Population-based demographic study of karyotypes in 1709 patients with adult Acute Myeloid Leukemia

Transcription

1 (2006) 20, & 2006 Nature Publishing Group All rights reserved /06 $ ORIGINAL ARTICLE Population-based demographic study of karyotypes in 1709 patients with adult Acute Myeloid RN Sanderson 1, PRE Johnson 2, AV Moorman 3,7, E Roman 3,8, E Willett 3,8, PR Taylor 4, SJ Proctor 4, N Bown 5, S Ogston 6 and DT Bowen 1,9* 1 Division of Pathology & Neuroscience, University of Dundee, Dundee, UK; 2 Department of Hematology, Western General Hospital, Edinburgh, UK; 3 Leukaemia Research Fund Epidemiology Unit, University of Leeds, Leeds, UK; 4 Department of Hematology, Royal Victoria Infirmary, Newcastle, UK; 5 Northern Genetics Service, Centre for Life, Newcastle, UK and 6 Division of Community Health Sciences, University of Dundee, Dundee, UK Few large demographic studies of acute myeloid leukemia (AML) are derived from population-based registries. Demographic and karyotypic data were provided for AML cases from two regional leukemia registry databases in Scotland and the Northern Region of England. A population-based dataset was compiled, comprising 1709 patients aged 416 years (1235 North England/474 Scotland patients). The most common cytogenetic abnormalities involved chromosomes 5 and/or 7 (17%). Patients with the following abnormal chromosome 5/7 combinations: 5, del(5q), 5/ 7 and del(5q)/ 7 represented a significantly older population (Po0.01, ANOVA). t(8;21) was the only favourable karyotype found in older age. Karyotypic complexity varied within chromosome 5/7 combination groups; those containing 5, 5/ 7, 5/del(7q), del(5q)/ 7 or del(5q)/ del(7q) combinations were significantly more frequently complex than those containing 7 and del(7q) (Po0.01, v 2 test). Additional recurring cytogenetic abnormalities within complex karyotypes containing chromosome 5/7 combinations included (in order of frequency), abnormalities of chromosomes 17, 12, 3 and 18. Complex karyotypes not involving chromosomes 5 or 7 represented 30% of all complex karyotypes, occurred in younger patients than those involving chromosomes 5 and 7, and frequently included additional trisomy 8 (26%). In conclusion, we describe subgroups within adverse karyotypes, with different demographics, degree of complexity and additional chromosome abnormalities. (2006) 20, doi: /sj.leu ; published online 19 January 2006 Keywords: acute myeloid leukemia; cytogenetics; populationbased Introduction The first major classification of acute myeloid leukemia (AML) was that of the French American British (FAB) group, who proposed criteria for defining AML by morphological subtype. 1 The subsequent identification of recurring cytogenetic abnormalities has informed the WHO classification. 2 The first prospective study of AML by chromosomal abnormality was the International Workshops on Chromosomes in in 1982, from which the Chicago karyotype classification was * Correspondence: Dr DT Bowen, Department of Haematology, Leeds Teaching Hospitals, Leeds LS1 3EX, UK. d.t.bowen@dundee.ac.uk 7 Leukaemia Department of Hematology, Leeds Technology Hospital, Leeds LS13EX, UK 8 Epidemiology and Genetics Unit, University of York, York, UK 9 Department of Haematology, Leeds Teaching Hospitals, UK Received 24 March 2005; revised 20 October 2005; accepted 25 October 2005; published online 19 January 2006 derived. 3 The long-term survival of patients identified at this workshop has been reported recently, and multivariate analysis shows that karyotype is an independent predictor of survival for all patients. 4 The Medical Research Council (MRC) AML 10 trial defines three prognostic groups based on cytogenetic abnormalities detected at presentation; favourable, intermediate and adverse. 5 The concept of classifying AML according to the pretreatment karyotype has been accepted by most leukemia investigators and is one of the most powerful independent prognostic indicators. 5 These cytogenetic classifications are now central to predicting outcome in AML and have been developed into risk-adapted therapeutic strategies. 