1  Changing Paradigms of Treatment Approaches for Acute Myeloid Leukemia

1.1 Session Overview

Session Details
Session Changing Paradigms of Treatment Approaches for Acute Myeloid Leukemia
Speaker Katherine Tarlock, MD
Affiliation Seattle Children’s Hospital, University of Washington, Seattle, WA, USA
Time Day 1, 9:00–9:30 a.m.

Dr. Tarlock walks through six practice-changing ASH 2025 abstracts that together illustrate how AML care is pivoting away from uniform intensive induction toward biology- and fitness-adapted regimens [slide p.1]. The narrative arc moves from chemotherapy de-escalation in fit adults (PARADIGM) and pediatric low-risk AML (AAML1831), through a retrospective allo-HCT outcomes comparison after HMA/venetoclax versus intensive induction, and then into targeted combinations anchored by the menin inhibitors revumenib (SAVE) and ziftomenib (KOMET-007) in NPM1-mutant and KMT2A-rearranged disease, closing with two novel immunotherapy platforms—CD7 universal CAR T cells and the logic-gated CD33/FLT3 NOT-EMCN CAR NK product SENTI-202 [slides p.4, p.9, p.13, p.17, p.22, p.27, p.31]. The three learning objectives frame the talk: understanding chemotherapy reductions in defined AML subsets, reviewing targeted combination strategies, and identifying novel immunotherapy targets [slide p.3].

1.2 Speaker Spotlight

Katherine Tarlock, MD is a pediatric hematologist-oncologist at Seattle Children’s Hospital and the University of Washington [slide p.1]. Her research focuses on molecularly targeted therapy for pediatric and young adult AML, with particular expertise in FLT3-mutated and KMT2A-rearranged disease. She has been a key contributor to Children’s Oncology Group (COG) AML trials—including AAML1831—and to translational studies linking genomic findings to clinical outcomes across the age spectrum.

1.3 What’s New in 2025–2026

1.3.1 Reducing Intensive Chemotherapy in Fit Adults: PARADIGM (Abstract #6)

PARADIGM is a phase 2 multicenter trial that randomized 172 newly diagnosed, induction-eligible adults with AML 1:1 to azacitidine + venetoclax (Aza-Ven) versus conventional 7+3 or CPX-351 intensive chemotherapy (IC), with allo-HCT allowed at any time after response [slide p.5]. Eligibility excluded APL, core-binding factor, FLT3-ITD/TKD (≥5% VAF), NPM1-mutant disease in patients <60 years, BCR::ABL1, and mixed-phenotype leukemias, so the enrolled population was enriched for ELN 2022 adverse-risk AML (72% adverse) with a median age of 64 years [slide p.6].

  • Efficacy: At a median follow-up of 21.9 months, median event-free survival was 14.6 months with Aza-Ven versus 6.2 months with IC, with no significant difference in overall survival [slide p.8].
  • Tolerability: Aza-Ven patients had lower grade ≥3 infection (15.1% vs 20.8%, p=0.04) and bleeding (1.3% vs 6.3%, p=0.01), zero 60-day mortality (vs 4.7% with IC), fewer hospital days in the first 6 months (41.0 vs 58.5, p<0.001), and no ICU admissions in the first 30 days (vs 9.6%, p=0.003) [slide p.7].
  • Take-home: The data support Aza-Ven as a lower-toxicity alternative to intensive induction in functionally fit adults with intermediate- or adverse-risk, FLT3-wildtype AML [slide p.8].

1.3.2 Pediatric Low-Risk AML: De-escalation with AAML1831 (Abstract #878)

The Children’s Oncology Group AAML1831 trial stratified pediatric patients by cytomolecular features and end-of-induction-1 (EOI1) MRD: Low Risk 1 (LR1) = favorable cytomolecular features with EOI1 MRD <0.05% (4 cycles), Low Risk 2 / Intermediate Risk = all others not meeting LR1 or HR (5 cycles), and High Risk = unfavorable cytomolecular features or EOI1 MRD positive without favorable features (HSCT) [slide p.10].

