The preferential occurrence of FLT3‐TKD mutations in inv(16) AML and impact on survival outcome: a combined analysis of 1053 core‐binding factor AML patients

Acute myeloid leukaemia (AML) is a heterogeneous haematological cancer characterized by abnormal growth and a block in myeloid differentiation. Chromosomal karyotype aberrations are central to AML risk stratification; for example Core binding factor (CBF) AML, which represents approximately 13% of adult de novo AML (Mrozek et al, 2008), has a good prognosis with 60–70% long-term survival (Dohner & Gaidzik, 2011). This group of AML is most commonly characterized by the presence of the translocation t(8;21)(q22;q22), or the inversion inv(16)(p13q22)/t(16;16)(p13q22), which result in the RUNX1-RUNXT1 and CBFΒ-MYH11 fusion proteins respectively (Muller et al, 2008). An important independent poor prognostic outcome factor in AML is mutation of the FMS-like tyrosine kinase 3 gene (FLT3), which is mutated in approximately 30% of AML cases (Weisberg et al, 2010). Overall, across AML cases the occurrence of substitutions in the tyrosine kinase domain (FLT3-TKD mutations; ~7%) is less frequent than internal tandem duplications (FLT3-ITD; 20–25%) (Weisberg et al, 2010), although there is considerable heterogeneity amongst subtypes. FLT3-TKD mutations are over-represented in PML-RARA and inv(16) AML subtypes and more generally in the intermediate cytogenetic subgroup, suggesting that FLT3-TKD mutations are not randomly distributed amongst AML subtypes (Mead et al, 2007; Bacher et al, 2008; Mrozek et al, 2008; Li et al, 2012). While FLT3-ITD mutation is a marker of poor outcome in normal karyotype AML (NK-AML) (reviewed in (Dohner & Gaidzik, 2011)) the association of FLT3-TKD mutations with outcome is controversial and more difficult to study due to the relative rarity and spread across subtypes (Mead et al, 2008). Therefore meta-analysis provides a tool for studies aimed at dissecting the impact of FLT3-TKD in AML. In CBF AML the frequency and prognostic outcome of the FLT3-TKD mutation remains unclear due to the low frequency of this subtype of AML. Here we have examined closely the association of the FLT3-TKD mutations with the CBF AML subtypes by performing a combined-analysis of previously published reports in the literature between the years 2000 and 2012. We limited this analysis to adult CBF leukaemia studies that reported screening for both FLT3 mutation types in each subgroup. Patients carrying both mutations (FLT3-ITD and FLT3-TKD), and those with unknown FLT3 mutation status, were excluded. We also filtered to exclude overlapping studies based on similar clinical trial cohorts. We then compared the frequency of the FLT3-TKD mutation to the frequency of FLT3-ITD in both subtypes of CBF AML. We analysed 21 adult AML studies, including our South Australian AML cohort, with a total of 1053 CBF leukaemia patients (447 inv(16) patients and 606 t(8;21) patients). The details for each study are listed in Table SI. Within the inv(16) group 16 were positive for FLT3-ITD and 56 for FLT3-TKD genotype with 37 FLT3-ITD and 20 FLT3-TKD positive genotypes in the t(8;21) group. We examined the association of the mutations in each group using the chi-squared test to determine whether the frequency of FLT3-ITD and FLT3-TKD mutations varied significantly compared to the expected frequency in adult AML (20% and 7% respectively; (Weisberg et al, 2010)). In the inv(16) subtype, the FLT3-TKD mutation was significantly overrepresented (13·3%) while the FLT3-ITD mutation was under-represented (4·9%) (P = 3·6 × 10−7 and P = 7·6 × 10−12, respectively, Fig 1). In contrast, we found that FLT3-ITD was more frequent than FLT3-TKD in t(8;21) AML (7·1% vs 3·6%), however both mutations were significantly underrepresented (P = 2·6 × 10−13 and P = 1·7 × 10−3, respectively) (Fig 1). Thus our combined analysis of a large number of patients confirms a previous finding that FLT3-TKD mutations occur preferentially with inv(16) AML (Mead et al, 2007; Mrozek et al, 2008). This association was not seen in the t(8;21) subtype, highlighting a key difference between the two groups of CBF AML. FLT3-ITD mutations were under-represented in both groups raising the possibility that particular signalling events that are unique or amplified by the FLT3-TKD mutant, but distinct from those generated by FLT3-ITD, may cooperate with the CBFΒ-MYH11 fusion in AML pathogenesis. We next investigated whether the FLT3 mutation has an impact on clinical outcome in CBF leukaemia patients. To do this, we compiled data from multiple published studies that provided clinical information for individual patients. Meta-analysis of 5 published studies to determine the prognostic clinical outcome of the FLT3-TKD mutation in inv(16) AML was performed. The weighted measures of odd-ratio (OR) was estimated by the Mantel-Haenszel method and as there was no significant heterogeneity observed (P = 0·29), a fixed effects model was applied. Evidence for publication bias was assessed qualitatively, through the assessment of funnel plots of the relationship between treatment effects expressed as the natural logarithm of the odds ratio (OR) against the standard error of the difference for each outcome, and was not found (data not shown). As expected, the number of patients with FLT3-ITD mutation was very low in both t(8;21) and inv(16) subtypes and we were unable to draw a conclusion on the clinical outcome (data not shown). However, we investigated outcome for 27 FLT3-TKD patients in the inv(16) subtype and, as shown in the forest plot (Figure S1), although we observed a trend towards better outcome for FLT3-TKD positive patients (n = 26) compared to the FLT3-TKD negative patients (n = 113), this did not reach statistical significance (OR=0·42; 95% confidence interval=0·17–1·05; I2 = 20%; P = 0·06). This study is the largest analysis of FLT3 mutations in CBF AML to-date. Our findings confirm that the FLT3-TKD mutation occurs preferentially in the inv(16) AML subtype, implying a biological significance, however the presence of FLT3-TKD does not appear to strongly affect prognosis in this subgroup. CHK was supported by a scholarship from The Queen Elizabeth Hospital Research Foundation. MP was supported by the Shahin Early Career Fellowship. ©Central Adelaide Local Health Network Inc 2012. This work is copyright. Apart from any use as permitted under the Copyright Act 1968, no part may be reproduced without prior written permission. Requests and enquiries concerning reproduction and rights should be directed in the first instance to John Wiley & Sons Ltd of The Atrium, Southern Gate, Chichester, West Sussex PO19 8SQ UNITED KINGDOM; alternatively to the SA Pathology, IP, Research Governance & Contracts Office, Level 3, IMVS South Building, Frome Road, Adelaide, SA 5000, Australia. The Work has been produced with the assistance of funding provided by The Queen Elizabeth Hospital Research Foundation (“TQEHRF”). TQEHRF does not give any warranty nor accept any liability in relation to the contents of this work. CHK, ALB and RDA designed and analysed the research and wrote the paper. CHK performed the research. IDL performed clinical review and edited the manuscript. MP and DGI performed critical review of the manuscript. The authors have no conflicts of interest financial or otherwise to declare. Table SI. Summary of studies included in the meta-analysis. Figure S1. Prognostic significance of the FLT3-TKD mutation in AML. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.