6 A refined classification of AML based upon detailed molecular, cytogenetic and phenotypic characteristics is necessary to begin to inform the epidemiological and etiological study of AML. Most demographic studies of AML derive from patients eligible for entry into randomised controlled trials. Few studies have used population-based registries and most are limited by relatively small numbers. 7 9 The aim of this study was to provide a more detailed demographic description of cytogenetic abnormalities in AML from a large cohort of population-based data. Materials and methods Patients Patient information was derived from two separate sources, namely regional leukemia registry databases, from Scotland and the Northern Region (North East) of England. The Northern Region Hematology database is a population based registry containing patient data collected prospectively by the Northern Region Hematology Group for patients with AML since 1983 from a geographical-based population of 3.01 million. 7,8 The Scotland Registry is a population based dataset containing patient data prospectively collected in Scotland for all patients with AML aged 16 and above since 1998 using similar methodology to the Northern Region data collection. Cytogenetic data Clonal chromosomal abnormalities were defined and described according to the International System for Human Cytogenetic Nomenclature (ISCN). 10 Cytogenetic analysis was considered successful if a clonal chromosomal abnormality was detected or a minimum of ten metaphases were present, otherwise it was classified as failed. Cases were assigned a modified Chicago classification, 3 by allocation to one of 21 cytogenetic groups in a hierarchical

2 fashion (i.e. placed in the first applicable category). Patients were also split into subgroups representing each combination of chromosome 5 and 7 abnormalities (5/7 subgroups). If a recurring abnormality did not fit any of these groups, a new group was created. Karyotypes with X5 abnormalities were classified as complex. Cases were also assigned a MRC prognostic group (favourable, intermediate and adverse) 5 and a mechanistic cytogenetic classification (normal, translocation, deletion, trisomy) described fully in Moorman et al. 11 Datasets The total number of AML cases studied was 1709, comprising 1235 patients from the Northern Region and 474 patients from the Scotland registry, all aged 416 years. This dataset represents an unselected population of AML patients. Data are presented only for karyotypic subgroups with nx10. De novo versus secondary status was known only for the Scotland registry patients (n ¼ 474), and given the relatively small numbers of secondary cases (n ¼ 50), these were not analysed separately. Statistical analysis Data for modified Chicago cytogenetic groups are presented as a percentage of the total number of cases successfully karyotyped. For each cytogenetic group, the sex ratio of observed number of cases was compared against the sex ratio of cases which would be expected to be seen when adjusted for age and sex by decade (according to the Scotland census 2001; uk/scrol/common/home.jsp) using binomial distribution. The mean age of cytogenetic subgroups was compared using oneway analysis of variance (ANOVA) test, with Scheffe post hoc testing to determine where differences between means existed. For comparison of prevalence and complexity between cytogenetic categories a w 2 test was used. Because of the large number of significance tests performed and the associated increased probability of obtaining conventionally significant (Po0.05) results by chance, only P-values of o0.01 are quoted. Results Frequency of cytogenetic abnormalities From 1709 patients 70% had successful cytogenetic analysis, with a median age of 62 years. Of the 517 patients without a successful cytogenetic analysis, 369 (71%) were not done, and 148 (29%) failed (Table 1). However, for older patients, cytogenetic analysis was not done for 28, 47 and 61% patients in decades 70 79, and years, respectively. For those patients successfully karyotyped, 45% had a normal karyotype, with