Prognostic Relevance of Recurrent Genetic Aberrations in Pediatric Acute Megakaryoblastic Leukemia

Genetic abnormalities and early treatment response are the main prognostic factors in pediatric acute myeloid leukemia (AML). Non-Down Syndrome (non-DS) acute megakaryoblastic leukemia (AMKL) is a rare subtype of AML, with a poor prognosis. These cases present with diverse cytogenetic aberrations, which might be important for risk-group stratification. A well-known genetic aberration in non-DS-AMKL is t(1;22)(p13;q13), resulting in expression of the RBM15/MKL1(OTT/MAL) fusion transcript. In recent years, targeted and genome-wide sequencing has identified novel, recurrent molecular abnormalities in pediatric AMKL. New fusions identified are inv(16)(p13.3q24.3) resulting in CBFA2T3/GLIS2 (Gruber et al, Cell, 2012; Masetti et al, Blood, 2013) and t(11;12)(p15;p13) resulting in NUP98/KDM5A (de Rooij et al, Leukemia, 2013) as specific aberrations in pediatric non-DS AMKL. Also KMT2A(=MLL)-rearrangements are recurrent in non-DS AMKL. The prognostic relevance of these novel abnormalities is not determined in a large cohort, since this disease is rare.

To assess frequencies, clinical characteristics and outcome parameters of recurrent cytogenetic aberrations of pediatric non-DS AMKL, databases of the BFM-SG, DCOG, AIEOP, and COG were combined. In this study, we analyzed 151 newly diagnosed pediatric non-DS AMKL cases diagnosed between 1998 and 2014 of whom a sample was available. All patients included were screened for NUP98/KDM5A, CBFA2T3/GLIS2 and RBM15/MKL1 with reverse transcriptase(RT-)PCR, and for KMT2A-rearrangements using split signal FISH and RT-PCR. To assess outcome, probability of event-free survival (pEFS) and probability of overall survival (pOS) were estimated by the Kaplan-Meier method, and groups were compared with the log-rank test; the cumulative incidence of non-response or relapse (pCIR) was analysed by the Kalbfleisch and Prentice method, and groups were compared with the Gray's test. A Cox regression analysis was done for EFS and relapse incidence with as co-variables cytogenetic subgroup, age, sex, WBC and stem cell transplantation (time dependent variable).

Patients were treated with different protocols; however, all protocols consisted of intensive chemotherapy using an anthracycline and cytarabine backbone for both induction and consolidation; 26% of patients received stem cell transplantation (SCT) in first remission. Median age was 1.6 years (range 0.1-17.1 years), with 62% of the patients younger than 2 years at diagnosis; 45% of patients were males, and median white blood cell count was 13.7x10⁹/L (range 1.1-378.5x10⁹/L). Translocation NUP98/KDM5A was identified in 9%; the CBFA2T3/GLIS2 translocation in 16%; RBM15/MKL1 in 12% and KMT2A-rearrangements in 9%, and hence in 54% none of these abnormalities were detected. All these aberrations were mutually exclusive. Comparing patients with NUP98/KDM5A, CBFA2T3/GLIS, RBM15/MKL1, or KMT2A-rearrangements with other pediatric AMKL patients, no significant differences in sex, age, and white blood cell count were found. Outcome between the included collaborative groups did not differ significantly. The 5-year pOS of the entire pediatric non-DS AMKL cohort was 56±4%, 5-year pEFS was 51±4%. NUP98/KDM5A, CBFA2T3/GLIS2 and KMT2A-rearranged positive patients showed a poor outcome (pOS 36±13%, 38±10%, and 38±13% respectively), compared to RBM15/MKL1 positive cases and other pediatric non-DS AMKL (70±11% and 65±5% respectively, p=0.04, Figure 1). Harboring NUP98/KDM5A, CBFA2T3/GLIS2 or a KMT2A-rearrangement resulted in an increased risk of experiencing an event (EFS; HR1.65, 95%CI;1.03-2.66, p=0.039) or relapse (RFS; HR2.14, 95%CI;1.12-4.11, p=0.022). SCT was not an independent factor for event or relapse free survival (HR1.20, 95%CI;0.65-2.20, p=0.565 and HR 1.13, 95%CI;0.57-2.24, p=0.717, respectively), nor did sex, age at diagnosis or WBC.

Our results indicate that non-DS AMKL is a heterogeneous group within pediatric AML. Although the overall survival is poor for non-DS AMKL in general with a 5-yr pOS of ~55%, the poor outcome is specifically determined by cases with NUP98/KDM5A, CBFA2T3/GLIS2 and KMT2A-rearrangements. Other variants, including those with RBM15/MKL1, are associated with a better outcome. These data show that international collaboration allows the identification of prognostic subgroups, which may lead to new and more refined risk-group stratification.