Molecular Features, Treatments and Outcomes for Pediatric AML Patients from APAL2020SC Pediatric Acute Leukemia (PedAL) Screening Trial

Background Survival for pediatric AML patients with relapsed (rAML) and primary refractory (prAML) disease remains low, and designing clinical trials is challenging. Randomized trials in this rare population are often not feasible and single-arm trials are limited by lack of adequate historical controls. There are few large, harmonized datasets which limits clinical trial design. APAL2020SC (NCT04726241) is an ongoing screening trial and registry for patients with relapsed/refractory acute leukemias, sponsored by the Leukemia & Lymphoma Society. APAL2020SC collects leukemic blast%, identifies cell surface targets, provides RNA/DNA sequencing, and collects treatment and outcome data for patients not enrolled on a therapeutic trial. The goals are to provide target identification for clinical trial eligibility and to establish historical control data that will inform and expedite early phase pediatric trials of novel agents. Here we present molecular features, re-induction regimens and available response data of patients with rAML and prAML. Methods Data were abstracted for 124 patients with confirmed rAML or prAML enrolled prior to Jan 2024. Central blast quantification and target identification were provided by Hematologics. Central next-gen sequencing (NGS) was provided by Foundation Medicine. Re-induction treatment and response data were entered by sites. Results The cohort included 79 rAML and 45 prAML patients. Central NGS results were available for 63 rAML (80%) and 35 prAML patients (78%). Fusions or copy number alterations (CNA) were identified in 87% of rAML and 74% of prAML patients. KMT2A fusions were enriched in rAML (36.5%) compared to 17% in prAML patients; the rate in de novo patients is about 20%. KMT2A fusion partners for rAML included MLLT3 (11=48%), MLLT10 and MLLT4 (3=13% each), ELL (2=9%), and 4 others found in 1 patient each. For prAML, there were 3 MLLT3 and 3 MLLT10 partners. RUNX1::RUNX1T1 fusions occurred in 19% of rAML and 17% of prAML patients; the the de novo rate is about 12%. CBFB::MYH11 fusions were detected in 1 rAML and 1 prAML patient. NUP98::NSD1 fusions were more common in prAML at 8.6% (3/35) v. 3% for rAML. CBFA2T3::GLIS2 fusions were identified in 3 (13%) prAML patients. FLT3-ITD was detected in 8 (13%) rAML and 3 (9%) prAML patients, less than the 15% rate at diagnosis. In pediatric AML single nucleotide variants and indels are often subclonal, and persistence between diagnosis and relapse is variable. Longitudinal central NGS results at diagnosis and relapse/refractory timepoints were available for 32 rAML and 14 prAML patients, enabling us to determine which mutations were stable, lost or gained between the two timepoints. RAS pathway mutations (NRAS, KRAS, PTPN11) were most common at both timepoints. Of 22 such mutations in rAML patients at diagnosis, 15 (68%) were lost at relapse. At relapse, 10 such mutations were identified in the same patients, 7 of which were persistent and 3 new. Six FLT3-ITD variants were identified in 2 rAML patients at diagnosis, only 1 of which persisted to relapse. For WT1 mutations, 4 of 5 at diagnosis persisted and 1 was new at relapse. The pattern of losses and gains was similar for prAML patients despite the shorter time frame. At diagnosis, there were 11 RAS pathway mutations in 14 prAML patients with NGS. At the refractory timepoint, 2 were new and 6 persisted from diagnosis. One prAML patient had a FLT3-ITD at diagnosis that was not detected in the refractory sample. A key goal of APAL2020SC is to collect treatment and outcome data for patients not treated on clinical trials. Treatment and response data were provided for the first re-induction cycle for 40 rAML patients. Overall, 21/40 reportedly achieved CR/CRi/CRp after one cycle. The most common regimens were combinations including fludarabine/cytarabine (FLA); 18/40 (45%) and 12/18 (67%) achieved remission. Combinations including venetoclax also were common in cycle 1, with 7 of 16 achieving CR/CRi/CRp. In the prAML group, 10 of 23 with cycle 1 treatment and response data achieved remission: 2 with a FLA regimen, 5 with a venetoclax regimen, and 3 with CPX-351. Conclusion The APAL2020SC protocol is collecting valuable real-world treatment and response data, as well as comprehensive molecular characterization, for children with relapsed and refractory AML. These data will fill gaps in knowledge regarding this patient population and will inform the design of future therapeutic trials.