Decitabine improves outcomes in older patients with acute myeloid leukemia and higher blast counts

The median age at diagnosis of patients with acute myeloid leukemia (AML) is 66 years, and most patients are > 60 years of age 1. Due to comorbidities and adverse-risk disease, many older patients may not benefit from intensive chemotherapy 2, 3. Hypomethylating agents such 5-azacytidine (AZA) and decitabine (DAC) have been used as “lower-intensity” therapy in older patients with AML and may be associated with outcomes similar to intensive therapy 4, 5. In a post hoc analysis of a randomized phase III trial of AZA vs. conventional care regimens (CCR) in patients with higher-risk myelodysplastic syndrome (MDS) and 20–30% blasts, AZA produced a response rate of 18% and was associated with an improvement in overall survival (OS) compared to CCR (24.5 vs. 16 months, P = 0.001) 6. Subsequent reports have confirmed the activity of AZA in older patients with AML and in those unfit for intensive chemotherapy 4, 7-9. A follow-up trial (AML-001) sought to address the question of efficacy for AZA in older patients with BM blasts ≥ 30% 10. Four hundred eighty-eight older patients with AML with BM blasts ≥30% and WBC < 15 × 109/L were randomized to treatment with AZA or CCR (CCR = best supportive care [SC], low-dose cytarabine [LDAC], or intensive chemotherapy [IC]). While the rates of complete remission (CR)/CR with incomplete recovery of counts (CRi) were similar between the two groups: 28% for AZA vs. 25% for CCR, there was a trend toward improvement in OS for patients on the AZA arm, compared to CCR (median 10.4 vs. 6.5 months, HR = 0.84; P = 0.08) 10. After applying a preplanned censoring of patients at the time of subsequent therapy, AZA was associated with a significant OS benefit (12.1 vs. 6.9 months, HR = 0.75; P = 0.01). DAC was studied in a randomized phase III trial (DACO-016) compared to treatment choice (TC: SC or LDAC) in older patients with newly diagnosed AML 11. The primary endpoint was OS. The study did not exclude patients with higher BM blast counts or higher WBC counts. Among 485 patients (median age of 73 years), DAC therapy resulted in a CR/CRi rate of 17.8% compared to 7.8% for TC. This translated into an improved OS of 7.7 vs. 5 months for the DAC arm, which did not reach statistical significance in the primary analysis (HR 0.85, P = 0.11), but did achieve significance in a later unplanned analysis with 446 deaths (HR 0.82, P = 0.037) 11. Several single-arm studies have further confirmed the activity of DAC in AML 12-14. Since the DACO-016 trial 11 provided a prospective, randomized data set in older patients with AML, we sought to examine the efficacy of DAC in patients with higher BM blasts. For this unplanned post hoc analysis, we selected the subset of patients with BM blasts ≥ 30% and WBC < 15 × 109/L. These criteria were chosen to define a subset of patients with disease characteristics similar to those described in a similar prospective data set in the AML-001 trial. We analyzed the clinical characteristics and outcomes of these patients and compared them alongside available data from the AML-001 trial. The OS data is based on the mature data set of DACO-016 with the clinical cutoff in October 2010, which had 446 deaths (92%) and a median follow-up of 30.7 months. All patients provided written informed consent and the trial was conducted in accordance with the Declaration of Helsinki. DAC was administered at the standard dose. TC was chosen prior to randomization (SC or LDAC). Further details of the trial have been previously described 11. Clinical characteristics were summarized using medians and percentages. Time-to-event variables were described using the Kaplan–Meier method. Hazard ratios and 95% confidence intervals were calculated using a Cox proportional hazards model stratified by age, cytogenetic risk, and ECOG performance status. Of the 485 patients treated on the DACO-016 trial, 271 (56%) met the criteria to be analyzed for the current study: 127 (47%) received DAC and 144 (53%) had TC. Patient characteristics are summarized in Supporting Information Table 1. For comparison, available data from the AML-001 trial are also included, acknowledging the possible dissimilarities between patient populations between