A phase II study of IDH1 inhibition with ivosidenib as maintenance therapy after ivosidenib, azacitidine, and venetoclax for Acute Myeloid Leukemia: Trial in progress Free

Background and Significance: Approximately 10% of cases of acute myeloid leukemia (AML) feature an isocitrate-dehydrogenase 1 gene mutation (IDH1m). For newly diagnosed (ND) patients with IDH1m AML ineligible for intensive induction, the hypomethylating agent (HMA) azacitidine and venetoclax (AZA/VEN) is one standard of care based on the VIALE-A study. In VIALE-A, AZA/VEN resulted in a composite complete remission (CRc) rate of 79%, and 40% of patients achieved measurable residual disease (MRD) negativity, but median overall survival (mOS) was only 10.2 months, highlighting the need for better approaches (Pratz, Am J Hematol 2024). Additionally, real-world studies showed frequent cytopenias, dose interruptions, and hospitalizations (Chin, Leuk Lymphoma 2025). Ivosidenib (IVO), a selective IDH1 inhibitor, has been combined with AZA for ND IDH1m AML, demonstrating a mOS of 29.3 months and reduced myelosuppression (Montesinos Blood 2022, Blood Advances 2025). However, the CRc rate was 54%, MRD negative rate was 30%, and 12-month event free survival (EFS) was 37%.

Early-phase studies have explored adding IVO to HMA/VEN for ND AML patients, achieving deep responses (CRc 86%, 81% MRD negative) and promising survival (72% 2-year EFS; DiNardo, JCO 2025). However, this triplet regimen often required prolonged cycles (cycle 1 range 27-79 days) and HMA/VEN durations were frequently reduced at physician's discretion (e.g. median 7 days of VEN by cycle 3). Grade ≥3 infections occurred in 37% of patients. These studies have included more fit patients, with 34% proceeding to allogeneic stem cell transplant.

In clinical practice, managing triplet therapy in less-fit, transplant-ineligible patients is challenging. Considering the cytopenias experienced with real-world HMA/VEN, the indefinite addition of a third agent can increase risks of prolonged cytopenia, infections, and transfusion dependence, especially without structured dose de-escalation. There remains an unmet need for strategies that achieve the deep remissions associated with triplet therapy while minimizing long-term toxicity.

Study Design and Methods: This study will be a multicenter, phase 2, investigator-initiated study evaluating IVO maintenance following induction with IVO/AZA/VEN in ND induction-ineligible, non-transplant AML patients with IDH1m. In this less-fit population, we aim to achieve rapid, deep remissions while minimizing long-term myelosuppression and toxicity. We hypothesize that patients will achieve superior EFS compared to historical controls.

Key eligibility criteria include age ≥60 years, ND IDH1m AML, and ineligibility for intensive induction chemotherapy due to age, comorbidities, or treating physician discretion. Patients must not be planned for allogeneic hematopoietic stem cell transplantation in first remission. Major exclusion criteria include prior therapy with IVO or VEN.

Patients will receive AZA 75 mg/m² (IV/SC per institutional preference, days 1-7), VEN 400 mg PO daily (days 1-14), and IVO PO 500 mg daily (days 15-28 for cycle 1, then days 1-28 in subsequent cycles). Patients that achieve CR, CRi, or MLFS after cycle 1, will transition immediately and indefinitely to IVO monotherapy until morphologic relapse, intolerable toxicity, or death. Patients that do not achieve CR/CRi/MLFS continue triplet therapy for maximum of 2 more cycles and can transition to IVO monotherapy if they achieve CR/CRi/MLFS at any time. Transition to maintenance is independent of MRD status, though serial MRD assessments will be performed.

The trial will enroll 45 patients across 4 U.S. centers. Bone marrow assessments and MRD testing will occur every 3 months for the first year of IVO maintenance and then every 6 months up to 2 years.

The primary endpoint is 12-month EFS from IVO/AZA/VEN initiation. Based on historical controls of less-fit subgroups (including AGILE, VIALE-A, and MSKCC data), we define an unpromising 12-month EFS rate as 30% and a promising rate as 50% for our cohort. Using a one-sided binomial test (α=0.1), the trial is powered at 90% to detect this difference. Secondary endpoints include CRc rate, MRD negativity, overall survival, duration of remission, safety, and rates of infection, transfusion dependence, and hospitalization. MRD will be assessed by multiparameter flow cytometry. IDH1 variant allele frequency will be serially assessed by NGS.

The study is funded and IRB-approved. First patient accrual is anticipated in fall of 2025.