2026 ASH®
Annual Meeting & Exposition December 12-15, 2026 ▪ New Orleans, Louisiana

Paper Details

Abstracts, authors, affiliations, and disclosures.

Sunday, December 13, 2026
09:30 a.m. - 11:00 a.m.
MCCNO - Great Hall BC
Technology and Procedures, Assays, Molecular Testing, Omics Technologies, Bioinformatics, Computational Biology, Emerging Technologies, Profiling, Treatment Considerations, Drug Development, Artificial Intelligence (AI), Machine Learning, Biological Processes, Molecular Biology, Diseases, Lymphoid Malignancies, Lymphoid Leukemias, Myeloid Malignancies, Acute Myeloid Malignancies, Measurable Residual Disease, ALL, Research, AML, APL, Translational Research, Clinical Research, Real-World Evidence, Study Population, Human, Adult
PAPER

499 Real-time epigenomic classification of acute leukemias demonstrates feasibility and concordance in clinical practice

Evan C. Chen, MD, Maria Reyes Reyes Capilla-Guerra, MSc2, Alexia C. Ferreira3, Kate Marinchenv4, Yael Flamand, MS5, Salvatore Benfatto, PhD6, Til Steinicke7, Jonathan H. Young, MD, PhD8, Frederick D Tsai, MD, PhD9, Benson M. George, MD, PhD10, Shai Shimony, MD, MPH11, Virginia O Volpe, MD12, Rahul Vedula, MD13, Jacqueline S. Garcia, MD14, Marlise Rachael Luskin, MD, MSCE15, Eric Stephen Winer, MD16, Martha Wadleigh, MD17, Richard Stone, MD18, Daniel J DeAngelo, MD, PhD19, Volker Hovestadt, PhD20, Gabriel Griffin, MD21
1 Dana-Farber Cancer Institute, Boston, Massachusetts, United States
2 Dana-Farber Cancer Institute, Boston, Massachusetts, United States
3 Dana-Farber Cancer Institute, Boston, Massachusetts, United States
4 Dana-Farber Cancer Institute, Boston, Massachusetts, United States
5 Dana-Farber Cancer Institute, Boston, Massachusetts, United States
6 Dana-Farber Cancer Institute, Boston, Massachusetts, United States
7 Dana-Farber Cancer Institute, Boston, Massachusetts, United States
8 Dana-Farber Cancer Institute, Boston, Massachusetts, United States
9 Dana-Farber Cancer Institute, Boston, Massachusetts, United States
10 Dana-Farber Cancer Institute, Boston, Massachusetts, United States
11 Dana-Farber Cancer Institute, Boston, Massachusetts, United States
12 Dana-Farber Cancer Institute, Boston, Massachusetts, United States
13 Dana-Farber Cancer Institute, Boston, Massachusetts, United States
14 Dana-Farber Cancer Institute, Boston, Massachusetts, United States
15 Dana-Farber Cancer Institute, Boston, Massachusetts, United States
16 Dana-Farber Cancer Institute, Boston, Massachusetts, United States
17 Dana-Farber Cancer Institute, Boston, Massachusetts, United States
18 Dana-Farber Cancer Institute, Boston, Massachusetts, United States
19 Dana-Farber Cancer Institute, Boston, Massachusetts, United States
20 Dana-Farber Cancer Institute, Boston, Massachusetts, United States
21 Dana-Farber Cancer Institute, Boston, Massachusetts, United States

Introduction: Newly-diagnosed acute leukemias, such as acute myeloid (AML) and acute lymphoblastic leukemia (ALL), require comprehensive diagnostic testing for initial treatment selection. Current standard-of-care (SOC) testing is complex, requiring modalities such as next-generation sequencing (NGS) that take multiple days to result and may not be broadly accessible. We previously developed MARLIN—a neural network-guided, DNA methylation-based classifier of acute leukemia—that can generate leukemia classifications in real-time, typically within 2-3 hours, and as a single assay (Steinicke*, Benfatto* et al, Nature Genetics 2025). Here, we report initial results of a prospective, non-interventional study testing the feasibility of real-time epigenomic acute leukemia classification using MARLIN in clinical practice, and its concordance with SOC methodologies.

