Phase I/II study of decitabine/cedazuridine (ASXT727), venetoclax, and gilteritinib for patients with FLT3-mutated Acute Myeloid Leukemia or high-risk myelodysplastic syndrome

Abstract Background: Patients (pts) with FLT3-mutated acute myeloid leukemia (AML) who are ineligible for intensive chemotherapy (IC) or with relapsed/refractory (RR) disease have poor outcomes. The triplet therapy with azacitidine, venetoclax (VEN), and gilteritinib (GILT) has shown encouraging survival in FLT3-mutated AML. Here, we evaluated the safety and efficacy of an all-oral combination of ASTX727, VEN, and GILT in both RR and ND FLT3-mutated AML, high-risk myelodysplastic syndrome (MDS) or myeloproliferative neoplasm (MPN). Methods: Pts with FLT3-mutated AML or high-risk MDS/MPN were enrolled in this phase I/II study. Phase I (dose-escalation) included RR pts. Phase II (dose-expansion) consisted of two cohorts: Cohort A enrolled newly diagnosed (ND) pts who were ineligible for IC, and Cohort B included RR pts. In cycle 1, pts received ASTX727 35/100mg on days 1-5, VEN 400mg (targeted dose) on days 1-28, and GILT on days 1-28. An amendment to the protocol recommended a bone marrow (BM) on day 14, and VEN and GILT were held if blasts <5% or if insufficient/aplastic. For cycles 2 and beyond, ASTX727 was given on days 1-5, VEN on days 1-14, and GILT on days 1-28. The number of treatment days could be reduced depending on the degree of myelossupression. During phase I, GILT dose ranged from 80mg (level 1) to 120mg (level 2) daily. Results: As of 10 June 2025, 39 pts were enrolled (Phase 1: n=12; Phase II: n=15 in Cohort A and n=12 in Cohort B). Although no DLTs was observed, the recommended Phase II dose of GILT was 80mg daily, based on the safety profile. 21 pts had AML, 2 had CMML, and 1 had aCML. The median age was 71 years (67 years in RR group, and 73 years in ND group). Among RR pts, the median number of prior lines of therapy was 1 (range, 1-12); 19 pts (80%) had prior HMA, 15 (63%) had prior VEN, 6 (25%) had prior FLT3 inhibitors, and 8 (33%) had undergone HSCT. 2 ND pts received prior therapy for MDS before progression to AML. 27 pts (69%) had FLT3-ITD, 8 (21%) had FLT3-TKD, and 4 (10%) had both mutations. The median FLT3-ITD allelic ratio (AR) was 0.23 (range, 0.01-10.6), with 9/31 pts (29%) having an AR ≥0.5. Comutations in DNMT3A were observed in 19 pts (49%), NPM1 in 16 pts (41%), with both DNMT3A and NPM1 in 11 pts (28%). NRAS and/or KRAS mutations were present in 7 pts (18%). 34 pts underwent BM assessment on day 14, and 18 (53%) met criteria to hold VEN and GILT to allow for count recovery (67% in the ND and 42% in the RR group). The median time to neutrophil (≥1x10⁹/L) and platelet (≥100x10⁹/L) recovery was 42 and 35 days, respectively. All pts had a reduction in therapy duration in subsequent cycles to avoid prolonged cytopenias. The most common grade ≥3 treatment-emergent adverse events (TEAEs) were febrile neutropenia (12 patients, 31%), pneumonia (12 patients, 31%), and sepsis (12 patients, 31%). The 8-week mortality rate was 8%. In the RR group, the best composite complete remission (CRc) rate was 46%, with 4 pts (17%) achieving complete remission (CR) and 7 (29%) achieving CR with incomplete hematological recovery (CRi). An additional 6 pts (25%) achieved MLFS, for an overall response rate (ORR) of 71%. Among evaluable responders, 5/13 (38%) achieved undetectable MRD by flow, and 6/13 (46%) cleared FLT3 mutation by PCR. The median number of cycles to best response was 1 (range, 1–2), and the median number of cycles received was 2 (range, 2–8). 3 pts (13%) were bridged to HSCT. With a median follow-up of 25.9 months, the median relapse-free survival (RFS) was 11.6 months (95%CI, 7.2–NR), and median overall survival (OS) was 7.7 months (95%CI, 5.2–13.7). In the ND group, CRc as best response was achieved in 12 pts (80%), including 10 with CR (67%) and 2 (13%) with CRi. 1 pt achieved PR. The 2 non-responding pts had S-AML and prior HMA exposure; one had also received VEN. Among responders, 9/12 (75%) achieved MRD negativity by flow, and all 7 with available PCR data cleared FLT3 mutation. The median number of cycles to best response was 1 (range, 1–2), and the median number of cycles was 3 (range, 1–14). Five patients (33%) were bridged to HSCT. Longer follow-up is needed to evaluate survival outcomes. Among the 10 pts who experienced relapse (7 in the RR group and 3 in the ND group), 3 (30%) were FLT3-negative at the time of relapse (2 in the RR group and 1 in the ND group). Conclusion: The combination of ASTX727, VEN and GILT is feasible but associated with myelosuppression, necessitating dose reductions.