the two trials. The median age of patients receiving DAC or TC was 73 years, with 35% and 44% of patients being ≥ 75 years of age in each group, respectively. Patients on the AML-001 trial were older (median age 75 years), had a higher median blast count (70–74%), a similar percentage of patients with an adverse karyotype (34–35%), a similar median WBC (2.3–3.1), and fewer patients with secondary AML (15–20%). There is an additional important difference to consider between the two trials. The options for treatment in the AML-001 trial included AZA, LDAC, SC, or IC, whereas the DACO-016 study did not include IC. Knowing the options at the time of randomization could affect the patient selection for enrollment onto either trial. For DACO-016, fewer “fit” patients (suitable for IC) would have been selected for enrollment. Conversely, more “fit” patients may have been selected for enrollment on the AML-001 trial. Indeed, 18% of patients randomized to the CCR arm of AML-001 were deemed suitable for IC prior to randomization. This difference in the number of “fit” patients and other unmeasurable factors could account for the difference in median OS in the control arms of AML-001 and the current analysis of the DACO-016 trial (6.5 vs. 4.7 months, respectively). Within this subgroup of DACO-016, the CR/CRi rate was 27% for patients receiving DAC vs. 11% for those on TC. This was associated with a significantly improved OS for patients receiving DAC (median 8.6 vs. 4.7 months, HR = 0.67; P = 0.0033) (Fig. 1). This represented an absolute improvement of 3.9 months or a relative improvement of 83% over TC. Applying a preplanned censoring at the time of subsequent disease-modifying therapy, DAC maintained a significant OS advantage compared to TC (9.3 vs. 5 months, HR = 0.69; P = 0.011). The relapse-free survival was 9.5 vs. 7.7 months and the 30-day mortality was 10% vs. 6% for DAC vs. TC, respectively. These results compare favorably with the outcomes reported on the AML-001 trial. DAC was associated with higher rates of grade 3/4 hematologic toxicities compared to the TC arm. It was more difficult to compare hematologic toxicities between AZA and CCR, given the variability of treatment-intensity on the CCR arm. However, relative to DAC, comparing across studies, AZA was associated with lower rates of anemia, thrombocytopenia, and febrile neutropenia (Supporting Information Table 1). Overall survival (Kaplan–Meier method) of a subgroup of patients from the DACO-016 trial with white blood cell count < 15,000 and bone marrow blasts > 30%. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.] Based on a subset analysis of older patients with higher blast counts treated on a phase III study, DAC appears to be effective and confers a significant OS benefit. Although several measured patient characteristics were similar between this cohort and those patients treated on the AML-001 trial, widely different outcomes in the control arms of both trials highlight the unmeasurable factors that make these comparisons difficult. Nonetheless, similar to AZA, DAC confers improved outcomes over TC in older patients with higher-blast-count AML and is an important treatment option. TMK and HK collected, analyzed, and interpreted the data. TMK, XGT, AD, AW, MM, CA, JD, FR, and HK contributed patients for the study. TMK and HK wrote the manuscript with input from the other authors. All authors approved the final manuscript. Tapan M. Kadia,1* Xavier G. Thomas,2 Anna Dmoszynska,3 Agnieszka Wierzbowska,4 Mark Minden,5 Christopher Arthur,6 Jacques Delaunay,7 Farhad Ravandi,1 and Hagop Kantarjian1 1Department of Leukemia, University of Texas MD Anderson Cancer Center, Houston, Texas 2Department of Haematology, Lyon-Sud Hospital, Lyon, France 3Department of Hematooncology and Bone Marrow Transplantation, Medical University of Lublin, Lublin, Poland 4Department of Hematology, Medical University of Lodz, Lodz, Poland 5Princess Margaret Cancer Center, Toronto, Canada 6Department of Haematology, Royal North Shore Hospital, St. Leonards, New South Wales, Australia 7Nantes University Hospital, Nantes, France Additional Supporting Information may be found in the online version of this article. Supporting Information Table 1. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.