See related editorial on pages 1650-2 and accompanying article on pages 1744-50, this issue.
Aberrant DNA methylation, which results in leukemogenesis, is frequent in patients with myelodysplastic syndromes (MDS) and is a potential target for pharmacologic therapy. Decitabine indirectly depletes methylcytosine and causes hypomethylation of target gene promoters.
A total of 170 patients with MDS were randomized to receive either decitabine at a dose of 15 mg/m2 given intravenously over 3 hours every 8 hours for 3 days (at a dose of 135 mg/m2 per course) and repeated every 6 weeks, or best supportive care. Response was assessed using the International Working Group criteria and required that response criteria be met for at least 8 weeks.
Patients who were treated with decitabine achieved a significantly higher overall response rate (17%), including 9% complete responses, compared with supportive care (0%) (P < .001). An additional 12 patients who were treated with decitabine (13%) achieved hematologic improvement. Responses were durable (median, 10.3 mos) and were associated with transfusion independence. Patients treated with decitabine had a trend toward a longer median time to acute myelogenous leukemia (AML) progression or death compared with patients who received supportive care alone (all patients, 12.1 mos vs. 7.8 mos [P = 0.16]; those with International Prognostic Scoring System intermediate-2/high-risk disease, 12.0 mos vs. 6.8 mos [P = 0.03]; those with de novo disease, 12.6 mos vs. 9.4 mos [P = 0.04]; and treatment-naive patients, 12.3 mos vs. 7.3 mos [P = 0.08]).
DNA methylation is an important regulator of gene transcription and appears to play a key role in the pathogenesis of hematologic malignancies.1 Methylation of DNA in the proximity of a gene promoter results in stable transcriptional silencing of its expression.2, 3 Hypermethylation has been associated with tumor progression and arrest of differentiation, and has been detected in myelodysplastic syndrome (MDS) and acute leukemia.1, 4 However, this process can be reversed during DNA synthesis, which makes it a potential therapeutic target.
Decitabine is a cytosine nucleoside analog that, once incorporated into DNA, inhibits further DNA methylation. As a consequence, aberrantly silenced genes, including tumor suppressor genes, can be reactivated and expressed. Decitabine has demonstrated activity in a broad range of hematologic disorders, including MDS, acute myelogenous leukemia (AML), chronic myelogenous leukemia (CML), and sickle cell anemia.5–10 In Phase II studies of decitabine in patients with MDS, overall response rates (ORRs) of 42% to 54% were achieved, including complete responses (CRs) in 20% to 28% of patients.11, 12 These promising results led to a multiinstitutional, Phase III randomized trial of decitabine versus supportive care in patients with MDS. Preliminary results of this open-label study were presented at the 2004 American Society of Hematology Annual Meeting in San Diego, California13; the final results are summarized in the current study.
MATERIALS AND METHODS
Adults age ≥ 18 years with a confirmed diagnosis of MDS (either de novo or secondary) who fit any of the recognized French–American–British (FAB) classifications14 and had an International Prognostic Scoring System (IPSS) score ≥ 0.5 as determined by complete blood count, bone marrow assessment, and cytogenetics were eligible for study participation (Table 1). Patients were required to provide written informed consent. Exclusion criteria included a diagnosis of AML (percent blasts ≥ 30%) or other progressive malignant disease.
Table 1. Patient Characteristics (in the Intent-to-Treat Population)
MDS: myelodysplastic syndrome; BM: bone marrow; IPSS: International Prognostic Scoring System; FAB: French-American-British; RA: refractory anemia; RARS: refractory anemia with ring sideroblasts; RAEB: refractory anemia with excess blasts; RAEB-t: refractory anemia with excess blasts in transformation; CMML: chronic myelomonocytic leukemia; WHO: World Health Organization; PS: performance status; RBC: red blood cell.
All comparisons were made using a 2-sided Fisher exact test, with the exception of age, weeks since myelodysplastic syndromes diagnosis, and percent bone marrow blasts, which were performed using a 2-sided Wilcoxon rank-sum test.
Nonevaluable patients were excluded from this calculation.
Patients were stratified by IPSS risk group and type of MDS (de novo or secondary), and randomized between July 24, 2001 and January 12, 2004 using a centralized call-in process. Patients were randomized on a 1:1 basis to receive decitabine plus supportive care or supportive care alone. Decitabine at a dose of 15 mg/m2 given intravenously was administered over 3 hours every 8 hours for 3 days; this constituted 1 treatment cycle, which was repeated every 6 weeks, depending on recovery from myelosuppression. Supportive care was given according to generally acceptable guidelines for both study arms. Supportive care measures included the use of red blood cell (RBC) transfusion in patients with a hemoglobin level < 8 g/dL and platelet transfusion in patients with a platelet count < 7.5 × 109/L. Hematopoietic colony-stimulating factors could be given as outlined in the treatment guidelines.15
Patients were withdrawn from the study in the case of disease progression, transformation to AML by FAB criteria, failure to achieve a partial response (PR) after 6 cycles of decitabine therapy, or failure to achieve a CR after 8 cycles of decitabine therapy. In addition, the design of the current study dictated that patients be removed from therapy after 2 cycles of a maintained CR.
Quality of life (QOL) was assessed at baseline, at the end of each dosing cycle, and at the end of treatment using the European Organization for Research and Treatment of Cancer (EORTC) QOL Questionnaire (QLQ) C30 (versions 1.0 and 3.0). Scale scores were computed according to EORTC standards published in the third edition of the EORTC QLQ-C30 scoring manual.16
The coprimary endpoints in the current study were ORR and time to AML transformation or death. Response was assessed according to the International Working Group (IWG) criteria (Table 2).17 A CR was defined as normalization of peripheral counts and bone marrow for at least 8 weeks with serial bone marrow blasts < 5% without dysplastic changes, hemoglobin >11 g/dL, a neutrophil count ≥1.5 × 109/L, and a platelet count ≥100 × 109/L. Criteria for a PR were the same as for a CR, with the exception of a reduction of ≥ 50% of blasts that remained above 5%, or a downgrade in the FAB category. Hematologic improvement (HI) was described by the magnitude of the response (major or minor) and the individual responsive cell lines. During the study, local pathologists evaluated the bone marrow aspirates and biopsies, and clinical investigators applied the MDS IWG criteria (investigator assessment dataset). In addition, a blinded central review of all bone marrow aspirates and biopsies was performed by an expert hematomorphologist (J.M.B.) to determine each patient's best hematologic response per the MDS IWG criteria (centrally reviewed dataset). The date of progression to AML was taken from either the independent reviewer (centrally reviewed data) or the investigator analysis, whichever provided the earliest diagnosis of AML.
Table 2. International Working Group Response Criteria for Myelodysplastic Syndrome17
• Other response criteria same as those for a complete response
• Downgrade in FAB category
Two analyses, one interim and one final, were planned using the stopping rules of O'Brien and Fleming. The overall type 1 error rate was maintained at a maximum of 5% by applying a Bonferroni correction for the coprimary endpoints at the final analysis. A maximum P value of .024 was required to establish statistical significance using a 2-sided analysis for either of the coprimary endpoints (ORR or time to AML or death).
Efficacy analyses were performed on the intent-to-treat (ITT) population. The Fisher exact test was used to compare treatment groups with respect to ORR. Time to AML or death was compared between treatment groups using the log-rank test. An additional analysis of time to AML or death was performed using a Cox proportional hazards model by adjusting for the stratifying factors of IPSS category and type of MDS. Differences in QOL were tested using the Wilcoxon rank-sum test.
Of the total of 170 patients in the current study, 89 were randomized to the decitabine arm and 81 to supportive care alone. Six patients who were randomized to the decitabine arm withdrew before receiving treatment. Baseline patient characteristics were well balanced between the 2 study arms (Table 1). The median patient age was 70 years (range, 30-85 yrs). The median duration of MDS was 29 weeks and 35 weeks, respectively, for patients in the decitabine and supportive care arms. Prior therapy was noted in 21% and secondary MDS in 14% of patients. An IPSS score of intermediate-2 or high-risk MDS was noted in approximately 70% of patients. The majority of patients were RBC transfusion-dependent (71%); 23% of patients were platelet transfusion-dependent.
