Longitudinal genomic and cytogenetic dynamics in myelodysplastic syndromes (MDS): Insights into clonal evolution and disease progression

Abstract

Abstract Background: Myelodysplastic syndrome (MDS) is a heterogeneous clonal hematopoietic disorder, with variable risk of progression to secondary acute myeloid leukemia (sAML). While clonal evolution is central to disease progression, the interaction between somatic mutations, cytogenetic changes, and treatment pressure remains incompletely defined. We queried longitudinal targeted sequencing, cytogenetic analyses, error corrected sequencing and clinical outcomes to characterize founding and emergent mutations, map disease trajectories, and evaluate subclone detection at diagnosis that may influence disease progression. Cohort description: We retrospectively analyzed 310 samples from 76 newly diagnosed MDS patients at our institute between 2006-2017, who underwent serial NGS and cytogenetic profiling on blood or marrow samples (up to 11 timepoints). Median age at diagnosis: 67 years; 67.1% of patients were male. By 2025, 28.9% (n=22) were alive, 42.1% underwent stem cell transplant (SCT), and 44.7% progressed to sAML. Results: To map clonal evolution in MDS, we profiled recurrent mutations longitudinally in 76 patients. The most common mutations at diagnosis were also found recurrently at later timepoints ASXL1 (n=20®25), TP53 (n=15®19), DNMT3A (n=13®16), and TET2 (n=12®14). Mutations in epigenetic regulators were stable in 82% of cases (41/50), whereas 53% (19/36) of signaling gene mutations, (e.g. KRAS, FLT3), were acquired later suggesting their emergence as subclonal drivers of disease. The most common cytogenetic abnormalities included (del(5q), monosomy 7, del(7q), del(20q), and trisomy 8, which showed the most dynamic acquisition, rising from 8% (6/76) at diagnosis to 20% (15/76) at follow-up. Patients were classified into cytogenetic subtypes: single abnormalities (36%, n=27), normal karyotype (23%, n=17), acquired abnormalities (23%, n=17), and complex karyotypes (16%, n=12). To delineate evolutionary trajectories, we analyzed 72 patients with both serial NGS and cytogenetic data. Linear evolution, marked by mutation gain, was the dominant pattern (38%, 27/72). A combined gain/loss pattern was observed in 28% (20/72), consistent with branched evolution under therapeutic pressure. Mutation stability (25%) and rare swap or loss patterns were also noted. Overlaying cytogenetic data revealed that mutation gain was most common in patients with single abnormalities (33%, 9/27), while stable profiles were enriched in those with normal karyotypes (35%, 6/17). Branched evolution (gain/loss) was most frequent in patients with acquired or single cytogenetic abnormalities, suggesting increased clonal plasticity in response to stress. Finally, we assessed if TP53 status influenced clonal dynamics. TP53-mutant patients (n=19) showed limited subclonal diversification and a high rate of complex karyotypes (42%, 8/19), reflecting early genomic instability. TP53-wildtype patients exhibited more stable cytogenetics and greater evolutionary flexibility. Most TP53-mutant cases (11/15) were classified as Very High or High risk by IPSS-R, aligning with their aggressive clinical course. TP53-mutant patients had a median overall survival of 4.1 months, with all patients deceased, compared to 4.9 months in TP53 wild-type patients, (22/57 alive). Notably, 83% (20/24) of patients who progressed to sAML harbored TP53 or signaling mutations. KRAS mutations were exclusively acquired at later timepoints, suggesting a role in subclonal evolution. These patients exhibited linear trajectories (mutation gain, 46%) and acquired cytogenetic changes (29%). In contrast, SCT patients (n=16) showed more branched evolution (56% with gain/loss) and complex karyotypes (31%). Those without SCT or sAML progression (n=18) had the most stable genomic architecture. Finally, we have optimized a high-sensitivity amplicon error corrected sequencing assay to detect low VAF mutations at diagnosis. These ongoing analysis (n=23), and single cell DNA sequencing (n=4) aim to resolve the timing of signaling mutation acquisition and evaluate the potential for sensitive NGS methods to offer prognostic information for MDS transformation to AML. Conclusion: TP53 and KRAS mutations drive leukemic transformation, while early epigenetic regulator remains stable. TP53 mutations are linked to poor prognosis, high-risk IPSS-R, and reduced survival. Integrating mutational, cytogenetic, and clinical data at time of MDS diagnosis allows for precise tracking of disease evolution.