A Phase II Study of Post–Remission Therapy with Azacitidine (AZA) in Patients with AML Post-MDS and High-Risk MDS: A GFM Group Study.

intensive chemotherapy (IC) in AML post-MDS (here AML) or High Risk-MDS (here MDS) is associated with lower CR rate, shorter response duration and poor overall survival compared to de novo AML. Upfront treatment of MDS with AZA delays time to AML evolution and prolongs overall survival (OS). The low toxicity profile of AZA should also allow its use in IC responders in an outpatient setting. Preliminary results with this approach by the Nordic MDS group in 23 pts (ASH 2008, abstr n°223) showed its feasibility. We used it in AML or MDS patients (pts), who had achieved response after IC.

Inclusion criteria were: (1) de novo MDS patients and IPSS int-2 or high risk (including CMML if WBC <13G/l), or AML preceded by a documented MDS phase (2) who had reached CR, CRi or PR after IC (IWG-AML criteria) with anthracyclin and AraC in the 28 days preceding study entry (3) ECOG ≤2, absence of ongoing infection or renal and liver toxicities from IC. (4) with no identified donor for allo SCT at entry. Dosing of AZA was 60mg/m²/d for 5 days, every 28 days for the first two cycles. In the absence of grade ≥ 3 toxicities, dosing of subsequent cycles could be increased to 75mg/m²/d. Dosing of AZA had to be reduced in case of grade ≥3 toxicities, persistent cytopenias beyond day 42 of each cycle, or if cytopenias were associated with re-hospitalization. AZA post-remission therapy was to be continued until relapse (according to IWG-AML and MDS criteria) or toxicity.

From July 06 to June 09, 51 pts (M31/F20) were included but 5 were not evaluable (2 IWG failures upon review, 1 protocol violation, 1 therapy-related MDS, 1 patient too early). The 46 evaluable pts had achieved CR (n=33), CRi (n=11) and PR (n=2) before study entry. Median age was 66y (55-78). Diagnosis at IC onset was MDS (N=13) and AML (N=33), IPSS cytogenetics was normal (N= 28), intermediate (N=10), high (N= 6), and failed (N=2). Median time from diagnosis of MDS to IC was 8 months (0.5-101). Median number of AZA maintenance cycles was 5.5 (0-32+)in CR pts (>23 cycles in 4 of them) and 2 (0-11) in CRi or PR pts. Median follow-up was 24 months (m). Median DFS was 6.5 m and 18-m DFS was 20% (95% CI 9-35); median and 18 month OS from response were 15 m and 43% (95% CI 27-58), respectively. 18-m-DFS and OS were 30 % (95% CI 13-48) and 56% (95% CI 35-72), in the 33 CR pts (median OS 24m), compared to O% and 17% (95% CI 3-41), in the 11 PR/CRi pts, ( P= .0001 and .04, respectively) . 21 of the 33 CR pts had relapsed and 3 had CR duration >24 months. None of the pts in CRi/PR after IC improved their response with AZA. Only initial response to IC (ie CR vs CRi/PR), but not cytogenetics (normal vs. others), diagnosis at IC onset (MDS vs AML), % bone marrow blasts (using 20 or 30 % cut-off), age or time from MDS diagnosis to treatment, significantly predicted DFS or OS in this cohort. AZA dosing in CR patients was escalated in 9 pts but had to be reduced in 6 other pts, due to GI tract intolerance (n=1) and cytopenias (n=5). Only 2/33 CR pts, developed severe infection or febrile neutropenia while in persistent response, compared to 4/11 pts in CRi or PR (including 1 fatal case). We finally compared outcome of the 22 AML study pts in CR to that of 46 similar AML post MDS pts from our ALFA group elderly AML 9803 study (Gardin, Blood 2007), in which pts in CR received DNR/IDA–AraC post-remission therapy. Cumulative incidence of relapse did not differ between the present AZA cohort in AML patients and the ALFA cohort (18-m CIR at 41% (95% CI 18-63) and 25% (95% CI 13-40), respectively; P=0.92). OS from CR was also identical between the 2 cohorts (18-m OS of 56% (95% CI 31-76) and 54% (95% CI 37-69), respectively; P=0.84)

In this study, post–CR therapy with AZA alone was associated with a 30% DFS rate and 56% OS at 18m and a median OS of 24m, while pts in CRi or PR did not benefit from AZA. Although selection biases may exist between cohorts, DFS and OS of AML post MDS in CR were identical to those observed in similar pts treated after CR by anthracyclin and AraC consolidation. Hematological toxicity of AZA, however, was minimal. Combinations of AZA with other drugs may be of further interest in those patients at high–risk of relapse.

Gardin:Celgene: Consultancy, Honoraria, Research Funding. Off Label Use: azacitidine as post-remission therapy in AML patients. Dombret:Celgene: Consultancy, Honoraria, Membership on an entity's Board of Directors or advisory committees, Speakers Bureau. Fenaux:Celgene: Honoraria, Research Funding; Roche: Honoraria, Research Funding; Ortho Biotech: Honoraria, Research Funding; Amgen: Honoraria, Research Funding; Cephalon: Honoraria, Research Funding; Merck: Honoraria, Research Funding; Novartis: Honoraria, Research Funding.