lines indicate the number of patients with this deletion. The dashed horizontal lines indicate the critical region, ie, the smallest overlapping region that was deleted in each of the patients. CLINICAUCYTOGENETIC FEATURES OF t-ANLL Table 7. Primary Disease, Primary Therapy, and Chromosome Abnormality Chromosome Abnormality (n = 61) Abnormality of No. 5 Abnormality of No. 7 Abnormality of Nos. 5 and 7 Other Abnormality, Not No. 5 or 7 Normal (n = 2) Primary disease Lymphorno Other hematologic malignant disease Solid tumor Renal transplant Primary therapy Chemotherapy only Radiation only Both modalities 4 2 (1) 8 (7) 0 6 (2) 5 (5) 3 (1) 11 3 a 2 0 2 14 14 (7) 1 2 (2) 0 4 (3) 2 11 (6) 3 0 3 0 2 2 2 1 0 0 1 2 0 0 Total 14 (8) 24 17 ( 9 ) 6 2 NOTE: Number in parentheses designates number of patients with a deletion of no. 5. 339 were there significant associations with patient age. The types of chromosomal abnormalities are shown in Table 7 with the various primary diseases and prima?-therapy categories. There were statistically significant differences in the distribution of abnormalities of no. 5 or 7 or both of these chromosomes based upon the primary disease. Specifically, of the 17 patients with abnormalities of both nos. 5 and 7. 15 iS8qc) had previously had a hematologic malipant disease ( P = .05). Sixteen of the 21 ( 7 6 q ) patients whose initial disease was a solid tumor had an abnormality of either chromosome no. 5 or 7. but not both. whereas 14 of the 33 ( 4 2 4 ) patients whose primary disease was lymphoma had an abnormal clone involving both nos. 5 and 7. When only patients with a del(5q) were considered, this rearrangement was observed more frequently in patients whose initial disease had been a solid tumor than in patients with a hematologic Primary disease. Thus, of ten patients with prior carcinoma and an abnormality of no. 5 . nine had a del(5q)and one had a - 5 . whereas only eight of2 I patients with a prior hematologic malignant disease and an abnormality of no. 5 had a deletion ( P = .07). There were no significant associations between primary-therapy subgroups and chromosomal abnormalities. although aberrations of chromosome no. 7 were seen someuhat more frequently in parients who had received chemo- therapy or both modalities than in those who had received radiotherapy (lower section of Table 7). The type of initial therapy was not associated with the subsequent occurrence of abnormalities of no. 5. nor with the nature of the abnormality (loss I' deletion). The intervals between the start of therapy for the primary disease and the subsequent development of bone marrow dysfunction were similar when patients were grouped according to the type of chromosomal abnormality (median. no. 5 only, 56 months: no. 7 only, 63 months; both no. 5 and no. 7, 48 months: other abnormalities, 4.4months). The clinical aspects of the secondary disorder were also analyzed with respect to the type of chromosomal abnormality observed (Table 8). Aberrations of no. 5, or 7 , or both, were more often associated with the presence of t-MDS. with or without subsequent t-ANLL, than with tANLL alone ( P = .03). Moreover, abnormalities of no. 5 were associated more commonly with t-MDS than with t-ANLL ( P = .02). The eight patients with a normal karyotype or with aberrations of chromosomes other than nos. 5 and 7 were more likely than the rest to have acute leukemia as opposed to t-MDS with or without leukemia ( P = .03);f9ur of the six patients with other abnormalities already had leukemia at initial diagnosis. The two patients with a N 15: I7)(q22:q2I ) had the characteristic clinical and morphologic features of acute promyelocytic I- _- 340 LE BEAU ET AL Table 8. Secondary Bone Marrow Disorder and Chromosome Abnormalities Secondary Disorder Chromosome Abnormalities (n 61) Abnormality of No. 5 Abnormality of No. 7 Abnormality of Nos. 5 and 7 Other Abnormalit) Not No. 5 or 7 t-MDS only t-MDGt-ANLL t-ANLL only 9 5 40 3 12 12 2 2 7 14 Tatnl 14 24 17 6 leukemia de novo and may reflect the development of acute leukemia unrelated to their prior cytotoxic therapies. Of the five patients who had a complete response to remission induction therapy for leukemia. two had an abnormality of no. 7 only, one had a - 7 plus additional complex changes, one had a 8. and the remaining pitient had a 8+ and a t( 15:17)(q21:q.!l). The tM'o patients with normal karyotypes were not treated. %ne of the 18 treated patients with an abnormality of no. 5 had a complete remission. The median survival from the time of onset of bone marrow dysfunction did not differ among the 6 I patients with the various chromosomal abnormalities 1 no. 5 , 6 months; no. 7 . 