  • In the LR2/intermediate-risk cohort, 5 cycles of chemotherapy (AAML1831 and AAML0531 arms) produced lower relapse risk (24% and 15% vs 44%), better DFS (74% and 78% vs 54%), and superior OS (87% and 90% vs 81%) compared with the 4-cycle AAML1031 historical arms—confirming that a fifth cycle is required in this subset [slide p.11].
  • In the LR1 cohort, AAML1831 (cumulative anthracycline only 150 mg/m² doxorubicin-equivalent, Mito:Doxo 10:1) maintained excellent outcomes versus AAML1031 (630 mg/m²): relapse risk 19% vs 22% (p=0.33), DFS 81% vs 75% (p=0.11), and OS 99% vs 92% (p=0.031)—demonstrating that anthracycline exposure can be dramatically reduced in favorable-risk pediatric AML without sacrificing survival [slide p.12].

1.3.3 Allo-HCT After HMA/Venetoclax vs Intensive Induction (Abstract #44)

A retrospective Johns Hopkins analysis compared first allo-HCT outcomes in CR1 between patients induced with intensive cytarabine-based chemotherapy (n=190) versus HMA/venetoclax (n=53), all using uniform Flu/Cy/400 cGy TBI conditioning with post-transplant cyclophosphamide GVHD prophylaxis between 2018 and June 2024 [slide p.14]. Recipient/donor age was older and sAML, TP53 mutation, and pre-allo MRD were more prevalent in the HMA/Ven group.

  • Adjusted 2-year OS was 81.6% (95% CI 75.6–87.6) after intensive chemo versus 73.5% (95% CI 62.4–84.5) after HMA/Ven (p=0.17); 2-year relapse incidence (15.4% vs 15.9%, p=0.89) and NRM (9.7% vs 13.8%, p=0.33) were not different [slide p.15].
  • Subgroup analyses showed no difference in OS, RI, or NRM by age ≥64, secondary AML, or IDH-mutant status; no difference in platelet or neutrophil engraftment; and nearly identical grade 2–4 acute and chronic GVHD [slide p.16].
  • Take-home: With a uniform transplant platform, HMA/Ven induction delivers allo-HCT outcomes comparable to intensive chemotherapy, reinforcing its role as a bridge to transplant [slide p.16].

1.3.4 Menin Inhibitors: Frontline SAVE and R/R KOMET-007

The menin–KMT2A interaction remains a key dependency in KMT2A-rearranged, NPM1-mutant, and NUP98-rearranged leukemias, and Dr. Tarlock highlighted two new combination datasets that extend menin inhibitor use into both the frontline and relapsed/refractory settings [slide p.18].

SAVE (Abstract #47) is an all-oral phase I/II combination of revumenib (SNDX-5613) + ASTX727 (decitabine/cedazuridine) + venetoclax in patients ≥12 years with KMT2A-r or NPM1-mutant AML (or myeloid MPAL) who are not eligible for intensive chemotherapy or have R/R disease [slide p.18]. In 21 newly diagnosed patients:

  • ORR 86%, CR/CRh 81% (CR 76%), with identical 86% response rates in both the NPM1-mutant (n=14) and KMT2A-r (n=7) subsets [slide p.20].
  • MRD negativity by MFC (10⁻⁴) was 86% overall and 100% of responders [slide p.20].
  • At median 9 months follow-up, 12-month DOR was 71% (NPM1mt) and 80% (KMT2A-r); 12-month OS was 53% and 69% respectively [slide p.21].
  • Grade ≥3 differentiation syndrome occurred in 10% and QTc prolongation (any grade) in 43% with no grade ≥3 events; early (30-day) death rate was 10% [slide p.19].

