AUTHORS
Matthew Epstein, Robert Yuan, MD2, David Cachia, MD3, Lucio Castilla, PhD4, Sakiko A Suzuki, MD5, Laurie K Pearson, MD6, Poorva Bindal, MD7, Andrew Gillis-Smith, MD8, Muthalagu Ramanathan, MD9, Jonathan Gerber, MD10, Salwa Khedr, MD11, Lloyd Hutchinson, PhD12, Jan Cerny, MD, PhD13, Shyam A Patel, MD, PhD14
AFFILIATIONS
1 UMass Chan Medical School, Worcester, Massachusetts, United States
2 UMass Chan Medical School, Worcester, Massachusetts, United States
3 UMass Chan Medical School, Worcester, Massachusetts, United States
4 UMass Chan Medical School, Worcester, Massachusetts, United States
5 UMass Chan Medical School, Worcester, Massachusetts, United States
6 UMass Chan Medical School, Worcester, Massachusetts, United States
7 UMass Chan Medical School, Worcester, Massachusetts, United States
8 UMass Chan Medical School, Worcester, Massachusetts, United States
9 UMass Chan Medical School, Worcester, Massachusetts, United States
10 UMass Chan Medical School, Worcester, Massachusetts, United States
11 UMass Chan Medical School, Worcester, Massachusetts, United States
12 UMass Chan Medical School, Worcester, Massachusetts, United States
13 UMass Chan Medical School, Worcester, Massachusetts, United States
14 UMass Chan Medical School, Worcester, Massachusetts, United States
ABSTRACT
Background: RUNX1 encodes a transcription factor that plays a master regulatory role in myeloid cell fate, and RUNX1 disruption within the hematopoietic stem/progenitor cell pool leads to impaired hematopoiesis and predisposition to myeloid neoplasm (MN). MNs harboring RUNX1 mutations are uniformly high-risk conditions. Specifically, evidence-based prognostication schemes have implicated RUNX1 mutations in high-risk clonal hematopoiesis (CH) (Weeks LD et al., 2023), high-risk myelodysplastic neoplasms (MDS) (Bernard E et al., 2022), and adverse risk acute myeloid leukemia (AML) (Dohner H et al., 2022). The mutational heterogeneity within RUNX1-mutant MNs may carry practical implications in the clinical setting, but additional study is needed to refine prognosis. We have recently shown that RUNX1 mutation in patients with CH confers the strongest risk for progression to overt MN and progression from therapy-related CH to overt therapy-related MN (Patel SA et al., 2026). To date, there are no targeted therapies effective against RUNX1-mutant cells. Here, we aim to discern the independent prognostic relevance of RUNX1 disruptions and cooperative mutations across the spectrum of RUNX1-mutant MNs.
Methods: We queried the UMass Chan Hematopathology and Molecular Pathology archives for patients with bone marrow aspirates/biopsies harboring any intragenic RUNX1 mutation between 2014 and 2026. We excluded patients with t(8;21)(RUNX1::RUNX1T1) given the distinct pathogenesis of core binding factor disruption in AML. Genomic DNA was purified from bone marrow aspirates, then a library was created using multiplex PCR targeting oncogene “hot spot” mutations and full-exon sequencing (CTMPv3 panel). PCR amplicons were isolated, followed by the ligation of a sequencing linker, and a barcode for patient identification. Real-time PCR was used to quantify the library. A sequencing template was loaded onto an IonTorrent PGM next-generation sequencer. For RUNX1 exome sequencing, the entire coding region from codons 1 through 453 was covered. Non-parametric analyses using the Kaplan-Meier method were used to discern the prognostic impact of various RUNX1 genomic aberrations and co-mutations.
Results: A total of 99 patients with intragenic RUNX1 mutations were identified: this included 45% with AML, 36% with MDS, 11% with chronic myelomonocytic leukemia (CMML), and 4% with clonal cytopenia of unknown significance (CCUS). Median age of diagnosis was 73.5 years (range 18-91). RUNX1 mutations most commonly had a VAF ≤50% (88%), with 12% exhibiting a VAF >50%. The most common intragenic RUNX1 mutations were missense mutations (40%), frameshift insertions (23%), nonsense mutations (22%), and frameshift deletions (11%). Median overall survival (OS) for these subsets was 18 months, 13.1 months, 13.3 months, and 25.7 months, respectively (log-rank p = 0.53). Stratification of patients by top-line hematopathologic diagnoses showed median OS for RUNX1-mutant AML, MDS, CMML, and CCUS of 13.4 months, 18.6 months, 19.8 months, and 18.4 months, respectively (log-rank p = 0.80). Cooperating mutations were found in 95 (96%) of 99 patients, and the most common co-mutations were SRSF2 (31%), TET2 (27%), and DNMT3A (25%). Median OS for each of these groups was 15.4 months, 16.1 months, and 16.1 months, respectively (log-rank p = 0.34). Stratification by ASXL1 co-mutation (17% of patients) demonstrated significantly inferior survival, with median OS of 7.8 versus 19.5 months for ASXL1-mutant and ASXL1-wild-type patients, respectively (log-rank p = 0.007).
Conclusions: In this real-world cohort of RUNX1-mutant MN, concurrent ASXL1 mutation imparted significantly adverse prognostic risk. The prognostic homogeneity among patients with various intragenic RUNX1 aberrations suggests that type of mutation is not a major determinant of clinical outcome. Furthermore, top-line hematopathologic diagnosis did not discern prognostically relevant subgroups, suggesting that the mere presence of RUNX1 mutation within MN confers sufficiently adverse risk irrespective of morphologic features. Downstream mechanistic studies might aim to perform single-cell analyses of patients harboring concurrent RUNX1 and ASXL1 mutations to understand the role of chromatin modifiers in RUNX1-mutant MN. The double-mutant RUNX1/ASXL1 group may represent an ultra-high-risk subset of MN, and future clinical trials may aim to develop targeted therapies for this subgroup.
DISCLOSURES
Matthew Epstein: No relevant conflicts of interest to declare.