median age 64 years. The most common cytogenetic abnormality was of chromosomes 5 and/or 7, with a frequency of 17% and median age 68 years. Of the good prognostic cytogenetic groups, the most common recurring abnormality was t(15;17), at 8%, (median age of 41 years), followed by t(8;21) at 4% (median age 53 years), then inv(16) at 2% (median age 47 years). The most common trisomy was þ 8, found in 6% (median age 68.5 years). By MRC prognostic classification, 5 14% had favourable cytogenetics (median age 44), 65% had intermediate cytogenetics (median age 64 years) and 21% had poor risk (median age 68 years). By Moorman classification, 11 20% had translocational karyotypes (median age 48 years), 22% a deletional karyotype (median age 66.4) and 10% a trisomic karyotype (median age 67 years). Age distribution There were significant differences between the mean ages of different cytogenetic groups (Figure 1). The t(9;11) population was significantly younger than any other, followed by t(15;17), then inv(16) (Po0.01,ANOVA). The 5 population was significantly older than any other, followed by del(5q) and 5/ 7, then del(5q)/ 7 and þ 11 (Po0.01,ANOVA). All other groups were not significantly different from each other. The percentage of patients in each cytogenetic group by decade is described in Table 1. By Moorman classification, the translocation population was significantly younger than any other (Po0.01,ANOVA), with no significant differences between the other groups; deletion, trisomy or normal. For the MRC classification, significant differences were seen between each group, with the favourable prognostic group significantly younger than the intermediate, which was younger than the adverse group (Po0.01,ANOVA). Sex distribution Three subgroups showed a strong absolute male predominance (ratio male: female X2), namely t(8;21), inv(16) and þ 11, while three subgroups showed a strong absolute female predominance (ratio male: female p0.5). namely t(9;22), 5q and þ 4 (Table 1). However, when sex ratios were adjusted for the expected population sex ratio by age decade, no significant sex predominance was seen for most cytogenetic subgroups. Where a sex predominance was identified, it was always male, and included t(8;21) (7th and 8th decades only, Po0.01), normal (7th and 9th decades only, Po0.01), þ 8 (8th decades only, Po0.01) and complex (3rd decade only, Po0.01) karyotypic subgroups. Proportion of patients with a complex karyotype Most cytogenetic groups had a low frequency of complex karyotypes (X5 abnormalities), with the exception of the chromosome 5/7 subgroups. Within these chromosome 5/7 subgroups, 68% of the patients also had complex karyotypes. As age increased, the proportion of patients with complex karyotypes and of 5/7 subgroups increased. However, the proportion of patients with a complex karyotype not involving chromosomes 5 or 7 decreased with increasing age beyond the fifth decade. Patients with the following combinations of chromosome 5 and 7 abnormalities, 5 alone, 5/ 7, 5/ del(7q), del(5q)/ 7 and del(5q)/del(7q) formed a group with a significantly greater percentage of patients with a complex karyotype, compared with 7 alone and del(7q) alone (Po0.01). Patients with del(5q) alone had an intermediate percentage of patients with complex karyotypes compared with these two distinct groups (Po0.01) (Figure 2). Although we analysed only karyotypes with X5 abnormalities within our complex group, a definition of complexity of X3 abnormalities would add a further 18% patients in our series. Patients with three abnormalities (n ¼ 26) had predominantly translocational karyotypes by the Moorman classification (50% translocation, 31% deletion, 15% trisomy, 4% unclassified), whilst those with four abnormalities (n ¼ 14) were predominantly deletional (71% deletion, 21% translocation and 7% trisomy). Common additional karyotypic abnormalities The frequency of common additional karyotypic abnormalities was compared within each subgroup of chromosome 5 and 7 combinations (complex or noncomplex) and also in the complex group that lacked 5 or 7 abnormalities (Table 2). The median number of additional abnormalities for complex 445

