Karyotype complexity and prognosis in acute myeloid leukemia

A complex aberrant karyotype consisting of multiple unrelated cytogenetic abnormalities is associated with poor prognosis in patients with acute myeloid leukemia (AML). The European Leukemia Net classification and the UK Medical Research Council recommendation provide prognostic categories that differ in the definition of unbalanced aberrations as well as the number of single aberrations. The aim of this study on 3526 AML patients was to redefine and validate a cutoff for karyotype complexity in AML with regard to adverse prognosis. Our study demonstrated that (1) patients with a pure hyperdiploid karyotype have an adverse risk irrespective of the number of chromosomal gains, (2) patients with translocation t(9;11)(p21∼22;q23) have an intermediate risk independent of the number of additional aberrations, (3) patients with ⩾4 abnormalities have an adverse risk per se and (4) patients with three aberrations in the absence of abnormalities of strong influence (hyperdiploid karyotype, t(9;11)(p21∼22;q23), CBF-AML, unique adverse-risk aberrations) have borderline intermediate/adverse risk with a reduced overall survival compared with patients with a normal karyotype.


INTRODUCTION
The term complex aberrant is designated to describe karyotypes with multiple unrelated cytogenetic abnormalities. In acute myeloid leukemia (AML), 10-14% of all AML patients, and up to 23% among older AML patients, have karyotypes with ⩾ 3 aberrations. [1][2][3][4] These karyotypes with ⩾ 3 aberrations are classified as adverse genetic risk according to the recommendations of the European Leukemia Net (ELN). 1 However, the UK National Cancer Research Institute Adult Leukaemia Working Group (abbreviated as MRC for Medical Research Council) classification requires ⩾ 4 abnormalities as an informative cutoff for adverse prognosis. 5 Beyond the 3 vs 4 cutoff discordance, the impact of the so-called pure hyperdiploid karyotype (HDK) without structural aberrations or monosomies has not been addressed conclusively. 6,7 Further complicating, the definition of unique adverse-risk abnormalities, which define adverse risk per se, is not fully congruent in both classification systems, with some abnormalities conferring adverse risk according to the ELN but not the MRC and vice versa. Given this heterogeneity, further exploration of complexity seems desirable for several reasons. First, it has been demonstrated that in the adverse-risk group, some patients with certain chromosomal abnormalities fare even worse than others when receiving standard treatment regimens for adverse-risk patients. [8][9][10][11][12] Second, better individual risk prognostication and uniformly defined adverse-risk group allocation are required in order to homogeneously compare treatment regimens at different institutions.
The aim of this study was to define the optimized cutoff of complexity in adult AML in the context of the number of unrelated aberrations (3 vs ⩾ 4) as well as to define the impact of the pure HDK within these groups. Therefore, we evaluated the survival of 417 intensively treated adult non-APL and non CBF-AML patients with complex aberrant karyotypes out of 3526 AML patients who were included in three prospective, randomized, multicenter treatment trials of the Study Alliance Leukemia.

Patient population
The databases of three prospective, randomized trials of the Study Alliance Leukemia, which enrolled a total of 3526 non-APL, intensively treated AML patients between February 1996 and November 2009, were reviewed for patients with multiple cytogenetic aberrations (⩾3) as well as normal karyotype (NK as a control group). The studies were approved by the institutional review boards of all participating centers of the Study Alliance Leukemia in agreement with the Declaration of Helsinki and registered with the National Clinical Trial numbers 00180115 (AML96 trial), 00180102 (AML2003 trial) and 00180167 (AML60+ trial). Written informed consent had been obtained from each patient.
At diagnosis, chromosome analyses were performed on bone marrow and/or peripheral blood samples using standard techniques, including short-term cultures as reported recently. 13 Karyotype description was performed in accordance with the International System for Human Cytogenetic Nomenclature criteria. 14 According to the definition of the MRC, a balanced translocation, for example, t(8;21)(q22;q22), was defined as a single abnormality, because the two breaks and fusions lead to one active chimeric fusion protein. A balanced translocation involving more than two chromosomes was also regarded as a single abnormality. Trisomies or monosomies were regarded as single abnormalities, whereas the gain of two chromosomes, even if they were identical (e.g., tetrasomy 8), was regarded as two abnormalities. Unbalanced translocations leading to gain and loss of chromosomal material were counted as two abnormalities. 5 For instance, a derivative chromosome der(7)t(1;7)(q21;q22) is characterized by a partial monosomy 7q as well as a partial trisomy 1q. In this manner, an isochromosome i(17)(q10) results in two aberrations, that is, monosomy 17p and trisomy 17q. The monosomal karyotype (MK) was defined by the presence of two or more distinct autosomal chromosome monosomies or a single autosomal chromosome monosomy in the presence of one or more structural chromosomal abnormalities. 8

