7074 Background: Hypoxia is prevalent in leukemia bone marrow (BM) microenvironment, suggesting its role as a therapeutic target. PR104 is a pro-drug activated by hypoxia-dependent reductases and by hypoxia-independent aldo-ketoreductase 1C3 (AKR1C3). Methods: Patients (pts) with relapsed/refractory AML (n=40) after 1 or 2 prior treatments; or ALL (n=10) after any number of prior treatments received PR104 as a 1-hr i.v. infusion q 2 weeks. Biomarkers for hypoxia and AKR1C3 were assessed. Results: Pts received PR104 at doses of 1.1 (n=6), 1.6 (n=1), 2.2 (n=1), 3 (n=20) and 4 gm/m2 (n=22) for a median of 1 cycle (range, 1 – 3 cycles). 17 pts had PR104 doses assigned using an adaptive method using toxicity-response trade-offs, patient age, and prior remission duration, and 33 pts were assigned doses by investigators for cohort expansion at 3 and 4gm/m2. Most common treatment-related grade 3/4 adverse events included myelosupression (anemia 44%, neutropenia 56%, thrombocytopenia 52%), febrile neutropenia (22%), infection (22%) and enterocolitis (14%). 3 (14%) PR104-related deaths occurred at the 4 gm/m2 dose level: hepatic failure (n=1), enterocolitis (n=1) and pneumonia (n=1). 49 of 50 pts were evaluable for response. No CRs were seen at doses < 3 gm/m2. 10 of 31 AML pts (32%, (0.19, 0.5)% CI) and 2 of 10 ALL pts (20%, (0.06,0.51)% CI) at 3 or 4 gm/m2 had responses (CR, n=1; CRp, n=5; morphologic leukemia-free state (MLFS), n=6). In AML pts with 1 prior treatment, responses were seen in 7 of 21 pts (CR, n=1; CRp, n=3; MLFS, n=3). Median overall survival of all pts treated at 3 and 4 gm/m2 was 72 days, and of pts who achieved CR/CRp/MLFS 143 days. 3 pts with AML (MLFS, n=2; CRp, n=1) and 1 with B-ALL (CRp, n=1) underwent allogeneic stem cell transplantation. Biomarker studies showed hypoxia in the BM; and levels of AKR1C3 in leukemic blasts did not correlate with responses (Benito et al., ASH 2012). Conclusions: PR104 administered at doses 3 to 4x the solid tumor MTD had moderate toxicity (most commonly myelosuppression and enterocolitis) in pts with relapsed/refractory AML and ALL. Evidence of activity supports continued evaluation of hypoxia-activated cytotoxins in acute leukemias. Clinical trial information: NCT01037556.