Azacitidine (AZA) Treatment Prolongs Overall Survival (OS) in Higher-Risk MDS Patients Compared with Conventional Care Regimens (CCR): Results of the AZA-001 Phase III Study.

Abstract

Abstract Background: A previous CALGB trial (JCO2002;20:2429) showed a positive OS trend with AZA vs best supportive care (BSC) in MDS. The objective of this Phase III, international, multicenter, randomized, prospective trial was to demonstrate the superiority of AZA + BSC for prolonging OS vs CCR + BSC. Design: Higher-risk MDS patients (pts), FAB-defined as RAEB, RAEB-T, or CMML (10–29% marrow blasts) with an IPSS of Int-2 or High by central pathology/cytogenetic review, were enrolled. Before randomization, investigators preselected pts to 1 of 3 CCR: BSC only (transfusions, antibiotics, and G-CSF for neutropenic infection); low-dose ara-C (LDAC, 20 mg/m2/d x 14d, q 28d); or standard chemotherapy (Std CT: conventional induction/consolidation). Pts were stratified by FAB/IPSS and randomized to AZA (75 mg/m2/d x 7d, q 28d) or CCR. This trial did not allow erythropoietin. All analyses used the ITT population. Results: In all, 358 pts (70% male), were randomized at 79 sites to AZA (N=179) or CCR (N=179): BSC only (N=105, 59%), LDAC (N=49, 27%), or Std CT (N=25, 14%). Median age was 69 yrs (38–88) and per treatment (TX): AZA (69 yrs); BSC only (70 yrs); LDAC (71 yrs); and Std CT (65 yrs). The AZA and CCR groups were comparable for baseline (BL) parameters. At BL, 95% of pts were higher risk: RAEB (58%), RAEB-T/WHO AML (34%), CMML (3%), and other (5%). By IPSS, 87% were higher risk: Int-2 (40%), High (47%), and 13% indeterminate/other. AZA was administered for a median of 9 cycles; LDAC for 4 cycles. Median followup for the OS analysis was 21.1 months (mo). AZA demonstrated statistically superior OS vs CCR (stratified log-rank p=0.0001). AZA showed a median Kaplan-Meier (KM) OS time of 24.4 mo vs 15 mo with CCR (Figure). The hazard ratio (HR, Cox Model) was 0.58 (95% CI: 0.43, 0.77) for a 74% OS improvement. At 2 yrs, there was a 2-fold OS advantage: AZA (51%) vs CCR (26%), 95% CI: 13, 36%, p<0.0001. Differences in OS KM medians (HR; log-rank p) between AZA and BSC, LDAC, and Std CT, respectively, were 12.9 mo (0.55; p=0.0003), 9.1 mo (0.60; p=0.016), and 8.7 mo (0.69; p=0.19). Median OS per IPSS cytogenetic subgroup showed similar results (Table). The 1, 2, and 3-mo survival rates did not differ between AZA and BSC only (p>0.20). AZA was well tolerated with safety data consistent with previous reports. Conclusion: These data confirm and extend previous CALGB findings. This AZA trial is the first MDS clinical study to demonstrate a significant OS advantage, thus altering the natural disease course. AZA should now be considered first-line therapy for higher-risk MDS pts. OS Analyses per IPSS Cytogenetic Group Group % (n/N) Pts AZA Median (Months) CCR Median (Months) HR (95%CI) Log-rank p Good 46 (166/358) Not reached 17.1 0.61 (0.39, 0.96) 0.030 Intermediate 21 (76/358) 26.3 17.0 0.43 (0.21, 0.88) 0.017 Poor 28 (100/358) 17.2 6.0 0.52 (0.32, 0.87) 0.011 Figure Figure