3 446 Karyotype t(9;11) t(15;17) inv(16) t(8;21) t(9;22) 11q23 Normal * Complex no 5/7-7q -7-5q/-7* -5q* -5/-7* -5* Age Figure 1 Box plot of age distribution of cytogenetic groups. *older subset, y younger subset (Po0.01). Complex no 5/7 ¼ patients with a complex karyotype but lacking a chromosome 5 or 7 abnormality. karyotypes with chromosome 5 or 7 abnormalities was 10 (range 5 31), and for those complex karyotypes without chromosome 5 or 7 abnormalities it was also 10 (range 5 25). Additional abnormalities in descending order of frequency involved chromosome 17 ( 17, add(17p) and del(17p)), followed by abnormalities of chromosomes 3 ( 3, del(3p), del(3q)), 12 ( 12, del(12p)), 18 ( 18), 8 ( þ 8) and 21 ( 21). Of all complex chromosome 5 abnormalities, 47% (52/110) had coexisting chromosome 7 involvement and 56% (62/110) had coexisting chromosome 17 involvement. All additional abnormalities had a similar distribution across different cytogenetic subgroups, with the exception of additional þ 8. Additional þ 8 was almost the exclusive common additional abnormality found in association with the 7 (noncomplex), or del(7q) (noncomplex) subgroups. Additional þ 8 was also the most common additional abnormality in the complex group lacking abnormalities of chromosomes 5 and 7 (26% cases)(table 2). Additional abnormalities found to recur in this group in descending order of frequency involved chromosomes 17 (21%), 21 (15%), 12 (15%), 19 (13%), 21 (9%), 18 (6%) and 3 (6%). Using the Moorman classification, % of complex karyotypes with chromosome 5 or 7 abnormalities were primarily deletional abnormalities, whilst complex karyotypes without 5 or 7 abnormalities were deletional in 38%, trisomic in 26%, translocational in 24%, and 12% were unclassifiable (as karyotype not written in full). Discussion This study provides a population-based demographic description of karyotypes in the adult AML population of the UK. Other unselected population-based studies are limited by an upper age limit, 12 or relatively small numbers. 9,13 We constructed a large Dataset from the Scotland and North of England population based registries. The median age at 65 years reflects the entire adult AML population, and is consistent with a median age of 67 seen in another population based study. 9 Favourable cytogenetic abnormalities, t(15;17), t(8;21) and inv(16) are known to occur more frequently in younger patients, 9,14,15 with a relatively constant incidence throughout life. 15,16 For each of these balanced translocation groups, a much larger proportion of patients are in the younger age decades, giving a more even age-related incidence, in contrast to the greater weighting towards the oldest age decades seen in trisomic and deletional abnormalities. Furthermore, t(15;17), t(9;11) and inv(16) represent a significantly younger subset in our dataset. Patients with chromosome 5 abnormalities were significantly older, and these abnormalities were rare below the age of 60 years. Absolute male:female sex ratios for AML patients have previously described a marked male excess in older age groups, 17 but with no sex difference within specific cytogenetic abnormalities based upon small numbers. 15,18 In an attempt to verify crude incidence data for cytogenetic groups, our data were age/sex adjusted by decade, and now few cytogenetic groups showed clear sex predominance. Those that did were male predominant. The most common balanced abnormality seen in both datasets was t(15;17) at 8%, which is less than in the MRC AML 10 clinical trial cohort for patients o55 years (12%), 5 but more than the 3.3% observed in another population-based study. 9 In the latter study, this proportion was increased to 8.9% when considering only younger patients (o55 years). The MRC AML 11 clinical trial of older adults found a frequency of 4%. 19 In our study, only 15% of the cases of t(15;17) were aged 459.