9 months: nos. 5 and 7.6 months: other chromosomes. 10 months). When patients were grouped according to the complexity of their karyotype (abnormality of nos. 5 andior 7 alone. or with additional rearrangements) and according to the total number of abnormalities. there were no significant differences in the type of primary disease. the type of initial therapy, the clinical course of the srcondary disorder. or the sunival time. Patients in whom both normal and abnormal ( A N ) metaphase cells were present at diagnosis did not have a significantly longer median sun.ival(9months) t h m did patients who had only abnormal IAA) metaphase cells ( 6 months. P = .68).The clinical characteristics and disease course of patients wirh ahnormditirs of chromosomes no. 1. 4. 12. 14. or 18 (the remaining chromosomes involved more frequently in therapy-related leukemia than in ANLL de novo) did not differ from those of patients with other abnormalities or with a normal karyotype. Two unexpected correlations of clinical interest were noted. however. First, 13 of the 16 patients with an abnormality of no. 8 % had received chemotherapy as the primary treatment. either alone (ten patients) or together with radiotherapy ( 3 patients). Of the ten patients with a + 8 , only a single patient (no. 2057) had received radiotherapy alone. Second. the 13 patients who had an abnormality of chromosome 21 had a shorter median survival from the time of initial bone marrow dysfunction ( 5 months) than did the patients without an aberration of this chromosome ( 9 months. P = ,021. DISCUSSION In this study. we have confirmed our earlier observations as well as those of other investigators that a secondary MDS is often the initial disorder seen in patients in whom therapy-related ANLL develops.Lh l').'b.i- Our data confirm that tMDSit-ANLL (ie. either alone or in sequence) is a distinctive therapy-related hematopoietic disorder that is characterized by the presence of recurring clonal abnormalities of chromosomes 5 and 7. This bone marrow disorder is not part of the natural history of any of the primary diseases for which these patients were treated. We have observed t-MDS and/or t-ANLL both in patients receiving treatment either with a single modality (radiotherapy or chemotherapy) and in those receiving combined or sequential treatment with these modalities. %'e want to emphasize that patients who had received radiotherapy alone. particularly to the pelvic area. shared an increased risk for developing leukemia because other investigators have suggested that t-ANLL does not occur in such patients."."Similar to our findings. the data of Greene et al indicated that the volume of irradiated marrow may be related to this late complication." Among our 2 1 patients who had received chemotherapy alone, all but one had undergone pro- CLINICAUCYTOGENETIC FEATURES OF t-ANLL longed treatment with alkylating agents. Our data, therefore, confirm the high leukemogenic potential of long-term alkylating agent therspy,w.1 l .I.