Impact of Cytarabine Pharmacogenomics on Survival of Adolescent and Young Adults with AML and Its Clinical Relevance in Black Patients

Cytarabine (also known as ara-C) has been the mainstay of AML chemotherapy for both pediatric and adult patients (pts) since 1970s, and when given in combination with anthracycline it induces remission in ~60% of adult and 80% of pediatric pts. However, almost ~40% of pts relapse. Cytarabine is a prodrug requiring activation to ara-CTP for its antileukemic effect. Our group has been investigating the relevance of cytarabine pharmacogenomics and recently reported a pharmacogenomics-based polygenic score (ACS10) that includes 10 SNPs spanning 9 genes within metabolic pathway of cytarabine. ACS10 is predictive of intracellular levels of ara-CTP as well as outcome in pediatric AML pts enrolled in multiple clinical trials led by St Jude Children's Research Hospital and Children's Oncology Group (Elsayed et al, JCO 2022;40:772-3). Overall, pediatric AML pts with ACS10 ^low (≤0), had inferior outcomes compared with ACS10 ^high pts. Thus far, the ACS10 score has not been tested for its performance in adult AML, especially adolescent and young adult (AYA) AML pts. We hypothesized that ACS10 might be predictive of chemotherapy resistance in this defined pt population that is commonly treated with intensive induction therapy. Moreover, as several SNPs included in ACS10 show ancestry-enrichments, we speculated whether this might help explain the high resistance to treatment recently observed in Black AYA AML pts.

A total of 406 newly diagnosed AYA AML pts (age 17-39y, median: 30y), similarly treated in 8 Alliance for Clinical Trials in Oncology frontline protocols were included in the study. All pts had centrally reviewed cytogenetics, targeted gene mutation profiling and SNP data (Illumina 2.5 omni array, see Walker et al, Leukemia 2019;33:771-5). ACS10 score was calculated for each pt using the equation defined by Elsayed et al (>0, ACS10 ^high; ≤0, ACS10 ^low) and Kaplan-Meier method and the log-rank test was used for survival analysis and Fisher's exact test was used for frequency comparison.

Two hundred eighty-three pts were ACS10 ^high and 123 pts were ACS10 ^low. Though not statistically significant, ACS10 ^low pts tended to have a shorter overall survival (OS) compared with ACS10 ^high pts (p=.11). Restricting the analysis to those who received chemotherapy only (excluding allogeneic transplant recipients), showed that ACS10 ^low group had significantly shorter OS compared with ACS10 ^high group (p=.03). Further, event-free survival (EFS) was shorter (3y rates, 32% vs. 42%, p=0.09) and number of early deaths (death within 30 days of treatment initiation) was higher (5% vs 1%, p=.07) in ACS10 ^low pts. Given that the majority of patients received etoposide as part of the induction regimen, analysis within this subset showed ACS10 ^high pts having higher complete remission (CR) rates (85% vs 72%, p=.05) and longer EFS and OS compared with ACS10 ^low pts (EFS, 3y rates, 41% vs 28%, p=.02; OS, 3y rates, 54% vs 32%, p=.007).

We have previously shown significant difference in abundance of ACS10 ^low by race/ethnicity in pediatric AML, 70% of Black pts being ACS10 ^low as compared with only 30% of White pts. Similarly, the current cohort showed significant differences in ACS10 score groups by race/ethnicity: of 32 Black pts included in the analysis, 27 (84%) were ACS10 ^low and only 5 (16%) were ACS10 ^high. Given the enrichment of the racial-ethnic survival disparity in AYA pts (Larkin et al, Blood Adv 2022;6:5570-81), we further investigated the impact of ACS10 on survival of Black AYA pts. Though limited by numbers, we observed worse OS (p=0.05) of ACS10 ^low group, which comprised 80% of Black pts.

We show the impact of cytarabine pharmacogenomics and its relevance in AYA AML pts. Consistent with the results in pediatric AML, pts with low ACS10 score, which is reflective of lower cytarabine activation and intracellular ara-CTP accumulation, had worse outcome than pts with high ACS10 score. Further, low ACS10 score was highly abundant in Black compared to White AYA AML pts. Given the low CR rates, high relapse rates and inferior survival of Black pts, our results imply cytarabine pharmacogenomics as a possible crucial contributor. In conclusion, given that cytarabine is the backbone of AML chemotherapy and is used as a frontline agent in essentially all induction regimens, our results highlight the need for in-depth evaluation of ACS10 score in AML that may open opportunities to improve treatment response in AYA AML.