Cytogenetic definitions
Out of the 3526 patients, a total of 2007 patients with either a complex karyotype or a normal karyotype were identified for further analyses (n = 1590 patients with NK; n = 417 patients with ⩾ 3 aberrations which accounted for 30% of the patients in the AML96 trial and 29% of the patients in the AML2003/60+ trials-referring to those patients for whom an aberrant karyotype was diagnosed). Patient characteristics are summarized in Tables 1A and B. The median follow-up time for all patients was 6.2 years (interquartile range, 4.5-8 years). Core-binding factor AML patients (CBF-AML, t(8;21)(q22;q22), inv(16)(p13q22), t(16;16) (p13:q22)) were excluded since additional chromosomal abnormalities even if they resulted in complex aberrant karyotypes have no or little impact on the outcome of patients with favorable-risk CBF-AML 5 and could be confirmed with our CBF-AML patients. 11,15 However, previous results demonstrated an independent influence of the pure HDK on patients' outcome worsening overall survival (OS) and event-free survival significantly. 15 The scoring criterion for pure HDK performed in our analyses was defined by (i) gains of whole chromosomes (e.g., trisomies, tetrasomies), (ii) no additional structural aberrations and (iii) no monosomies.

Treatment protocols
Detailed treatment descriptions of the three trials were reported previously. 13,17,18 In brief, the AML96 trial enrolled adult patients without age restriction, whereas the AML2003 trial included patients up to 60 years of age, and the AML60+ trial patients above the age of 60 years. Apart from double induction chemotherapy administered to patients aged ⩽ 60 years, all three protocols involved a risk-adapted consolidation strategy, including HLA-compatible related or unrelated allogeneic hematopoietic  Abbreviations: NK, normal karyotype (control group); HDK, hyperdiploid karyotype; CK3, complex aberrant patients with three unrelated abnormalities without HDK, without t(9;11), and without adverse-risk abnormalities; CK4, complex aberrant patients with four or more unrelated abnormalities without HDK, without t(9;11), and without adverse-risk abnormalities; CK+adv, complex aberrant patients with three or more aberrations of which at least one aberration was at specific adverse risk per se with exclusion of patients with t(9;11) or HDK; CK3+MK, patients with MK and with three unrelated aberrations; CK3 − MK, patients with three unrelated aberrations without MK; CK4+MK, patients with four or more unrelated aberrations with MK; CK4 − MK, patients with four or more unrelated aberrations without MK; WBC, white blood count; LDH, lactate dehydrogenase. To determine the prognostic influence of the distinct cytogenetic groups independent of age, WBC, serum lactate dehydrogenase levels at baseline, and type of AML (de novo AML, AML with preceding myelodysplastic syndrome, therapy-related AML) as covariates, a stratified multivariable Cox regression analysis for OS was performed. Stratification variable again was study generation. Because of its informative character, allogeneic hematopoietic stem cell transplantation was not censored. All statistical analyses were performed using SPSS version 19.0.1 (SPSS Inc, Chicago, IL, USA) and the R environment for statistical computing version 2.15.3. 21

