Pharmacogenomics, Race, and Treatment Outcome in Pediatric AML (2024)

Key Points

Question Can tailoring the intensity of induction therapy mitigate racial disparities in the outcome of pediatric patients with acute myeloid leukemia (AML)?

Findings In this comparative effectiveness analysis of 86 Black and 359 White patients with AML, Black patients exhibited a higher prevalence of low cytarabine pharmacogenomic 10–single-nucleotide variant (ACS10) scores, which were associated with an unfavorable outcome after initial treatment with low-dose cytarabine-based induction therapy. Augmentation of induction therapy was associated with improved outcome for patients with low ACS10 scores and with no disparity in outcome between Black patients and White patients.

Meaning This study suggests that discrepancies in pharmacogenomics may explain the differences in outcomes between Black and White patients with AML, providing a strategy to overcome these disparities.

Importance Disparities in outcomes exist between Black and White patients with acute myeloid leukemia (AML), with Black patients experiencing poorer prognosis compared with their White counterparts.

Objective To assess whether varying intensity of induction therapy to treat pediatric AML is associated with reduced disparities in treatment outcome by race.

Design, Setting, and Participants A comparative effectiveness analysis was conducted of 86 Black and 359 White patients with newly diagnosed AML who were enrolled in the AML02 trial from 2002 to 2008 or the AML08 trial from 2008 to 2017. Statistical analysis was conducted from July 2023 through January 2024.

Interventions Patients in AML02 were randomly assigned to receive standard low-dose cytarabine-based induction therapy or augmented high-dose cytarabine-based induction therapy, whereas patients in AML08 received high-dose cytarabine-based therapy.

Main Outcomes and Measures Cytarabine pharmacogenomic 10–single-nucleotide variant (ACS10) scores were evaluated for association with outcome according to race and treatment arm.

Results This analysis included 86 Black patients (mean [SD] age, 8.8 [6.5] years; 54 boys [62.8%]; mean [SD] leukocyte count, 52 600 [74 000] cells/µL) and 359 White patients (mean [SD] age, 9.1 [6.2] years; 189 boys [52.6%]; mean [SD] leukocyte count, 54 500 [91 800] cells/µL); 70 individuals with other or unknown racial and ethnic backgrounds were not included. Among all patients without core binding factor AML who received standard induction therapy, Black patients had significantly worse outcomes compared with White patients (5-year event-free survival rate, 25% [95% CI, 9%-67%] compared with 56% [95% CI, 46%-70%]; P = .03). By contrast, among all patients who received augmented induction therapy, there were no differences in outcome according to race (5-year event-free survival rate, Black patients, 50% [95% CI, 38%-67%]; White patients, 48% [95% CI, 42%-55%]; P = .78). Among patients who received standard induction therapy, those with low ACS10 scores had a significantly worse 5-year event-free survival rate compared with those with high scores (42.4% [95% CI, 25.6%-59.3%] and 70.0% [95% CI, 56.6%-83.1%]; P = .004); however, among patients who received augmented induction therapy, there were no differences in 5-year event-free survival rates according to ACS10 score (low score, 60.6% [95% CI, 50.9%-70.2%] and high score, 54.8% [95% CI, 47.1%-62.5%]; P = .43).

Conclusions and Relevance In this comparative effectiveness study of pediatric patients with AML treated in 2 consecutive clinical trials, Black patients had worse outcomes compared with White patients after treatment with standard induction therapy, but this disparity was eliminated by treatment with augmented induction therapy. When accounting for ACS10 scores, no outcome disparities were seen between Black and White patients. Our results suggest that using pharmacogenomics parameters to tailor induction regimens for both Black and White patients may narrow the racial disparity gap in patients with AML.

