Myelodysplastic syndromes are a heterogeneous group of clonal bone marrow hematopoietic stem cell disorders characterized by ineffective hematopoiesis and peripheral cytopenias. Chromosomal abnormalities and gene mutations have been shown to have essential roles in pathogenesis and correlate with prognosis. Molecular markers, however, are not integrated into currently used prognostic systems. The goal of this study is to identify plasma microRNAs useful for classification and risk stratification of myelodysplastic syndromes. We applied a novel, high-throughput digital quantification technology (NanoString) to profile microRNA expression in plasma samples of 72 patients with myelodysplastic syndromes and 12 healthy individuals. We correlated these results with overall survival. In patients with myelodysplastic syndromes associated with a diploid karyotype, we identified and validated a 7-microRNA signature as an independent predictor of survival with a predictive power of 75% accuracy (P=0.008), better than those of the International Prognostic Scoring Systems and the MD Anderson Prognostic Lower Risk Prognostic Model. We also identified differentially expressed plasma microRNAs in patients with myelodysplastic syndromes versus healthy individuals and between patients with myelodysplastic syndromes associated with different cytogenetic features. These results validate the utility of circulating-microRNA levels as noninvasive biomarkers that can inform the management of patients with myelodysplastic syndromes. Our findings also shed light on interactions of gene regulation pathways that are likely involved in the pathogenesis of myelodysplastic syndromes.