Methods: Eligible patients were aged ≥18 years with clinical suspicion of new acute leukemia. Bone marrow aspirate samples with sufficient DNA (≥1µg) were sequenced by nanopore and underwent DNA methylation-based classification with MARLIN. Samples with adequate blasts (≥20%) were retained for analysis. Co-primary endpoint 1 was MARLIN’s feasibility to produce a leukemia classification within 5 weekdays of sample receipt (the typical timeframe for SOC testing) and at a pre-defined confidence threshold of ≥0.8. Co-primary endpoint 2 was the concordance of confident MARLIN predictions with SOC testing in terms of (i) blast lineage (AML, B-ALL, T-ALL, and “other”) and (ii) specific genetic subtypes of AML (CEBPA-mutated [mut], core binding factor-mut, PML::RARa, NPM1-mut, KMT2A-rearranged [r], IDH1/2-mut, TP53-mut/aneuploidy, and Chromatin/spliceosome-mut) and B-ALL (Philadelphia chromosome [Ph]-positive, Ph-like, and KMT2Ar).

Results: A total of 35 consecutive patients were enrolled, of whom 26 had aspirate samples with sufficient DNA and blasts to be evaluable. Of these, 21/26 (81%) received a high-confidence MARLIN prediction at a median of 2.6 hours (range 2.2-5.4) from time of sample receipt. In comparison, SOC testing required a median of 1 day (range 0.5-3) for flow cytometry, 5 (range 3-14) for in-house myeloid NGS panel testing, and 6 (range 2-36) for in-house cytogenetics.

Of the 21 patients who received a confident MARLIN prediction, 21/21 (100%) showed lineage concordance (18 AML, 3 B-ALL) and 19/21 (90%) showed genotype concordance with SOC testing (4 NPM1-mut, 3 KMT2Ar, 3 TP53-mut, 3 IDH1/2-mut, 2 CEBPA-mut, 2 AML with myelodysplasia-related changes, 1 PML::RARa, 1 NUP98r). In 1 concordant case of acute promyelocytic leukemia, for which rapid diagnosis is crucial, MARLIN produced a confident PML::RARa classification in 2.7 hours from sample receipt. In 3 concordant cases, suggestive MARLIN classifications prompted additional SOC testing that confirmed cryptic KMT2Ar (KMT2A::AFF1), NUP98r (NUP98::NSD1), and CEBPA mutations.

Of the 2 cases with genotype discordance, the first was an AML with biallelic CEPBA-mut (but not bZIP in-frame indels) that MARLIN classified within the aneuploidy class, likely reflecting co-occurring KRAS-mut, BCOR deletion, and complex karyotype. The second case was an AML with complex karyotype, TET2-mut, WT1-mut, and very low variant allele fraction TP53 alteration (<5%) for which MARLIN’s real-time prediction showed atypical vacillation that ultimately resulted in the CEBPA methylation class at a borderline confidence threshold. NGS testing and additional nanopore DNA sequencing of this case did not reveal any genetic alterations in the gene body or regulatory elements of CEBPA, nor evidence of CEBPA promoter hypermethylation. Thus, a CEBPA alteration could not be identified in this unusual case.

Conclusion: In a prospective study, rapid epigenomic classification of acute leukemias using MARLIN was feasible in clinical practice within a median of 2.6 hours from the time of bone marrow aspirate receipt, with high lineage and genotype concordance with SOC testing (100% and 90%, respectively). In addition to its speed and one-assay simplicity, nanopore coupled with MARLIN provided insights into disease biology such as by indicating the presence of cryptic genomic alterations. In an era of increasing AML treatment complexity, real-time DNA methylation-based classification has significant potential to complement SOC testing and improve patient care. Further validation across molecular subtypes of acute leukemias is ongoing.

Evan C. Chen, MD: Syndax - Research Funding; Arcellx - Research Funding; Takeda - Research Funding; AbbVie - Research Funding; Merck & Co., Inc. - Consultancy (Includes expert testimony); BMS - Consultancy (Includes expert testimony); Syndax - Consultancy (Includes expert testimony); Autolus - Consultancy (Includes expert testimony)