Although the outcome of patients with hematologic malignant neoplasms has improved, racial and ethnic disparities, some of which can be attributed to differences in social determinants of health, persist.¹ Disparities in outcomes among individuals with acute myeloid leukemia (AML) have been observed across age groups, with Black patients experiencing poorer prognosis compared with their White counterparts among children,^2-4 adolescents and young adults,⁵ and adults.^6-8 For instance, Black children treated in the CCG2961 trial from 1996 to 2002 had a significantly lower overall survival rate compared with White children (45% compared with 60%; P = .007).² Similar patterns emerged from the Surveillance, Epidemiology, and End Results (SEER) database, where survival rates from 2001 to 2007 stood at 46% for Black children and 67% for White children (P < .01).⁹ Furthermore, in a study of 1122 children treated for AML from 2004 to 2014, Black patients had a notably higher mortality rate than White patients during the initial course of therapy (4.9% compared with 1.9%; P = .04).⁴ A comprehensive analysis of 814 pediatric patients with AML (108 Black, 154 Hispanic, and 552 non-Hispanic White) treated in cooperative group studies from 1996 to 2013 revealed that Black race was associated with inferior survival in the entire cohort as well as in analyses limited to patients with KMT2A rearrangements or core binding factor leukemia (RUNX1::RUNX1T1 or CBFB::MYH11).³

A recent comparison of 89 Black and 566 White adolescent and young adult patients with AML demonstrated that Black patients, especially those 18 to 29 years of age, experienced worse outcomes.⁵ These outcomes included a higher early mortality rate (16% compared with 3%; P = .002), lower complete remission rate (66% compared with 83%; P = .01), and worse overall survival (22% compared with 51%; P < .001), despite a higher prevalence of the favorable RUNX1::RUNX1T1 fusion transcript among Black patients (22% compared with 10%; P = .002). These findings emphasize the need to address and rectify the disparities in AML among different racial and ethnic groups.

Racial and ethnic disparities in treatment outcomes are undoubtedly complex, associated with a combination of socioeconomic and biological factors. The increased induction mortality observed among Black patients may be associated with their greater acuity at diagnosis and increased reliance on intensive care support.⁴ Moreover, disparities persist in clinical trial enrollment, with Black patients with AML less likely to enroll in clinical trials compared with White counterparts, a factor potentially associated with the observed discrepancies within the SEER database.¹⁰ Despite these socioeconomic considerations, the persistent inferior survival of Black patients relative to White patients, even when treated in the same clinical trial, underscores the potential effect of biological differences. To unravel the multifaceted causes of outcome disparities and to develop treatment strategies to overcome these disparities, we conducted a comparative analysis of the outcomes of Black and White pediatric patients with AML treated across 2 consecutive multi-institutional clinical trials. Our focus centered on exploring the associations between race, treatment modalities, and cytarabine pharmacogenomics.

This comparative effectiveness study used the existing clinical trial databases of AML02¹¹ (NCT00136084), conducted from 2002 to 2008, and AML08¹² (NCT00703820), conducted from 2008 to 2017, in which pediatric patients with newly diagnosed AML were enrolled. The report adheres to the International Society for Pharmacoeconomics and Outcomes Research (ISPOR) reporting guideline. As previously described, patients enrolled in AML02 were randomly assigned to receive standard induction therapy consisting of low-dose cytarabine, daunorubicin, and etoposide or augmented induction therapy with high-dose cytarabine, daunorubicin, and etoposide. All patients treated in AML08 received augmented induction therapy with high-dose cytarabine, daunorubicin, and etoposide or clofarabine and high-dose cytarabine as initial therapy. Details regarding therapy, risk classification, and outcomes have been previously reported.^11,12 Given the similarities in supportive care measures, risk classification, and outcomes between AML02 and AML08, the data from both trials were combined for the present analysis. Patients underwent bone marrow aspiration for the assessment of morphologic response and minimal residual disease (MRD) on day 22 of therapy, the results of which were used for risk classification and treatment assignment. Minimal residual disease was determined by flow-cytometric assessment of leukemia-specific immunophenotypes that were identified in diagnostic specimens, with MRD positivity defined as 0.1% or more. The protocols were approved by the institutional review boards of all participating institutions, and written informed consent was obtained from the patients’ guardians or parents, and assent from the patients, in accordance with the principles of the Declaration of Helsinki.¹³

Information on race and ethnicity, as defined by patients or their legal guardians, was collected at the time of enrollment in the AML02 and AML08 clinical trials. Race and ethnicity categories were African American or Black, American Indian or Alaska Native, Asian, Native Hawaiian or Other Pacific Islander, White, and not specified. This information was collected to establish the baseline demographic characteristics of the patients in the AML02 and AML08 studies to examine any associations with outcome.