4 Table 1 Population demographics; median age, sex ratio, frequency and percentage by decade 447 Number of cases (%) by age decade Cases Median age Sex ratio Total Males 900 (53) (62) 43 (51) 54 (46) 76 (48) 119 (50) 237 (58) 228 (55) 106 (49) 5 (28) Females 809 (47) (38) 41 (49) 63 (54) 82 (52) 118 (50) 175 (42) 185 (45) 112 (51) 13 (72) Cytogenetic analysis Successful 1192 (70) (81) 65 (77) 99 (85) 123 (78) 182 (77) 307 (75) 265 (64) 103 (47) 6 (33) 1.08 Failed 148 (9) 63 6 (12) 11 (13) 12 (10) 20 (13) 18 (8) 33 (8) 34 (8) 13 (6) 1 (6) 1.74 Not done 369 (22) 74 4 (12) 8 (10) 6 (5) 15 (9) 37 (16) 72 (17) 114 (28) 102 (47) 11 (61) 1.04 Chicago group Normal 539 (45) (31) 22 (34) 41 (41) 44 (36) 88 (48) 146 (48) 135 (51) 49 (48) 1 (17) 0.91 t(15;17) 95 (8) (26) 16 (25) 19 (19) 19 (15) 15 (8) 14 (5) 1 (0) 0 (0) 0 (0) 0.76 t(8;21) 49 (4) 53 6 (14) 7 (11) 5 (5) 4 (3) 11 (6) 7 (2) 7 (3) 2 (2) 0 (0) 2.27 inv(16) 21 (2) 47 1 (2) 4 (6) 2 (2) 8 (7) 2 (1) 4 (1) 0 (0) 0 (0) 0 (0) 2.5 t(9;22) 11 (1) 52 0 (0) 1 (2) 0 (0) 2 (2) 3 (2) 4 (1) 0 (0) 1 (1) 0 (0) 0.38 t(9;11) 10 (1) 38 1 (2) 2 (3) 2 (2) 3 (2) 0 (0) 2 (1) 0 (0) 0 (0) 0 (0) 1 t(11q23) 14 (1) 49 2 (5) 1 (2) 2 (2) 2 (2) 3 (2) 2 (1) 2 (1) 0 (0) 0 (0) (2) 70 0 (0) 0 (0) 0 (0) 0 (0) 2 (1) 11 (4) 9 (3) 4 (4) 1 (17) (3) 65 2 (5) 0 (0) 4 (4) 2 (2) 6 (3) 7 (2) 12 (5) 2 (2) 0 (0) / 7 30 (3) 70 0 (0) 0 (0) 2 (2) 2 (2) 4 (2) 7 (2) 9 (3) 6 (6) 0 (0) 1.31 del(5q) 56 (5) 74 1 (2) 1 (2) 0 (0) 3 (2) 7 (4) 10 (3) 20 (8) 13 (13) 1 (17) 0.51 del(7q) 20 (2) 61 0 (0) 1 (2) 1 (1) 3 (2) 4 (2) 5 (2) 4 (2) 2 (2) 0 (0) 0.82 del(5q)/ 7 15 (1) 66 0 (0) 0 (0) 0 (0) 0 (0) 2 (1) 8 (3) 4 (2) 1 (1) 0 (0) 1.5 5/7 total 197 (17) 68 3 (7) 2 (3) 8 (8) 12 (10) 27 (15) 53 (17) 60 (23) 29 (28) 3 (50) (1) 65 0 (0) 0 (0) 1 (1) 0 (0) 1 (1) 6 (2) 4 (2) 0 (0) 0 (0) (6) 69 1 (2) 3 (5) 4 (4) 5 (4) 5 (3) 20 (7) 22 (8) 8 (8) 2 (33) (1) 65 0 (0) 0 (0) 1 (1) 2 (2) 1 (1) 2 (1) 3 (1) 1 (1) 0 (0) 0.43 Complexity Noncomplex 1019 (85) (90) 61 (94) 90 (91) 113 (92) 160 (88) 254 (83) 222 (84) 77 (75) 4 (67) 1.08 Complex 173 (15) 67 4 (10) 4 (6) 9 (9) 10 (8) 22 (12) 53 (17) 43 (16) 26 (25) 2 (33) 1.06 Complex 5 57 (33) 72 1 (25) 0 (0) 0 (0) 0 (0) 5 (23) 17 (32) 19 (44) 13 (50) 2 (100) 0.48 Complex 7 13 (8) 62 1 (25) 0 (0) 1 (11) 2 (20) 1 (5) 3 (6) 2 (5) 3 (12) 0 (0) 0.63 Complex 5 and 7 50 (29) 68 0 (0) 0 (0) 3 (33) 2 (20) 8 (36) 19 (36) 12 (28) 6 (23) 0 (0) 1.26 Complex non 5/7 53 (31) 60 2 (50) 4 (100) 5 (56) 6 (60) 8 (36) 14 (26) 10 (23) 4 (15) 0 (0) 2.53 Moorman Trans 238 (20) (50) 31 (48) 35 (35) 45 (37) 41 (23) 43 (14) 17 (6) 5 (5) 0 (0) 1.11 Deletion 268 (23) 67 5 (12) 8 (12) 12 (12) 24 (20) 39 (21) 71 (23) 69 (26) 37 (36) 3 (50) 1.14 Trisomy 125 (11) 67 3 (7) 4 (6) 10 (10) 8 (7) 9 (5) 38 (12) 39 (15) 12 (12) 2 (33) 1.78 Normal 539 ( (31) 22 (34) 41 (41) 44 (36) 88 (48) 146 (48) 135 (51) 49 (48) 1 (17) 0.91 MRC Favourable 165 (14) (43) 27 (42) 26 (26) 31 (25) 28 (15) 25 (8) 8 (3) 2 (2) 0 (0) 1.2 Intermediate 778 (65) (48) 34 (52) 61 (62) 75 (61) 116 (64) 216 (70) 186 (70) 67 (65) 3 (50) 1.05 Adverse 247 (21) 68 4 (10) 4 (6) 12 (12) 16 (13) 37 (20) 66 (21) 71 (27) 34 (33) 3 (50) 1.09 The next most common recurring balanced abnormality was t(8;21) (4%), similar to other population-based studies, with 3.3% 9 and 4.6%. 18 t(8;21) was the only good prognosis abnormality seen to recur over seventy years of age. We found inv(16) in 2% of patients, consistent with the results from previous studies. The large numbers of patients in our dataset enabled subclassification of chromosome 5/7 combinations, and the demonstration that these subgroups have different demographic characteristics, and frequencies of karyotypic complexity. It is clear that abnormalities of chromosome 5 occur in a considerably older population than all other karyotypic abnormalities. The median age of patients with chromosome 5 abnormalities is considerably higher than reported in the MRC AML 11 cohort. 19 This is not surprising, given that the AML 11 clinical trial offered intensive chemotherapy only; a therapeutic option considered inappropriate for many older patients, in which these adverse risk karyotypes are more common. We have used a definition of complexity as X5 abnormalities. 