For many years there were no approved therapies for patients with myelodysplastic syndromes (MDS). Prior to 2004, only general supportive care, off-label use of medications such as hematopoietic growth factors or thalidomide, clinical trial enrollment, or, rarely, allogeneic hematopoietic stem cell transplantation were available. Then, between 2004 and 2006, three drugs were approved by the Food and Drug Administration for MDS: azacitidine (AZA), lenalidomide, and decitabine. Since then, MDS drug development has entered another era of stagnation from which it has yet to emerge, with no new disease-modifying therapies approved in more than five years.
During this comparatively quiet period, a multicenter trial formally demonstrated that AZA treatment confers a survival advantage compared to supportive care for patients with higher-risk MDS or oligoblastic acute myeloid leukemia (AML with < 30% blasts).1 But this survival advantage is relatively modest (median ~9 months), and AZA is not curative. Although some patients experience an excellent and durable treatment response with AZA, the drug eventually fails, as do all of the other agents now available to treat patients with MDS; the median duration of AZA benefit is less than two years.2
Thomas Prébet and his colleagues in Le Groupe Francophone des Myélodysplasies (GFM) recently underscored the pressing need for new approaches in MDS by analyzing 435 patients with higher-risk MDS or oligoblastic AML in whom AZA had failed. The GFM investigators reviewed outcomes from four patient cohorts: 138 patients enrolled in the AZA001 randomized international study of AZA monotherapy,1 27 participants in Johns Hopkins University-led pilot AZA-combination studies, and 270 patients registered in a French AZA compassionate use program. Types of treatment failure included inability to tolerate AZA therapy (9% of patients), lack of any clinical response (“primary failure”: 55% of patients), or relapse after an initial response (“secondary failure”: 36% of patients).
The median overall survival after AZA failure for these 435 patients was just 5.6 months, and only 15 percent of the patients were still alive two years later. These results are almost identical to an M.D. Anderson Cancer Center review of 87 patients with MDS for whom decitabine failed; that cohort had a median post-failure survival of only 4.3 months.3 In a multivariate analysis, characteristics of the AZA-failing patients who did most poorly in the GFM study included older age, male sex, high-risk cytogenetics, a bone marrow blast count ≥10 percent, and primary rather than treatment failure.
None of the GFM patients who were subsequently treated with decitabine achieved a response, suggesting that switching from one failing hypomethylating agent to another is an ineffective general strategy. For 270 patients from whom information was available on subsequent treatments after AZA failure, the subset of 37 patients (8.5% of the overall cohort) who were young and healthy enough to undergo allogeneic stem cell transplantation had the best outcomes (median survival 19.5 months), followed by those who participated in investigational trials (13-month survival) and those who were treated with low-dose or high-dose chemotherapy (7.3 or 8.9 months median survival, respectively). But the largest subgroup, the 45 percent of patients who received only supportive care, lived just 4.1 months on average. These post-AZA survival differences likely represent selection bias in how clinicians triaged patients for subsequent treatment approaches, yet they also highlight the poor outcomes for most patients in whom AZA treatment becomes ineffective.
In Brief
Several ongoing clinical trials are enrolling patients for whom the hypomethylating agents have failed, including studies of various deacetylase inhibitors and other novel targeted agents, as well as a multicenter randomized trial comparing rigosertib (a PI3 kinase inhibitor) with low-dose cytarabine or supportive care. However, markedly better treatment options will likely emerge only after improved understanding of molecular pathobiology of MDS allows us to move beyond the current empiric approaches.
References
Competing Interests
Dr. Steensma indicated no relevant conflicts of interest.