Heterogeneity of IKZF1 genomic alterations and risk of relapse in childhood B-cell precursor acute lymphoblastic leukemia

Abstract

Genomic alterations of IKZF1 are common and associated with adverse clinical features in B-ALL. The relationship between the type of IKZF1 alteration, disease subtype and outcome are incompletely understood. Leukemia subtype and genomic alterations were determined using transcriptome and genomic sequencing and SNP microarray in 688 pediatric patients with B-ALL in St. Jude Total Therapy 15 and 16 studies. IKZF1 alterations were identified in 115 (16.7%) patients, most commonly in BCR::ABL1 (78%) and CRLF2-rearranged, BCR::ABL1-like B-ALL (70%). These alterations were associated with 5-year cumulative incidence of relapse (CIR) of 14.8 ± 3.3% compared to 5.0 ± 0.9% for patients without any IKZF1 alteration (P < 0.0001). IKZF1 deletions of exon 4-7 (P = 0.0002), genomic IKZF1 ^plus with any IKZF1 deletion (P = 0.006) or with focal IKZF1 deletion (P = 0.0007), and unfavorable genomic subtypes (P < 0.005) were independently adversely prognostic factors. Associations of genomic IKZF1 ^plus and exon 4-7 deletions with adverse outcomes were confirmed in an independent cohort. Genomic IKZF1 ^plus with any IKZF1 deletion, IKZF1 deletion of exon 4-7, and unfavorable subtype confer increased risk of relapse. The type of IKZF1 alteration, together with the subtype, are informative for risk stratification and predict response in patients with B-ALL.

Figure 1. Frequency of IKZF1 alterations within B-ALL genetic subtypes, in eligible patients with B- ALL in the Total15 and 16 study group (n=688).

Color-coded stacked horizontal bar graphs for which segments within each bar represent the proportion of patients (frequency shown on X axis) within respective subtypes (indicated on the Y axis), with color key below each graph, are shown. A-C, Focal IKZF1 deletions and mutations are included. D-F, Focal IKZF1 deletions and −7/del(7p) chromosome losses are included. Genomic IKZF1plus includes all CRLF2r and DUX4r in IKZF1plus profile (A, C, and D-F). In MLPA-based definition for IKZF1plus, PAR1 deletion serves as surrogate for P2RY8::CRLF2 fusions and ERG deletions are a surrogate of DUX4r. B-Other (n = 34) and rare subtypes (BCL2/MYC, CDX/UBTF, IKZF1 N159Y, TCF3::HLF, and ZEB2/CEBP; n = 6) detected in 2 or fewer patients are not included. Mut: missense, nonsense, or frameshift mutations.

Figure 2. Outcomes of patients with or without any IKZF1 alterations in the Total 15 and 16 study group.

A, Event-free survival (EFS) and B, Cumulative Incidence of Relapse (CIR) for patients based on presence or absence of IKZF1deletion or mutations in studied patients. IKZF1 deletion includes focal IKZF1 deletions, −7/del(7p). mut: missense, nonsense or frameshift mutations.

Figure 3. Outcomes based on type of IKZF1 alterations, including IKZF1^plus (genomic, any ΔIKZF1) and sequence mutations in the Total 15 and 16 study group.

A, Event-free survival (EFS) and B, Cumulative Incidence of Relapse (CIR). P values for pairwise comparisons for 5-year EFS compared to the no IKZF1 alteration group are < 0.0001 for IKZF1^plus group, 0.0006 for IKZF1 del only, and 0.19 for IKZF1 mut. Alteration groups are mutually exclusive; data are shown for patients with only one type of alteration. IKZF1 deletions (DIKZF1) are defined as focal IKZF1 deletions or −7/del(7p). Mut: missense, nonsense or frameshift mutations.

Figure 4. Outcomes based on presence or absence of focal IKZF1 deletions (IKZF1 Δ4–7 or not) or sequence mutations.

A, Event-free survival (EFS) and B, Cumulative Incidence of Relapse (CIR). Pvalues for pairwise comparisons for 5-year EFS compared to the no IKZF1alteration group are < 0.0001 for IKZF1 Δ4–7 group, 0.0002 for other IKZF1 focal deletions, and 0.3899 for IKZF1 mut. Alteration groups are mutually exclusive; data are shown for patients with only one type of alteration. Mut: missense, nonsense or frameshift mutations.

Figure 5. Estimated hazard ratios for event-free survival (EFS) and cumulative incidence of relapse (CIR) from Multivariable Cox Proportional Hazards Model in the Total 15 and 16 study group.

A and C, Event-free survival (EFS) and B and D, Cumulative Incidence of Relapse (CIR) adjusting for IKZF1 alteration status, age, genetic subtype group and EOI MRD. Comprehensive genomics-based definition of IKZF1^plus is used and IKZF1 deletion is defined as focal IKZF1 deletions or −7/del(7p) (A and B). All models include a time interaction term to reflect non-proportional hazards effect by the covariate IKZF1 alteration status. P values for time interaction terms: A, 0.06; B, 0.11; C, 0.009; D, 0.011. WBC: presenting white blood cell count/μl, EOI MRD: end of induction minimal residual disease, Mut: nonsense, missense or frameshift mutations.

Figure 6. Outcomes based on type of IKZF1 alteration (IKZF1 Δ4–7 or not) and subtype group in the Total 15 and16 study group.

A, Event-free survival (EFS) and B, Cumulative Incidence of Relapse (CIR) for patients based on type of focal IKZF1 deletions (IKZF1 Δ4–7 or not) or other IKZF1 alterations (missense or frameshift mutations or −7/del(7p)), and presence or absence of unfavorable subtype group, among eligible patients with B-ALL. Five-year EFS and CIR are shown on Table S17. Alteration groups are mutually exclusive; data are shown for patients with only one type of alteration.