Abstract BACKGROUND AND AIMS Approximately 30% of patients with myelodysplastic syndrome (MDS) progress to acute myeloid leukemia (AML) within a few months up to several years. So far, blast counts and cytogenetic abnormalities are known to be major determinants of the risk for AML transformation; however, little is known about the molecular genetic events associated with MDS progression. Alongside, recent advances in DNA sequencing have enabled the identification a large number of new mutated genes in 90% of MDS patients. We focused on the identification of prognostic factors for patients who progress, namely, creating a specific mutational profile predicting disease progression. METHODS We examined paired samples from 23 patients by TruSight Myeloid Sequencing Panel containing 54 genes (Illumina) for formation of mutational profile. Patients were evaluated at the time of diagnosis and during progression to more advanced MDS subtype or to AML. Selected mutations were confirmed by Sanger sequencing. The survival distributions were estimated using the Kaplan-Meier method and prognostic factor analysis was performed with Cox`s multivariate regression method. RESULTS A total of 43 mutations at diagnosis and 56 mutations in progression across 54 genes were identified in the patient cohort. Patients carried at average 2.4 mutations, two patients carried five mutations and only one patient had no detectable mutation. The variant allele frequency (VAF) of mutations detected at the time of diagnosis increased during disease progression, except of mutations in SFR3B1 and TET2 genes. The most common acquired mutation at the time of progression was found in PTPN11 gene. 57% of patients harbored mutations in RUNX1, TP53, PTPN11 or NRAS gene at the time of diagnosis and 13 % of patients evolved these mutations in the course of disease. The median overall survival (OS) was 9.9 months in patients with this combination of mutations compared to 34 months (HR 4.0; p=0.01) in patients without mutations above mentioned. Multivariate analysis demonstrated that the combination of RUNX1, TP53, PTPN11 or NRASmutations was the strongest independent prognostic factor for OS (HR: 20.9; p=0.003) at the time of diagnosis. CONCLUSIONS Comparison of mutational profiles of MDS patients at diagnosis and during progression suggested candidate genes involved in AML transformation. Our study demonstrates that mutated RUNX1, TP53, PTPN11 and NRAS are most associated with the disease progression. Mutations in these genes are often detectable with low VAF already at the time of diagnosis and in some patients are detectable only at the time of progression. On the contrary, individual SF3B1 or TET2 mutations are associated with delayed progression and longer OS, which corresponds to a good prognostic value of these mutations. In conclusion, the results indicate that routine monitoring for mutations in MDS should be performed to refine the risk prediction of disease progression. Supported by AZV grants (16-33485A and 16-31689A) and the project for conceptual development of research organization (00023736) from the Ministry of Health of the Czech Republic. Disclosures No relevant conflicts of interest to declare.