For pharmacogenomics studies, genotype information and the polygenic cytarabine pharmacogenomic 10–single-nucleotide variant (ACS10) score for patients in the AML02 trial were derived from a combination of 10 single-nucleotide variants (SNVs) and were computed as previously reported.¹⁴ For the AML08 trial cohort, genomic DNA was genotyped using the Illumina Omni 2.5M Exome Beadchip (Illumina Inc) at Hussman Institute for Human Genetics, University of Miami, Miami, Florida. Six of the 10 SNVs that are part of the ACS10 score were typed on the Illumina 2.5 Omni array. Two (rs17103168 and rs11030918) of the remaining 4 SNVs were not represented in the array and were thus genotyped using TaqMan allele discrimination assay using QuantStudio 5 real-time polymerase chain reaction system (Thermo Fisher Scientific Inc). For the other 2 SNVs, genotype data for SNVs (rs1890005 in linkage disequilibrium with rs10916819 and rs10805074 in linkage disequilibrium with rs4643786) that occurred in high linkage disequilibrium (D′ = 1.0 and r² = 1.0) were used to calculate the ACS10 score.

Statistical analysis was conducted from July 2023 through January 2024. Patients’ characteristics were computed as mean (SD) values for continuous variables and as frequency and percentage for categorical variables by race. Event-free survival was defined as the time from study entry to induction failure, relapse, secondary malignant neoplasm, or death, with event-free patients censored on the date of last follow-up. Overall survival was defined as the time from study enrollment to death, with living patients censored on the date of last follow-up. Event-free and overall survival probabilities were estimated using the Kaplan-Meier method.¹⁵ Cox proportional hazards regression models¹⁶ were used to associate ACS10 score with event-free and overall survival. Multivariable Cox proportional hazards regressions were also used to model event-free and overall survival with ACS10 groups, risk group assignment, MRD, race, leukocyte count at diagnosis, and age as factors potentially associated with outcomes. The 95% CI of hazard ratios was calculated to quantitatively measure the association with clinical outcome. The Kruskal-Wallis test and the Wilcoxon rank sum test were used to compare the median values of continuous variables across groups. The χ² test and the Fisher exact test were used to evaluate the association among pairs of categorical variables. All P values were from 2-sided tests, and results were deemed statistically significant at P < .05. All statistical analyses were performed in R Statistical Software, version 4.1.0 (R Project for Statistical Computing).

Of 515 patients treated in AML02 or AML08, 86 (16.7%) were Black (mean [SD] age, 8.8 [6.5] years; 54 boys [62.8%]; mean [SD] leukocyte count, 52 600 [74 000] cells/µL), and 359 (69.7%) were White (mean [SD] age, 9.1 [6.2] years; 189 boys [52.6%]; mean [SD] leukocyte count, 54 500 [91 800] cells/µL); 70 patients (13.6%) with other (American Indian or Alaska Native, Asian, and Native Hawaiian or Other Pacific Islander) or unknown racial and ethnic backgrounds were not included in this analysis (eTable 1 in Supplement 1). There were no statistically significant differences between Black and White patients in age, sex distribution, or leukocyte count at the time of enrollment, but a higher prevalence of core binding factor leukemia was observed among Black patients than White patients (27 of 86 [31.4%] and 72 of 359 [20.1%], respectively; P = .04) (eTable 1 in Supplement 1).

Complete remission rates after 2 courses of induction therapy were comparable between Black and White patients (92.6% [76 of 82] and 95.0% [321 of 338], respectively; P = .63), as were the rates of MRD negativity after 1 course of therapy (55.8% [48 of 86] and 55.4% [199 of 359], respectively; P = .85). The event-free survival distributions of Black and White patients were similar, with 5-year estimates of 58.3% (95% CI, 48.8%-69.8%) and 58.2% (95% CI, 53.2%-63.6%; P = .94), respectively (Figure 1A). Likewise, overall survival did not differ significantly between Black and White patients, with 5-year estimates of 63.8% (95% CI, 54.3%-74.8%) and 69.4% (95% CI, 64.7%-74.5%; P = .24), respectively (Figure 1B). In addition, the estimated 5-year cumulative incidence of relapse did not differ by race (26.0% among Black patients and 26.1% among White patients; P = .99), nor did the incidence of treatment-related mortality (7.0% among Black patients and 6.8% among White patients; P = .78).

