Clinical Application for Modification of the 2022 European Leukemianet Genetic Risk Stratification in Pediatric Acute Myeloid Leukemia

The application of the 2022 European LeukemiaNet (ELN) genetic risk stratification has been validated for the prognostic significance in several adult AML studies. However, the prevalence and significance of recurrent genetic abnormalities differ between pediatric and adult AML patients. Therefore, the prognostic significance of the 2022 ELN genetic risk stratification on pediatric AML is unclear. We aim to assess the correlation of the 2022 ELN genetic risk groups with treatment outcome, and to investigate the applicability of modification of the genetic risk on outcome prediction in pediatric patients with AML. We enrolled 199 pediatric patients consecutively diagnosed with de novo AML between 2011 and 2023 in Taiwan. Patients were treated with Taiwan Pediatric Oncology Group AML protocols (TPOG-AML-97A, 97B, and 2021). The decision of allogeneic hematopoietic stem cell transplantation in first remission was decided by the treating physician. Common recurrent genetic abnormalities in AML were screened by karyotyping and/or RT-PCR. For those without common recurrent genetic abnormalities, PCR with capillary electrophoresis or NGS were used for detecting FLT3-ITD and NPM1 mutations, other mutational analysis was performed by NGS. In the modified risk allocation for pediatric patients, those with RBM15::MKL1 were categorized as favorable-risk. Patients with wild-type NPM1 and concomitant FLT3-ITD mutation, KMT2A-PTD, and the presence of NRAS, KRAS or WT1 without favorable or adverse cytogenetic/molecular abnormalities were categorized as high-risk. The prevalence of recurrent genetic abnormalities was compared with an adult cohort of Taiwan AML study group (n=2193) during the same period. Compared to adult patients, the prevalence rates of RUNX1::RUNX1T1 (26.6% vs 5.9%, P < 0.001), MLLT3:KMT2A (7.5% vs. 1.2%, P < 0.001), other KMT2A arrangement (7.0% vs. 1.8%, P < 0.001) were significantly higher in pediatric patients. The prevalence rates of mutated NPM1 without FLT3-TID (4.0% 10.7%, P = 0.006), mutated NPM1 with FLT3-ITD (1.0% vs. 8.4%, P < 0.001), MDS-related cytogenetic changes (1.0% vs. 13.3%, P < 0.001), and MDS-related molecular changes (5.5% vs. 12.6%, P = 0.006) were significantly lower. According to the 2022 ELN risk stratification, patients were classified into favorable (n = 101, 50.8%), intermediate (n = 56, 28.1%), and adverse risk groups (n = 42, 21.1%). The 5-year event-free survival (EFS) differed significantly among genetic risk groups (P = 0.001); favorable 75.0% vs. intermediate 42.2%, (P < 0.001), favorable vs. adverse 52.5%, (P = 0.01), and intermediate vs. adverse (P = 0.4). The 5-year overall survival (OS) among genetic risk groups were also significantly different (P < 0.001); favorable 86.5% vs. intermediate 52.7% (P < 0.001), favorable vs. adverse 63.3%, P = 0.03), and intermediate vs. adverse, (P = 0.5). The survival difference mainly resulted from the superior outcome of the favorable-risk group compared to other 2 groups in which adverse-risk group exhibited a paradox of better survivals than the intermediate-risk group. Per our strategy of risk modification, 22 patients (11.1%) were reclassified to other risk groups. Consequently, patients were classified into favorable (n = 105, 52.8%), intermediate (n = 35, 17.6%), and adverse risk groups (n = 59, 29.6%). The 5-year EFS remained significantly different among modified genetic risk groups (P = 0.002); favorable vs. intermediate (P = 0.1), favorable vs. adverse (P < 0.001), intermediate vs. adverse (P = 0.3). The 5-year OS also had significant differences (P < 0.001); favorable vs. intermediate (P = 0.02), favorable vs. adverse (P < 0.001), intermediate vs. adverse (P = 0.4). Treatment outcomes were more reasonably distinguishable between the intermediate-risk group and adverse-risk group although statistical significance was not reached. Our results demonstrate that the prevalence rates of recurrent genetic abnormalities are significantly different between pediatric and adult AML patients. The 2022 ELN risk stratification can reliably assign those with favorable outcome but does not distinctly stratify intermediate from adverse risk. Risk reallocation based on our modifications had a trend of better distinguish between the two non-favorable groups, highlighting the unmet need of development of a new risk stratification for pediatric AML patients.