Clinical responses in pediatric patients with relapsed/refractory leukemia treated with azacitidine and venetoclax

Despite optimization and escalation of risk-based chemotherapy regimens, children with relapsed or chemotherapyrefractory acute leukemias, particularly acute myeloid leukemia (AML), remain difficult to cure.^1,2 Traditional intensive cytotoxic chemotherapy salvage regimens require extended inpatient hospitalization due to infectious risk during severe myelosuppression and are accompanied by therapy-related morbidity and deleterious impact upon quality-of-life.^3,4 The combination of the hypomethylating agent azacitidine with venetoclax, an oral selective BH3-mimetic inhibitor of the anti-apoptotic protein B-cell lymphoma 2 (BCL-2), is effective at remission induction in elderly or intensive induction chemotherapy-ineligible adults with previously-untreated AML and is Food and Drug Administration-approved for this indication.^5,6 The recent VENAML phase I clinical trial demonstrated safety and activity of venetoclax combined with idarubicin and/or high-dose cytarabine in children and adolescents/young adults (AYAs) with relapsed refractory AML (clinicaltrials gov. Identifier: NCT03194932).⁷ However, clinical experience with the azacitidine/venetoclax regimen in children has been limited to small case series.^8,9 Herein, we report clinical characteristics and outcomes of 37 pediatric patients with relapsed/refractory acute leukemias treated at our institution with commercially-available azacitidine/venetoclax therapy.

We analyzed data from patients aged 0-21 years with multiply-relapsed/refractory acute lymphoblastic leukemia (ALL), AML, or mixed-phenotype acute leukemia (MPAL) treated with azacitidine/venetoclax without or with the CD33 antibody-drug conjugate gemtuzumab ozogamicin (GO) at the Children’s Hospital of Philadelphia from January 1^st, 2018 to March 31^st, 2022. Institutional Review Board exemption was obtained for retrospective chart review. Clinical data were abstracted from electronic medical records on baseline patient demographics, leukemia-associated immunophenotyping and genetic characteristics, therapy administration, clinical response, and post-azacitidine/venetoclax outcomes with clinical follow-up through June 30, 2022.

Azacitidine 100 mg/m² daily was given intravenously on days 1-5 of each 28-day cycle. Venetoclax was given once daily as oral tablets (swallowed intact or crushed) with 3-day ‘ramp-up’ dosing during cycle 1 to minimize tumor lysis syndrome (TLS) risk with body surface area-adjusted adult exposure-equivalent dosing (AED) of 200 mg (day 1), 400 mg (day 2), and 800 mg (days 3-28).¹⁰ Subsequent cycles utilized venetoclax at full 800 mg AED for 28 days without ramp-up. Patients receiving concurrent moderate or strong CYP3A inhibitors, including azole-class anti-fungal mediations, received a 50% dose reduction of venetoclax.¹⁰ Some patients with AML treated with azacitidine/venetoclax also received GO 3 mg/m²/dose during cycle 1 usually given on days 4, 5, or 8.

Of the 37 pediatric patients with relapsed/refractory acute leukemias treated with azacitidine/venetoclax-based therapy during the study period, 27 (73%) had AML, seven (19%) had B-ALL, one (3%) had T-ALL, and two (5%) had MPAL (Table 1). The majority of patients (n=29, 78%) had multiply-relapsed/refractory leukemia. Eight children (22%) had primary chemotherapy-refractory leukemia defined as never achieving bone marrow measurable residual disease (MRD) negativity by flow cytometry at thresholds of <0.01% for ALL and <0.1% for AML. Sex was evenly balanced between female and male patients, and the median age at initial leukemia diagnosis was 8 years. Many children had high-risk cytomolecular genetic alterations, including 11 patients with KMT2A-rearranged leukemia (n=3 B-ALL, n=7 AML, n=1 MPAL), four with CBFA2T3::GLIS2 acute megakaryoblastic leukemia (AMKL), one with NUP98::KDM5A AMKL, one with ETV6::EP300 AML, and one with FLT3-ITD AML (Table 2). Of note, four patients with refractory AML treated with azacitidine/venetoclax were initially diagnosed with B-ALL (n=3 KMT2A-rearranged, n=1 TCF3::ZNF384) and had relapsed with myeloid lineage switch following CD19-directed (n=3) or CD22-directed (n=1) chimeric antigen receptor T-cell immunotherapy (CAR T). Prior to initiation of azacitidine/venetoclax therapy, the median marrow MRD of patients who completed at least one 28-day cycle azacitidine/venetoclax was 10.5% (range, 0.01–91.5%). Three patients with AML and one with MPAL were in MRD-negative/low remission prior to starting azacitidine/venetoclax, which was administered as bridging therapy prior to allogeneic hematopoietic stem cell transplantation (HSCT).

