Background: The acute myeloid leukemia (AML)-M4 subtype is frequently associated to eosinophilia and/or to the cytogenetic alteration inv(16)/t(16;16). The presence of these features is generally associated with good prognosis, but the studies concerning their exact role are hampered by the low number of cases. We retrospectively analyzed patients with AML-M4 enrolled in two consecutive GIMEMA studies to assess the influence of eosinophilia and of the inv(16) cytogenetic abnormality on the prognosis of acute myelomonocytic leukemia (M4) and acute myelomonocytic leukemia with abnormal eosinophils (M4eos).

Setting: A retrospective study, conducted over 9 years in patients affected by AML, admitted to 35 Italian hematological divisions.

Patients and methods: Between December 1993 and December 2002, 1686 patients aged over 15 years with a diagnosis of AML were admitted to the EORTC-GIMEMA AML10 and AML 99p trials; of these, 400 patients (355 M4 and 45 M4Eo) were studied. The diagnosis of M4 and M4eos was first established at each institution and subsequently centrally reviewed at the time of study entry. The following parameters were evaluated: morphology, immunophenotype, cytogenetics performed at the onset of the disease, complete remission achievement and duration, overall survival (OS) and event-free survival (EFS) from AML diagnosis. Patients with M4eo were younger and more frequently associated with inv(16) compared to M4. Cytogenetic analisis failed or was not carried out in 40% of cases, while it was successfully analyzed in 240 cases; inv(16) was found in 17% of them.

Results: Concerning the probability of obtaining a CR after standard treatment, at univariate analysis M4Eo had a non significant advantage compared to M4, while presence of inv(16) was significantly correlated to a higher CR probability; multivariate analysis showed a significant advantage only of M4Eo+ inv(16) compared to M4-without eosinophilia and without inv(16). DFS was not different in univariate analysis between patients carrying or not inv(16), while a borderline advantage of M4Eo was observed with respect of M4, not confirmed at multivariate analysis. OS curves showed at univariate analysis a significant advantage both of the presence of eosinophilia (P=0.004) and of inv(16) (P=0.01); at multivariate analysis, patients with M4Eo+ inv(16) had a highly significant advantage compared to M4 without eosinophilia and without inv(16) (P=0.004), but also compared to M4+ inv(16) (P=0.045), and M4Eo-without inv(16) (P=0.076).

Conclusions: AML-M4 with or without eosinophilia represent 23.7% of AML. The presence of eosinophilia and of inv(16)/t(16;16) can be both considered favorable prognostic factors; however, only the association of both features allows a highly significant advantage in terms of CR and OS.

Author notes

Corresponding author