( For example. none of our patients developed therapy-related bone marrow dysfunction after receiving only a standard 6-month course of MOPP. Our study included a single patient who had received only an antimetabolite drug. 5-FL'. orally for 5 years. Our patients Lvho received both radiotherapy and chemotherapy were a heterogeneous group. Sixteen received combined-modality therapy concurrently as initial treatment. whereas the other 15 received sequential courses of the two modalities, in several cases on multiple occasions. We \cere unable to show an association between the type of primary therapy and the interval between initial treatment and the onset of bone marrow dysfunction. Thus. the latency period may depend primarily on the intensity of mutapenic damage and on the time required for clonal expansion and not on the precise modality that inflicts the damage. The influence of possible genetic predisposition is unknown. Among our 37 most recent patients. there was a striking increase in the number of patients with solid tumors and a decrease in the number of those with Hodgkin's disease compared with the relative frequency among our 26 previously published cases. There are two possible reasons for this observation. The first is the growing recognition of the late complications of intensive cytotoxic therapy and. thus, the more judicious use of such treatment for patients with Hodgkin's disease. The second is the success of recent therapeutic trials with solid tumors. which has resulted in longer survival of these patients. Our data do not allow us to draw any conclusions regarding optimal treatment for patients with therapy-related bone marrow dysfunction. No consistent treatment approach was taken with the patients in our series because the treatments were given at several institutions over a 1?-year period. Recently. both intensive chemotherapy and bone marrow transplantation have been shown to be useful in such Among ouf five patients who attained a complete remission. only one had had a myelodysplastic preleukemic phase. and none of the five had an abnormality of chromosome no. 5 . Some authors have suggested that patients receiving a 341 single primary treatment modality have a lower incidence of abnormalities of nos. 5 andlor 7 as well as a better response of their leukemia to subsequent therapy for remission induction.2o In general. leukemia patients with abnormalities of nos. 5 and/or 7 do very poorly.'".'' In our series, 8 7 9 of all patients had abnormalities of nos. 5 and.'or 7 . including a number who had initially received only single-modality treatment. The nonrandom involvement of chromosomes no. 5 and 7 in t-ANLL was first recognized by Ruwley et al ( 1977)and was confirmed in severaI subsequent investigations ,4.5,'k'2.27 In the present study. 55 of 61 (90%)patients with karyotypic aberrations had an abnormality of one or bod?of these chromosomes. Moreover, we have idenrified several additional chromosomes that may be preferentially involved in aberrations in this therapy-related disorder. namely. nos. 1. 4. 12, 14, and 18. The frequency of t-ANLL patients with abnormal karyotypes and with aberrations of nos. 5 and 7 has varied among different studies, some of which were based on nonbanded chromosome preparations. Pedersen-Bjergaard et al reported on an updated series of 55 patients?'; of 39 patients studied with banding techniques, 31 (79%) had clonal abnormalities, and 27 ( 6 9 9 ) had abnormalities of chromosomes no. 5 and/or 7. The frequency of abnormalities of chromosomes no. 5 and:or 7 was somewhat lower ( 3 5 % )among 20 patients reported by Sandberg et ai.:' They noted the frequent involvement of two other chromosomes. namely, nos. 3 (four cases) and 17 (five cases). and they suggested that these two chromosomes are also preferentially involved in chromosomal abnormalities in t-ANLL. In addition, these authors have identified 206 previously reported cases of t-ANLL in which cytogenetic analyses were described.'' Chromosome no. 7 was abnormal in 60 cases (29%).no. 5 in 38 (19%).no. 17 in 34 (17%), no. 3 in 24 (12%). no. 21 in 21 ( I O % ) , and no. 8 in 19 cases (9%). Our data have not confirmed their suggestion that nos. 3 and 17 are preferentially involved in tANLL. In our series, ten patients had abnormalities of no. 3, and I5 had clonal rearrangements of no. 17. When we compared the frequency of involvement of each chromosome in patients with t-ANLL with that noted in our concurrent series of patients with XNLL de novo.??we de- 342 [ermined that the frequency of involvement of no. 3 ( P = .I4)or