5 Although a recent study reported that only 6% complex karyotypes show complexity X3 but o5 abnormalities, 20 the proportion in our study was 20%. We observed a relatively low frequency of complexity in patients with karyotypes containing only 7 or del(7q) (but lacking involvement of chromosome 5), which contrasted with the almost universal complexity of karyotypes involving 5, irrespective of other combinations, as previously reported. 19 Complex karyotypes lacking involvement of chromosomes 5 and 7 appeared to be a distinct but numerically significant (30% all complex karyotypes) biological entity with a lower median age and a high proportion of þ 8 abnormality. Common additional abnormalities have been elegantly combined with primary abnormalities to devise cytogenetic

5 /27 27/30 5/6 14/ /4 70 Complexity (%) / /35 4/ /-7-5/del(7q) del(5q) del(5q)/-7 del(5q)/del(7q) -7 del(7q) Modified Chicago Group Figure 2 Frequency (above histogram bars) and percentage (y-axis) of chromosome 5 and 7 subgroups, which can also be classified as complex, as defined by X5 abnormalities. Table 2 Additional abnormalities associated with chromosome 5 and 7 subgroups. A. for all combinations of abnormalities of chromosomes 5 and 7; B. for complex abnormalities involving chromosomes 5 and/or 7 versus complex abnormalities lacking involvement of chromosomes 5 and 7. Note that many patients have 41 additional abnormality and the total number of cases in columns 3 16 may therefore exceed that in column 2 Cytogenetic subgroup Total cases 3 del (3p) del (3q) del (12p) 17 Del (17p) Add (17p) (A) All 5/7 combinations complex / / 7 complex del(5q) del(5q) complex del(5q)/ del(5q)/ 7 complex complex del(7q) del(7q) complex del(5q)/del(7q) del(5q)/del(7q) complex /del(7q) /del(7q) complex /7 Other Total 5/7 (%) (12) 8 (4) 3 (2) 29 (15) 26 (13) 10 (5) 51 (26) 5 (3) 12 (6) 27 (14) 3 (2) 11 (6) 18 (9) (B) Complex 5/7 vs non 5/7 5/7 complex (%) (20) 8(7) 3 (3) 18 (15) 25 (21) 9 (8) 48 (40) 4 (3) 12 (10) 25 (21) 3 (3) 10 (8) 17 (14) Non 5/7 complex (%) 53 2 (4) 1 (2) 0 (0) 14 (26) 8 (15) 0 (0) 8 (15) 1 (2) 2 (4) 3 (6) 7 (13) 8 (15) 5 (9) and molecular pathways to the development of therapy-related MDS/AML. 21 The frequency of these additional abnormalities within our 5/7 combinations and complex karyotypic subsets, indicate that the spectrum of abnormalities is broadly similar across all complex karyotypes, irrespective of the primary abnormality (e.g. 5, 7 etc.). Noncomplex 7 or del(7q) groups largely lacked these additional abnormalities, while in the noncomplex del5(q) they were present, albeit at a lower frequency than in the complex karyotype groups. Our data differ from Pedersen Bjergaard in that complex karyotypes with 7, and del(7q) (but lacking 5 or del(5q)) have a similar spectrum of additional abnormalities to complex karyotypes with chromosome 5 involvement. Abnormality of chromosome 17 was overwhelmingly the most frequently observed additional karyotypic change. As previously observed, coexistence of chromosome 5 and 17 abnormalities was equally as common

6 as coexistence of abnormalities of chromosomes 5 and Whilst the spectrum of additional abnormalities was similar for complex karyotypes lacking chromosome 5 or 7 involvement, the frequency of these abnormalities differed, with a high proportion of trisomy 8 the most notable. A major limitation of our study is the exclusive dependence upon G-banding karyotypic analysis. Recent analyses of complex karyotypes by a combination of multicolour fluorescence in situ hybridisation (FISH) and G-banding demonstrated that 480% cases showed a deletion of chromosome 5q. 20,23 In our study using G-banding alone, the corresponding figure was 57%. Thus, it may be that many of our patients lacking the 5 or del(5q) abnormality by G-banding, would be reclassified using M-FISH. A high frequency of associated submicroscopic deletions at commonly deleted loci (including p53 on chromosome 17p, and NF1 on chromosome 17q has also been reported in patients with del(5q), although the vast majority of these patients fulfilled our definition of complexity. 