Low Dose Decitabine Versus Best Supportive Care in Elderly Patients with Intermediate or High Risk MDS Not Eligible for Intensive Chemotherapy: Final Results of the Randomized Phase III Study (06011) of the EORTC Leukemia and German MDS Study Groups

Introduction: In 2002 the EORTC and the German MDS Study Group initiated a randomized phase III study comparing low dose Decitabine to supportive care in patients (pts) of 60 years or older with primary or secondary MDS or CMML. MDS patients with either 11–20% BM blasts or ≤ 10% blasts and poor cytogenetics could be included. Pts with a BM blast count between 21–30% without signs of disease progression for at least one month were also candidates for the study.

Methods: Patients were centrally randomized; stratification factors were cytogenetics risk group, IPSS, MDS (primary vs secondary) and study centre, The treatment schedule was 15 mg/m² Decitabine i.v. over 4 hours every 8 hours for the first 3 three consecutive days, of every 6 week-cycle, for a maximum of 8 cycles. Results were evaluated every 2^nd cycle. When a complete remission was reached at least another 2 courses were given. The primary endpoint of the study was Overall Survival. AML free survival, Progression Free Survival (PFS), response rate, toxicity and QoL were secondary endpoints. A total of 185 deaths were required to detect a hazard ratio (HR) of 0.66 (alpha=5%, beta=20%). Intent-to-treat analysis was used.

Results: Between 10.2002 and 5.2007 a total of 233 pts (149 male and 84 female) were recruited from 40 centres. The median age was 70 (60–90 years); RAEB-t was diagnosed in 32% of the pts. Most pts had an IPSS Intermediate-2 (55%) or high risk (38%). Poor risk cytogenetics was found in 46% of the patients. Prior therapy for MDS (not being intensive chemotherapy) was given in 20% of pts. The randomized groups were well balanced regarding stratification factors, age and FAB classification. The median follow up was 2.5 years. Time to Off Study was 180 (Decitabine) vs 112 days (SC arm). The median number of cycles given to the patients was 4 with 40%getting no more than 2 cycles. In a significant number of pts, subsequent treatment, consisting of transplant (10%) or induction chemotherapy (11%), was given.

The distribution of best response in Decitabine vs SC arm was CR (13% vs 0%), PR (6% vs 0%), HI (15% vs 2%), SD (14% vs 22%), PD (29% vs 68%), hypoplasia (14% vs 0%), inevaluable (8% vs 8%). The 18 pts on Decitabine with a HI showed the following responses: 3-lineage (n=7), 2-lineage (n=5) and 1-lineage (n=6). The median time to response (CR/PR/HI) was 0.32 yrs and the response duration was 0.72 years. Median OS was 0.84 (Decitabine) vs 0.71 years (SC arm), estimated HR was 0.88, 95% CI 0.66–1.17, p=0.38 (logrank 2-sided). The PFS was significantly (p=0.004) longer in Decitabine vs SC arm: median was 0.55 vs 0.25 years, HR=0.68 (95% CI 0.52–0.88). Time to AML or Death was not significantly improved (p=0.24): median was 0.73 vs 0.51 years (HR=0.85, 95% CI 0.64–1.12).

Toxicity. The toxicity was mainly cytopenia related toxicity that was either disease related or hematotoxicity; CTC grade 3–4 febrile neutropenia was 26% (Decitabine) vs 7% (SC arm) and Grade 3–4 infection was 59% vs 47%. Differences in non hematologic toxicities were mainly gastrointestinal: grade 1–2 nausea (28% vs 16%) and grade 1–2 vomiting (16% vs 9%). During the study period, 29 (Decitabine) vs 25 (SC arm) patients died: due to either progression to MDS/AML (7 vs 20), toxicity (9 vs 0), progression and/or toxicity (10 vs 1), other reasons (3 vs 4).

Conclusions. Decitabine was found to be an effective drug in these high risk MDS patients with a overall RR of 34%, (similar to earlier studies), leading to a significant PFS improvement as compared to SC arm. The difference Decitabine vs SC arm regarding time to AML or Death was not significant. Due to shorter treatment duration (not being continued beyond 8 cycles) and maybe also due to subsequent treatments administered after disease progression, the difference regarding OS was lower (HR=0.88) and not statistically significant.