of no. 17 ( P = .88)did not differ between these two groups of patients. Recently. Arthur and Bloomfield suggested that abnormalities of nos. 5 and 7 may be related to a particular form of therapy administered for the primary disease.'O These authors proposed that abnormalities of nos. 5 andior 7 may be particularly frequent in patients who initially received combination Chemotherapy. with or without radiotherapy. and that patients who received less intensive therapy were more likely to have abnormalities characteristic of ANLL de novo. In their series of 70 patients with t-ASLL. nine had been treated less heavily: overall. ten of 19 ( 5 3 9 ) patients with abnormal kaTotypes and only two of the nine with less therapy had abnormalities of no. 5 or no. 7. In contrast. the eight patients in our series who had normal kapotypes or other abnormalities had not received less intensive primary therapy than had the 55 patients with abnormalities of nos. 5 and,or 7 . Moreover. nos. 5 andlor 7 were abnormal in 18 of the 21 (86%) patients nho had initially been treated with chemotherapy alone and in nine of the 11 (82%) patients who had initially been treared with radiotherap! alone. Our present findings as uell as our observation that abnormalities of nos. 5 and 7 occur frequently in patients with XNLL de novo who had significant occupational exposure have led us to believe that aberrations inidsing these chromosomes niay be indicators of mutagen-induced malignant disease.'5..:sThis association. first noted by Xiitelman et al and confirmed by Golomb et a1 and later bj. the Fourth U'orkshop. was based on a correlation betueen the occupation of adult patients with ANLL de novo and the karyo- '-type of the leukemic cells.'" The early find- ings on this association have been summarized by Golomb et Of 236 patients. 65 (19%) were classitied 3s having a history of significant exposure to chemicals. petroleum solvents. pesticides. or industrial metals. The frequency of abnormal karyotypes was higher in the exposed ( 5 1 of 68. 75% 1 than in the nonexposed patients (60of 168. 36%).\lore importantly. abnormalities of chromosomes no. 5 and.or 7 were obsemed in 37% ofthe exposed patients and in only 12% of the nonevposed patients.:' :h Recent evidence suggests that proto-onco- \ LE BEAU ET AL genes. the cellular homologues of the transforming sequences of retroviruses. may play a role in human tumorigenesis. Several proto-oncogenes have been mapped to chromosoqe regions that are the breakpoints in the specific chromosomal rearrangements in hematologic malignant diseases. In at least one human tumor. Burkitt's lymphoma. transcriptional deregulation of a proto-oncogene ( c - m , ~o)ccurs as the result of a chromosomal translocation, t( 8;14)(qlJ:q32').In a second case. c-ubl is consistently translocated from the distal long arm of chromosome no. 9 to the long arm of chromosome no. 22. resulting in the Philadelphia chromosome that is characteristic of chronic myelogenous leukemia. As a consequence of this rearrangement. transcription of the fused c-nbl and bcr (on no. 2 2 ) genes results in a larger rnRNA3' and therefore a larger protein.%foreover. this altered protein has tyrosine kinase activity.'" Since abnormalities of nos. 5 and 7 are characteristic of t-MDSit-ANLL. it is appropriate to consider genes that are present on these chromosomes and their potential role in leukemogenesis. In this regard. analysis of the deleted segments of chromosome no. 5 has enabled us to identify the critical reZion (bands 5q23 to q32) that is likely to contain the genes whose altered function may result in neoplasia. The results of a similar analysis of the critical region on 7q have been more uncertain because only four patients in this series had a del(7ql and it has been difficult to determine the precise breakpoints. However. if we consider the deletions observed in 3 additional patients with t-MDS or t-.4NLL ascertained subsequent to the preparation of this manuscript. the critical region of no. 7 appears to be either in bands 7q32 to q33 or in bands 7q34 to q35. Two