Decitabine (DAC) is another nucleoside analog with the capacity to induce DNA hypomethylation through the inhibition of DNMT. The mechanism of hypomethylation of DNA is the same as seen in AZA. Since AZA but no other treatment including chemotherapy with anthracyclins and Ara-C showed clinical efficacy for MDS, DNA hypomethylation was thought to be very important for the treatment of MDS. Considering the mode of action of DAC in which 100% of DAC theoretically works as DNMT inhibitor, DAC was expected to show higher efficacy than AZA of which about 20% would be incorporated into DNA to inhibit DNMT. The clinical trials with the middle or the high dose of DAC were conducted initially. However, there was no preeminence with many adverse events as compared with other drug, and then clinic development was discontinued. After, the demethylation activity of DAC drew attention, and clinical trials with low-dose DAC were conducted. Wijermans et al. conducted a phase II clinical study with decitabine for 66 patients of MDS to test the efficacy and toxicity of DAC [25]. DAC was administered at a dose of 15 mg/m² infusion over a 4 h period every 8 h (45 mg/m²/day) for 3 consecutive days. DAC treatment courses were repeated every 6 weeks up to six courses. Primary endpoint was response rate. There were eight cases of RA or RARS, 29 of RAEB, 20 of RAEB-T, and nine CMML cases, and most of them (76%, 50 out of 66 cases) were included in the higher-risk group of IPSS category. Among them, 32 patients (49%) showed clinical response, including CR (n = 13, 19.7%), PR (n = 3, 4.5%), and improvement (n = 16, 24.2%) of cytopenia as defined by the protocol (reduction of at least 50% of transfusion requirements, etc.). Response rates by the risk category of IPSS were as follows: 25% in the intermediate (Int)-1 category, 48% in Int-2, and 64% in high-risk category. The median time of response duration was 31 weeks. These suggested the clinical efficacy of DAC over the risk of MDS predicted by IPSS. The treatment-related mortality (toxic death) was 7.6% (n = 5), and non-hematological toxicity of grade 3–4 was observed at only 13.6% (n = 9). There were 17 cases (26%) of disease progression. These results were comparable or even better than those of AZA trials mainly conducted by CALGB (mentioned above).

Based on these, a randomized trial was performed in which response rate and time to AML or death were compared between DAC with BSC and BSC alone in MDS at 18 years or older [26]. In this study, the dose of decitabine was 15 mg/m² and administered intravenously over 3 h every 8 h for 3 days (at a dose of 45 mg/m²/day and 135 mg/m² per course) and which repeated every 6 weeks depending on the recovery from myelosuppression, almost the same schedule as in the previous phase II study. Response was assessed according to the International IWG criteria [19]. A total of 170 patients with MDS (median age was 70 years old) were randomized to receive either DAC or BSC. Fifty-two patients (31%) were categorized into Int-1 risk group of IPSS, 74 patients (44%) into Int-2 risk group of IPSS, and 43 (25%) in high-risk group of IPSS. Overall response rate was 17% (9% with CR and 8% with PR) in DAC and 0% in BSC, and clinical improvement that included CR, PR, and HI was obtained in 30% of cases in DAC arm and 7% in BSC arm. There were statistically significant differences both in overall response rate and clinical improvement rate (both P < 0.001). Although DAC group showed a trend toward a longer median time to AML or death than BSC group, there was no significant difference between the two groups (12.1 months for DAC and 7.8 months for BSC, respectively, P = 0.16) (Table 19.3). These results led to the approval of DAC for MDS in the USA.

A similar phase III clinical study of DAC for MDS was conducted among European countries by European Organization for Research and Treatment of Cancer Leukemia Group (EORTC) [27]. The administration method of DAC was similar to the prior phase III study, and the primary endpoint was set as overall survival (OS). Two-hundred thirty-three patients were enrolled; 92% of patients had high-risk diseases (IPSS Int-2 and high) with 32% of AML by WHO criteria, and 53% had poor-risk cytogenetics. These patients were randomly assigned to decitabine arm (n = 119) and the BSC arm (n = 114), with a balanced distribution for age, sex, performance status, and risk profile (IPSS subgroup and FAB subtype). DAC (15 mg/m²) was given intravenously over 4 h three times a day for 3 days in 6-week cycles. The maximum number of treatment cycle was limited to eight (or ten in case CR was achieved after eight courses). As the primary endpoint, OS prolongation in DAC arm was not shown over BSC with statistical significance (median OS, 10.1, and 8.5 months, respectively, P = 0.38). Progression-free survival (PFS), defined as time from randomization to progression, relapse after CR or PR or death, prolonged significantly by DAC (1-year PFS rate 27% and 15%, in the DAC and BSC arm, respectively, P = 0.004). There was no statistical difference in AML-free survival (AMLFS, defined as time from randomization to AML transformation or death) between the DAC and BSC arms (median AMLFS, 8.8 versus 6.1 months, respectively, P = 0.24), although AML transformation was significantly (P = 0.036) reduced at 1 year (from 33% with BSC to 22% with DAC). There was no difference in cumulative incidence of death without AML, either between two arms (P = 0.17). Additional analyses suggested the advantage of DAC over BSC in PFS among several subgroups. In terms of toxicity, grade 3 or 4 infections were most frequent AEs followed by gastrointestinal events. Neutropenia and fatigue were also frequently observed, as seen in other DAC studies.