First Clinical Results of a Randomized Phase 2 Dose-Response Study of SGI-110, a Novel Subcutaneous (SC) Hypomethylating Agent (HMA), in 102 Patients with Intermediate (Int) or High Risk (HR) Myelodysplastic Syndromes (MDS) or Chronic Myelomonocytic Leukemia (CMML)

Background: SGI-110 is a second generation HMA formulated as a dinucleotide of decitabine (DAC) and deoxyguanosine delivered as a small volume SC injection that yields longer half-life and more extended DAC exposure than DAC IV infusion. A phase 1 clinical trial of SGI-110 demonstrated a differentiated pharmacokinetic profile from DAC IV, potent hypomethylation, and clinical responses in previously treated patients with MDS and AML (Kantarjian et al. 2012). Phase 1 established 60 mg/m² QDx5 as the biologically effective dose (BED), and 90 mg/m²QDx5 as the Maximum tolerated dose (MTD) in MDS patients given in 28-day cycles. Phase 2 is conducted to evaluate dose response between the BED and MTD in MDS patients.

Methods: Int, or HR MDS, and CMML patients who were either treatment-naïve or previously treated were randomized to either 60 mg/m² or 90 mg/m²QDx5 every 28 days. Efficacy was evaluated by the clinical responses of CR, PR, marrow CR (mCR), and Hematological Improvement (HI) based on the International Working Group Criteria 2006 as well as transfusion-independence. Adverse events (AEs) as graded by the CTCAE v4 criteria and pharmacodynamic biological activity as measured by Long Interspersed Nucleotide Element LINE-1 (an index of global DNA methylation) were also assessed. Patients are being followed for overall survival.

Results: The study completed target enrolment with 102 patients: 53 who were refractory or have relapsed on prior treatment, and 49 previously untreated. Fifty three patients were randomized to 60 mg/m² and 49 patients to 90 mg/m² QDx5 with a median follow up of 8.2 months (range 1-21). Most baseline patient characteristics were well balanced between the 2 treatment arms with no significant differences in median age (71.7 vs. 72.5 y); male gender (70 vs. 61%); ECOG PS 2 (15 vs. 12%); HR MDS (28 vs. 37%); transfusion-dependence (57 vs. 55%), median baseline Hb (9.25 vs. 9.3 g/dL); platelets count (42.5 vs. 45 x10⁹/L), and neutrophils count (1.19 vs. 1.16 x10⁹/L) for the 60 and 90 mg/m² QDx5 respectively. However, there were 15 CMML patients (28%) randomized to 60 mg/m² vs. only 7 CMML patients (14%) randomized to 90 mg/m²(p=0.097). All but 2 patients in the previously treated group received one or more HMA.

At the time of the data cutoff, 38 of 102 patients (37%) were still on treatment. Median number of treatment cycles was 4 for previously treated patients, and 5 for the treatment naïve group (range 1-22 cycles). CR+mCR were observed in 10/53 (19%), and 11/49 (22%) in the 60 and 90 mg/m² arms respectively (p=0.8). CR was observed in 7/49 treatment naïve patients (14%) while CR+mCR were observed in 11/53 previously treated patients (21%) with no significant differences between 60 and 90 mg/m² arms. Transfusion independence for at least 8 weeks was reported in 32% and 24% for platelets and RBCs respectively with no difference between the 2 treatment arms. Treatment naïve patients benefited from higher rate of transfusion-independence (58% and 46% for platelets and RBCs respectively). Potent demethylation was achieved in both treatment arms with a mean LINE-1 DNA demethylation of 27.3 and 30.5% for 60 and 90 mg/m² respectively (p=0.12). Overall incidence of Grade ≥3 AEs regardless of relationship to treatment was reported in 81 vs. 88% for 60 and 90 mg/m² treatment arms respectively (p=0.42).There was a slightly higher but non-significant difference in Grade ≥3 thrombocytopenia (51 vs 38%) and pneumonia (20 vs. 13%) for the 90 mg/m² arm compared to 60 mg/m². Early 8-week any-cause mortality occurred in 3/102 patients treated (3%), 2 at 60 and 1 at 90 mg/m². Follow up for survival is still ongoing.

Conclusions: SC SGI-110 is a well-tolerated novel HMA with biological and clinical activity in the treatment of Int and HR MDS, and CMML patients with particularly promising activity in patients previously treated with azacitidine or decitabine. Clinical responses, transfusion independence, DNA demethylation as assessed by LINE-1, and Safety did not significantly differ between the 2 treatment doses.