Document wD3opK9pJjQ9LRJv3xwz3VQz4

Scand J Haematol 1986; 36,Suppl45: 82-90 Key words cnromosomes . secondary acute nonlymphocytic leukaemia . secondary myelodysplastic syndromes Chromosome abnormalities in secondary myelodysplastic syndromes Clara Derber Bloomfield Section of Medical Oncology, Department of Medicine, University of Minnesota Health Sciences Center. Minneapolis. Minnesota, U S A and the Department of Medical Genetics, University of Helsinki. Helsinki, Finland Specific chromosome abnormalities occur in most cases of secondary myelodysplastic Syndromes (MDSI, broadly defined to include acute nonlymphocytic leukaemia (ANLL). To explore their relation with the type of prior cancer and its treatment, and their use in predicting response of the secondary MDS to therapy, 216 patients from the literature were analyzed. Abnormal karyotypes were grouped into 10 nonoverlapping categories, depending first upon the presence of abnormalities of chromosomes 5 and 7 ( - 5 . 5q-, -7, 7q-), then upon specific rearrangements seen in de novo ANLL. and finally upon other structural or numerical abnormalities. Karyotype correlated with type of prior cancer (p = .002). -5, -7 and additional whole chromosomes were associated primarily ( 6 - 4 6 TO) with prior haematologic malignancies, and the specific rearrangements and 5q- groups (70-71 ma) with prior solid tumors. Karyotype also correlated with prior rherapy (p = .0009). - 5 , -7 and 7q- were highly ( > 85 "0) associated with chemotherapy (alone or with radiotherapy); the 5q-, specific rearrangements, normal. and "other" structural abnormalities groups were often (29-60 mo) associated with radiotherapy alone. Finally, karyotype also predicted achievement o f complete remission ( C R ) ; Sq- a n d -5 were associated with the lowest CR rate (13 To), specific rearrangements and other structural abnormalities with the highest (65 ma). Prospective studies are required to determine if karyotype is an independent risk factor. .4s more patients with cancer have been successfully treated with radiation and/or chemotherapy, there have been increasing numbers of reports of secondary myelodysplastic syndromes (MDS), including acute nonlymphocytic leukaemia (ANLL), occurring a m o n g such patients (1-2). To date, relatively fen of these patients have had cytogenetic studies donelwhen they developed the secondary MDS. Only five sizeable series of banded chromosome studies have been published (3-10). I n these studies it has been suggested that kar!otype may be related t o type of prior malignancy, and 'or type of prior treatment. It has also been suggested that karyotkpe may be the major predictor of response t o treatment of the secondary MDS. Because of the small numbers of patients studied in each series thi. conclusions ha\ e been tenuous. This review colisiders the relation of karyotype to the type of prior disease. the nature of prior treatment, and th: response t o therapy of the secondary MDS in 216 patients from the above 5 published series. Materiak and methods Fo: pU;poSeS of this re\ le\+,the 5 studies, compribing 2.'; parientc \\ith iecondar! \IDS following rherap! for I. primary disease, were carefully exar;' cases reasonably successfully anal) z: niques were included. Cases which Fourth international Workshop on < kemia (31hlWCL) from the Unibr: University of .Minnesota, the Finst.:were listed with their respe<ti\e S . (Table I). TABLE I Porienr charocrerisr,cs in .five studies II ~~~~ ~ Chi<.. References Number of cases 36! To male 4Y Age at secondary 51DS o'rs) Median Range .<'. 