proto-oncogenes (one of which codes for the receptor of a hematopoietic growth factor), one transforming sequence. and one _geneencoding a growth factor have been localized to chromosome no. 5 or 7 . C j h s (McDonough feline sarcoma virus) has been mapped to 5q34. a band that is just below the distal break point on the deleted chromosome"; the protein encoded by this gene has recently been shown to have substantial homology with macrcphage colonystimulating factor (CSF-I ).'2 The proto-oncogene c-rrb B (avian erythroblastosis virus) and CLINICAUCYTOGENETIC FEATURES OF t-ANLL 343 the transforming sequence met are located on no. 7. The former gene has recently been shown to code for a protein that is a truncated form of the receptor for epidermal growth factor.'? C-erb B has been localized to 7p or proximal 7q" and. therefore. appears to be at some distance from the deletion of no. 7 in t-MDS.t-ANLL. Met has k e n localized to 7q22 tM.M. Le Beau. unpublished observations) and. thus. is closer than cerb B to the critical region on no. 7 . The growth factor gene. GM-CSF. has recently been mapped to the critical region on no. 5.u Other noteworthy genes on no. 5 are the glucocorticoid receptor (5q15 to qter) and the &adrenergic receptor ( 5 centromere to q21), both of which are known to influence cell division in specific cell types." How are abnormalities of nos. 5 and'or 7 related to a potential role for oncogenes in leukemogenesis'?At present. it is unknown whether c-erb B or rnrt are involved in the pathogenesis of tMDS!t-ANLL; however. data obtained recently suggest that c-fms (CSF-I receptor) and GMCSF may be relevant genes in this disease. There are several mechanisms by which chromosomal abnormalities might influence the function of growth factors or their receptors and. thus. lead to malignant transformation. Loss of a whole chromosome or part of a chromosome (deletion) rnay result in the loss of certain genes that regulate cell growth. Alternatively. loss of genetic material (which occurs both in cells with deletions or with loss of a whole chromosome) may allow expression of a recessive mutant gene on the homologous chromosome, as is currently proposed for retinoblastoma and Wilms' tumor.45 As mentioned above. c-fms was previously mapped to band q34 of chromosome 5 : recent data, however, suggest that this gene rnay be more proximal. That is, by fusing human bone marrow cells from two patients with a del(Sq) with rodent cells. Nienhuis et al found that clfms was within the deleted segment of chromosome no. 5%: the distal breakpoint on chromosome 5 was thought to be proximal to band q34. By using in situ chromosomal kjbrjdization. Le Besu et a1 ha1.e hcafi7ed the c-fins gene to band qZ?: G,M-CSF has been localized to 5q33 to Sq? I .zBoth of these genes are deleted in several patients with a del(Sq)." As nored earlier. the cfins gene encodes the receptor for CSF- I . Thus, the products of both the CSF-I and GM-CSF genes play an integral role in the proliferation and differentiation of cells of the mononuclear phagocytic lineage."The variability in the break points in the interstitial deletions of chromosome 5 observed in patients with t-ANLL sugest5 that the juxtaposition of identical gene sequencer. an event that niay occur as a result of rearrangements such as translocations. is not a consistent genetic consequence of these chromosomal abnormalities. Therefore, a more likely mechanism is that previously outlined. namely, the loss of critical gene(s) resulting from whole chromosome loss or a deletion and the expression of a mutant allele on the remaining chromosome. A s phvsicians. we have the primary responsibility to reduce the incidence of this essentially fatal complication; as scientists, we should use the information provided by studies on these patient:, to advance our understanding of the critical ,oenes and of the mechanisms involved in mutagen-induced leukemia. The insights obtained in this system will be of great value in elucidating the carcinogenic process in general. ACKNOWLEDGMENT The author, u ish to thank the man} physicians in the Chicago area u ho referred patients and clinical specimens to the L'niverjlt! of Chicago for this study. We thank Margaret Johnson and Xlar.iorie lsaacson for data management. Elisabeth Lanzl for ediiorial suggestions. and Deborah Gifford for secretarial assis- tance REFERENCES 1 . Rlisner F. Grunwald H: Hodgkin's diseahe and acute Ieukemia Report of right case<and revieu of the literature. .Am J Med J8:.;39-353. 