When investigations for myelodysplastic syndrome (MDS) reveal sub-threshold dysplasia and a normal karyotype, patients may be labeled with equivocal/suspicious/idiopathic cytopenia or dysplasia, discharged and lost to follow-up, or subjected to serial bone marrow investigations and diagnostic delays. Although not part of the current WHO classification, targeted somatic mutation profiling increases clonality detection in MDS and clonal hematopoiesis of indeterminate potential (CHIP).


The application of targeted DNA sequencing to suspicious MDS cases may improve diagnostic yield, inform the natural history of MDS-like CHIP, and shift the paradigm to earlier intervention.


With IRB approval, frozen BM mononuclear cells or air-dried smears were obtained from Kingston General & Sunnybrook Hospitals. These included: a) 16 age-matched controls (8 negative lymphoma staging; 8 non-MDS cytopenia), b) 18 BM suspicious for MDS, c) 20 diagnosed MDS & d) 16 MDS/MPN. Genomic DNA was subjected to mutation profiling of 589 coding regions in 48 recurrently-mutated genes using a 1,662-amplicon Ion Torrent sequencing panel at Queen's University (QGLO). Variant calling was performed with Ion Reporter v4.6 (Life Technologies) & Integrative Genomics Viewer (Broad Institute). Selected variants were confirmed by PCR/Sanger-sequencing. Statistical analysis was conducted using Prism GraphPad software.


Clinical characteristics of suspicious MDS: Compared to age-matched controls (n=16), suspicious MDS cases (n=18) were associated with significantly reduced mean hemoglobin (119.6 ± 3.7 vs 96.0 ± 3.1 g/l; p<0.0001), platelet count (227.8 ± 20.2 vs 157.9 ± 25.5; p=0.04), & increased MCV (88.1 ± 1.6 vs 97.0 ± 2.2 fl; p=0.003). No significant differences were detected for these parameters between suspicious MDS & diagnosed MDS (n=20).

Mutation profiles: We found suspected mutations in 14/18 (78%) of suspicious cases, 17/20 (85%) of MDS & 16/16 (100%) of MDS/MPN. While the mean number of mutations per suspicious patient (1.33 ± 0.3; most commonly in SF3B1, TET2, DNMT3A and ASXL1) was significantly lower than MDS (2.15 ± 0.3; p=0.03) & MDS/MPN (3.75 ± 0.4; p<0.0001), there were no significant differences in the average variant allele frequencies (VAF): suspicious 42%, MDS 41%, MDS/MPN 41%. These results suggested that CHIP is common in suspicious MDS cases, with similar clonal size but lesser mutational burden than diagnosed MDS.

CHIP in control subjects and suspicious cases: CHIP was also detectable in 31% of control marrows (5/16) drawn from negative lymphoma staging & non-MDS cytopenia, albeit with significantly lower mutational and clonal burdens than suspicious MDS (0.31 ± 0.12 mutations/patient, p=0.002; 23 ± 0.1% avg. VAF, p=0.049). Among control & suspicious cases, CHIP (n=19) was associated with a 13.3 g/l reduction in mean hemoglobin (p=0.03). SRSF2, SETPB1, CBL & PTNPN11 mutations were statistically absent or under-represented in CHIP (as compared to MDS+MDS/MPN), with similar trends for EZH2, RUNX1, & TP53 mutations.

CHIP versus low-grade MDS: Importantly, given the diagnostic uncertainty between CHIP & low blast count, normal karyotype MDS (i.e. hinges on the presence of 10% dysplasia), we next stratified our control/suspicious & diagnostic low-grade MDS cases by the presence or absence of detectable mutations. The control/suspicious group (n=34) included 19 patients with & 15 patients without mutations; low-grade MDS (n=12) included 9 mutant & 3 non-mutant. Ignoring whether 10% subjective dysplasia was present in this cohort, using Kaplan-Meier analysis we found the presence of a mutation was associated with reduced median survival (677 days, n=28) versus the absence of a mutation (median survival undefined, n=18) (p=0.045 by Gehan-Breslow-Wilcoxon test).


CHIP is commonly detected in suspicious but non-MDS-diagnostic cases (78%). While the average clone size (42% VAF) is indistinguishable from MDS & MDS/MPN, CHIP-positive suspicious marrows have a reduced & more restricted mutation profile. CHIP is associated with increased anemia & worse outcome. These findings suggest clinical utility for mutation profiling in suspicious MDS cases & call for larger prospective studies of the natural history of MDS-like CHIP.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.