Myelodysplastic Syndrome with Excess Blasts and Secondary Acute Myeloid Leukemia: Same Disease with Different Blast Count

Introduction: The World Health Organization (WHO) diagnoses acute myeloid leukemia (AML) if ≥20% myeloid blasts are present in peripheral blood or bone marrow. Consequently a patient with even 19% blasts is often ineligible for an "AML study". A less arbitrary means to define "AML" and myelodysplastic syndromes ("MDS") emphasizes biologic features. Here, focusing on patients with WHO-defined MDS with excess (5-19%) blasts (MDS-EB) or AML with myelodysplasia-related changes (AML-MRC) or therapy-related (t-AML) (WHO defined secondary AML), we compared morphologic blast percentage (MBP) with the frequency of mutations in genes belonging to different functional groups, and with the variant allele frequency (VAF) for individually mutated genes.

Methods: 328 adults with WHO-defined AML (de novo and secondary; n=149) or MDS (n=179) and with mutational analysis by next-generation sequencing (NGS) performed at the University of Washington Hematopathology Laboratory between 2015-2017 were included. Of these, 86 had MDS-EB and 49 had secondary AML. Mutational analysis was performed using a customized, amplicon-based assay, TruSeq Custom Amplicon (Illumina, San Diego, CA). Custom oligonucleotide probes targeted specific mutational hotspots in ASXL1, CBL, CEBPA, CSF3R, EZH2, FBXW7, FGFR1, FLT3, GATA1, GATA2, HRAS, IDH1, IDH2, JAK2, KIT, KMT2A, KRAS, MAP2K1, MPL, NOTCH1, NPM1, NRAS, PDGFRA, PHF6, PTEN, RB1, RUNX1, SF3B1, SRSF2, STAG2, STAT3, TET2, TP53, U2AF1, WT1, and ZRSR2. VAF ≥5% was required to identify point mutations. Spearman's correlation coefficient was used to examine the relation between VAF of individually mutated genes and MBP. The Mann Whitney test served to compare the distribution of VAF in AML (≥20% blasts) vs. MDS (<20% blasts), before and after exclusion of subgroups as described below. Fisher's exact test was used to compare incidence of mutations.

Results: 96% of cases had ≥one mutation in the 36 genes tested using NGS. Considering all 328 patients, mutations in tumor suppressor and cohesin complex genes were similarly frequent in MDS and AML, whereas spliceosomal genes, in particular SF3B1 and SRSF2, were more frequently mutated in MDS than in AML (46% vs. 26%, p<0.001). Mutations in epigenetic modifiers were more common in AML than MDS (54% vs. 42%, p= 0.035) as were transcription factor mutations (52% vs. 28%, p<0.001). However comparisons limited to MDS-EB vs. AML-MRC/t-AML, indicated the differences observed when comparing all MDS and all AML were less apparent, both statistically and more perhaps importantly with respect to observed frequencies. For example, spliceosomal gene mutations were found in 35% in MDS-EB and 27% in AML-MRC/t-AML (p=0.34) vs. 46% and 26% in all MDS and all AML. NPM1 mutations were detected in only 8% of AML-MRC/t-AML vs. 3% in MDS-EB but 29% for all AML. Results were analogous with FLT3 ITD, FLT3 TKD, and JAK2 mutations. Examining 20 individually mutated genes detected in ≥ 10 patients only with SRSF2 (p=0.04), did distribution of VAF differ statistically according to whether blast percentage was <20% versus ≥20%.

Conclusions: The similar prevalence of mutations in different functional categories in MDS-EB and AML-MRC/t-AML suggests these entities are two manifestations of the same disease. We believe it appropriate to combine these WHO entities allowing patients in each to be eligible for both AML and MDS trials.