Cytarabine Pharmacogenomics and Outcomes Among Children and Young Adults With Acute Myeloid Leukemia

Importance Therapeutic responses in acute myeloid leukemia (AML) demonstrate considerable variability both across and within established risk stratifications and age groups. Moreover, significant racial disparities persist, with Black patients experiencing inferior survival outcomes compared with their White counterparts. Objective To validate the association of the previously reported 10 single nucleotide variant (SNV)–based ara-C pharmacogenomics score (ACS10) with survival outcomes in a large cohort of pediatric AML patients; to evaluate whether ACS10 remains relevant in an adolescent and young adult (AYA) population of patients with AML treated with similar intensive induction chemotherapy protocols; and to assess the association of ACS10 with race and treatment outcomes in both cohorts. Design, Setting, and Participants This cohort study included patients from the Children’s Oncology Group’s AAML1031 trial, a multicenter, open-label randomized clinical trial that enrolled pediatric patients with newly diagnosed, treatment-naive primary AML from June 2011 to July 2017 (aged 0 to 29.5 years) and from the Alliance for Clinical Trials in Oncology frontline protocols, which included AYA patients from 9 different trials that enrolled patients with newly diagnosed AML from 1992 to 2010. Data were analyzed from September 2022 to March 2025. Exposures Patients in the AAML1031 trial were randomized to 2 arms, standard chemotherapy alone or standard chemotherapy with the addition of bortezomib. Patients in the Alliance for Clinical Trials in Oncology cohorts were treated with similar intensive induction chemotherapy protocols. Main Outcomes and Measures ACS10 scores were evaluated for association with outcomes according to race, treatment arm, and hematopoietic stem cell transplant (HSCT) status. Results The study included 1086 patients with AML. There were 717 patients from the pediatric AML cohort (median [range] age, 9.6 [0.04-29.2 years]; 379 [53%] male; 33 [5%] Asian, 84 [12%] Black, and 522 [73%] White) and 369 AYA patients with AML from the Alliance for Clinical Trials in Oncology group (median [range] age, 30 [17-39] years; 196 [53%] male; 7 [2%] Asian, 32 [9%] Black, and 288 [78%] White). Within the standard treatment arm of AAML1031, patients in the low ACS10 group had significantly worse event-free survival (EFS) compared with those in the high ACS10 group (all patients: hazard ratio [HR], 1.42; 95% CI, 1.05-1.95; P = .02; non-HSCT cohort: HR, 1.48; 95% CI, 1.06-2.07; P = .02). The ACS10 score remained significantly associated with EFS in multivariable analysis after adjusting for age, race, risk group and white blood cell count, within the standard treatment arm (HR, 1.44; 95% CI, 1.03-2.02; P = .03). In the Alliance for Clinical Trials in Oncology AYA non-HSCT cohort, the low ACS10 score group had significantly inferior overall survival (OS) and a higher point estimate for EFS compared with patients with a high ACS10 score (OS: HR, 1.50; 95% CI, 1.05-2.14; P = .03; EFS: HR, 1.32; 95% CI, 0.95-1.83; P = .10). A higher number of early deaths was observed in the low ACS10 group compared with the high ACS10 group, but the difference was not statistically significant (death within 30 days of treatment initiation: 6 of 112 [5%] vs 2 of 257 [1%]; P = .07). Across both cohorts, a low ACS10 score was significantly more abundant in Black patients compared with White patients (eg, in Alliance for Clinical Trials in Oncology cohort, 27 of 32 Black patients [84%] had low ACS10 scores compared with 64 of 288 White patients [22%]; P < .001) and inferior survival was observed in Black patients (eg, OS of Black compared with White patients in AAML1031 cohort: HR, 1.47; 95% CI, 1.02-2.13; P = .04). In the AAML1031 cohort, there were no significant differences in EFS or OS between Black and White patients receiving augmented treatment, suggesting that the addition of bortezomib was associated with benefit for Black patients. Conclusions and Relevance In this study of 717 pediatric and 369 AYA patients with AML, the ACS10 score was associated with EFS in pediatric and AYA patients when treated with a standard induction regimen. There was a higher abundance of low ACS10 scores in Black patients, and Black patients treated with augmented therapy (ie, the addition of bortezomib) seemed to have improved outcomes. Integrating the ACS10 score into a prospective clinical trial to personalize induction therapy based on an individual’s genetic profile has the potential to improve treatment outcomes.