KOMET-007 (Abstract #764) assessed ziftomenib 600 mg + venetoclax + azacitidine in 83 R/R NPM1-mutant or KMT2A-rearranged AML patients [slide p.23]:

  • NPM1-mutant (n=48): CRc 48%, ORR 65%, MRD negativity 60% [slide p.25].
  • KMT2A-rearranged (n=32): CRc 28%, ORR 41%, MRD negativity 43% [slide p.25].
  • Prior venetoclax exposure blunted responses substantially (NPM1mt CRc 70% → 28%; KMT2A-r CRc 60% → 14%), underscoring the impact of upstream BCL2-inhibitor pressure [slide p.26].
  • No ziftomenib-related QTc prolongation, only 2% discontinuations for ziftomenib-related AEs, and one grade 3 differentiation syndrome (1%) that resolved with protocol mitigation [slide p.24].
Menin Inhibitor Differentiation Syndrome

Differentiation syndrome remains a class effect of menin inhibitors. Across SAVE, grade ≥3 events occurred in ~10% of treated patients [slide p.19], and in KOMET-007 one patient had grade 3 differentiation syndrome that resolved with protocol-specified mitigation and allowed ziftomenib resumption [slide p.24]. Early recognition and prompt corticosteroids remain essential. Clinicians should also monitor QTc with revumenib-containing regimens, where any-grade QTc prolongation was reported in 43% of SAVE patients [slide p.19].

1.3.5 Novel Immunotherapy Platforms

CD7 universal CAR T cells (Abstract #816) target aberrant CD7 expression, which is seen in 10–20% of AML and absent from normal myeloid lineages [slide p.28]. The off-the-shelf product uses CRISPR/Cas9 knockout of CD7 (to prevent fratricide), TCR (GvHD), and CIITA (rejection), plus proprietary ANSWER (ANS5) anti-rejection technology [slide p.28]. In 10 heavily pretreated R/R AML patients (ages 3–70, bone marrow blasts ≥5%, CD7-positive):

  • Doses up to 1×10⁹ CAR⁺ T cells were tolerated without DLTs or treatment-related deaths; CRS occurred in 100% but all grade ≤2, ICANS in 10%, no GvHD or HLH [slide p.29].
  • Best ORR (CR/CRi) was 62.5% in patients with high CD7 expression, all MRD-negative, with median CAR-T persistence 31 days (range 7–640+) [slide p.30].
  • CD7-negative expansion was seen in most non-responders or those losing response, and hematopoietic stem cell support (transplant or boost) was crucial to sustaining remission [slide p.30].

SENTI-202 (Abstract #1044) is a first-in-class, off-the-shelf logic-gated CAR NK product with a bivalent CD33/FLT3 activating CAR, an EMCN (endomucin) NOT-gated inhibitory CAR to protect normal HSC/HSPCs, and calibrated-release IL-15 to support persistence [slide p.32]. The phase I SENTI-202-101 trial enrolled 20 R/R AML patients (median 2 prior lines, 85% prior venetoclax, 75% prior HMA, 60% refractory to most recent regimen) across 2 dose levels (1×10⁹ and 1.5×10⁹ CAR⁺ NK) and 2 schedules [slide pp.32–33]:

  • At RP2D (1.5×10⁹), ORR 50% and CR/CRh 42%; overall ORR 50% and CR/CRh 39% [slide p.34].
  • 100% of CR patients were MRD-negative by multiparametric flow (sensitivity ≤10⁻⁴) [slide p.34].
  • Grade ≥3 AEs were driven by cytopenias; CRS/infusion reactions occurred in 35% (all grade 1–2); peripheral blood expansion peaked within 14 days and cleared by ~2 weeks [slide pp.33–34].
Clinical Pearl: Match the Regimen to Biology and Fitness

Dr. Tarlock’s case for paradigm change rests on matching therapy to disease biology and patient fitness rather than defaulting to 7+3. PARADIGM shows Aza-Ven is viable even in fit adults with adverse-risk, FLT3-wildtype disease [slide p.8]; AAML1831 shows that favorable pediatric AML can be treated with dramatically less anthracycline without losing OS [slide p.12]; and SAVE shows that an all-oral menin-containing triplet can deliver 86% ORR and 86% MRD negativity in frontline KMT2A-r/NPM1-mutant disease [slide p.20].

Asia-Pacific Perspective

Access to revumenib, ziftomenib, and novel cell therapy platforms (CD7 UCAR-T, SENTI-202) remains limited across much of the Asia-Pacific region. Aza-Ven is increasingly available and—based on PARADIGM—now has randomized evidence supporting its use even in fit adults, which may be particularly relevant where ICU capacity and prolonged inpatient chemotherapy stays strain health systems [editorial]. KMT2A-rearranged and NPM1-mutant AML frequencies vary regionally, and access to rapid molecular diagnostics remains the prerequisite for deploying any menin-containing regimen [editorial].