24 We may therefore have underestimated the additional abnormalities within complex groups and perhaps within the del(5q) noncomplex group, which demonstrated a similar spectrum of additional abnormalities. It therefore remains unclear how this spectrum of additional abnormalities relates to specific abnormalities of chromosomes 5 or 7, or simply to complexity. Our data would indicate the latter, given the distinct demographic differences between complex karyotypic groups lacking abnormalities of chromosomes 5 and 7, compared with complex karyotypes involving these chromosomes at the G-banding level. Additional trisomy 8 was the exceptional additional abnormality, represented across all combinations of chromosome 5 and 7 abnormalities, whether complex or not. An excess of trisomy 8 was seen in the complex karyotypic groups lacking chromosome 5 and 7 involvement. The biological and clinical significance of trisomy 8 in AML remains unclear. 5,25 27 In MDS, trisomy 8 is present only in progenitors and is lacking in long-term repopulating cells capable of engraftment in NOD- SCID mouse systems. In contrast del (5q) is present in the most primitive stem/progenitor cell population. 28 In conclusion, this study describes in detail the demographics of AML in a large population-based series. Prognostically favourable balanced abnormalities occurred less frequently in the older AML population than in the typically younger clinical trial population, with a more even age-related incidence when compared to other cytogenetic groups. We were able to demonstrate subgroups within the adverse risk karyotypes, with different demographics, degrees of complexity and frequency of additional abnormalities and suggest that cytogenetic pathways previously described for therapy-related AML also exist in both de novo AML, and AML secondary to an antecedent hematological disorder. Acknowledgements The Hematology Units in Scotland, and the North East of England, who contributed invaluable data to the respective registries. The assistance of Christine Maguire and Jo White in the Scotland Leukaemia Registry is gratefully acknowledged. We are grateful to Dr. David Grimwade for critical review of the manuscript. PRT is supported by the Newcastle upon Tyne Hospitals Regional Research and Development Fund. The Scotland Leukaemia Registry was generously supported by a grant from the Lloyds/ TSB Research Foundation. References 1 Bennett JM, Catovsky D, Daniel MT, Flandrin G, Galton DA, Gralnick HR et al. Proposals for the classification of the acute leukaemias. French-American-British (FAB) co-operative group. Br J Haematol 1976; 33: Bennett JM. World Health Organization classification of the acute leukemias and myelodysplastic syndrome. Int J Hematol 2000; 72: Rowley JD, Golomb HM. The 4th International Workshop on chromosomes in leukemia a prospective study of acute nonlymphocytic leukemia, Chicago, Illinois, USA, September Cancer Genetics And Cytogenetics 1984; 11: Bloomfield CD, Shuma C, Regal L, Philip PP, Hossfeld DK, Hagemeijer AM et al. Long-term survival of patients with acute myeloid leukemia: a third follow-up of the Fourth International Workshop on Chromosomes in. Cancer 1997; 80 (11 Suppl): Grimwade D, Walker H, Oliver F, Wheatley K, Harrison C, Harrison G et al. The importance of diagnostic cytogenetics on outcome in AML: analysis of 1,612 patients entered into the MRC AML 10 trial. The Medical Research Council Adult and Children s Leukaemia Working Parties. Blood 1998; 92: Wheatley K, Burnett AK, Goldstone AH, Gray RG, Hann IM, Harrison CJ et al. A simple, robust, validated and highly predictive index for the determination of risk-directed therapy in acute myeloid leukaemia derived from the MRC AML 10 trial. United Kingdom Medical Research Council s Adult and Childhood Leukaemia Working Parties. Br J Haematol 1999; 107: Taylor PR, Reid MM, Stark AN, Bown N, Hamilton PJ, Proctor SJ. De novo acute myeloid leukaemia in patients over 55-years-old: a population-based study of incidence, treatment and outcome. Northern Region Haematology Group. 