Distinction of Early and Late Molecular Events In Patients with Myelodysplastic Syndromes (MDS) Who Progressed to Acute Myeloid Leukemia (AML),

Abstract 3566

MDS patients (pts) are destined to either progress to more aggressive disease or die of MDS-related complications. The International Prognostic Scoring System (IPSS) and other systems assess the survival and risk of evolution to acute myelogenous leukemia (AML), based in part on specific cytogenetic abnormalities. However, heterogeneity still exists even within cytogenetic groups, in part due to cryptic lesions, making predictions imperfect. In this respect, we and others have demonstrated that single nucleotide polymorphism array (SNP-A) karyotyping may improve the diagnostic yield, and that previously cryptic lesions can improve prognostic resolution of cytogenetic testing when compared to metaphase cytogenetics (MC) alone. Despite the recent identification of a variety of somatic mutations in MDS, their therapeutic value is unproven, and because of conflicting findings as to the risk for many mutations, they have not been routinely introduced into diagnostic algorithms.

Mutational events may constitute suitable biomarkers that could strengthen current prognostic schemes. In addition, malignant progression is associated with acquisition of new molecular defects, which may impact our understanding of the molecular pathogenesis of MDS and allow for distinction of ancestral from facilitating defects. We hypothesized that serial cytogenetic testing for a comprehensive spectrum of mutations and cytogenetic defects would reveal which molecular event is associated with progression.

We have analyzed a cohort of 76 MDS pts, 38 of whom experienced progression to AML (median age=68.5, IPSS = 0.5) matched to 38 pts (median age=69, IPSS =0.5) who did not experience progression. In addition to MC, we compared these groups for the presence of selected mutations, including TP53, RUNX1, TET2, ASXL1, DNMT3A, IDH1/2, NPM1, CBL and EZH2. The progressor group (PG) consisted of 20 (53%) low and intermediate-1 risk patients; the non-progressor group (NPG) had 22 (62%). In 60% of PG and 49% NPG at presentation, normal cytogenetics were present, while 16% and 14% had poor-risk cytogenetics, respectively. In addition, in 13% of PG and 11% of NPG, del7/7q was present. CBL, TP53, NPM1, ASXL1, IDH1, IDH2, RUNX1, DNMT3A, TET2 and EZH2 mutations were detected in (for PG vs. NPG): 0 (0/33) vs. 9% (3/35) p=1; 0 (0/29) vs. 5% (1/20) p=.4; 0 (0/33) vs N/A, 20% (6/30) vs. 18% (6/34) p=1; 9%(3/32) vs. 0 (0/39) p=.08; 0 (0/25) vs 0 (0/36) p=1; 0% (0/26) vs. 16% (1/6) p=.18; 7% (2/27) vs. 11% (4/37) p=1; 19% (4/21) vs. 37% (13/35) p=.23; and 5% (1/19) vs. 6% (2/32) p=1, respectively. We conclude that these pt groups had similar rates of mutations. We subsequently compared the cytogenetic and mutational status between initial presentation and at AML progression in the PG group. Progression was associated with occurrence of new cytogenetic lesions in 41% (14/34) while it remains stable in 53% (18/34) of pts. The most common new cytogenetics defects associated with evolution included trisomy 8, monosomy 7/del7q and monosomy 18. In 1, 2, and 1 pts, progression was associated with acquisition of CBL, TP53, and RUNX1 mutations, respectively. In contrast in 5, 3, 2 and 4 pts, ASXL1, IDH1, DNMT3A and TET2 mutations were present already at presentation. New mutations were detected upon leukemic evolution in 10/14 patients with abnormal MC and 3/14 with normal MC.

We conclude that ASXL1 and TET2 (and perhaps IDH1 and DNMT3A) constitute early events while acquisition of TP53, CBL and possibly RUNX1 were events leading to progressive, advanced disease.