Clinical experience with 5-azacytidine and decitabine treatment of myelodysplastic syndromes (MDS), complemented by biological and pharmacological studies, has revealed compelling mechanism of action differences compared with traditional myeloid cancer treatment mainstays such as cytarabine. For example, 5-azacytidine and decitabine produce remissions and better overall survival in MDS with high-risk chromosome abnormalities at a surprisingly high rate, consistent with experimental observations that noncytotoxic DNA methyltransferase depletion by 5-azacytidine/decitabine can trigger cell cycle exit independently of p53, thus circumventing a basis for resistance to apoptosis-based DNA-damaging therapy. That responses cut across the chaotic genomic landscape of MDS highlights common threads in disease, such as high expression in myeloblasts of differentiation-driving transcription factors yet paradoxical epigenetic suppression of proliferation-terminating late-differentiation genes. Less toxic regimens (lower dosages but more frequent administration) of 5-azacytidine/decitabine have been more successful, underscoring the importance of preserving functionally normal stem cells, which are rendered more precious by attrition from age, previous cytotoxic treatments, and the disease process and are needed to relieve cytopenias, the cause of morbidity and mortality. Also emphasized is that there can be no therapeutic benefit, regardless of mutation or cytogenetic subtype, if DNA methyltransferase is not depleted by sufficient overlap between intracellular drug half-lives and S-phase entries of malignant cells. Improved understanding of mechanism-of-action differences demands new approaches, from historic (but not scientific) more-is-better and one-size-fits-all empiricism to pharmacodynamic-based designs and combinations directed not solely at suppressing malignant clones, but at improving therapeutic indices.