1995; 9: Proctor SJ, Taylor PR. Age cohort subgroups in adult acute myeloid leukaemia studies the population perspective. 2001; 15: Preiss BS, Kerndrup GB, Schmidt KG, Sorensen AG, Clausen NA, Gadeberg OV et al. Cytogenetic findings in adult de novo acute myeloid leukaemia. A population-based study of 303/337 patients. Br J Haematol 2003; 123: Mitelman F. ISCN: An international system for human cytogenetic nomenclature. Basel, Switzerland: Karger, Moorman AV, Roman E, Willett EV, Dovey GJ, Cartwright RA, Morgan GJ et al. Karyotype and age in acute myeloid leukemia Are they linked? Cancer Genet Cytogenet 2001; 126: Schoch C, Kern W, Krawitz P, Dugas M, Schnittger S, Haferlach T et al. Dependence of age-specific incidence of acute myeloid leukemia on karyotype. Blood 2001; 98: Mauritzson N, Johansson B, Albin M, Billstrom R, Ahlgren T, Mikoczy Z et al. A single-center population-based consecutive series of 1500 cytogenetically investigated adult hematological malignancies: karyotypic features in relation to morphology, age and gender. Eur J Haematol 1999; 62: Moorman AV, Roman E, Cartwright RA, Morgan GJ. Patients entered into MRC AML trials are biologically representative of the totality of the disease in the UK. Clin Lab Haematol 2002; 24: Moorman AV, Roman E, Cartwright RA, Morgan GJ. Age-specific incidence rates for cytogenetically-defined subtypes of acute myeloid leukaemia. Br J Cancer 2002; 86: Vickers M, Jackson G, Taylor P. The incidence of acute promyelocytic leukemia appears constant over most of a human lifespan, implying only one rate limiting mutation. 2000; 14: Mertens F, Johansson B, Mitelman F. Age- and gender-related heterogeneity of cancer chromosome aberrations. Cancer Genet Cytogenet 1993; 70: Cartwright RA, Gurney KA, Moorman AV. Sex ratios and the risks of haematological malignancies. Br J Haematol 2002; 118: Grimwade D, Walker H, Harrison G, Oliver F, Chatters S, Harrison CJ et al. The predictive value of hierarchical cytogenetic classification in older adults with acute myeloid leukemia (AML): analysis of 1065 patients entered into the United Kingdom Medical Research Council AML11 trial. Blood 2001; 98:

7 Schoch C, Haferlach T, Bursch S, Gerstner D, Schnittger S, Dugas M et al. Loss of genetic material is more common than gain in acute myeloid leukemia with complex aberrant karyotype: a detailed analysis of 125 cases using conventional chromosome analysis and fluorescence in situ hybridization including 24-color FISH. Genes Chromosomes Cancer 2002; 35: Pedersen-Bjergaard J, Andersen MK, Christiansen DH, Nerlov C. Genetic pathways in therapy-related myelodysplasia and acute myeloid leukemia. Blood 2002; 99: Castro PD, Liang JC, Nagarajan L. Deletions of chromosome 5q133 and 17p loci cooperate in myeloid neoplasms. Blood 2000; 95: Van Limbergen H, Poppe B, Michaux L, Herens C, Brown J, Noens L et al. Identification of cytogenetic subclasses and recurring chromosomal aberrations in AML and MDS with complex karyotypes using M-FISH. Genes Chromosomes Cancer 2002; 33: Crescenzi B, Lastarza R, Romoli S, Beacci D, Matteucci C, Barba G et al. Submicroscopic deletions in 5q- associated malignancies. Haematologica 2004; 89: Paulsson K, Sall T, Fioretos T, Mitelman F, Johansson B. The incidence of trisomy 8 as a sole chromosomal aberration in myeloid malignancies varies in relation to gender, age, prior iatrogenic genotoxic exposure, and morphology. Cancer Genet Cytogenet 2001; 130: Wolman SR, Gundacker H, Appelbaum FR, Slovak ML. Impact of trisomy 8 (+8) on clinical presentation, treatment response, and survival in acute myeloid leukemia: a Southwest Oncology Group study. Blood 2002; 100: Schoch C, Haase D, Fonatsch C, Haferlach T, Loffler H, Schlegelberger B et al. The significance of trisomy 8 in de novo acute myeloid leukaemia: the accompanying chromosome aberrations determine the prognosis. German AML Cooperative Study Group. Br J Haematol 1997; 99: Nilsson L, Astrand-Grundstrom I, Anderson K, Arvidsson I, Hokland P, Bryder D et al. Involvement and functional impairment of the CD34+CD38-Thy 1+ hematopoietic stem cell pool in myelodysplastic syndromes with trisomy 8. Blood 2002; 100: