Azacytidine (AZA) In Relapsed MDS and AML After Allogeneic Stem Cell Transplantation (allo-HSCT): Results of the French ATU Program.

Abstract

Abstract Abstract 1293 Background: Relapse is the most common cause of treatment failure after allo-HSCT for myeloid malignancies, and generally carries a very poor prognosis (median survival of 3–4 months without active treatment) (Savani et al., BMT 2009). Those patients (pts) are rarely candidates for chemotherapy due to the limited efficacy, risk of myelosuppression and subsequent loss of transplant. AZA improves survival in higher risk MDS (including pts with 20–29% marrow blasts), with less myelosuppression than chemotherapy (Lancet Oncol 2009). AZA has so far been evaluated in only 9 AML and MDS pts who had relapsed after allo-HSCT (Jabbour et al., Cancer 2009). We report 29 pts treated by AZA for AML or MDS relapsing after allo-HSCT. Methods: An AZA compassionate program (ATU) was opened in France between Dec 2004 and Dec 2008 for higher risk MDS, and for AML not candidates or refractory to intensive chemotherapy. We retrospectively analyzed, in that program, the outcome of AML or MDS pts having received at least 1 cycle of AZA for relapse after allo-HSCT. Results: Our study included 29 pts (M/F: 14/15; median age: 53y, range 22–66). At initial diagnosis, 6 pts had MDS, including RAEB-1 in 1, RAEB-2 in 4 and CMML in 1. IPSS was int-1 in 1, high in 4 and undetermined in 1. Twenty three pts had AML, including (WHO 2008) AML-NOS in 13, therapy related AML in 3, AML secondary to myeloid neoplasms in 6 (3 following MPD and 3 following MDS) and 1 AML with mutated NPM1. 2 pts had 20–29% marrow blasts (AML/RAEB-t). According to European LeukemiaNet (ELN) recommendations, cytogenetic and molecular genetic related prognosis was favorable in 1 pt, intermediate-II in 9, adverse in 11 and undetermined in 2. 14 pts received a myeloablative conditioning and 15 a reduced intensity one. 23 pts received PBMCs transplant (matched unrelated donor in 8, mismatched donor in 1 and related donor in 14), 3 pts BM transplant and 3 pts received cord blood transplant. 5 of 25 evaluable pts had experienced GVHD prior to relapse. Relapse had occurred at a median of 228 days post allo-HSCT (range: 66–1489). Median WBC count at relapse was 2.91 G/L (range: 0.2–6.3). Treatment was according to FDA-EMEA approved schedule (AZA 75 mg/m2/d × 7 d every 4 weeks) in 24 pts (83%) and a less intensive schedule (5d/4w) in 5 pts. With a median follow-up from relapse of 21 months, pts received a median of 3 AZA cycles (range: 1–15). Sixteen (55%) pts received less than 4 cycles because of progressive disease in 8 pts, infectious complications in 4 pts, early death in 2 pts, pt decision in one case, and undetermined reason in 1 pt. With the 5d-schedule, only 1/5 pts had less than 4 cycles (because of infectious complication) compared to 15/24 pts treated with the 7d-schedule. Febrile neutropenia was reported in 63% pts. Only 1 case of GVHD exacerbation was observed. In MDS, overall response rate (ORR, IWG 2006 criteria) was 4/6 pts (66%), including CR in 1, marrow CR in 2 and stable disease with hematological improvement (HI) in 1. In AML, ORR according to ELN criteria was 5/23 pts (22%), including CR in 3, CRi in 1 and PR in 1. Using FAB classification (ie including pts with 20–29% marrow blasts in MDS) ORR was significantly better in MDS than in other pts (75% vs 17%, p=0.015). There was a trend for better ORR with the 5d-schedule vs the 7d-schedule (80% vs 25%, p=0.06). Median response duration was 8 months (range: 3–27). Median overall survival (OS) from relapse after allo-HSCT was 10.5 months, significantly higher in responders (16.8 months) compared with non responders (5.5 months) (p=0.02). There was a trend for better OS in FAB classified MDS (14 months vs 7 months, p=0.13). Progressive disease was the most common cause of death (14/24 died pts). Conclusion: Considering the very poor prognosis of MDS and AML relapsing after allo-HSCT, AZA appears as an interesting therapeutic option, especially in pts with <30% marrow blasts at diagnosis. The 5d-schedule may be associated with a better response rate possibly related to a less toxicity, allowing more prolonged treatment. Further reduction of AZA schedules in this context, already tested by Jabbour et al. (Cancer 2009) should be further explored in relapses after allo-HSCT. Disclosures: Fenaux: CELGENE, JANSSEN CILAG, AMGEN, ROCHE, GSK, NOVARTIS, MERCK, CEPHALON: Honoraria, Research Funding.