6-- Mos from initial treatment'diagnnsis Median 56 Range IO- First disease (cancer) Hodgkin's disease Malignant lymphoma Multiple myeloma Other haematologic ca. Breast Ovary Uterusicervix Lung Colorectal Other solid tumor Benign Prior therapy ("70) Radiation Minimal (R) Major (RR) Chemotherapy Single (C) Combination (CC) Combined modality R+C R + CC RR RR + - C CC 316 8 6 1 6 3 8 6 16 I1 21 5 .7.' Treatment for secondar!. MDS ( o b ) Intensive Minimal 3: - _7 - None L: Response io intensire treatment no Complete remission Durarion (mos) Median Range -41 3 b--: Survival (mor) .Median Range (1-4 Onl! secondar) 4 ~ i . L . * * Origi1.i ?j 1986; 36.Suppl45: 82-90 78s . secondary acute rnia . secondary mes lary f f ield hcology, Department of 3f Minnesota Health Sciences Minnesota, U S A . and the :al Genetics, University of inland astic *LL). r use ature mes, * 5q-, other ancer 6 ob) oups xapy done .tural ially. u ere :Up nt :major predictor of response condary MDS.Because of the ients studied in each series the :n tenuous. This review conkaryotype to the type of prior of prior treatment, and the of the secondary MDS in 216 )eve 5 published series. __ methods IC\\, the 5 ttudies. comprising 232 ) \1DS follouing therapy for a CHRO.MOSO.LlES IN SECONDARY LEUK.4E.LII.A 83 Primary ditease. were carefully examined. Only those 216 ;aces reasonablv successfully analyzed with banding tech- niques here included. Cases which uere included in the Founh International Workshop on Chromosomes in Leubsmia (3'"IWCL) from the University of Chicago, the Cni\ersity of Minnesota. the Finsen Institute and France \sere listed with their respective submitting institution (Table I). Patient characteristics which were available for most of the cases are summarized in Table 1. The series differed in the type of patients included. The WlWCL and Minnesota series included only cases meeting the criteria for a diagnosis of ANLL (11). In each series there was a slight predominance of females, and the median age ranged from 55 to 63. The median time from the primary thdisease (or its treatment) to the diagnosis of e secondary IABLE I poirent .haracrerrsrics in five studies of 5anded chromosome anolvses in secondorv UDS Chicano French Finsen 4thlWCL. Minnesota' L _ _ ~~~~~~~~~~ ~ ~~~ ~~ ~~~~~~~~ ~ Reicrences \umber of cases 3-5 6 7.8 9 IO 63 55 42 36 20 'TUmale 19 49 45 44 u) ~ g aei secondary 51DS (yrs) \ledian Range 55 6-76 60 13-89 63 22-75 55 18-3 59 22-76 \los from initial treatment/diagnosis to secondary MDS Sledian 56 72'' Range 10-192 24-288 48 12-120 58 7-204 76 35-182 F i x disease (cancer)(ob) Hodgkin's disease 37 -7 .I 36 22 20 \lalignanr lymphoma 16 II 21 II IO Slultiple myeloma 8 13 5 I1 5 Other haematologic ia. 7 4 2 3 IO Breast 6 16 10 22 15 O\ ar) 3 < I4 6 15 Uterus ?ervi.\ 6 4 5 05 Lung Colorecral 3 7 53c 8 4 000 Orher solid iumor 6 15 7 I I 15 Benign 5 5 0 II 0 Prior therapy ("'0) Radiation llinimal I R ) 4lajor t R R ) Chemotherapy Single (C) 3 I6 I1 Combination (CC) Combined modality R- C R - CC RR - C R R - CC 'I 3 8 c 35 T:eaiment ior secondary \IDS ( o b ) Inrensi\e \linimal Vone 44 Rcjponse IO intensive treatment vo Complere remission Duration cmo,) 30 \ledtan Range -20 6-43 S u r ~ i i a(l m o r ) 18." u n hnou n IO 5 I4 '9 0 5 7 36 45 12 43 41 5 l*-34+ 19 8 17 14 0 0 6 36 53 13 28 26 1.5 6-l9+ IO IO 25 5 IS < 0 30 75 IO IO 53 3 2-11 \ledian Range -7 0-43 3 1-16 9 0-39 31 0-2 I 5 0-39 * Only =:ondar? .4NLL. O*I riginal data pro\ided in years, * * * Amount of prior therapy not specified. Total 216 47 No. ?o 62 29 31 14 19 9 73 28 13 16 7 94 73 73 20 9 IO 5 I O 59.. 15 7 17 8 20 9.24 11 31 I4 2s 12-9 42 84 63 56 26 55 25". 73 34 28 13 58 27 38 - - - 84 BLOOMFIELD disorder ranged from -18 to 76 months: these differences, in part, probably represent differences in the date at which the diagnosis of the secondar! \IDS !