AUTHORS
Ozgur Can Eren, MD, PhD, Andriy Derkach, PhD2, Mingfei Yan, MD, PhD3, Rakshit Tyagi, MD4, Zachary Barbati, MD5, Razie Amraei, MD, PhD6, Elizabeth Corley, MD7, Alessandro Baldi, PhD8, Mina L Xu, MD9, Sophie X. Song, MD10, Yi Xie, MD, PhD11, Hammad Tashkandi, MD12, Torsten Haferlach, MD, PhD13, Xueyan Chen, MD14, Wenbin Xiao, MD, PhD15
AFFILIATIONS
1 Memorial Sloan Kettering Cancer Center, New York, New York, United States
2 Memorial Sloan Kettering Cancer Center, New York, New York, United States
3 Moffitt Cancer Center, Tampa, Florida, United States
4 UCLA, Los Angeles, California, United States
5 UCSF Medical Center, San Francisco, California, United States
6 Yale University, New Haven, Connecticut, United States
7 Memorial Sloan Kettering Cancer Center, New York, New York, United States
8 MLL Munich Leukemia Laboratory, München, Germany
9 Yale University, New Haven, Connecticut, United States
10 UCLA, Los Angeles, California, United States
11 UCSF Medical Center, San Francisco, California, United States
12 Moffitt Cancer Center, Tampa, Florida, United States
13 MLL Munich Leukemia Laboratory, München, Germany
14 Fred Hutchinson Cancer Research Center, Sunnyside, Washington, United States
15 Memorial Sloan Kettering Cancer Center, New York, New York, United States
ABSTRACT
Background: Mixed-phenotype acute leukemia (MPAL) poses diagnostic and management challenges. Although a mixed-phenotype immunoprofile is a defining feature of MPAL, it also occurs in other entities, including chronic myeloid leukemia in blast phase, myeloid/lymphoid neoplasm with eosinophilia and tyrosine kinase fusion (MLN-TK), and acute myeloid leukemia (AML) with fusions such as RUNX1::RUNX1T1, recognition of which changes treatment. We recently identified a subset of AML with a history of MDS or MDS/MPN, MDS-defining cytogenetic abnormalities, and/or prior cytotoxic therapy that presents with a mixed-phenotype immunoprofile, termed AML-MP, which has inferior response to lymphoid-based regimens and worse outcomes than MPAL. AML-MP significantly overlaps with AML with myelodysplasia-related (MR) gene mutations (AML-MR). Whether MR gene mutations alone (in the absence of such history and MDS-defining cytogenetic abnormalities) warrant reclassifying MPAL as AML-MP remains unclear.
Methods: In this multicenter retrospective study, we identified patients from 7 academic medical centers meeting the following criteria: newly diagnosed acute leukemia (≥20% blasts in BM and/or PB), a mixed-phenotype immunoprofile confirmed by flow cytometry per WHO classification, and targeted next-generation sequencing covering common myeloid driver genes.
Results: In total, 376 patients were identified; 20 were excluded for CML in blast phase, AML with fusions (PML::RARA, RUNX1::RUNX1T1, MYH11::CBFB, MECOM), or MLN-TK. The remaining 356 patients (female/male 160 [44.9%]/196 [55.1%]; median age 59 years, IQR 40-70) were included: 199 (56%) B/myeloid, 143 (40%) T/myeloid, 8 (2.2%) B/T/myeloid, and 6 (1.7%) B/T. 81 patients (22.7%) had progressed from MDS or MDS/MPN, 43 (12.1%) were therapy-related, and 232 (65.2%) were de novo.
Among all 356 patients, 62 (17.4%) had BCR::ABL1 or KMT2A rearrangements, 9 (2.5%) had rearrangements involving 14q32 (BCL11B locus), and 123 (34.6%) had MDS-defining cytogenetic abnormalities. In addition, 79 (22.2%) harbored MR gene mutations, 75 (21.1%) had RUNX1 mutations, and 57 (16.0%) had TP53 mutations. Treatment was lymphoid-based in 142 (39.9%), myeloid-based in 147 (41.3%), and hybrid in 27 (7.6%); by backbone, 120 (33.7%) received 7+3, 148 (41.6%) ALL-type induction, 26 (7.3%) AZA/VEN, and 23 (6.5%) palliative care.
On univariate analysis of the full cohort, age, prior history, treatment regimen, MDS-defining cytogenetic abnormalities, TP53 mutations, and MR gene mutations were each associated with inferior OS. On multivariable analysis, only age, treatment regimen (AZA/VEN HR 2.0 vs 7+3, p=0.005; ALL HR 0.6 vs 7+3, p<0.001), prior history (MDS/MDS-MPN HR 1.8; therapy-related HR 2.3; p=0.005 vs no prior history), and MDS-defining cytogenetic abnormalities (HR 1.3, p=0.04) remained independent predictors; MR and RUNX1 mutations did not. Median OS was 15 months for MDS/MDS-MPN history and 16 months for therapy-related disease, versus 95 months for de novo (p<0.001).
To determine whether MR and/or RUNX1 mutations independently affect outcome, we restricted analysis to a subgroup without MDS/MDS-MPN or therapy-related history, MDS-defining cytogenetic abnormalities, or an MPAL-defining fusion (n=129; female/male 56 [43.4%]/73 [56.6%]; median age 52 years): 58 (45.0%) B/myeloid, 65 (50.4%) T/myeloid, 4 (3.1%) B/T, and 2 (1.6%) B/T/myeloid. Fifteen patients (11.6%) had MR gene mutations and 26 (20.2%) had RUNX1 mutations. Treatment included 7+3 (n=48, 37.2%), ALL-type induction (n=61, 47.3%), AZA/VEN (n=7, 5.4%), hybrid regimen (n=4, 3.1%), and palliative care (n=9, 7.0%). On both univariate and multivariable analysis, only age and induction regimen (AZA/VEN HR 7.8 vs ALL, p=0.002) were associated with inferior OS; median OS did not differ by MR mutation, RUNX1 mutation, or co-mutation status.
Conclusion: This multicenter study shows that among patients with mixed-phenotype leukemia, a prior history of MDS/MDS-MPN or therapy-related disease, and MDS-defining cytogenetic abnormalities, are strongly associated with survival, supporting classification of such patients as AML-MP. In patients without this history or these cytogenetic abnormalities, MR and/or RUNX1 gene mutations alone did not correlate with worse outcome, arguing against reclassifying MPAL as AML-MP based on gene mutations alone. This finding is based on a limited number of MR-mutation only patients (n=15) and warrants confirmation in larger cohorts.
DISCLOSURES
Ozgur Can Eren, MD, PhD: No relevant conflicts of interest to declare.