RESULTS
The following cytogenetic subgroups were analyzed for their influence on OS.
The median OS for these patients was 4.6 months (95% CI, 0-17.4) (Figure 2a). The multivariable Cox regression including age, WBC, lactate dehydrogenase and type of AML showed that HDK was an independent prognostic factor for OS (HR 2.2; 95% CI, 1.4-3.5; P = 0.001). There was no influence of the number of trisomies or tetrasomies on survival. Patients with three trisomies and patients with four or more trisomies/tetrasomies had a similar probabilities of OS (P = NS, data not shown). Furthermore, we compared HDK patients with patients with cytogenetic adverserisk criteria according to the ELN/MRC classifications (Figure 2b). OS did not differ significantly (HR 0.6; 95% CI, 0.4-1.1; P = 0.082). For the adverse control group (CK+adv, n = 333) no further distinction between CK3+adv (n = 35) and CK4+adv (n = 298) was performed since HDK patients with 3 or ⩾ 4 aberrations had similar survival.
The MRC data demonstrated that the level of karyotype complexity has little impact on the outcome in patients already having at least one of the independent abnormalities conferring favorable or adverse risk. Additionally, the MRC reported that in patients lacking any of these independent adverse-risk abnormalities, the presence of ⩾ 4 unrelated changes was found to provide the most informative cutoff, predicting a significantly inferior prognosis. 5 The ELN classification scheme allocated patients with the recurring aberration t(9;11)(p22;q23) to the intermediate-II genetic risk group. 1 Our data confirm this stratification showing that the t(9;11)(p21-22;q23) confers an intermediate risk even with an accompanying complex karyotype.
Our study implies that AML patients with an HDK, specifically those without additional monosomies or structural aberrations, should be allocated to an adverse-risk category because of the significant influence on survival in comparison with NK patients   (OS HR, 2.2; 95% CI, 1.4-3.5; P = 0.001). Recently, the impact of hyperdiploidy in AML patients was published independently, emphasizing the impact of this category. 6 Although the authors identified a similar, obviously non-random pattern of chromosomal gains comparable to our data, they applied a different approach by also including patients with monosomies in their HDK group with numerical changes, whereas we addressed the impact of pure hyperdiploidy separately without including patients with structural abnormalities and patients with loss of chromosomes. An analysis of the French Groupe Francophone de Cytogenetique Hematologique investigated 38 AML patients with high HDKs restricted to karyotypes with only high hyperdiploidy with 49 or more chromosomes. 7 Because of the inclusion of children in the analysis by Luquet et al., a different methodology and different statistical methods being applied, a direct comparison with our results is not possible.
To date, it is not clear whether additional monosomies or structural aberrations occur earlier in the development of AML or whether one or the other abnormality acts as the driver or the passenger aberration. Although gains of additional chromosomes appear to represent the result of clonal evolution caused by the failure of the mitotic machinery rather than an initiating event in AML, we suggest classifying patients with pure HDK as a distinct category, excluding monosomies and structural abnormalities.
The introduction of the MK category by Breems et al. offered the application of a further criterion in risk stratification of AML patients showing that MK identifies a subset of patients with very poor prognosis, which has been confirmed by other groups. 5,8,22 Our study, which was restricted to the distinct population of complex aberrant patients, showed a relevant influence on prognosis for the presence of an MK only in the CK4 situation with an increased risk of death for patients with CK4+MK as compared with patients with CK4 − MK. In patients with CK3 − MK or with CK3+MK the risk of death was superimposable. Additionally, our data confirm that distinct cytogenetic features that accompany other abnormalities have a strong influence on outcome and must be considered independently, for example, patients with t(9;11) conferring intermediate risk.
A consistent definition of adverse-risk complex aberrant karyotype AML appears to be warranted. Here, we confirm a karyotype with ⩾ 4 aberrations and a pure HDK as impressive adverse-risk abnormalities in AML. Patients with three unrelated aberrations fare worse than NK, too, but with an OS classifying between the ⩾ 4 patients and the intermediate NK. This is an important finding that may help to stratify patients to individual optimized treatment strategies and may therefore lead to improved individual survival prognostication. Therefore, based on our findings, we suggest the following re-classification of cytogenetic risk: (1) favorable risk: CBF-AML; (2) intermediate risk: normal karyotype, t(9;11); (3) adverse risk: three aberrations without specific adverse-risk abnormalities, without HDK; (4) very adverse risk: ⩾ 4 aberrations, HDK, specific adverse-risk abnormalities, as defined by the ELN and MRC.