1.4 Clinical Pearls

Five Key Takeaways
  1. Aza-Ven is now a randomized option for fit adults. PARADIGM showed median EFS 14.6 vs 6.2 months favoring Aza-Ven over 7+3/CPX-351 in fit intermediate/adverse-risk, FLT3-wildtype AML, with lower infections, bleeding, hospital days, and 60-day mortality [slide p.8].
  2. Pediatric AML needs less anthracycline—but not fewer cycles. AAML1831 proved that favorable (LR1) pediatric AML tolerates a 4-fold reduction in anthracycline exposure with superior OS, while LR2/intermediate patients still require 5 cycles [slides pp.11–12].
  3. HMA/Ven is a legitimate bridge to allo-HCT. Under a uniform PTCy-based transplant platform, post-HCT OS, relapse, NRM, and GVHD were equivalent after HMA/Ven and intensive induction [slides pp.15–16].
  4. Menin-based triplets are delivering deep frontline responses. Revumenib + decitabine/cedazuridine + venetoclax (SAVE) achieved 86% ORR, 81% CR/CRh, and 86% MRD negativity in newly diagnosed KMT2A-r and NPM1-mutant AML; ziftomenib + Ven/Aza (KOMET-007) extended activity into the R/R setting, with responses blunted by prior venetoclax exposure [slides p.20, p.25].
  5. Immunotherapy in AML is finally moving. CD7 UCAR-T produced 62.5% CR/CRi in CD7-high R/R AML, and the logic-gated CD33/FLT3 NOT-EMCN CAR NK product SENTI-202 achieved 50% ORR / 42% CR/CRh at RP2D with 100% MRD negativity in CR, providing two mechanistically distinct off-the-shelf platforms for heavily pretreated disease [slides p.30, p.34].

1.5 Key References

  1. Fathi AT, Perl AE, Fell GG, et al. Results from PARADIGM—a phase 2 randomized multi-center study comparing azacitidine and venetoclax to conventional induction chemotherapy for newly diagnosed fit adults with AML. Blood. 2025;146(Suppl 1): Abstract 6 [slide p.4].
  2. Pabari R, Glasser C, Alonzo T, et al. Optimized treatment for pediatric patients with low-risk acute myeloid leukemia: a report from the Children’s Oncology Group study AAML1831. Blood. 2025;146(Suppl 1): Abstract 878 [slide p.9].
  3. Claiborne JP, Kaduluri VPS, Tsai H-L, et al. Post-allogeneic hematopoietic cell transplantation outcomes are similar after HMA/venetoclax and cytarabine-based remission induction for AML when using uniform transplant techniques. Blood. 2025;146(Suppl 1): Abstract 44 [slide p.13].
  4. Jen W-Y, DiNardo CD, Short NJ, et al. Phase II study of the all-oral combination of revumenib with decitabine/cedazuridine and venetoclax (SAVE) in newly diagnosed AML. Blood. 2025;146(Suppl 1): Abstract 47 [slide p.17].
  5. Issa GC, Fathi AT, Zeidan AM, et al. Ziftomenib in combination with venetoclax and azacitidine in relapsed/refractory NPM1-m or KMT2A-r acute myeloid leukemia: updated phase 1a/b safety and clinical activity results from KOMET-007. Blood. 2025;146(Suppl 1): Abstract 764 [slide p.22].
  6. Hu Y-X, Zhu X, Mei H, et al. CD7-targeted universal CAR T-cell therapy in relapsed or refractory (R/R) acute myeloid leukemia (AML). Blood. 2025;146(Suppl 1): Abstract 816 [slide p.27].
  7. Farhadfar N, Strickland S, Bajel A, et al. Promising results from an ongoing phase I multicenter study of SENTI-202, a first-in-class, CD33 AND/OR FLT3 & NOT endomucin (EMCN), selective off-the-shelf logic-gated CAR NK cell therapy in adults with R/R AML. Blood. 2025;146(Suppl 1): Abstract 1044 [slide p.31].