CONCURRENT DELETIONS OF IKZF1 AND PAX5, CDKN2A, CDKN2B OR PAR1 (IKZF1-PLUS) CONFER A VERY POOR PROGNOSIS IN PEDIATRIC ACUTE LYMPHOBLASTIC LEUKEMIA

ABSSUB-4904

Background: Recent technical advances in the field of genomic analyses have stimulated a large number of discovery studies on etiological and clinical endpoints in acute lymphoblastic leukemia (ALL). Recently, GATA3was described as a new susceptibility locus for pediatric ALL (Migliorini et al. Blood 2013;122:3298-3307). The variant GATA3 allele (rs3824662) was found to be specifically associated with B-other ALL (B-lineage ALL lacking recurrent chromosomal translocations and hyperdiploidy) and to confer a dismal outcome.

Aims: Analysis of somatic genetic aberrations describing B-other ALLs associated with rs3824662 to characterize the leukemia genetics underlying the prognostic germline GATA3 ALL risk-allele.

Methods: Clinical epidemiological analysis of 500 patients treated on trial AIEOP-BFM 2000 having information on rs3824662 genotype and leukemia genetic characteristics including MLPA (SALSA kit P335-A3; MRC-Holland) and ERG deletion. Validation of results by employing an independent cohort of 599 patients from trial AIEOP-BFM 2000. Treatment outcome was analyzed excluding BCR/ABL1-positive patients.

Results: GATA3 rs3824662 genotype was negatively associated with hyperdiploidy and positively associated with prednisone poor-response, higher loads of minimal residual disease (MRD), and deletions of IKZF1, PAX5,CDKN2A and CDKN2B. Out of the 476 patients from the discovery cohort, 88 (18.5%) bore an IKZF1 deletion, 98 (20.6%) a PAX5 deletion, 61 (12.8%) a heterozygous and 69 (14.5%) a homozygous deletion of CDKN2A, 61 (12.8%) a heterozygous and 60 (12.6%) a homozygous deletion of CDKN2B. Similar numbers were observed in the validation cohort. When single marker analyses were conducted, IKZF1 deletion was the strongest determinator of outcome. In CDKN2A- and CDKN2B-deleted patients, the prognostic impact was restricted to those with homozygous deletions. When IKZF1 deletions were analyzed in the discovery and validation cohorts in combination with PAX5, CDKN2A and CDKN2B deletions, patients displaying an additional deletion to that ofIKZF1 had the worst event-free survival (EFS) and highest cumulative incidence of relapse (CIR). We also analyzedIKZF1 deletion in association with PAR1 deletions leading to P2RY8-CRLF2 fusion, which we previously described as prognostically relevant in our patient population (Cario et al. Blood 2010;115:5293-5397) and detected similar combinatorial effects as described above. Consequently, we defined a group by presence of IKZF1 deletion and at least one additional deletion in PAX5, CDKN2A, CDKN2B or PAR1. This group comprising 6% of B-lineage ALL patients was termed IKZF1plus and had a very poor clinical outcome: 5y-EFS 55%±0.07 compared to 85%±0.01 in IKZF1plus-negatives (p<0.0001); 5y-CIR 42%±0.07 compared to 11%±0.01 (p<0.0001). In multivariate analyses including MRD, slow early response, prednisone response, ETV6/RUNX1 status, and WBC (≥100.000/µl),IKZF1plus displayed the highest hazard ratio for relapse (3.29; 95% CI 2.01-5.38; p<0.001).

Summary/Conclusion: The previously described effect on clinical outcome of GATA3 SNP rs3824662 is due to its association with distinct prognostic leukemic lesions. Combinatorial analysis of deletions in IKZF1, PAX5,CDKN2A, CDKN2B and PAR1 allowed description of a very poor prognostic subgroup of ALL – termedIKZF1plus – with significantly worse outcome compared to the use of IKZF1 deletion or others as a sole marker. The definition of IKZF1plus is likely to aid in the practical implementation of newly detected markers for risk stratification in childhood ALL.

Grant support: TRANSCALL (EU FP7 - TRANSCAN, FKZ01KT1312), GACR-P302/12/G101, UNCE204012

Keywords: Acute lymphoblastic leukemia, Childhood, Ikaros, Prognosis

http://www.ehaweb.org/