In an ITT analysis of the total study population (n = 170), the centrally reviewed ORR (CR + PR) of patients who received decitabine (15 of 89 patients [17%]) was superior to the ORR of patients on the supportive care arm (0 of 81 patients [0%]) (P < .001) (Table 3). Among patients who responded to decitabine therapy, 8 patients (9%) achieved a CR and 7 patients (8%) achieved a PR. All responses were durable, with a median duration of 10.3 months (range, 4.1 mos-13.9 mos) for patients with a confirmed date of disease progression. The median time to first response (PR or CR) was 3.3 months (range, 2.0 mos-9.7 mos), or after 2 cycles of decitabine therapy. An HI was observed in 13% of patients on the decitabine arm and 7% of patients who received supportive care alone, resulting in overall improvement rates (CR + PR + HI) of 30% and 7%, respectively (P < .001). On the decitabine treatment arm, 12 patients had major HI, including 2 with major erythroid HI, 4 with major platelet HI, 1 with major neutrophil HI, 3 with major erythroid plus platelet HI, and 2 with major neutrophil plus erythroid HI. On the supportive care arm, 6 patients had an HI: 5 with major HI (3 with neutrophil, 1 with platelet, and 1 with both HIs) and 1 with a minor platelet HI.
Table 3. Intent-to-Treat Analysis of Response to Decitabine
CR: complete response; PR: partial response; HI: hematologic improvement; IPSS: International Prognostic Scoring System; FAB: French–American–British; RA: refractory anemia; RARS: refractory anemia with ring sideroblasts; RAEB: refractory anemia with excess blasts; RAEB-t: refractory anemia with excess blasts in transformation; CMML: chronic myelomonocytic leukemia; MDS: myelodysplastic syndrome.
Determined using the Fisher 2-sided exact test.
Includes 2 major erythroid hematologic improvements, 4 major platelet hematologic improvements, 1 major neutrophil hematologic improvement, 3 major erythroid plus platelet hematologic improvements, and 2 major neutrophil plus erythroid hematologic improvements.
Includes 3 major neutrophil hematologic improvements, 1 major platelet hematologic improvement, and 1 major both hematologic improvement; 1 minor platelet hematologic improvement.
Analysis did not include prior treatment with hematopoietic growth factors.
In an analysis of centrally reviewed response data by patient subgroups (Table 3), responses were observed across all IPSS risk groups. Similar response rates were observed in patients who had received prior therapy for MDS (not including the use of growth factors [15%]) and previously untreated patients (17%). Response rates among patients with de novo disease (17%) and secondary MDS (16%) were also similar. Furthermore, response was noted regardless of karyotype with respect to the presence or absence of 5q and 7q deletions. Among patients with 5q or 7q deletions, 13% (2 of 16 patients) and 21% (4 of 19 patients), respectively, achieved a response. In patients without 5q or 7q deletions, the response rates were 16% (11 of 67 patients) and 14% (9 of 64 patients), respectively.
Nine patients on the decitabine arm and 3 patients on the supportive care arm were found retrospectively to have AML at baseline by central review. Of these 9 patients treated with decitabine, 5 (56%) achieved an objective response versus no response noted on the supportive care arm.
Time to AML or Death
Patients treated with decitabine had a trend toward a longer median time to AML or death compared with patients treated with supportive care alone (12.1 mos vs. 7.8 mos; P = .16 by the log-rank test) (Fig. 1A). When patient subgroups were analyzed, patients on the decitabine arm experienced a longer median time to AML or death than those who received supportive care, if they were treatment-naive (12.3 mos vs. 7.3 mos; P = .08) (Fig. 1B), had an IPSS score of intermediate-2/high risk (12.0 mos vs. 6.8 mos; P = .03) (Fig. 1C), were classified based on IPSS score as high risk (9.3 mos vs. 2.8 mos; P = .01), or had de novo MDS (12.6 mos vs. 9.4 mos; P = .04).
In an exploratory analysis using the Cox proportional hazards model and accounting for IPSS category and type of MDS, patients on the supportive care arm had an approximately 2-fold greater risk of AML progression or death compared with those on the decitabine arm (hazards ratio = 0.580; 95% confidence interval, 0.37-0.91).
Patients with clonal abnormalities at baseline who underwent follow-up cytogenetic evaluations at the end of the study were evaluable for a cytogenetic response (26 patients on the decitabine arm and 21 patients on the supportive care arm). Cytogenetic CRs were observed in 35% of patients treated with decitabine versus 10% of those who received supportive care alone. In addition to the cytogenetic CRs reported, 1 patient who received decitabine had a minor cytogenetic response (reduction of > 50%), 1 patient who originally was randomized to supportive care achieved a cytogenetic CR 3 months after crossover to the decitabine arm, and 1 patient on the supportive care arm (who did not have a baseline assessment but had an abnormal cytogenetic assessment 2 weeks after randomization) subsequently achieved a cytogenetic CR.