\as made. The series differed in two important respeits. The first diseace was primarily (63 ob-63 " 0 ) haematologic in the Chicago and Finsen studies. The first disease \vas most often a solid tumor or a nonneoplastic disorder (herein abbreviated benign) in the 3`>IWCL and Irlinnesota series. Prior therap! tended to be more intensive in those series including primarily haematologic malignancies (Table I ) . In an attempt to correlare chromosome findings of the secondary MDS ~ i t thheir response to therap!, onl! those cases treated with currently acceptable induction chemotherapy for ANLL 1i.e. combination chemotherapy uhich included cytarabine and an anthracysline or high dose cytarabine) Here Considered. The frequency of such therapy differed among the series. and onl! a total of 7 3 cases iould be used for this analvsis. Results Karyotype classification An abnormal karyotype was reported in 185 cases (86 40).Abnormal karyotypes uere grouped into I O nonoverlapping categories. Sixty-six percent (142) of the cases had one of the abnormalities of chromosomes 5 or 7 (i.e. -5, 5q-, -7 , 7q-) frequently reported in secondary MDS. Each of these cases was assigned to one of 7 groups: -5, 5q-, -7, 7q-, -5 and -7, 5q- and -7, other combinations of 5 and/or 7 abnormalities (misc. 5 8:7 )(Table 2 ) . Patients wirh abnormal karyotypes not including one o f these abnormalities were classified into one of 3 groups: a "specific" group (i.e. those containing o n e of the reciArring specific rearran,w n e n t s associated with de n o w ANLL), an additions, "adds". group (i.e. a karyotype with only additions of whole chromosomes), and an "other" strucrural abnormalities group. The largest category uab the rnonozomy 7 (-7) group which comprised 24 To of all patients. Interestingly, among the 52 cases in this group. 21 had 1 single abnormal clone nith -7 as the only abnormality. The smallest karyotype group (misc. 5 & 7) consisted of 9 cases. The "specific" rearrangements group was comprised of 5 cases with t(15;17), 2 uith t(6;ll), and one each uith t(9;Il). t(11;19) and t(8:21). Relation between karyotype, prior disease, and prior therapy Each kar!otype group included patients with a wide spectrum of prior diseases, but the relatite freqaency of the different diseases varied (Table 2). A though the incidence of specific haematologic malignancies (i.e. Hodgkin's disease ys. malignant lymphoma \ s. multiple myeioma vs. other) or solid tumors (breast vs. oLary. vs. other) did not var! signiiicantl! cp > .09)among chromosonie groups. whether the prior disease \!as a haematologic malignancy, a solid rumor or nonneoplastic did (p = ,002). In particular, the karyotype groups including a - 5 or c? -7 tended to be largely composed of patients \\ i!h a prior haematologic malignanq (67 O-ob--S6 `yo). as did the karyotype group \qith only additions of u hole chromosomes (75 co). The specific rearrangements and 5q- karyotype group` were primaril! (70-71 Eo) comprised of patien!< who initial!! had solid tumors. The kar!-onpe groups also baried significantl! accordinp io :he type of prior therapy the patient< comprising :hem had received ( p = .0009). 0 ~ t . r 85 "0 of the karyotypes including a - 5 , -7 , or 7 q \ w e seen in patients who had received prior chemotherapy alone or in combination with radiotherapy. In contrast, 60 Eo of the patients with 3 specific rearrangement had received only radiotherapy, as had nearly one-third of patients with normal karyotypes, a 5q- or other structural abnormalities. In this analysis, significant difference. among karyotb-pe groups according to the exrenr of radiation (minimal vs. major) or chemotherap! (single agen: \s. combined) the patient had recei\ed could nor be sho\Vn. However, detailed description. of prior trea:rnenr \\ere often not available and thu. an accurate assignment to extent of radiotherap! or chemo!h:iaF!, may not have been possible. The -? group, hoire\er. tended to be associated with combination as opposed to single agent chemotherap! (among patients recei\ing chemotherap! . E6 "0 \ s . 1-1 ro respectikely). Karyotype groups did differ ( p = ,031 in their incidence of patients nhL1 ha3 re:ei\ed limited as compared to more euen,i\ i' single modali:! therapy (Table 2 ) . The -5 and l h e -- hhr!Ot!.pe groups \+ereprimaril) (83-8s rul a`- karyotype, prior disease, ii. 