Patients received up to six cycles of azacitidine/venetoclax during the data collection window (median 2 cycles, range 0-6) and were followed for a median of 4.9 months (range, 0.1-32.6) after initiation of therapy (Table 1; Figure 1A). GO was given with azacitidine/venetoclax in nine of the 27 patients with AML. Thirty-one patients (84%) completed at least one full cycle of azacitidine/venetoclax, while six children (16%) received fewer than 28 days of therapy. Early discontinuation of azacitidine/venetoclax was due either to severe TLS (n=2) or rapidly progressive disease detected in peripheral blood (n=4). Treatment-free intervals (1 week, n=2; 2 weeks, n=4; ≥3 weeks, n=2) were required for seven patients with AML and one patient with B-ALL between cycles 1 and 2 to facilitate resolution of severe neutropenia or thrombocytopenia. Three patients with cycle 2 delay had concurrent infection (n=1 viral, n=1 bacteria, n=1 unknown etiology), and two others had persistent leukemia. Two patients proceeded to cycle 2 with continued cytopenias. During azacitidine/venetoclax therapy, six patients had acute bacteremia treated successfully with antibiotics, and two patients had biopsy-proven fungal infections.

Response data at first post-azacitidine/venetoclax marrow assessment were available for 29 patients. Median MRD level was 0.5% (range, 0.01-90%), and 14 patients (n=12 AML, n=1 B-ALL, n=1 MPAL) achieved a complete response (CR) with MRD-negative remission (38%, 14 of 37 treated patients). Of the 25 patients with AML who completed cycle 1, 12 (48%) were in MRD-negative remission, including several children with high pretreatment disease burden (Table 2; Figure 1B). Eight of these patients remained in MRD-negative remission with further cycles of azacitidine/venetoclax, while three treated with palliative intent experienced disease progression after cycle 2 (n=2) or cycle 4 (n=1). Children with AML who received azacitidine/venetoclax with GO (n=9) had a similar CR rate (4/9, 44%) to those treated without GO. One patient with B-ALL and one with MPAL also achieved MRD-negative CR (Figure 1C). Of note, six of the 14 azacitidine/venetoclax-induced CR occurred in patients with primary chemorefractory leukemia (n=5 AML, n=1 B-ALL). Three patients with residual leukemia (n=2 AML, n=1 MPAL) after cycle 1 continued azacitidine/venetoclax therapy, and the one with MPAL achieved MRD-negative CR after cycle 2.

Importantly, responses occurred in patients with high-risk leukemia genetics. Eight of 11 patients with relapsed/refractory KMT2A-rearranged leukemias (n=2 B-ALL, n=5 AML, n=1 MPAL) completed cycle 1 of azacitidine/venetoclax, and two patients achieved MRD-negative CR (n=1 infant B-ALL, n=1 AML). Patients with ALL-to-AML lineage switch relapse (n=4) had particularly aggressive leukemias with two children unable to complete cycle 1 due to disease progression and the other two experiencing persistent MRD after cycle 1. Strikingly, three of four children with CBFA2T3::GLIS2 AML achieved MRD-negative remission. Successful remission induction with azacitidine/venetoclax +/- GO was surprisingly HSCT-enabling for 11 patients with AML and one patient with MPAL. HSCT complications in these children included graftversus-host-disease of skin (n=3, grade 2-3 acute), liver (n=1, grade 1 acute), or lung (n=1, grade 3 chronic). Post-transplant death occurred in five of 12 patients (n=3 relapse, n=2 infection in leukemia remission) at a median of 177 days. Two patients with persistent relapsed/refractory B-ALL received subsequent autologous CD19-directed CAR T, and one patient with refractory KMT2A-rearranged AML after infant B-ALL lineage switch relapse received allogeneic CD123-directed CAR T. At last follow-up, 12 patients (32%) were alive, all of whom received at least one cycle of azacitidine/venetoclax (Figure 1A).

Table 1.Demographics, clinical characteristics, and outcomes of the study cohort treated with azacitidine and venetoclax (N=37 patients).

Table 2.Patient characteristics, leukemia-associated cytomolecular genetic alterations, and treatment outcomes following azacitidine and venetoclax therapy in study cohort (N=37).

