EHA Library - The official digital education library of European Hematology Association (EHA)

The optimal treatment decision in older patients (pts) with AML remains controversial, especially in patients pts with comorbidities, non-fit to intensive therapy or with AML adverse biologic features. Recently decitabine was approved in Italy in AML pts unfit to chemotherapy aged > 65 years (y) and could be adopted in a population based setting.

To evaluate efficacy and toxicity of decitabine in a consecutive series of elderly AML pts (no M3), considered unfit to chemotherapy (CT) according to Ferrara et al (Leukemia, 2013) and treated at 6 centers of the Hematological Lombardy Network (REL).

Between Dec 2015 and Dec 2016, 46 (F/M: 22/24) newly diagnosed AML pts received decitabine (20mg/mq/daily for 5 days every 4 weeks) as outpatients. Median age was 76 y (69-85), ECOG performance status (PS) was >3 in 10.8%. According to “fitness”, 41 pts (89.1%) were defined unfit to intensive CT, 1 frail and 4 fit. Unfitness causes were age >75y (58.5%), PS ECOG>3 unrelated to leukemia (12.2%), and comorbidities (29.3%). AML was “de novo” in 25 pts (54.3%), therapy-related in 3 and secondary to antecedent hematological disorders in 18 pts. WBC count at diagnosis was 4.4 x10^3/mL (0.46 to 63), marrow blasts were 51% (<30% in 19.5% of pts). Karyotype (K) was normal (NK) in 43%, t(8;21) in 4.5%, intermediate in 20.5%, adverse (adv) in 32% of pts, according to ELN (Doehner, 2017). In 2 pts it was not evaluable. Molecular analysis was available in 17/19 NK, NPM1 was mutated in 5 cases, with (2) or without (3) FLT3-ITD mutation.

The total number of cycles administered was 231 (median 3.5; range 1-20). In 37/46 evaluable pts (2 ongoing, 1 early and 6 aplastic deaths), overall response rate (ORR) and complete remission (CR) rate were 51.% and 32%, respectively. Partial response and hematological improvement were achieved in 5.5% and in 13.6%, stable disease in 29.9% and failure in 19% of pts, respectively. Median time to best response was 3.5 months (range 1-8.5). Median response duration was 5.3 months (1-18+ ms). Relapse/disease progression was observed in 42% of responders. ORR was 21.4%, 47.3% and 77% in adv, NK and intermediate K, respectively (P=0.0289). After a median follow-up of 6.5 months, median survival was 8.4 months and projected OS at 1 y and 2 y was 43%+/-9 (SEM) and 30%+/-12% (SEM). Treatment was fairly well tolerated except for a high incidence of infections (46 episodes in 231 cycles) particularly during the first 3 cycles (29% vs 11%) (p 0.0072). Pneumonia was the most frequent infection (46%), followed by sepsis (28%). It was more frequent during the first 3 cycles [14% vs 4% : p 0.012)] when 44% of cases were of suspected fungal origin (3 probable aspergillosis and 4 possible IFI). Death occurred in 24 pts (52.2%): 12 (50%) of disease progression, 1 of early CNS hemorrhage and 11 (45.8%) of infection. In the first 3 months, infections were responsible for 46.7% of deaths. Pulmonary IFI were fatal in 57% of cases. These figures are higher than those reported by Cashen (JCO 2010) where the frequency of pneumonia was 11%.

These preliminary data confirm, in a population based setting, the high efficacy of decitabine and its longer time to response (more than 3 cycles) compared to CT. However infections complications were more frequent than expected and often fatal, particularly early during treatment. Since pneumonia, especially IFI, was the major cause of death, the adoption of routine antimicrobial prophylaxis may be considered in order to reduce early mortality and further improve the results.