'ed patients with a wide asc - u t the relative frequenseases varied (Table 2). AI>f specific haematologic makin's disease vs. malignant le myeloma vs. other) or soavary, vs. other) did not vary I among chromosome groups, .me \\as a haematologic mamor or nonneoplastic did lar, the karyotype groups in.ided to be largely composed )r haematologic malignancy J the karyotype group with !e chromosomes (75 To). The its and 5q- karyotype groups .I 4 )comprised of patients .d tumors. ., ? : .' ' ups also varied significantly of prior therapy the patients received ( p = .0009). Over ws including a -5, -7, o r -qts who had received prior )r in combination with radio60 "0 of the patients with a it had received only radioi hird of patients u.ith 5q- or other structural abilysis, significant differences ups according to the extent ~ 3 m. ajor) or chemotherapy nzd) the patient had received .NeLer. detailed descripticns often not available and thils i t to extent of radiotherapy not ha\ e been possible. The nded to be associated with red to single agent chemoits receiving chemotherapy, bel!). Karyotype groups did ir incidence of patients who compared IOmore extensive y (Table 2). The -5 and the :re primarily (53-85 0%) as- C H R O X I O S O \ l E S IN S E C O N D A R Y LEL:K.AE\II.-\ 85 N J - s =N-=N I 1F, I 86 BLOOMFIELD sociated with extensive therapy while patients with normal karyotypes or specific rearrangements had often (63-67 070) received more limited single modality therapy. In a n attempt to evaluate the interaction of type of prior disease and therapy on the chromosome abnormality, karyotype was correlated with type of prior therapy separately for patients with prior haematologic malignancies and solid tumors (Table 3). In both instances, type of therapy tended to associate with karyotype group ( p = .06). Among patients with haematologic malignancies, chemotherapy alone had been given to most (64-83 To) patients in the -5 and 5q- karyotype groups, while combined modality therapy had been received by most patients with a -7, a 7q-, other structural abnormalities or a normal karyotype (54-93 %). In contrast among patients with prior solid tumors, only in the -7 group had a majority of patients (57 To) received combined modality therapy; radiation alone was received by almost half of the patients in the normal karyotype, the 5q-, and the "other" structural abnormality groups (42-47 070). a n d in almost all patients (86 Po) in the "specific" abnormality group. Chemotherapy alone had been received by most patients in the 7q- group. Similarly, karyotype group was correlated with the type of prior disease separately for patients treated with radiation, chemotherapy or combined modality approaches (Table 3). Among patients treated with only radiation, type of initial disease significantly correlated with karyotype group (p = .03). A prior haematologic cancer was frequently found only in the -7, -7 and -5, and the additions o f whole chromosome groups. In all other groups most patients had had solid tumors. Among patients treated only with chemotherapy, type of initial disease did not correlate with karyotype group. Among patients treated with combined modality approaches, type of first disease correlated only weakly (p = .06) with karyotype: most patients had previously had haematologic neoplasms; the major ekeption was the 5q- group in which 6 of 7 patients had previously had a solid tumor. Although the explanation for these results is unknown o n e hypothesis is that the type of radiation administered differs more between solid tumors and haematologic malignancies than does the type of chemotherap?. Certainly, for many solid tumor5 radiation is more localized than it usually is for Hodgkin's disease or malignant lymphoma - the 2 major prior haematologic malignancies in thecs series. Relation between karyotype, patient characteristics, response to treatment and survival .4lthouph the se?;distribution (medians 75-58 ro male). age at diagnosis of the secondary MDS (medians 53-66 years) and