Clinical Benefits in Decitabine Responders
An evaluation was performed to determine whether response was correlated with clinical benefit. The response group was defined as patients achieving a CR or PR (decitabine-treated patients only); all other patients were considered to be nonresponders (regardless of whether they received decitabine or supportive care). The median time to AML progression or death was 17.5 months for responders versus 9.8 months for nonresponders (P = .01). Furthermore, 100% of responders were RBC transfusion-independent and platelet transfusion-independent based on MDS IWG criteria during the time of the response. All 8 responders who had a cytogenetic abnormality and were evaluable for a cytogenetic response achieved a cytogenetic response (7 major responses and 1 minor response).
Number of Courses of Decitabine
The median number of decitabine courses was 3 (range, 0-9 courses). A total of 43 patients randomized to treatment with decitabine received ≤ 2 courses. Six patients never received decitabine, 19 patients received only 1 course, and 18 patients received 2 courses. Reasons for treatment discontinuation for the 37 patients who received only 1 or 2 courses were disease progression in 14 patients (38%), death in 10 patients (27%), adverse events in 3 patients (8%), withdrawal of consent in 1 patient (3%), other causes in 14 patients (38%) (including 9 patients reported as having blood count–related causes [ie, persistent cytopenias], 3 patients who decided to withdraw, and 2 patients who were withdrawn at the discretion of the investigators), and an unknown reason in 1 patient (3%). Therefore, only 46 of 89 patients (52%) received ≥ 3 courses of decitabine, and only 23 patients (26%) received > 6 courses. Of the 15 patients who responded after decitabine treatment, the median number of courses was 6 (range, 2-8 courses).
Growth Factor and Transfusion Requirements
During the current study, a higher percentage of patients on the supportive care arm received erythropoietic growth factors compared with patients on the decitabine arm (41% vs. 20%). Over time, the percentage of patients receiving decitabine who became RBC transfusion-independent increased, whereas the percentage of patients on the supportive care arm who required RBC transfusions did not change significantly (Fig. 2).
The ITT analysis of the investigator dataset indicates that median survival was not significantly different between the patients treated with decitabine and those who received supportive care (14.0 mos vs. 14.9 mos; P = .636). Survival was extended among decitabine responders compared with nonresponders (23.5 mos vs. 13.7 mos; P = .007) (Fig. 3).
Quality of Life
According to the evaluations that were completed at the end of each treatment cycle, decitabine resulted in a statistically superior QOL compared with best supportive care in several QOL parameters. Improvements in global health status (P < .05 at the end of Cycles 2 and 4) (Fig. 4), fatigue (P < .05 at the end of Cycles 2, 4, 5, and 6), and dyspnea (P < .05 at the end of all 6 cycles) were sustainable effects observed during decitabine treatment.
Safety data were evaluated for 83 patients treated with decitabine and 81 who received supportive care. Overall, decitabine was found to be well tolerated, with a manageable toxicity profile. Temporary dose reductions or delays occurred in 35% of patients on the decitabine arm, primarily due to adverse events. The overall incidence of patients who died on the study was less for the decitabine arm (14%) compared with those patients who received supportive care (22%). Causes of death were similar to those generally observed in MDS. Serious adverse events were experienced by 69% of decitabine patients compared with 56% of patients receiving supportive care. The most common adverse events were neutropenia, thrombocytopenia, anemia, febrile neutropenia, leukopenia, pyrexia, hyperbilirubinemia, and pneumonia (Table 4). Of note, neutropenia, thrombocytopenia, anemia, and leukopenia appeared to diminish in incidence over the first 4 cycles of decitabine treatment; however, these toxicities remained frequent, most likely due to the continuing presence of underlying disease and myelosuppression. Gastrointestinal toxicities were generally mild and occurred infrequently (< 5% of patients).
The results of the current study confirm the efficacy of decitabine in the treatment of MDS. Decitabine produced a higher ORR (including CR) compared with best supportive care. Furthermore, decitabine demonstrated a benefit across patient subgroups, including all IPSS subgroups; the majority of FAB classes; those patients with de novo, secondary disease; those who had received prior therapy for MDS; and in chemotherapy-naive patients. All decitabine responders became transfusion-independent, and the need for RBC transfusion was found to be reduced significantly among patients treated with decitabine. Importantly, from the patient perspective, decitabine-treated patients had an improved QOL compared with those receiving supportive care only.
