PII-046 - GENETIC VARIANTS PREDICTIVE OF FEBRILE NEUTROPENIA DURING INDUCTION I IN PEDIATRIC ACUTE MYELOID LEUKEMIA PATIENTS

S. Nittu¹, R. Marrero², X. Cao³, H. Wu⁴, R. Ribeiro⁵, J. Rubnitz⁶, S. Pounds⁷, J. Lamba⁸; ¹University of Florida, FL, USA, ²epartment of Pharmacotherapy andTranslational Research, College of Pharmacy, University of Florida, Gainesville,FL, FL, USA, ³College of Nursing Department of Health Promotion/Disease Prevention, Memphis, TN, ⁴St. Jude Children's Research Hospital Graduate School of Biomedical Sciences, Memphis, USA, ⁵Department of Oncology St. Jude Graduate School of Biomedical Sciences, Memphis, USA, ⁶St. Jude Children's Research Hospital, Memphis, USA, ⁷Member, St. Jude Faculty Director of Biostatistics Courses, St. Jude Graduate School of Biomedical, Memphis, USA, ⁸Department of Pharmacotherapy and TranslationalResearch, University of Florida, Gainesville, FL, GNV, USA.

Background: Febrile neutropenia (FN) is a common complication of chemotherapy in pediatric acute myeloid leukemia (AML), often leading to severe infections and contributing to treatment-related mortality (TRM). While clinical risk factors have been studied, the influence of genetic variants on FN severity remains unclear. This study aimed to identify single nucleotide polymorphisms (SNPs) associated with grade 3 or higher FN during induction I chemotherapy in pediatric AML patients.

Methods: Patients from the AML02 [NCT00136084, N=134] and AML08 [NCT00703820, N=176] trials with genotype and toxicity data generated using the Illumina 2.5 Omni array and exome chip were included. FN was defined as ANC < 1000/mm³ with a temperature >38.3°C or ≥38°C for more than one hour. Patients were divided into grade 3 or higher (toxicity) and grade 1 or 2 (non-toxicity) groups. After quality control, 1,132,935 variants were analyzed. Logistic regression, adjusted for ancestry, was conducted using PLINK, with genome-wide significance set at P < 5 x 10^-8 and a suggestive threshold of P < 1 x 10^-4.

Results: Of 311 patients, 234 experienced ≥ grade 3 FN, and 77 did not. Ninety-three SNPs were associated with ≥ grade 3 FN (P < 1 x 10^-4), with 65 independent SNPs across 58 genes identified after linkage disequilibrium pruning. Notably, a SNP in TFPI (OR=0.41, p=5.07 x 10^-5), a gene involved in coagulation and infection response, conferred decreased FN risk. A SNP in PLA2G4E (OR=2.39, p=3.76 x 10^-5) was linked to increased FN incidence, aligning with its role in inflammation. Additionally, a SNP in ELF2 (OR=2.78, p=9.01 x 10^-5), a gene associated with immune function, was associated with greater FN incidence.

Conclusion: Despite the limited sample size, this study is the first to identify common SNPs linked to chemotherapy-induced toxicity in pediatric AML during induction. Although no SNPs achieved genome-wide significance, several biologically relevant genes were identified, indicating their potential role in FN risk. Future validation in larger cohorts is necessary, along with multi-omics approaches, to further elucidate the genetic mechanisms contributing to chemotherapy-induced toxicity.