time between initial and secondae diseases (medians 18-70 months) t aried s o m w h a t among chromosome groups, overall the differences icere not significant (Table I ) .When individual groups were compared, time to the secondary >IDS was significantly longer in the group with additional *hole chromosomes as compared to the group with a normal karyotype (70 vs. 48, p = .04). Only 73 of the 216 patients received intensike treatment for their secondary MDS, limiting detailed analysis of :he effect of karyotype on response. Houever, kar!otype appeared to correlate with achieLsment of complete remission @ = .04, Tabls 5). The response rate was lowest in patients with abnormalities of chromosome 5 (5q- or -5) regardless of whether abnormalities in Chromosome 7 were also present; only 13 To responded. Similarly. a lotv response rate in patients with denovo ANLL with abnormalities of chromosome 5 was seen in the 4"IIwCL (9). The best response rate (65 vo) mas seen in patients with specific rearrangement. or other structural chromosome abnormalities. Numbers of cases were too small to statisticall! compare durarion of remission (Table 4). .Among the 7 3 patients receiving intensive treatment for their secondary XfDS, the rate of complete remission \\as not correlated with whether the first disease was haematologic (,40 ro), a solid tumor (39 Eo) or benign ( 2 5 Fo).H o u e v e r , response tiif` highl) correlated with type of prior treatmen: (p = ,005). .Among patients treated only u I r h radiation 67 achieved Lomplete remission, KXT~pared to 24 %o treated uith only chemotherap! 3 r d 28 0-0 treated uith both modalities. Similar re.ul:. sncies than does the type of inly, for many solid tumors alized than it usually is for IT rlant lymphoma - the -0. malignancies in these caryotype, patient .ponse to treatment and ribution (medians 25-58 TO lsis of the secondary .MDS ) and time between initial and :dims 48-70 months) varied mosome groups, overall the gnificant (Table 4).When inornpared, time to the second.intly longer in the group with mosomes as compared to the ~ o t y p (e70VS. 48, p = .M). 6 patients received intensive :ondary MDS, limiting detailc t of karyotype o n response. appeared to correlate with ete remission (p = .04,Table -' was lowest in patients with m o s o m e 5 (5q- or - 5 ) regardrmalities in chromosome 7 -1- responded. Similarly, pa .s with de novo ANLL chromosome 5 was seen in ' best response rate (65 TO) :ith specific rearrangements hromosome abnormalities. re too small to statistically remission (Table 4). nts receiving intensive treat.-y \-IDS, the rate of complete elated wirh whether the first ogic (40 To), a solid tumor "0). However, response was .h type of prior treatment patien:s treated only with :d complete remission, com\\ i t h only chemotherapy and h modalities. Similar results 4-, CHRO\lOSO\lES IN SECONDARY LEL'K.-\E.LlIA 87 - vc .w7 5 zf= 88 BLOOMFIELD i have been reported by others ( was correlated with karyotype y patients with different first di. difference \vas seen only fo. ( p = .003,Table 5 ) . tVhen rrsr with karyotype groups sepa: previously treated with radiat. or both, a trend toward signifi. for patients previously rt (p = .07, Table Cj. The data L to suggest that karyotype ma! prognostic factor, especially u has been less extensive. Median survival ranged fror the various karyotype groups 1 sults were not statistically d i f i i -7 group tended to survive Ion group. p = .06; -7 vs. 5q-, p = when survival was considered tients in the -7 group, those abnormality survived a medial those with -7 plus other abnorr. 6 months (p = .096). The lasentially identical to that for pa with -7 and - 5 , and -7 and 5q- I Discussion This analysis of 216 published TABLE 5 Response to inrensiie treatment by h - Karyor!pe group, - 5 : 5q.: -7 6; - 5 ; .- & sq. .