Given the higher prevalence of core binding factor leukemia among Black patients compared with White patients and its association with a favorable outcome, we conducted a subgroup analysis by examining outcomes based on race among patients with or without core binding factor leukemia (eFigure 1 in Supplement 1). There were no significant differences in event-free survival rates between Black and White patients with core binding factor leukemia (84.6% [95% CI, 67.1%-100%] and 88.6% [95% CI, 78.6%-99.7%], respectively; P = .70) or between Black and White patients without core binding factor leukemia (53.7% [95% CI, 43.4%-66.5%] and 54.9% [95% CI, 49.6%-60.7%], respectively; P = .77).

We next evaluated outcomes according to race and initial induction treatment (standard, low-dose cytarabine-based therapy vs augmented therapy). Patients with core binding factor leukemia had excellent outcomes independent of race or treatment regimen. However, among patients without core binding factor AML who received standard induction therapy, Black patients had significantly worse outcomes compared with White patients (5-year event-free survival rate, 25% [95% CI, 9%-67%] compared with 56% [95% CI, 46%-70%]; P = .03) (Figure 2). This disparity was not found among those treated with augmented induction therapy, with 5-year event-free survival rates of 50% (95% CI, 38%-67%) for Black patients and 48% (95% CI, 42%-55%) for White patients (P = .78).

To discern the causes of the aforementioned outcome disparities among patients treated with standard therapy, we further evaluated outcomes by race, treatment, and pharmacogenomics, focusing on ACS10, the recently defined cytarabine pharmacogenomics–based polygenic score.¹⁴ Among the 70 of 86 Black patients (81.4%) and 273 of 359 White patients (76.0%) for whom pharmacogenetic data were available, there was a significant difference in the distribution of ACS10 scores according to race. Specifically, low scores were present for 51 of 70 Black patients (72.9%) compared with 82 of 273 White patients (30.0%) (P < .001). When considering patients across all treatment regimens, there were no significant differences in event-free survival distributions between Black and White patients with low ACS10 scores (eFigure 2A in Supplement 1) or between Black and White patients with high ACS10 scores (eFigure 2B in Supplement 1). Among patients who received standard induction therapy, those with low ACS10 scores had significantly worse 5-year event-free survival rates compared with those with high scores (42.4% [95% CI, 25.6%-59.3%] and 70.0% [95% CI, 56.6%-83.1%]; P = .004), a finding consistent with a previous study.¹⁴ By contrast, no significant differences in outcome by ACS10 score were observed in Black or White patients who were treated with augmented induction therapy (low score, 60.6% [95% CI, 50.9%-70.2%] and high score, 54.8% [95% CI, 47.1%-62.5%]; P = .43). Neither race nor ACS10 score was associated with outcome among patients who received augmented induction therapy (Figure 3).

Multiple studies have shown that Black patients with AML have worse outcomes compared with non-Hispanic White patients.^2-8 For example, in a recent study of adolescent and young adult patients with AML, Black race was independently associated with poor outcome.⁵ This discrepancy was particularly pronounced among patients 18 to 29 years of age without core binding factor leukemia, with 5-year survival rates of 12% for Black patients compared with 44% for White patients (P < .001). In light of these reports, we assessed the association of race with patient outcomes within the context of the multisite AML02 and AML08 clinical trials. We combined the trials because they were conducted at largely the same sites, used similar supportive care measures, included similar risk stratification based on genomics and MRD, and produced comparable outcomes. Our pharmacogenomic analyses focused on SNVs in genes involved in cytarabine metabolism because activation of cytarabine to cytarabine-triphosphate and intracellular accumulation of cytarabine-triphosphate are important factors associated with treatment response.^17-22 Low ACS10 scores, which are associated with lower levels of intracellular cytarabine-triphosphate,¹⁹ were markedly more prevalent among Black patients, accounting for 72.9% of cases, in contrast to just 30.0% among White patients. This divergence was associated predominantly with 3 SNVs exhibiting dissimilar allele frequencies among Black and White patients. Single-nucleotide variant rs4643786 in DCK, which is associated with lower intracellular cytarabine-triphosphate levels and worse outcome, was markedly more abundant among Black patients than White patients (minor allele frequency, 0.48 for Black patients and 0.04 for White patients).^14,19 By contrast, variant alleles rs1044457 in CMPK1 and rs17343066 in SLC28A3, associated with favorable outcomes and higher intracellular cytarabine-triphosphate levels, were notably less common among Black patients compared with White patients (rs1044457, 0.11 among Black patients and 0.5 among White patients; rs17343066, 0.15 among Black patients and 0.53 among White patients).¹⁹ The distinctive distribution of ACS10 scores according to race takes on particular significance in light of a recent finding that patients with low ACS10 scores had worse outcomes compared with patients with high ACS10 scores when treated with standard, low-dose cytarabine–based induction therapy in the St Jude AML02 trial or the Children’s Oncology Group AAML0531 trial.¹⁴ In that report, the event-free survival rates for patients with low ACS10 scores were approximately 10 percentage points higher among patients who received augmented induction therapy with high-dose cytarabine in the AML02 trial or with gemtuzumab in the AAML0531 trial compared with those who received standard induction therapy.

