Venetoclax and azacytidine in childhood primary advanced myelodysplastic syndromes, refractory/relapsed Acute Myeloid Leukemia and therapy-related myeloid diseases

Abstract

Abstract Introduction. Refractory or relapsed acute myeloid leukemia (r/r-AML), myelodysplastic syndromes with excess blasts (MDS-EB), and therapy-related MDS/AML (t-MDS/AML) in children represent an unmet medical need, due to resistance to conventional therapies and high rate of treatment failure. While venetoclax combined with azacytidine (ven/aza) has reshaped the treatment landscape for adult myeloid malignancies, pediatric data are limited and derived from heterogeneous cohorts and treatment settings. Methods. This AIEOP retrospective multicenter study evaluated 50 patients (M/F = 1.8/1) with r/r-AML (n=32), primary MDS-EB (n=10), or t-MDS/AML (n=8) treated with ven/aza, consisting of venetoclax 360 mg/m²/day orally (days 1-14 to 28) and azacytidine 75 mg/m²/day intravenously (days 1-7). Response was assessed by bone marrow (BM) evaluation and classified as complete response (CR; BM blasts ≤5%), partial response (PR; 6-20% blasts or >50% decrease), or non-response (NR; blasts increase/persistence ≥20% or <50% reduction). Adverse events (AEs) were graded using CTCAE v5.0. Disease burden at ven/aza initiation was categorized as low (BM blasts ≤5%) or high (>5%). Results. Median age at the ven/aza start was 11 years (range 1-20). Patients received a median of 2 ven/aza cycles (range: 1–7). Forty-six patients had received previous cytotoxic therapy lines (median 3, range 1-8) and 22 a previous allogeneic hematopoietic cell transplantation (HCT). No treatment-related deaths occurred. Grade ≥3 AEs were recorded in 17 patients (34.0%), including febrile neutropenia (n=7), fungal infection (n=3), and prolonged pancytopenia (n=7), requiring venetoclax interruption in 9 patients. Best overall response achieved was CR in 29 patients (58%) [15 FCM-MRD and 7 PCR-MRD-negative], PR in 9 (18%), and NR in 12 (24%). Notably, all 25 patients who achieved CR already after cycle 1 maintained their response after cycle 2. Among 13 patients with PR after cycle 1, 4 maintained PR, 4 converted to CR, and 2 progressed to NR (3 received only 1 cycle). Regarding specific subgroups, among 12 patients with KMT2A-rearranged AML, 7 achieved CR, 2 PR and 3 NR. Two of these NR cases were rescued with revumenib achieving CR. Of 6 patients with FLT3-ITD AML, 4 received FLT3-inhibitors combined with ven/aza, achieving 2 CR and 2 PR. A cytofluorimetric CR was also reported in 2 DEK-NUP214 AML patients and one CBFA2T3-GLIS2 AML patient who receives ven/aza with STRO002. In the MDS-EB cohort (n=10), 7 achieved cytofluorimetric CR and 2 PR (defined as a reduction in BM blasts from 15% to 7% and a stable disease (<5% blasts), respectively). In the t-MDS/AML cohort (n=8), 7 patients reached a response (5 CR, 2 PR). Among the 6 AML patients with extramedullary disease 3 CR (defined as negativity at PET-FDG, CT scan and biopsy) and 3 PR (defined as reduction or stability of lesions at PET-FDG) were reported. Median blast percentage before treatment was 4% (range 0-92%) with a low and high disease burden in 29 and 21 patients, respectively. CR rates were comparable in patients with low (58.1%) and high (57.1%) disease burden. Thirty-two patients (64%) were bridged to HCT, after a median of 97 days (range: 33 – 933) from ven/aza start. Median follow-up was 366 days (range 18-1309) from ven/aza start and 342 days (range 39-1215) from HCT. The 1-year event-free survival (EFS) from the start of therapy was 63.0% (CI: 55.8 – 70.2) for the whole cohort, 83.2% (CI: 76.3 – 90.1) for transplanted patients and 75.9% (CI 67.2-84.6).) for patients achieving CR. Achieving CR and being transplanted were significantly associated with better EFS in Cox regression analysis (p=0.031 and p<.0001, respectively). EFS in MDS-EB and t-MDS/AML were 88.9% (CI: 78.4 – 99.4) and 75% (CI: 59.7 – 90.3). Causes of death included 13 disease progression before HCT, 4 relapses after HCT, and three transplant-related complications. Conclusions. . The study shows good efficacy of ven/aza in pediatric high-risk myeloid diseases with better outcome in patients bridged to HCT and a manageable toxicity. Our cohort is the largest reported to date in the difficult-to-treat subgroups of pediatric MDS-EB and t-MDS/AML, supporting future prospective protocols using ven/aza. Further biological characterization of specific genetic subgroups, such as KMT2A-rearranged malignancies, may help refine the role of Bcl-2 inhibition and improve the prediction of treatment response.