Cytogenetic, Molecular and Clinical Features Associated with Rare CBFB-MYH11 Fusion Transcripts in Patients (Pts) with Acute Myeloid Leukemia (AML) and inv(16)/t(16;16)

Approximately 8% of de novo AML pts have inv(16)(p13q22) or t(16;16)(p13;q22) [inv(16)] and this cytogenetic group is associated with high complete remission (CR) rates and a relatively favorable outcome after cytarabine/daunorubicin (ara-c/dnr)-based induction and high-dose ara-c (HiDAC) consolidation therapy. However, ∼40% of these pts still relapse. The molecular mechanisms that impact on the prognosis of inv(16) pts remain to be fully elucidated. We and others reported that presence of the mutated KIT (mutKIT) gene is associated with worse outcome, but other contributing factors may play a role. Molecularly, inv(16) results in disruption of the MYH11 gene at 16p13 and CBFB gene at 16q22, creating a CBFB-MYH11 fusion gene. Since the genomic breakpoints within CBFB and MYH11 are variable and the fusion transcripts depend on the exons fused, at least 11 different sized CBFB-MYH11 fusion transcript variants have been found. The frequency of each fusion varies, with ∼85% being type A and ∼5% each types D and E; types B, C, and F-K were reported in single cases (hereafter we refer to non-type A fusions as “rare”). To our knowledge, only one study (Schnittger et al. Leukemia 2007;21:725) has examined the biological and clinical significance of rare CBFB-MYH11 fusions in AML. Fusion type was not found to be prognostic, but the cohort included pts with therapy-related and secondary AML. Thus, further studies are warranted. Accordingly, we analyzed 149 de novo CBFB-MYH11 positive AML pts aged <60 years (y; median, 40 y; range 18–59) receiving ara-c/dnr-based induction and HiDAC consolidation therapy (Cancer and Leukemia Group B protocols 9621, 19808, 10503). Among 149 pts, 129 (87%) had type A fusion, 17 (11%) type E, 2 (1%) type D, and 1 (<1%) type I. Pts with rare fusions had lower white blood cell counts (P=.02; median, 18.9 v 27.5 ×10⁹/L), less often skin infiltration (P=.04; 0% v 18%) and tended to have FAB subtype M4Eo less often than the type A fusion pts (P=.13; 67% v 85%). Eighteen rare and 100 type A fusion pts had at least 1.7 y follow-up and thus could be analyzed for outcome. No significant differences in CR rates (P=1.00; 94% v 92%), cumulative incidence of relapse (CIR; P=.16; 5 y rates, 30% v 47%) or overall survival (OS; P=.30; 5 y rates, 78% v 60%) were found between rare and type A fusion pts. However, a trend toward longer event-free survival (EFS) was observed for rare fusion pts compared with type A pts (P=.07; 5 y rates, 66% v 42%; Figure). In an analysis of secondary cytogenetic abnormalities (abns) we found trisomies 8, 13 and 21 were more frequent in rare fusion pts than in type A pts (P=.03; P=.07; P<.001, respectively). Of the rare fusion pts, 44% had at least one of +8, +13 or +21 compared to only 11% of the type A pts. No rare fusion transcript pt harbored +22, which was present in 24 (19%) pts with type A fusion (P=.05; Table]. Strikingly, no rare fusion pt had mutKIT compared with 29% of type A fusion pts (P=.01). Given that +22 has been associated with improved and mutKIT with worse outcome, we also assessed the impact of fusion transcript type by restricting our analysis first to pts with no +22, then to pts without mutKIT, and finally excluding both +22 and mutKIT. No significant differences in CR rates, CIR, OS or EFS were observed between the two groups. We conclude that the type of fusion transcript does not appear to affect significantly the prognosis of inv(16) pts, although a favorable trend for pts with the rare transcript exists. Type A and rare fusion pts differ with regard to distribution of accompanying genetic or cytogenetic alterations, such as KIT mutation or +22, which did not occur in any of the rare fusion pts. Further studies are required to elucidate if the rare fusion transcripts themselves created the necessary biologic conditions that exclude the concurrent presence of the genome aberrations.