Phase II Study of the all-oral combination of revumenib (SNDX-5613) with decitabine/cedazuridine (ASTX727) and venetoclax (SAVE) in newly diagnosed AML

Abstract

Abstract Background The menin–KMT2A interaction is a dependency in acute leukemia caused by NPM1 mutations (NPM1m) or rearrangements of KMT2A (KMT2Ar) or nucleoporin 98 (NUP98r). The combination of the menin inhibitor revumenib and venetoclax is synergistic in preclinical models of these subtypes. NPM1m, KMT2Ar and NUP98r are observed in 30%, 5% and 1% of newly diagnosed (ND) AML, respectively. We designed a phase 2 study to evaluate the all-oral combination of decitabine/cedazuridine, venetoclax and revumenib (SAVE, NCT05360160) in these AML subtypes. Methods Pts aged ≥12 years with ND NPM1m, KMT2Ar or NUP98r AML or mixed-lineage acute leukemia (MPAL) who were not candidates for high intensity chemotherapy were eligible. Decitabine/cedazuridine was administered at 35mg/100mg QD D1-5, venetoclax at 400mg (target dose with ramp up) QD D1-14, and revumenib at 270mg (without strong CYP3A inhibitors) or 160mg (with strong CYP3A inhibitors) Q12h D1-28. All pts underwent bone marrow (BM) assessment on C1D14, in addition to end of cycle assessment, with revumenib held after D21 if BM blasts were <5%. Revumenib monotherapy following hematopoietic stem cell transplant (SCT) was resumed for 1-year maintenance. Measurable residual disease (MRD) was assessed by flow cytometry (FC, sensitivity 10-4) in all pts. Pts with NPM1m also had bidirectional paired-end NGS (sensitivity 5×10-5, Invivoscribe, USA), while those with KMT2Ar had a qPCR assay (sensitivity 5 copies/1µg RNA, Pupil Bio, USA). MEN1 mutation testing using ddPCR was performed at relapse. Results At data cut on 7/22/25, 17 pts have been treated, receiving a median of 3 (range, 1-10) treatment cycles. Eleven pts (65%) had NPM1m while 6 (35%) had KMT2Ar, including 1 pt with NUP98::KMT2A. The median age was 68 years (range, 60-83). 24% had secondary or therapy-related AML. By ELN2022 criteria, 53% had favorable, 12% intermediate and 35% adverse risk disease. Cytogenetics were complex in 12%. Signaling (FLT3, KIT, KRAS, NRAS, KIT) co-mutations occurred in 73% and 50% of the NPM1m and KMT2Ar groups, respectively. Myelodysplasia-related mutations occurred in 36% of NPM1m. One pt had a FLT3-ITD mutation with a low allele ratio of 0.03. The overall response rate (CR + MLFS) was 94% (16/17 pts, 95% CI, 73-99). The single non-responder was a KMT2Ar AML transformed from pre-treated myelofibrosis and withdrew consent on D3 of treatment to proceed to hospice. One pt with idiopathic pulmonary fibrosis died of diffuse alveolar hemorrhage during the third week of treatment, after a D14 BM showing MLFS. One pt has not yet reached end of cycle 1 (EOC1) but day 14 marrow demonstrated MLFS. Among evaluable pts, the CR rate was 88% (14/16 pts, 95% CI, 59-94), with MRD negative rate by FC of 100%. All KMT2Ar were also negative by FISH. NGS MRD results were available for 9/9 NPM1m pts with CR at data cut. The NPM1m MRD negative rate by NGS was 44% (4/9 pts). Six pts were evaluable for NPM1m NGS MRD at the EOC2 timepoint, with 4 (67%) attaining undetectable MRD, 1 with MRD above the limit of detection but below the 5×10-5limit of quantification (underwent SCT), and 1 with MRD persistence who relapsed 10 months after the start of therapy. At a median follow-up of 6 months (range, 1-14), the median OS and EFS were not reached. SCT has been performed in 5 (29%) pts; 2 (18%) NPM1m and 3 (50%) KMT2Ar. Relapse occurred in 2 pts (1 NPM1m and 1 KMT2Ar). Neither underwent SCT in CR1 and both had detectable MEN1 M327I mutations at relapse. Results of correlative studies will be presented at the meeting. The median cycle length was 41 days (range, 31-54) for C1 and 45 days (range, 30-64) for C2. The most common adverse event was infection in 9 pts (53%), all grade 3. QTc prolongation occurred in 8 (47%) pts; 3 were grade 2 (18%), the rest were grade 1 (29%). Differentiation syndrome (DS) occurred in 4 (24%) pts; 2 pts had grade 3 DS which promptly resolved with steroids. Two pts resumed revumenib as maintenance post-HSCT, 1 had detectable FLT3-ITD and NPM1m by NGS MRD at day 60 post-SCT which became undetectable without molecular recurrence after resuming revumenib maintenance (currently 9 monthspost-SCT). Conclusion SAVE, an all-oral combination, shows promising activity in older adults with ND NPM1m or KMT2Ar AML who are ineligible for intensive chemotherapy. Ongoing enrollment and longer follow-up are required to establish the durability of response.