-;iq. Normal Additions Other structural. specific rearrangrrnenls CR = compie:e :erni,r:on r" P fi: CHROLtOSOLlES IN SECONDARY LEUK.AE\!IA 89 have been reported by others ( 12). %'hen response it as correlated with karyotype groups separately for patients with different first diseases. a significant difference was seen only for the solid tumors ( p = ,003.Table 5 ) . When response u a s correlated with karyotype groups separately for patients previously treated with radiation, chemotherapy, or both, a trend toward significance was seen only for patients previously receiving radiation (p = .07, Table 5). The data could be interpreted to suggest that karyotype may be an independent prognostic factor, ?specially when prior treatment has been less extensive. Median survival ranged from 3 to 9 months for the various karyotype groups (Table 4). These results were not statistically different. However, the -7 group tended to survive longer (-7 group vs. -5 group, p = .06; -7 vs. 5q-, p = .08). Interestingly, ivhen survival was considered separately for patients in the -7 group, those with -7 as their only abnormality survived a median of 10 months and those with -7 plus other abnormalities a median of 6 rnonrhs ( p = .096). The latrer survival is essentially identical to that for patients in the groups ~ i t -h7 and - 5 , and -7 and 5q- (median 5 months). Discussion This analysis of 216 published cases of secondary LIDS on whom banded chromosome analyses were obtained confirms i n d extends previously suggested correlations between karyotype a n d type of prior disease and prior therapy. In particular, 5q- and -7 have previously been correlated respectively with prior solid tumors (5) and prior haematologic malignancies (6),and a -7 has been previously correlated with more extensive prior thelapy (6, IO). The interaction of these two factors requires further study with prospective collection of detailed data regarding primary therapy and sophisticated multivariate analyses of results. Karyotype as classified here also correIated with response to intensive chemotherapy. Patients with abnormalities of chromosome 5 , especially a 5q-, rarely achieved remission with conventional induction chemotherapy. Patients with specific rearrangements and other structural abnormalities seen in de novo ANLL frequently achieved remission and some of these patients were long-term diseasefree survivors. Kantarjian et al., in a preliminary report, also found that karyotype significantly correlated with response to intensive induction chemotherapy and that the specific rearrangements seen in de novo ANLL were most often associated with response (13). More patients must be studied prospectively to determine if karyotype is an independent prognostic factor. Moreover, it is not clear that the karyotype classification used here is the best one and other approaches should be tried. Ka:)oi)pe groups __ _--?: >q-; -7 -5: 6; iq.-: -9. \orm31 .Additions Other siructurai: qxcific rearranycmcnrs * CR = iornplsre renilision. 411 paiients so. caw CR` 16 13 ob 19 32 ob 1.4 43 rQ 6 50 r a 1- 65 ob p = ,036 ~ Paiients with only prior radiarion NO. cases CR 30 2 50 4 6 61 ma 1 100 4 9 89 "`0 p = ,068 Paiienis with prior solid !umors NO. cases CR 80 6 33 m0 to 20 ob I 100 13 77 r o p = ,003 90 BLOOMFIELD Although the results from this analysis of previously published series are provocative, the limitations of such an evaluation are obvious. The only way to clearly answer questions regarding the relation of specific chromosome abnormalities t o exposure to prior radiation and chemotherapy, to type of prior disease a n d to response to treatment of the secondary MDS is a large prospective trial where all patients have good chromosome studies, detailed information o n prior therapy, and optimal uniform treatment of the secondary MDS. Acknowledgements This work was peiformed during a sabbatical at the Department of Medical Genetics, University of Helsinki and supported in pan by the Coleman Leukemia Research Fund. The author is indebted to Professor Janet D. Rowley for providing prepublication data from the University of Chicago series and to Professor Albert de la Chapelle for critical review of the manuscript. References I . Coltman CA. Jr. Treatment related leukemia. In: Bloomfield CD, ed. .Adult leukemias I . Boston: Martinus hijhoff Publishers, 1982; 61-108. 