Azacitidine is the first drug to demonstrate a survival benefit for patients with MDS. However, only half of patients respond and almost all patients eventually relapse. Limited and conflicting data are available on predictive factors influencing response. We analyzed 128 patients from two institutions with MDS or AML treated with azacitidine to identify prognostic indicators. Genetic mutations in ASXL1, RUNX1, DNMT3A, IDH1, IDH2, TET2, TP53, NRAS, KRAS, FLT3, KMT2A-PTD, EZH2, SF3B1, and SRSF2 were assessed by next-generation sequencing. With a median follow up of 5.6 years median survival was 1.3 years with a response rate of 49%. The only variable with significant influence on response was del(20q). All 6 patients responded (p = 0.012) but survival was not improved. No other clinical, cytogenetic or molecular marker for response or survival was identified. Interestingly, patients from poor-risk groups as high-risk cytogenetics (55%), t-MDS/AML (54%), TP53 mutated (48%) or relapsed after chemotherapy (60%) showed a high response rate. Factors associated with shorter survival were low platelets, AML vs. MDS, therapy-related disease, TP53 and KMT2A-PTD. In multivariate analysis anemia, platelets, FLT3-ITD, and therapy-related disease remained in the model. Poor-risk factors such as del(7q)/-7, complex karyotype, ASXL1, RUNX1, EZH2, and TP53 did not show an independent impact. Thus, no clear biomarker for response and survival can be identified. Although a number of publications on predictive markers for response to AZA exist, results are inconsistent and improved response rates did not translate to improved survival. Here, we provide a comprehensive overview comparing the studies published to date.
Keywords: azacitidine; hypomethylating agents; myelodysplastic syndromes; response prediction; targeted therapy.