To our knowledge, we report the largest retrospective cohort of children and adolescents/young adults with multiply-relapsed/refractory acute leukemias treated with azacitidine/venetoclax to date. Within a heavily-pretreated and/or highly chemorefractory cohort of patients, we describe an encouraging 38% MRD-negative CR rate. Our observations are similar to those of Winters and colleagues, who described MRD-negative CR in three of six (50%) pediatric patients with relapsed/refractory AML treated with azacitidine/venetoclax and a lack of response in one child with ALL-to-AML lineage switch relapse.⁸ Our data are also concordant with several retrospective case series of adults with relapsed/refractory AML in whom activity has been reported across numerous leukemia genetic backgrounds, including those with poor-risk cytogenetics.^11-13 We observed remarkable rates of MRD-negative CR with azacitidine/venetoclax therapy in our cohort of children with multiply-relapsed/refractory leukemias, including many with high-risk cytomolecular genetic features such as CBFA2T3::GLIS2 AMKL. Importantly, our pediatric patients largely received azacitidine/venetoclax therapy in the outpatient oncology clinic with inpatient admission for tumor lysis or myelosuppression monitoring not routinely required, comparable with experience in adult AML.¹⁴ As many children had spent significant time inpatient for receipt of prior frontline or salvage therapies, the ability to receive outpatient chemotherapy anecdotally provided significant quality of life benefit for patients and their families, an area that merits future formal study. Finally, our data suggest that clinical responses of children with relapsed/refractory leukemia, particularly AML, to the azacitidine/venetoclax regimen are rapid and binary with achieved MRD-negative CR typically occurring after cycle 1 therapy or not at all.

Figure 1.Clinical course and responses of pediatric patients with multiply-relapsed/refractory acute leukemia treated with azacitidine and venetoclax. (A) Swimmer plot depicting the clinical course of each patient who received ≥1 complete cycle of aza citidine and venetoclax (aza/ven; N=31) is displayed. Each colored bar (grouped by leukemia subtype) represents the time after initiation of aza/ven for a single patient. Interior horizontal thick black bars represent aza/ven treatment duration. Squares represent patients with measurable residual disease (MRD)-negative remission at first evaluation (end of cycle 1 or cycle 2). Triangles represent treatment with chimeric antigen receptor T-cell immunotherapy (CAR T), including autologous CD19-directed CAR T (B-cell acute lymphoblastic leukemia [B-ALL]) or allogeneic CD123-directed CAR T (acute myeloid leukemia [AML]). Additional treatment modalities and outcomes are denoted by the symbols in the legend. (B) Clinical response assessment of pediatric patients with relapsed/refractory AML receiving ≥1 cycle of aza/ven. Prior to the initiation of aza/ven, patients had varying AML disease burden in the bone marrow as quantified by flow cytometric analysis of measurable residual disease (MRD). Lines depict response of individual patients from pretreatment (white circles) to first evaluation (blue circles, end of cycle 1 or 2). Many patients achieved MRD-negative remission at a threshold of <0.1% (dotted grey line) at first evaluation, including several patients with high pre-aza/ven treatment disease burden. (C) Clinical response assessment of pediatric patients with relapsed/refractory B-ALL (yellow squares) or mixed phenotype acute leukemia (MPAL) (green triangles) receiving ≥1 cycle of aza/ven as in (B). Lines depict flow cytometric analysis of bone marrow MRD for individual patients from pre-treatment (white symbols) to first evaluation (yellow or green symbols, end of cycle 1 or 2). One patient with B-ALL and 1 patient with MPAL achieved MRD-negative remission at a threshold of <0.01% (dotted grey line). HSCT: hematopoietic stem cell transplantation.

Although our study is limited by its retrospective nature, a relatively small number of patients treated at a single institution, and short duration of follow-up, it highlights the robust potential of azacitidine/venetoclax as an effective salvage regimen for pediatric patients with multiply-relapsed and highly-chemorefractory acute leukemias. In our cohort, azacitidine/venetoclax was often initially administered with palliative intent, but was subsequently HSCT-enabling in many patients given their excellent response rates. Given the surprisingly high efficacy in a very chemorefractory population, favorable toxicity rate, and appreciable improvement in quality of life for patients, formal clinical trial investigation of azacitidine/venetoclax-based therapies in children with relapsed or refractory acute leukemias, ideally at an earlier stage of relapse, is warranted.

Footnotes

• Received December 23, 2022
• Accepted March 28, 2023

Correspondence

S. K. TASIAN - tasians@chop.edu

Disclosures

SKT receives/d research funding for unrelated studies from Beam Therapeutics, Incyte Corporation, and Kura Oncology, has consulted for bluebird bio, has received travel support from Amgen, and serves on scientific advisory boards of Aleta Biotherapeutics, Kura Oncology, and Syndax Pharmaceuticals. The remaining authors have no conflicts of interest to disclose.

Contributions

LMN contributed to study design, performed data collection, analyzed and interpreted data, and wrote the manuscript. PC performed data collection and analyzed and interpreted data. CD analyzed and interpreted data. SKT conceived and directed the study, interpreted data, and wrote and edited the manuscript. All authors reviewed, edited, and approved the final manuscript.

Data-sharing statement

De-identified data without identifying patient health information are available upon reasonable request from the corresponding author.

Funding

Acknowledgments

We are grateful to the physicians, advanced practice providers, and nurses of the Hematologic Malignancies Program at the Children’s Hospital of Philadelphia (CHOP) for their excellent clinical care of the patients whose data are described in this report. We also kindly acknowledge Drs Bianca Goemans and Uri Ilan at the Princess Máxima Centrum in Utrecht, the Netherlands for helpful discussion and collaboration.