Abstract 1696 The standard treatment concept and approach in MDS aims for apoptosis-induction in expanding malignant clones in the hope that this will permit recovery by functionally normal hematopoietic stem cells (HSC). This concept is fundamentally undermined, however, by frequent genetic deletion in MDS/AML cells of key apoptosis genes. Hence, therapy selects for relapse with the most apoptosis-resistant malignant subclones (e.g., p53-null) while destroying crucial normal HSC (which have intact apoptosis-pathways). Observations regarding biology of sustained proliferation (MYC activity) in MDS/AML suggested an alternative approach: key late-differentiation MYC-antagonist genes (e.g., CEBPE) are repressed epigenetically even though lineage-specifying transcription factors (TF) (e.g., CEBPA) that drive expression of these genes are highly expressed. The basis for this paradox has been exposed: MDS/AML associated genetic abnormalities (e.g., RUNX1 mutation) favor recruitment of corepressors (e.g., DNMT1) instead of coactivators to lineage-specifying TF, thereby repressing differentiation target genes. Thus, in pre-clinical MDS/AML models, depleting DNMT1 by non-cytotoxic methods triggers MYC-antagonist expression and irreversible cell cycle exit by p53/p16-independent differentiation pathways (CEBPE-p27). The same treatment spares normal HSC, which unlike AML stem cells, do not express high levels of lineage-specifying TF. To translate these observations, this NIH-supported clinical trial (NCT01165996) incorporated the following novelty: (i) Lower dose of the cytidine analogue decitabine (DAC, Eisai) than in any previous trials (3.5–7 mg/m2) since hemoglobinopathy clinical trials demonstrated that these low doses deplete DNMT1 in normal HSC without cytotoxicity/apoptosis; (ii) SC instead of IV administration to avoid high peak levels that cause apoptosis; (iii) Use of the lack of toxicity to administer drug frequently 1–3X/week, to catch more MDS cells in S-phase via greater exposure times and distribution than in previous MDS trials (FDA-approved schedules only treat MDS cells entering S-phase in the first few days of multi-week cycles); (iv) Scientific correlates to measure non-cytotoxic, epigenetic-differentiation mechanism of action. Sample size = 25 as planned, 10/25 (40%) were relapsed/refractory after 5-azacytidine and/or lenalidomide, median age = 72y (range 46–85y), median disease duration = 900d (range 60–4680d). Anti-emetics were not needed, and cytotoxic side-effects e.g., hair loss did not occur. Neutropenic fever (NF) occurred in 11 subjects (5 had NF and 8 had ANC Differentiation therapy transformed the outlook for M3 AML. Not surprisingly, a worthy goal has been to extend this paradigm. Biology of sustained MYC activity in MDS/AML guided logical repurposing of DAC for non-cytotoxic differentiation therapy even in MDS/AML with complex cytogenetics. Tolerance permits ready application despite age or comorbidities. Treatment exposure time and surrogates of normal HSC reserve are more important predictors of CR/HI than disease-mutations, and some patients may require separate measures (e.g., cytokines) to boost normal HSC diminished by age and previous insults. Disclosures: Reu:Celgene: Research Funding. Maciejewski:NIH: Research Funding; Aplastic Anemia&MDS International Foundation: Research Funding. Saunthararajah:Cleveland Clinic Innovation: patent application for oral THU-decitabine., patent application for oral THU-decitabine. Patents & Royalties. Off Label Use: lower dose of decitabine, subcutaneous administration, metronomic administration.