1975 2 . Cadnun EC. Capizzi RL. Bertino JR. .Acure nonl!mpho- cytic leuhrmia A delayed complication of Hodskin', disedse therapy- .Analysis of 109 caseb. Cancer 10:1280-1296. 19'- 3 . Coleman CN. william^ CJ. Flint A . et al. Hemurologic neoplasia in pirients treated for Hodgkin'q dizraze U Engl I Med 297-I14Y-l252. 1977 J Roule! ID. Golonth HM. Vardiman JW: Nonrandom c h r ~ ~ m o w mablnormalities in acute nonlymphocytic leukemia in parisnth trcatrd for Hodgkin's disease and non-Hodgkln's I\rnphoma. B h d 50:759-770. 1977 I RouIe! JD. Golomh HM. Vardiman JW. Nonrandom ihrmn~wlmeahnormditiea in acute leukemia and dysmyelopoletic s!ndromcs in patient, H ith previousl! treated malignant d!\tdZs\ Blood 58.759-767. 198I 0 \ d i m a n J W . Golomb HM. Rowley JD. et al. Acute 344 LE BEAU ET AL nonl) mphocytic leukemia in malignant I! mphoma: .A morphologic study Cancer 12:129-212. I978 7. Rosner F. Grunwald H\V. C>totoxic d ~ andg leu~kemo- genesis. Clin Haematol 9:663-681. 1980 X. Bergsagel DE. Bailey 4J. Lmgle! GR. et 31: The chemo- therap) of plasma cell iii)eloma and the incidence of acute leukemia. N Engl J Sled 301:?13-'48. 1979 9. Zarrabi 41H. Rosner F. Bennett JSl: Non-Hod,okin's lym- phoma and Jciite ni~eloblasticleukemia. Cancer 44: 1070-1080. I979 IO. Fouc3r K . McKenna RW. Bloomiield CD. et 31: Ther- apy-related leukemia: .A panmyelobis. Cancer 13:1285-1296. I979 1 I . Berk PD. Goldberg ID. Silberstein MN. et 31: Increased incidence of acute leuhernia in polyc! themia vera associated with chlorambucil therapy N Engl J Med 304:1JI-447. 1981 I?. Zarrabi MH. Rosner F. Acute m!eloblastic leukemia fol- lowing treatment for non-hematopoietic cancers: Report of 19 cases and review of the literature. Am J Hcniatol 7 3 7 - 3 6 7 . 1979 I?. Kyle RA: Second malignancies associated uith chemotherapy agents Semi, Oncol 9:I 3 I- 112. 1982 14. Greene 51H. Young RC. Sicrriil 1.11. et al: Etidence of a treatment dose respnnbe i n acute mml) mphocytic leukemias H hich oxcur after therap) ut' nun- Hodgkin'j I! cnphorna. Cancer Re, 13:1891-1898. I983 IS. Albain KS. Le Beau 5151. Vardiman JW. et al: D e ~ e l o p - ment of d)smyelopoietic syndrome in a hairy cell leukemia patient treated with chlorambucil: C>to:enetic and morphologic evaluation. Cancer Genet Cytogenet S:107-1I S . 1983 16. Pedersen-Bjergiard 1. Ldrsen SO: Incidence of acute nonlymphoqtic leukemia. preleukemia. and acute myeloprolif- erative syndrume up to 10 years after treatment of Hadgkin's Jiaease. N Engl 1 \led 307.965-971. 1982 17. Anderson RL. Bagby CG. Richen-Boe K. et 11: Therapy- related preleukemic >!ndroinc Cancer 47.11167- I87 I . I981 18. Blcwmfield CD. Preisler H. Cuttner I . et ai: Treatinent- induced acute non-l? mphcx)tic leukemia. Rcbponse to cytara- hine-anthrac!cline t h e r a p . Blcrod 60: 152. I982 ( w p p l I 1 cabsir) 19. Pcderwn-Bjerprd I . Philip P. Slortensen BT. et 31: Acute nonl!mphcqtic leukemia. preleukemia. and Acute mye- Iopn.itift.r~ii\c,>ncirr*me,czunJ.in tu treatment oiother malig- nant Jiwases: Clinical m d c y t o g e m i c characteristics and re- d f s d HI b i r r o culture of bone marrow 3rd HL.4 [!ping. Blood j7::1:-7:-:. 1~))8[ 20. hnhur DC. Bloomiield CD: Banded chromosome analy- sis in patients H ith ireatmrnt-A,soci~tedacute ncin-l>niphocytic leukemia. Cancer Genet C) togenet I2 : 189- 199. I981 !I. Sdndberg .\,A. .%be S. Kowrllczyk JR. et SI: Chronio- homes and causation of human c m z e r and leukemia. L. C)toge- netics o!' Ieukr'miah complicating (1thr.r di\ea\es Cmcer Genet Cytogenet -.95- 1.76. 148.: 2 2 . Pcdcr,en-Bjergaard J. Philip P. Pcdcrscn NT. et al: Acute nod!. mphoc! tic kukernia. preleuhemia. and acute ni)- eloprolifcratibe >> ndrorne v x o n d a n to tremnent of orher ma- lignant di.rea2r.c Cancer 51 152-1h3. 