AUTHORS
Maya Rose Ball-Burack, MD, Jack H. Peplinski, MSc2, Rhonda E. Ries, MA3, Katherine Tarlock I, MD4, Danielle Kirkey5, Soheil Meshinchi, MD, PhD6
AFFILIATIONS
1 Fred Hutchinson Cancer Research Center, Seattle, Washington, United States, Seattle Children's Hospital, Seattle, Washington, United States
2 Fred Hutchinson Cancer Research Center, Seattle, Washington, United States
3 Fred Hutchinson Cancer Research Center, Seattle, Washington, United States
4 Fred Hutchinson Cancer Research Center, Seattle, Washington, United States, Seattle Children's Hospital, Seattle, Washington, United States
5 Fred Hutchinson Cancer Research Center, Seattle, Washington, United States, Seattle Children's Hospital, Seattle, Washington, United States
6 Fred Hutchinson Cancer Research Center, Seattle, Washington, United States, Seattle Children's Hospital, Seattle, Washington, United States
ABSTRACT
Core-binding factor (CBF) AML comprises two fusion-driven leukemias initiated by RUNX1::RUNX1T1 (t(8;21)) and CBFB::MYH11 (inv(16)). Both fusions can arise in utero, yet they show different age distributions: inv(16) AML is more frequent <3 years of age compared to t(8;21) AML (18% vs 2%), while t(8;21) is more common after age 5, suggesting distinct developmental trajectories after fusion acquisition. Here we define fusion-specific genomic alterations, variant molecular pathways, and their distinct prognostic effects in CBF AML.
Molecular characterization was performed on children and young adults with t(8;21) or inv(16) AML enrolled on Children’s Oncology Group trials. Of 657 evaluable patients with CBF AML, 370 had t(8;21) and 287 had inv(16). Cooperating alterations were grouped into pathways: RTK/RAS signaling, chromatin regulation, transcription factor regulation, cohesin biology, DNA methylation, RNA splicing and tumor-suppressor.
Despite disrupting the same transcriptional complex, the two subtypes demonstrated distinct pathway specific alterations when evaluated with Fisher’s exact test. t(8;21) AML was enriched in alterations involving chromatin regulation (34% vs 10%; Padj < 0.0001), notably ASXL1/2 and KDM6A; transcription factor regulation (17% vs 7%; Padj < 0.0001), including SPI1, GATA2, IKZF1, NCOR1 and MGA; DNA methylation (12% vs 2%; Padj < 0.0001), including IDH1/2 and TET2; RAD21, a cohesin complex subunit (6% vs 1%,) and RNA splicing (6% vs 0%; Padj = 0.0026). Cytogenetic abnormalities were more frequent in t(8;21) AML (63% vs 28%; Padj < 0.0001) particularly del(9q), X and Y. Conversely, inv(16) AML had preferential acquisition of RTK alterations, including NRAS (53% vs 28% in t(8;21); Padj < 0.0001), and recurrent KRAS, FLT3, KIT exon 17, PTPN11, JAK3 and PTEN alterations.
To identify fusion-specific prognostic factors, we ran Cox regression analyses of recurrent alterations and predefined biologic pathways. 5-year event-free survival (EFS) were estimated by Kaplan-Meier methods. Adverse EFS in t(8;21) AML was associated with SPI1, IDH1/2, trisomy 4, ASXL1, CCND2 and the DNA methylation pathway, while adverse EFS in inv(16) AML was dominated by FLT3-ITD, CSF3R, KIT exon 17 and cohesin alterations.
Cohesin alterations provided clear evidence of fusion-dependent clinical significance—they were more frequent in t(8;21) AML (12% vs 6%; P=0.12) and associated with superior EFS (83% vs 64% without cohesin alterations; P=0.02), driven predominantly by RAD21. Strikingly, cohesin alterations in inv(16) AML identified a subgroup with markedly inferior EFS (15% vs 62% without cohesin alterations; P < 0.0001).
A fusion-dependent pattern was also observed for RTK alterations. In inv(16) AML, FLT3-ITD had EFS of 7% vs. 58% in patients without FLT3-ITD (P < 0.0001), whereas FLT3-ITD was not associated with worse outcomes in t(8;21) AML (EFS 50% vs 64%; P=0.3316). KIT exon 17 alterations were also associated with inferior EFS (32% vs 65%; P=0.0002) in inv(16) AML, but not t(8;21) AML (60.6% vs 68.1%; P=0.3377). Conversely, alterations of methylation and transcriptional regulation—particularly SPI1, IDH1/2, ASXL1 and CCND2—identified adverse subsets in t(8;21) AML but were uncommon or non-prognostic in inv(16) AML.
Composite groups constructed using alterations associated with fusion-specific inferior EFS identified high-risk subsets. In t(8;21) AML, grouped patients with subtype-specific adverse alterations (DNA methylation, SPI1, IDH1/2, CCND2, ASXL1, trisomy 4) had EFS of 44% vs 69% (P<0.0001). In inv(16) AML, grouped patients with subtype-specific adverse alterations (cohesin family, CSF3R, KIT Exon 17, FLT3-ITD) had EFS of 27% vs 65% (P<0.0001).
Historically grouped and classified as favorable-risk, t(8;21) and inv(16) AML represent distinct leukemias with characteristic pathway-level genomic architectures that modify biology and prognosis. t(8;21) AML is enriched for alterations in chromatin, epigenetic, and cohesin regulation, whereas inv(16) AML is characterized predominantly by RTK/RAS signaling alterations. This large cohort reveals divergent clinical implications of cohesin alterations, FLT3-ITD and KIT exon 17, showing that CBF fusion alone is insufficient for risk prognostication. These findings establish a framework for fusion-specific molecular risk stratification and identify biologically defined pathways for precision therapeutic investigation in children and young adults with CBF AML.
DISCLOSURES
Maya Rose Ball-Burack, MD: No relevant conflicts of interest to declare.