Evidence for Selective Benefit of Sequential Treatment with Azanucleosides in Patients with Myelodysplastic Syndromes (MDS)

Azanucleosides (AZN) remain the mainstay of therapy in Myelodysplastic syndrome (MDS). Azacitidine (AZA) and decitabine (DAC) have distinct differences in nucleic acid specificity and mechanisms of cellular resistance. The only published experience with sequential treatment described 14 patients, with a 28% response with DAC treatment after failure or lack of response to AZA. The response rate to AZA after DAC is unknown. Sequential use of alternative AZNs is common practice given the limited alternatives. To investigate the potential benefit of this approach, we reviewed cases of sequential AZN treatment.

Patients who received treatment with both AZNs were identified through the Moffitt Cancer Center (MCC) MDS database. Two groups were identified; group one who received DAC after AZA failure, and group 2, who received AZA after DAC failure. AZN failure was defined as lack of response, loss of response or discontinuation due to adverse events or disease progression. The primary objective was to estimate response rates according to the International Working Group (IWG) 2006 criteria.

A total of 39 MDS patients were identified. Complete records were available in 31 patients, including 21 patients in group 1 (DAC after AZA), and 10 patients in group 2 (AZA after DAC). Table-1 summarizes baseline characteristics. In Group 1, 21 patients received DAC after AZA, 57% were originally red blood cell transfusion dependent (RBC TD). The median time from diagnosis to AZA treatment was 10 month (0. 2–52). The mean number of AZA treatment cycles was 8 (1–20). IWG responses to AZA treatment were 2 (10%) CR, 4 (20%) marrow CR (mCR), 7 (33%) hematological improvement (HI), 1 (5%) stable disease (SD) and 6 (28%) progressive disease (PD). Reasons for AZA discontinuation were 7(33%) no response, 1 (5%) progression, 9 (43%) lost response, and 4 (19%) toxicity. The median interval between end of AZA and initiation of DAC was 118 days (21–948). The mean number of DAC cycles was 4 (1–18). IWG responses to DAC were 1 (5%) mCR, 3 (14%) HI, 14 (67%) PD, and 3 (14%) unknown. The best overall response was 19%. The median OS from time of DAC initiation was 17. 8 month (95%CI 14–21). The rate of AML transformation was 29%.

In Group 2, 10 patients received AZA after DAC, and 70% were originally RBC TD. The median time from diagnosis to DAC treatment was 2. 5 month (0–56). The mean number of DAC cycles was 4. 3 (1–9). IWG responses to DAC treatment were 2 (10%) mCR, 4 (40%) HI, 2 (20%) PD. The reasons for DAC discontinuation were no response (n=1), loss of response (n=3), PD (n=2), toxicity (n=3), and physician choice (n=1). The median interval between end of DAC and initiation of AZA was 179 days (19–448). The mean number of AZA cycles was 6 (2–12). The best responses to AZA were 2 (20%) mCR, 2 (20%) HI, 2 (20%) SD, 4 (40%) PD. The best overall response was 40%. The median OS from time of AZA start was 22 month. The rate of AML transformation was 20%.

The median OS for Group 1 from diagnosis was 48 month and for group 2 was100 month (p=0. 7). Table-2 summarizes key differences between the two groups.

Sequential use of AZNs after failure of first line may be an effective alternative outside the context of clinical trials. Despite limitations of cohort size and retrospective nature of the analysis, our findings suggest that response rate may be higher in patients who receive AZA after DAC, which could reflect the dual RNA/DNA drug actions. Sequential use of HMA should be considered in context of randomized clinical trial of novel agent as the control arm.

List:Celgene: Consultancy. Komrokji:Celgene: Honoraria, Speakers Bureau.