2. Koeffler HP. Rowley JD. Therapy-related acute nonlymphocytic leukemia. In: Wiernik P H . Canellos G P , Kyle RA.Schiffer CA. eds. Neoplastic diseases of the blood. New York: Churchill LiLingstone, 1984: 357-81. 3. Rowley .ID, Golomb H M , Vardiman J . Nonrandom chromosomal abnormalities in acute nonlymphocylic leukemia in patients treated for Hodgkin disease and non-Hodgkin lymphomas. Blood 1977; SO: 759-70. 4. Rowley JD. Golomb HM, Vardiman J W . Nonrandom chromosome abnormalities in acute leukemia and dysmye- lopoietic syndromes in patients uith previous11 treated malignant disease. Blood 1981; 58: 759-67. 5 . LeBeau 41M.Albain KS, Vardiman JU'. Blough R. Golomh HM, Rowle) J D . Therapy-related acute nonlymphoc>Ticleukemia rANLL) and dysmyelopoietic syndrome (DMPSI: Clinical and c)Togenetic correlations. (Personal commun.. cation. 1985). 6 . Groupe Francais de Cvlogenerique Hematologique. Chromosome analysis of 63 cases of secondar) nonlymphuid blood disorders: ,A cooperatne study. Cancer Genet Cytogenet 1981: 12: 95-104. 7 . Pedersen-Bjergaard J . Philip P. Mortensen BT et al. Acute nonl>mphoc)-ric leukemia. preleukemia and acute myeloproliferailbe syndrome secondary to treatment of othc: malignant diseases: Clinical and cytogenetic characteristics and results of in l'irro culture of bone marruu and HLA tkping. Blood 1981; 57: 712-23. 8. Pedersen-Bjergaard J . Philip P. Pedersen NT, et al. .Asure nonlymphoiytic leukemia. preleukemia, and acute m)eloproliferatne syndrome secondary to treatment of other malignant diseases. 11. Bone marrow cytology, c!Togenetisr. results of HLA typing. response to antileukemic ihemotherap! and survi\al in a total series of 5 5 patients. Cancer 1981: 54: 151-62 9. Fourth International Workshop on Chromosomes in Leukemia. 1982: A prospective study of acute nonlymphocytii leukemia. Cancer Genet Cylogenet 1984; 1 I:219-360. IO. .Arthur DC. Bloomfield CD. Banded chromosome analw. in patients uith treatment-associated acute nonlymphoc!tiL leukemia. Cancer Genet Cytogenet 1984; 12: 189-99. 11. Bennet: 1\1, Catoxsky D. Daniel M.-T.. et al. Proposals IC" the classification of the acute leukaemias. Br J Haematd 19'6; 33. 151-58. 12. Bloomiield CD, Preisler H. Cuttner J. et al. Treatmen!. induced acute non-lymphoi!ric leukemia (t-..ZNLL): R i spons: to ~)~arabine-anrhrasvclinteherapy. Blood 1982: M . suppl I : IJ?a. 13. Kantarjian H. Keating S1. %'alters R. et al. Clinical. iV*genetic and prognostic features of secondary leukemlns. P:~L Am Soc Clin Oncol 1985 (in press). Correspondence to: Clara D. Bloomfield, 51.D. Box 277 Uni\ersit) of Minnesota Hospitals !vlinneapnlis. Minnesota 55155 USA t Translocation (1 non-lymphocyt i association A reciprocal apparently balanced [ @21;q23), was the sole chromosomr of suspected acute leukaemia. The posure to mutagenic agents, prese' lasting 17 months before ANLL-\I irradiated after surgical removal of a with .4NLL-M1. Both patients hac We have found 4 patients with t ' had a myelodysplastic syndrome at r the others survived after different Comparison of the cytogenetic f tion was apparently identical. The t The clinical features were somewha: included male sex, age over 50, . haematopoietic cells in a preleukaer posure to irradiation occurred in 1 presently be listed as a non-random myeloid cells. It has become increasingly e\ ic types of neoplastic cells are char chromosomal abnormalities, t numerical (1). Acute non-lymphocytic leu1 characterized by highly non-rar abnormalities (2). Karyotypic valuable diagnostic parametet between the different leukaemi particular, between several o morphologic subtypes of the dih significance of karyotyping in