CD123 Expression Is Associated With High-Risk Disease Characteristics in Childhood Acute Myeloid Leukemia: A Report From the Children's Oncology Group

Abstract

Purpose: Increased CD123 surface expression has been associated with high-risk disease characteristics in adult acute myeloid leukemia (AML), but has not been well-characterized in childhood AML. In this study, we defined CD123 expression and associated clinical characteristics in a uniformly treated cohort of pediatric patients with newly diagnosed AML enrolled on the Children's Oncology Group AAML1031 phase III trial (NCT01371981).

Materials and methods: AML blasts within diagnostic bone marrow specimens (n = 1,040) were prospectively analyzed for CD123 protein expression by multidimensional flow cytometry immunophenotyping at a central clinical laboratory. Patients were stratified as low-risk or high-risk on the basis of (1) leukemia-associated cytogenetic and molecular alterations and (2) end-of-induction measurable residual disease levels.

Results: The study population was divided into CD123 expression-based quartiles (n = 260 each) for analysis. Those with highest CD123 expression (quartile 4 [Q4]) had higher prevalence of high-risk KMT2A rearrangements and FLT3-ITD mutations (P < .001 for both) and lower prevalence of low-risk t(8;21), inv(16), and CEBPA mutations (P < .001 for all). Patients in lower CD123 expression quartiles (Q1-3) had similar relapse risk, event-free survival, and overall survival. Conversely, Q4 patients had a significantly higher relapse risk (53% v 39%, P < .001), lower event-free survival (49% v 69%, P < .001), and lower overall survival (32% v 50%, P < .001) in comparison with Q1-3 patients. CD123 maintained independent significance for outcomes when all known contemporary high-risk cytogenetic and molecular markers were incorporated into multivariable Cox regression analysis.

Conclusion: CD123 is strongly associated with disease-relevant cytogenetic and molecular alterations in childhood AML. CD123 is a critical biomarker and promising immunotherapeutic target for children with relapsed or refractory AML, given its prevalent expression and enrichment in patients with high-risk genetic alterations and inferior clinical outcomes with conventional therapy.

FIG 1.

Prevalence of select (A) high-risk and (B) low-risk genetic features on the basis of the CD123 quartile. *P < .05 for Q4 versus Q1-3 by proportional comparison. HAR, high allelic ratio; HR, high-risk; Q, quartile.

FIG 2.

Correlation of clinical outcomes with CD123 expression on the basis of protocol-defined risk classification: (A) EFS (n = 1,040), (B) RR (n = 772), (C) OS (n = 1,040) from study entry stratified by CD123 expression quartiles, (D) EFS from study entry for low-risk (LR) patients, (E) RR for LR patients, (F) OS for LR patients, (G) EFS from study entry for high-risk (HR) patients, (H) RR for HR patients, and (I) OS for HR patients. Q4 versus 1-3 comparison by log-rank and Gray's tests. EFS, event-free survival; OS, overall survival; Q, quartile; RR, relapse risk.

FIG 3.

Correlation of clinical outcomes with CD123 expression on the basis of the contemporary AAML1831-based cytogenetic and molecular risk factors: (A) EFS from study entry for non–high-risk (non-HR) patients, (B) RR for non-HR patients, (C) OS for non-HR patients, (D) EFS from study entry for high-risk (HR) patients, (E) RR for HR patients, and (F) OS for HR patients. Q4 versus 1-3 comparison by log-rank and Gray's tests. EFS, event-free survival; OS, overall survival; Q, quartile; RR, relapse risk.

FIG A1.

CONSORT diagram.

FIG A2.

Flow cytometric quantification of surface CD123 expression on diagnostic AML bone marrow specimens from children, adolescents, and young adults enrolled on AAML1031. Data are displayed as the median number of molecules/cell in ascending order of CD123 expression (n = 1,040 specimens). x-axis depicts the definition of each quartile (n = 260 specimens/quartile) and relative specimen numbers. AML, acute myeloid leukemia; MPC, molecules per cell.

FIG A3.

Correlation of clinical outcomes with CD123 expression quartiles: (A) EFS (n = 1,040), (B) RR (n = 772), and (C) OS (n = 1,040) from study entry stratified by CD123 expression quartiles. P < .001 for all Q4 versus Q3 versus Q2 versus Q1 by log-rank and Gray's tests. EFS, Event-free survival; OS, overall survival; Q, quartile; RR, relapse risk.