Masked Hypodiploidy: Hypodiploid Acute Lymphoblastic Leukemia (ALL) in Children Mimicking Hyperdiploid ALL: A Report From the Children's Oncology Group (COG) AALL03B1 Study.

Abstract 1580

Poster Board I-606

Hyperdiploidy with greater than 50 chromosomes is usually associated with a good prognosis in childhood acute lymphoblastic leukemia (ALL). By contrast, hypodiploidy with 44 or fewer chromosomes is a recurring abnormality detected in approximately 1% of children with ALL and is associated with an extremely poor prognosis with a projected event-free survival of less than 35% (Nachman et al Blood 110:1112, 2007; Harrison et al Br J Haematol 125:552, 2004; Raimondi et al Cancer 98:2715, 2003). Three distinct subgroups of hypodiploidy are recognized; near-haploidy (24-31 chromosomes), low hypodiploidy (32-39 chromosomes), and high hypodiploidy (40-44 chromosomes). It is common for leukemic cells with near-haploid or low hypodiploid chromosome numbers to undergo an exact or nearly exact doubling of the hypodiploid clone that results in a modal chromosome number in the hyperdiploid range, which might be misconstrued as indicating a good prognosis. In these cases of hypodiploid doubling, most chromosomes will be disomic or tetrasomic, but not trisomic. These cases are usually found to be mosaic with both hypodiploid and hyperdiploid (doubled) clones visible by standard cytogenetics, DNA Index (DI) by flow cytometry, and/or fluorescence in situ hybridization (FISH).

We reviewed the cytogenetics and presenting clinical features of children with ALL enrolled on the COG AALL03B1 study between 12/29/03 and 6/30/09. Among 8091 eligible patients, 92 (1.1%) had abnormal cytogenetics that were suggestive of hypodiploidy with fewer than 44 chromosomes. Of the 92 hypodiploid patients with visible chromosome abnormalities 54 were near-haploid, 32 low hypodiploid, and 6 high hypodiploid. Thirty-three patients (36%) had only the hypodiploid clone, 36 (39%) were mosaics with both a hypodiploid and a doubled hypodiploid clone identified, and most importantly, 23 (25%) patients displayed only the doubled hypodiploid clone. While the few previously described cases with only the doubled clone have always been shown to have evidence of hypodiploidy by DI, we found that only 10 of these 23 cases had a DNA index that suggested the presence of a hypodiploid clone. The remaining 13 cases had ploidy and FISH results consistent with the presence of a single, hyperdiploid (doubled) clone, thus effectively masking the underlying hypodiploidy. However, these cases all demonstrated a unique chromosome pattern characterized primarily by tetrasomy rather than trisomy, which distinguished them from a “typical” hyperdiploid case with an equivalent number of chromosomes. Tumor and germline DNA from all 13 masked hypodiploid patients were evaluated for loss of heterozygosity (LOH) using a microsatellite panel that tested 15 distinct loci on 13 different chromosomes. In all 13 patients LOH at multiple loci confirmed that the hyperdiploid cells had arisen by way of a doubling of a hypodiploid clone. Although numbers are too small for outcome analysis, there were no statistically significant differences in NCI Risk group or response to induction therapy (marrow morphology at day 15 or 29, and day 29 minimal residual disease levels) among masked hypodiploid cases and others in which the hypodiploid clone was visible. We conclude that a significant proportion (15-25%) of blast cell hypodiploidy may have been overlooked in children with ALL in previous studies due to the presence of a doubled hypodiploid clone and the absence of hypodiploid interphase/metaphase cells. Heightened awareness among cytogeneticists and clinicians about the unique karyotypic “signature” of a doubled hypodiploid clone coupled with the coordinated use of DI, FISH and LOH studies when indicated are important for the identification of patients with masked hypodiploidy so that they can be assigned to appropriate very high-risk treatment strategies.