1984 23 Larson RA. Le Beau Mhl. Vardiman JiS. et al: The predictive value of initial cyroprtnetic m d i e s in I18 adults with x u t e nonl)inphc+tic leukemia. Cancer Genet Cktogenet 10.219-236. 1983 \ 24. Mitelman F. Nilsson PG. Brandt L. et al: Chromosome pattern. occupation. and clinical features in patients with acute nonlynphocytic leukemia. Cancer Genet Cytogenet 1:197- 214. 1981 25 Golomb HM. Alimena G. Rowley JD. et al: Correlation of occupation and karyotype in adults with acute ni)nlymphwytic leukemia. Blood 60:404-11 I . I982 16. Le Beau MM. Larson R.4. Bitter M A . et al: .Association of an inversiun of chromowme 16 with abnomial ntartou eosinophils in acute myelomonocytic leukemia: .\ unique c::to$net- ic-clinicopaIholo_eicalassociation. N Engl J MeJ 309:b30-636. 1983 27, Fourth International Workshop on Chromosomes in Leu- kemia ( 1987).Cancer Genet Cytogenet I1:249-360. 1984 28. Bennett JM. Catovsky D. Daniel MT. et al: The French- American-British Co-operarive Group. Proposal5 for the classification of the myelodysplastic syndromes. Br J Haematol 5 1:189- 199. I 982 29. Vardiman JW. Coelho A. Golumb HM.et dl: Slorpho- logic and cytochemical obsenations on the ocen leukemlc phase of therapy-relared leukemia. Am J Clin Pathol 79525530. 1983 30. McKcnna RW. Parker JL. Brunning RD: Cltrastructural observations in therapy-related paninyelohis. in Pcilliack .4fed,. Human Leukemia Buston. Martinur Nijhoif. 1983. pp 123-15: 3 I. ISCN ( 1978): ,An international system for human cytogenetic nomenclature Cytogenet Cell Genet 21 :309-401. 1 9 3 32. First International Workshop on Chromowmes in Leukemia: Chromosomes in acute nonlymphocytic leukemia. Br I Haematol 3931 1-116. 1978 33. Pagano M. Halvorsen KT: An algorithm for finding the exact significance lecels of r x c cunringcncy tables. J Am Statist ;\ssoc 76.931-934. 198 I 31. Larson R.4. K-ondoK . Vardiman JW. et ai: Evidence for a 15:17 tran+xution in e b e r y patient with acute promyeloc>tic leukemia. Am J \led 76.827-840. 1981 35. Jacob5 RH. Curnbleet $1.4. Vardiman JW, et al: h o p nostic implications of morphology and k a n o t y w in prim* myelodysplaatic y d r o m e s . Blood ( i n press) 36. Prei\ler HD. Early ,4P. Ram A. et al: Therapy of w o n d a n acuie nonl)mphocytic leukemia with cytarabine. N En$ J Sied 308:11-24. 19x3 37 Vaughan WP. Karp JE. Burke PJ. Effective chemorherapy of acute m?elocytic leukemia occurring after alk! laring agenr or radiation therapy ior prior malignancy. J Clin Oncol 1:203-207. 1983 38. Rowle) JD: Chromosome change5 in leukemic cells as indicators riimutagt.nic e y p o w e . in Rouley I D and Lltniann JE ~ e d sC~h:romoronicr and Cancer. From Sloleculcs to Sian. Orlando. Fla. Academic. 1983. pp I40-lh! 39. Shrivelman E. Litshitz 9. Gale RP. et 31: Fused tranzcript o l cihl and b u gent\ in chronic niyelogen<u leukemia. Nature 3 15.550-551. 19x5 40. Konopka JB. Watanahe Ski. Witte ON: ,An alteration of the human c-uhl protcin in K562 leukemia cells unmask5 a s x ~ i ated p r o ~ i n ekina\e clctlcit) Cell 37: 1035-1042. 1984 1I . Human Gene Slapping 7 . Seventh Intrrnatwnal N'orkshop on Human Gene Mapping. Cytagenet Cell Genet 37:l-666. 1984 42. Shcrr CJ. Rettcnmier CW. Sdcca R. e! 31: The i : h s CLINICAUCYTOGENETIC FEATURES OF t-ANLL proto-oncogene product is related to the receptor for the mononuclear phagocyte growth factor, CSF-I. Cell 11:665-676. 19x5 43. Downward J. Yarden Y. Mayes E. et al: Close similarity of epidermal growth factor receptor and v-erb-Eoncogene prorein sequences. Nature 307:511-577. I981 44. Le Beau Mhl. Westbrook C.4. Diaz MO. ?t al: Evidence 345 for the in\ol\ement of GM-CSF andfins in the deletion (5q) in npeloid disorders. Science (in press) 15. Sturphree AL. Benedicf W'F: Retinoblastoma: Clues to human onco_eenesis.Science 223:1028-1033. 1984 16. Nienhuis AW. Bunn HF. Turner PH. et al: Expression of the human c-fms proto-oncogene in hematopoietic cells and its deletion in the 5q- syndrome. Cell 12:121-428. 1985 T