Treatment With Decitabine (DAC) After Azacitidine (AZA) Failure In High Risk Myelodysplastic Syndrome (MDS) and Advanced Chronic Myelomonocytic Leukemia (CMML)

Outcome of patients with high risk MDS and CMML who failed treatment with AZA remains poor with a median survival of 5-6 months (Prebet; JCO 2011 29:3322). No established therapy is available for the majority of those patients at this stage except allogeneic bone marrow transplantation in a few eligible patients. The rationale for DAC treatment after AZA failure is based on somewhat different pharmacological properties between those 2 hypomethylating agents (HMAs) (Hollenbach; Plos One 2010 5:e9001). Conflicting results for DAC salvage after AZA failure have been reported with 0-28% response rates (RR) in small patient series, and an overall survival up to 11 months (Borthakur; Leuk Lymphoma 2008 49:690; Prebet; JCO 2011 29:3322; Bhatnagar, ASH 2012 Abstr #3858). We retrospectively reviewed a larger cohort of 36 patients in this setting.

Characteristics, overall response rate (ORR) and outcome were studied in high risk MDS and CMML patients who received DAC after AZA failure from June 2007 to April 2013 in three French hematology departments. Response criteria were based on IWG 2006 for MDS and CMML with WBC <13G/L and also included for CMML with WBC >13 G/L evolution of WBC, splenomegaly (SMG) and extramedullary disease (EMD) (based on Wattel; Blood 1996 88:2480).

Median age of the 36 patients was 70.5 years (range 53-84), and M/F: 21/15. Median time from diagnosis to AZA onset was 5.7 months and all patients at AZA onset had IPSS ≥ int-2 MDS/CMML or CMML-2 (EU label for AZA). Median number of cycles of AZA received was 8 (3-41) and 8 patients had received less than 6 cycles (2-5) including 1 who received HSCT, 1 with progressive disease (PD) and 6 in whom AZA was considered ineffective after less than 6 cycles. Overall, 20 patients had primary AZA failure, 1 PD and 15 had relapsed after achieving CR (n=9), PR (n=1), mCR (n=1), mCR+HI (n=2) or SD+HI (n=2), 3 of whom had been allografted. Apart from the 3 last patients, responders were still receiving AZA when relapse occurred. Median number of treatments received after AZA and before DAC was 0 (range 0-3), including low dose cytarabine (n=5), intensive chemotherapy (n=5), clofarabine (n=5) and rigosertib (n=3). Twenty patients received DAC immediately after AZA failure and median time from AZA to DAC treatment for all patients was 3.6 months. At onset of DAC, 22 patients had AML post MDS (including 11 RAEB-t), 8 had RAEB-2, 1 RAEB-1, 1 CMML-2 and 4 CMML-1.Median marrow blast count was 23% (range 7-82). Karyotype was normal in 15 (41.7%) patients, 9 (25%) patients had unfavorable cytogenetics including 7 complex karyotypes and 2 monosomy 7, five (13.9%) patients had +8, three (8.3%) del 20q, 3 (8.3%) other anomalies and 1 (2.8%) cytogenetic failure. IPSS-R was intermediate for 3, high for 11 and very high for 6 patients and could not be evaluated for 16 patients, mainly those with overt AML. 2 CMML patients had SMG and 1 skin involvement. Median number of DAC cycles administered was 3 (1-27) with 12 patients receiving at least 6 cycles. Seven patients (19.4%) were responders to DAC according to IWG 2006 criteria including 3 marrow CR, 2 stable disease (SD)+HI-E, 1 SD+HI-P and 1 SD+HI-N. In a CMML patient with SD, specific skin lesions resolved upon treatment with DAC and a patient with marrow CR underwent HSCT. Median OS from onset of DAC was 7.3 months without significant difference between responders and non-responders. Responses were short lived (2.5-6 months) with 2 responders currently ongoing with short follow up (2.5+ and 3+ months respectively) and 2 non-responders had prolonged SD (without HI) of 21 and 27 months duration respectively. Of note, 6 of the responses were seen in patients who did not receive DAC immediately after AZA.

Treatment of DAC after AZA failure yielded modest ORR (19.4%) and no CR was achieved in this patient cohort. Responses were generally short lived but 2 patients had prolonged stable disease for 21 and 27 cycles respectively. OS remained poor. Those results do not support absence of cross resistance between AZA and DAC, at least in most higher risk MDS.

No relevant conflicts of interest to declare.