Low-dose oral decitabine and cedazuridine among patients with low-risk myelodysplastic syndromes

Abstract Background: Parenteral DNA methyltransferase inhibitors (DNMTi) such as azacitidine and decitabine (DEC), including an attenuated 3-day regimen, are recommended in the NCCN Guidelines for patients (pts) with lower-risk myelodysplastic syndromes (LR-MDS) with clinically significant cytopenias. However, parenteral administration requires frequent clinic visits, which imposes a substantial burden on pts, particularly those requiring long-term therapy. Oral DMTIs offer more flexible dosing and maintenance of pts' autonomy. Oral DEC-cedazuridine (DEC-C) is a fixed-dose combination tablet containing DEC (35 mg) and cedazuridine (100 mg). ASTX727-03 (NCT03502668) is a randomized, open-label, multicenter, phase 1/2 clinical trial evaluating safety, efficacy, and pharmacokinetics (PK) of various DEC-C doses and regimens in pts with LR-MDS. Phase 1 identified a 5-day regimen of DEC 10 mg/cedazuridine 100 mg as the optimized low-dose (LD) regimen. Phase 2 compares this LD regimen with the attenuated standard-dose (SD) regimen. Here, we report phase 2 results on efficacy, safety, PK, and pharmacodynamic (PD) data. Methods: Adults (aged ≥18 years) with International Prognostic Scoring System (IPSS) low-risk or Intermediate-1 MDS, ECOG PS 0−2, and ≥1 cytopenia or red blood cell (RBC) transfusion dependence were randomized 1:1 to receive oral DEC-C LD (10 mg x 5 days) or SD (35 mg x 3 days). The primary endpoint was hematologic response. Secondary endpoints included PK, PD (by LINE-1 demethylation), time to bone marrow blasts >5%, leukemia-free survival (LFS), overall survival (OS), and safety. Results: As of October 31, 2024, 81 pts were treated (LD arm: n=40; SD arm: n=41) and completed a median of 8.8 months on treatment. At cutoff, 17 (42.5%) and 12 (28.6%) LD and SD pts, respectively, were still continuing the study; 9 (22.5%) and 6 (14.6%) pts respectively, had discontinued treatment due to hematopoietic stem cell transplant. Median age was 70 (LD) and 75 (SD) years; 75.0% (LD) and 68.3% (SD) had IPSS Int-1 risk MDS and 60.0% vs 41.5% had IPSS-Molecular Lower-Risk MDS (very low, low, or moderate-low). Low-risk prognostic scoring system risk distribution was similar: Category 1 (2.5% vs 4.9%), Category 2 (47.5% vs 46.3%), and Category 3 (50.0% vs 48.8%). After a median follow-up of 29.3 months, median OS was 23.9 months (95% confidence interval [CI]: 13.0, not estimable [NE]) in the LD arm vs 26.0 months (95% CI: 19.4, 28.5) in the SD arm. Median LFS was 23.8 months (95% CI: 13.0, NE) and 25.7 months (95% CI: 17.8, 26.5), respectively. Hematologic improvement per International Working Group 2006 criteria was achieved in 27.5% (LD) and 26.8% (SD) of pts. Among RBC transfusion–dependent pts, 52.4% (LD) and 37.5% (SD) achieved ≥56-day independence and 47.6% (LD) and 20.8% (SD) achieved ≥112-day independence. Pts received a median of 10 (LD) and 9 (SD) cycles. Delayed cycles occurred in 72.5% (LD) vs 82.9% (SD) of pts; dose reductions occurred in 40.0% (LD) and 46.3% (SD) of pts. Both treatment regimens caused a decrease in blood counts, reaching a nadir before recovering in each cycle. Neutropenia was more pronounced during early cycles and was more severe in the SD arm. Blood counts across all lineages remained stable or improved through ≥12 cycles with the LD arm, suggesting a more favorable safety and tolerability profile. Grade ≥3 adverse events (AEs) occurred in 85.0% and 90.2% of pts in the LD and SD arms, respectively; treatment discontinuation due to AEs occurred in 2.5% and 17.1% of pts, respectively. Most commonly reported treatment-emergent AEs were anemia (42.5% LD, 39.0% SD), fatigue (32.5% LD, 43.9% SD), and thrombocytopenia (37.5% LD, 39.0% SD). Three deaths occurred in the trial: 2 in the LD arm, which were both unrelated to study treatment, and 1 in the SD arm (Pseudomonal bacteremia in Cycle 1), which was treatment related. Total AUC0–24 was 199 ng*h/mL in the LD arm and 465 ng*h/mL in the SD arm; AUC for all cycles of the LD arm was approximately half of the SD arm. Maximal % LINE-1 demethylation rate was lower with LD than SD, but narrowed by Cycle 3. Conclusion: LD oral DEC-C demonstrated comparable clinical benefit to SD oral DEC-C, with improved safety and tolerability, including fewer dose modifications and less severe myelosuppression. These findings support LD DEC-C as a potential optimal regimen for pts with LR-MDS. Clinical trial registration: NCT03502668