Background. A recent large retrospective study has observed that the presence of cytogenetic clonal heterogeneity, as detected by metaphase karyotyping, was associated with poor prognosis in patients (pts) with AML (Bochtler et al., J Clin Oncol 2013, 31: 3898-905).

Objectives. The aims of the current study were to reassess the impact of cytogenetic clonal heterogeneity on outcomes in younger AML pts (age 15-60 yrs) and to determine how clonal heterogeneity modified the effects of the type of remission-induction chemotherapy and of having a donor in an independent sample of AML pts < 60 years of age with an abnormal karyotype.

Methods. In the AML-10 trial pts were randomized to receive either daunorubicin (DNR), mitoxantrone (MTX), or idarubicin (IDA) in addition to standard-dose cytarabine and etoposide for induction chemotherapy. In the AML-12 trial, pts were randomized between SDAC and high dose cytarabine (HiDAC) induction, both with DNR and etoposide. In both trials, a second cycle of the same induction chemotherapy was administered in patients who achieved a PR. Patients who achieved a CR or a CRi after 1 or 2 courses of induction chemotherapy received a consolidation chemotherapy with the same anthracycline as in the induction course plus intermediate dose cytarabine. Thereafter, patients were scheduled to receive in first CR/CRi an allogeneic HCT if they had an HLA-identical donor or an autologous HCT otherwise. Cytogenetic analyses were performed at diagnosis and centrally reviewed. For the current analysis, cytogenetics were re-classified using the MRC classification. The definition of a (sub)clone was according to the international System of Human Cytogenomic Nomenclature (2016). Abnormal karyotypes were classified as no subclone (cytogenetic abnormality present in a single clone), subclones (presence of multiple (2, 3 or 4) clones) and composite karyotypes (CK) when clonal heterogeneity was too complex to allow enumeration of individual subclones.

Results. In the AML-10 trial, 2157 pts were randomized to receive DNR, MTX or IDA. The current analysis was performed in a subgroup of 684 pts with an abnormal karyotype: 555 of them had no subclones, 109 at least 1 subclone and 20 a CK. In the AML-12 trial, 1942 pts were randomized between HiDAC and SDAC. The current analysis was performed in a subgroup of 606 pts with an abnormal karyotype: 470 of them had no subclones, 117 at least 1 subclone and 19 a CK (Table 1). Pts with a CK (n=39) were older than 25 years and 84.6% had adverse MRC cytogenetic features. In comparison to pts with no subclones, those with subclones had a similar probability of achieving a CR and had similar DFS from CR/CRi and OS from randomization (Table 1). However, in comparison to pts with no subclones, those with a CK had a similar probability of achieving a CR/CRi but a shorter DFS (HR=2.10, 95% CI: 1.39-3.16) and OS (HR= 1.54, 95% CI: 1.07-2.21). In multivariate Cox model stratified by protocol and adjusted for age, cytogenetic risk group, secondary versus primary AML and PS, neither the presence of subclones (HR 1.11, 95% CI: 0.92-1.33) nor CK (HR 0.93, 95% CI: 0.64-1.34) were associated with shorter OS (P=0.49). Comparable findings were obtained after restricting the analyses to pts with adverse/intermediate cytogenetic features.

Among pts included in the AML-10 trial, the impact on the OS of the type of anthracycline used (IDA vs MTZ vs DNR) was not significant, and its magnitude was not statistically different according to the clonal heterogeneity status (no subclones versus subclones versus CK)(P=0.24). Further, among pts included in the AML-12 trial, the difference regarding OS between HiDAC and SDAC was limited (HR=0.94), and its magnitude was not impacted by the clonal heterogeneity status (P=0.54). Finally, in the whole cohort, DFS from CR/CRi was significantly longer in the subgroup of pts with a donor (n=250) than in those without a donor (n=449) (HR=0.72, 95% CI: 0.58-0.89, P=0.002); the impact on DFS of having a donor vs no donor was similar in the 3 clonal subgroups (P-value for interaction from the Cox model: P=0.26).

Conclusions. The influence of clonal heterogeneity on outcome was evaluated in an independent sample of AML patients. Clonal heterogeneity as defined by the presence of subclones showed no effect on OS as compared to cytogenetic abnormal patients without clonal subclones. The dismal outcome in the pts with a CK could be explained by the known predictors of poor prognosis.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.