Patient outcome in MDS is highly variable. Several classification systems, including the FAB and IPSS systems, were developed to define heterogeneity in MDS, to develop risk-oriented therapeutic strategies, and to evaluate the benefit of novel approaches within defined risk groups.14, 18–21 Both the FAB and IPSS classifications have been shown to be prognostically relevant. It is noteworthy that in the current study, the time to AML or death was found to be delayed significantly among IPSS high-risk and intermediate-2 patients treated with decitabine.
The optimal use of hypomethylating agents for maximum efficacy is likely to include the use of prolonged therapy. Decitabine is cell cycle-dependent and may require prolonged administration to achieve maximum benefit. The design of the current study dictated that patients who maintained a CR for 2 cycles be removed from therapy. As a result, the median number of cycles delivered was 3 (range, 0 cycles-9 cycles). In addition, 6 patients who were assigned to treatment with decitabine received no therapy, 19 patients received only 1 course, and 18 patients received only 2 courses and had their treatment interrupted for myelosuppression-related side effects. Therefore, 43 of 89 patients (48%) received no or only minimal (and possibly ineffective) therapy. In 2 previous Phase II decitabine studies in which the median number of decitabine cycles was slightly higher (median of 4 cycles), the response rates (CR + PR) were approximately 30%.12, 22 It is possible that a greater benefit may have been observed in the current study if the study design had allowed patients to continue receiving therapy for a longer period of time.
Alternative decitabine dosing schedules are currently being investigated, including intravenous infusion over 1-hour schedules given for 5 or 10 days and intravenous versus subcutaneous schedules. The optimal selected schedule appears to be decitabine at a dose of 20 mg/m2 given intravenously over 1 hour daily for 5 days.23 This dosing schedule provided timely and frequent delivery of decitabine in an effort to improve patient response. Decitabine courses were administered every 4 weeks (rather than 6 weeks) regardless of blood counts, provided that evidence of persistent disease was present on repeat bone marrow evaluation and no prohibitive or life-threatening toxicities had occurred. Time for recovery was allowed only after every 3 courses, and response was not evaluated until at least 3 courses of therapy had been completed. At this lower dose and alternative, convenient schedule, preliminary results have demonstrated that decitabine produced CRs in 40% of patients with MDS. Further studies currently are ongoing to confirm the optimal dosing schedule for this agent.
It is relevant to compare the efficacy of decitabine versus the only other currently available hypomethylating agent, 5-azacitidine. In preclinical models, decitabine has exhibited approximately 10-fold more potency than azacitidine; only 10% of azacitidine is metabolized to its active metabolite.24 In a randomized trial of 5-azacitidine treatment versus supportive care in patients with MDS, 5-azacitidine produced results similar to those observed with decitabine, with a higher CR rate (5.6% vs. 0%) and ORR (15.7% vs. 5.6%; P = .0001) compared with supportive care alone.25, 26 However, several factors potentially favored the 5-azacitidine arm when compared with the decitabine arm in the current study. Specifically, patients treated in the 5-azacitidine study had the opportunity to remain on treatment longer (median of 9 courses) compared with those treated in the current decitabine trial. The number of courses of 5-azacitidine needed to achieve a CR ranged from 5 to 15 courses, again emphasizing the need for persistence in therapy to achieve the best response. In addition, response criteria in the azacitidine trial were less rigorous (ie, requiring that a CR or PR last for at least 4 weeks and not requiring the disappearance of dysplastic changes).26 In the current study, response was determined using the more rigorous IWG criteria,17 with a blinded analysis of best response by independent, expert hematopathologists.
Analysis of methylation patterns demonstrating that methylation in MDS increases progressively with time on the supportive care arm and that patients randomized to decitabine treatment appear to have this methylation pattern reversed will be the subject of a subsequent report. However, it must be emphasized that although hypomethylation is the proposed mechanism of action of decitabine, the drug may work through other mechanisms or perhaps through a direct cytotoxic effect, particularly at higher doses.
The results of the current study have demonstrated the beneficial anti-MDS activity of decitabine compared with supportive care in relation to time to AML transformation or death, response, transfusion needs, and supportive care parameters. Patients treated with decitabine were found to experience durable responses that were clinically relevant. The efficacy of decitabine in patients with MDS was found to be even greater in Phase II studies, in which it was delivered for longer time periods.7, 12, 27 Decitabine also may be effective in patients with other hematologic malignancies such as AML and CML, and in various malignancies in which hypermethylation and tumor suppressor gene silencing may play a role.