Considering the higher prevalence of low ACS10 scores among Black patients and a previous demonstration that the outcomes of patients with low ACS10 scores can be improved by augmenting induction therapy,¹⁴ we explored potential interactions between race, induction therapy, pharmacogenomics, and outcome in the present study. Overall, we observed no significant differences in outcome between Black and White patients treated in the St Jude AML02 and AML08 trials. In contrast to recent reports,^3,5 the outcome of Black and White patients with core binding factor was excellent, with no significant differences according to race. However, among patients without core binding factor leukemia who received standard induction therapy, Black patients had significantly worse outcomes compared with White patients. Because both Black patients and White patients with low ACS10 scores had significantly worse outcomes compared with patients with high scores when treated with standard induction therapy, the high prevalence of low ACS10 scores likely explains the racial disparity in outcome for this treatment group. These disparities in outcome were nullified when induction therapy was augmented with high-dose cytarabine or with clofarabine and high-dose cytarabine, after which neither race nor ACS10 score had an association with outcomes. The improved outcome after intensified induction therapy suggests that the poor outcome after standard induction therapy was associated with relative underdosing rather than excessive toxic effects. In addition, the similarity in consolidation therapy regimens among Black patients and White patients further supports the association of early intensification with outcomes.

We propose that the high prevalence of low ACS10 scores among Black patients is also associated with the poor outcomes reported in other studies, which primarily included patients who underwent standard, low-dose cytarabine–based induction therapy in cooperative group trials.^3,5 This hypothesis is supported by results reported by investigators from the Children’s Oncology Group.^3,14 Among patients with AML who were treated in the Children’s Cancer Group 2961 trial or the Children’s Oncology Group AAML03P1 or AAML0531 trials, the event-free and overall survival rates for Black patients were only 35% and 52%, respectively, compared with rates of 50% and 71%, respectively, for White patients.³ In a multivariable analysis of patients treated in the standard arm of the AAML0531 trial, Black race was identified as independently associated with worse event-free and overall survival.¹⁴ However, in an analysis limited to patients with low ACS10 scores, treatment in the standard arm, but not race, was associated with inferior event-free survival. In addition, a multivariable analysis of patients treated in the augmented (gemtuzumab) arm of the AAML0531 trial indicated that Black race was no longer associated with a worse outcome.¹⁴ An analysis of patients with core binding factor leukemia also showed poorer prognosis for Black patients when low-dose cytarabine-based induction was administered, with event-free survival rates for Black patients of 44% and for White patients of 64%.³ In this subgroup analysis, gemtuzumab again had a differential association with outcomes among Black patients and completely eliminated the racial disparity in outcome, resulting in event-free survival rates of 69% for both Black and White patients.³ A recent analysis of patients treated in the Children’s Oncology Group AAML1031 trial further suggests that augmentation of induction therapy with bortezomib may have similar associations, with worse outcomes for Black patients who were treated on the standard arm but no disparities among patients in the augmented (bortezomib) arm.²³ We postulate that, in the present study, the overall lack of significant disparities in outcome according to race or ACS10 score is likely associated with administration of augmented therapy for 78% of patients.

This study has some limitations, including the availability of pharmacogenomic data for 343 patients (77.1%) and the small numbers of patients in certain subgroups, such as Black patients with high ACS10 scores who received standard induction therapy and Black patients with genetic alterations other than core binding factor leukemia, thereby hampering our ability to perform additional multivariable analyses. In addition, we did not directly measure intracellular cytarabine triphosphate levels and were thus unable to explore associations between intracellular cytarabine levels, ACS10 scores, race, and outcome. Moreover, although other studies have demonstrated similar results, we do not have a formal validation cohort to directly test our hypothesis. Given that current clinical trials for pediatric AML do not incorporate high-dose cytarabine or clofarabine during induction therapy, it is not feasible to conduct a validation study.

In this comparative effectiveness study of pediatric patients with AML, Black patients without core binding factor leukemia had significantly worse outcomes compared with White patients after treatment with standard induction therapy. However, among patients treated with augmented therapy, we observed no racial disparities. The associations between race, pharmacogenomics, treatment intensity, and outcome suggest that racial disparities in outcome are associated with variations in pharmacogenomics. We propose that future studies should include the nonrandomized tailoring of induction regimens to pharmacogenomic parameters to improve the outcome of Black and White patients and serve as a pivotal bridge in addressing the racial gap in AML treatment outcomes.

Accepted for Publication: March 15, 2024.

Published: May 16, 2024. doi:10.1001/jamanetworkopen.2024.11726

Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2024 Lamba JK et al. JAMA Network Open.

Corresponding Author: Jeffrey E. Rubnitz, MD, PhD, Department of Oncology, St Jude Children’s Research Hospital, 262 Danny Thomas Pl, Memphis, TN 38105-2794 (jeffrey.rubnitz@stjude.org).

Author Contributions: Dr Rubnitz had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Lamba, Marrero, Pounds, Rubnitz.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Lamba, Marrero, Wu, Pounds, Rubnitz.

Critical review of the manuscript for important intellectual content: Lamba, Marrero, Cao, Parcha, Karol, Inaba, Kuo, Degar, Heym, Taub, Lacayo, Pui, Ribeiro, Pounds.

Statistical analysis: Lamba, Wu, Cao, Pounds.

Obtained funding: Lamba.

Administrative, technical, or material support: Lamba, Inaba, Lacayo, Pui.

Supervision: Lamba, Kuo, Pounds, Rubnitz.

Conflict of Interest Disclosures: Drs Lamba, Cao, and Pounds reported a having patent for Pharmacogenomics Score to Make Decisions on Therapy Augmentation in AML pending 18/683,969. Dr Inaba reported receiving personal fees from Servier, Jazz Pharmaceutical, and Amgen outside the submitted work. Dr Pui reported receiving personal fees from Novartis outside the submitted work. Dr Pounds reported receiving grants from Gateway for Cancer Research and the American Cancer Society outside the submitted work and having a patent for Leukemia Diagnostic Based on Gene Expression pending, a patent for Methods for Predicting AML Outcome pending, and a patent for AML Risk Stratification Using OS iScore pending. No other disclosures were reported.

Funding/Support: The study was supported by National Cancer Institute award CA021765, American Lebanese Syrian Associated Charities, grant R01-CA132946 from the National Institutes of Health (Drs Lamba and Pounds), UF Opportunity seed grant AGR DTD 04-26-2018 (Dr Lamba), and the American Cancer Society and St Baldrick’s Foundation (SAP-21-061-01 SBF-ACS). Dr Marrero is supported by an institutional research training award from the National Institute of General Medical Sciences of the National Institutes of Health (T32HG008958). Research reported in this publication was also supported by the University of Florida Health Cancer Center, supported in part by state appropriations provided in Fla. Stat. § 381.915 and the National Cancer Institute of the National Institutes of Health under Award P30CA247796.

Role of the Funder/Sponsor: The funding sources had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Additional Contributions: We gratefully acknowledge the support and collaboration of all the research centers and clinicians involved in the treatment of patients and the collection of data and we thank the patients and families who participated in the studies.

Data Sharing Statement: See Supplement 2.