A Comparison of 1 or 2 Courses of High Dose Cytarabine As Consolidation in Younger Patients with AML: First Results of the UK NCRI AML17 Trial

On behalf of the UK NCRI AML Working Party

Introduction

Outcomes in younger patients with AML have steadily improved over the last 4 decades. However, the question of the optimal chemotherapy schedule remains open. As outcomes improve attention turns from survival to survivorship: can similar – or better – outcomes be obtained with less cost to the patient. We have previously shown in our MRC AML12,15 trials that a fifth course of chemotherapy did not improve outcomes in these younger patients. A retrospective analysis of patients who received 3 or 4 courses of treatment showed that, while patients with poor risk disease had better outcomes with four courses, any effect of a fourth course was much less clear for patients with good or standard risk disease.

Methods

The UK NCRI AML17 trial (ISRCTN: 55675535) was designed for patients with AML or high risk MDS up to the age of 60. Following their first course of treatment, patients were allocated a risk group based upon a validated score (comprising cytogenetics, WBC, age, secondary disease and blast response to course 1)[1]; additionally, patients who failed to achieve at least a 50% reduction in blasts were considered poor risk. A protocol amendment included patients who were otherwise standard risk, but were FLT-3 ITD mutant/NPM1c wild type in the poor risk group. Remaining adults who were either good or standard risk could be randomised after 2 courses of treatment (DA or ADE, with or without GO at 3/6 mg/m²) to receive 1 or 2 courses of consolidation (3 or 4 courses in total). Before mid-2010 patients were randomised between MACE/MidAC vs MACE; based on the results of AML15 [2], a protocol amendment changed consolidation to 1 or 2 courses of Ara-C 3g/m²/d for 5 days. Follow-up is complete to 1^st March 2015, with median follow-up of 28.8 months (range 0.1-68.2).

Results

From July 2009 to June 2015, 1017 patients were randomised. Median age was 48 (range 16-72); 45% were male; 24% had core binding factor leukaemia; 76% intermediate risk disease; 1% had secondary disease, and 3% high risk MDS; WHO PS was 2+ in 4% of patients. Overall survival at 5 years was 57%.  An additional course of consolidation reduced relapse (CIR 53% vs 57%; HR 0.81 (0.67-0.99) p=0.04), with no difference in death in remission (7% vs 6%; HR1.10 (0.61-2.00) p=0.8), leading to some evidence of improved relapse free survival (41% vs 37%, HR 0.84 (0.70-1.01) p=0.06). Survival however, did not differ significantly between the arms (58% vs 55%, HR 0.90 (0.71=1.14) p=0.4), reflecting the effect of salvage: 3 year survival from relapse was 32% vs 31% (HR 0.97 (0.75-1.26) p=0.8). In stratified analyses, there was no difference in the effect of the extra course by cytogenetics (p-values for interaction between cytogenetics and treatment p=1.0 for survival); while relapse was reduced more in CBF leukaemias (HR 0.63 (0.40-1.00), p=0.05) than in others (HR 0.91 (0.73-1.13) p=0.4), the test for interaction was not significant (p=0.16). Information on minimal residual disease (by immunophenotype to a sensitivity of up to 1 in 10⁴) was available for 353 patients post course 1 and 265 patients post course 2. There was no evidence of a differential effect of an extra course of consolidation by MRD status after either course (p=1.0 for survival post course 1; p=0.3 for survival post course 2). In stratified analyses, there was no interaction between treatment and age, sex, performance status or diagnosis. However, there was a significant benefit for FLT3-ITD WT patients (OS: 63% vs 54%, HR 0.75 (0.57-0.99) p=0.04), which was not present for ITD mutant patients (41% vs 58%; HR 1.38 (0.84-2.26) p=0.2; p=0.04 for interaction), and evidence of greater benefit for 4 courses in patients with lower WBC at diagnosis (p=0.04 for trend). There was no heterogeneity by induction therapy (p=0.7), in particular the use of gemtuzumab ozogamicin or not (p=0.2).

Conclusions

These data in good and standard risk patients (as defined comprising 63% of patients diagnosed) suggests no overall survival benefit of a 4^th course of chemotherapy in younger patients with AML, although relapse was significantly reduced. With respect to subgroups, there was a suggestion that 4 courses of treatment was beneficial for FLT3 wild type patients. Attention now turns to whether the use of more intensive induction regimens such as FLAG-Ida can improve outcomes while delivering fewer courses.

[1] Burnett et al. Blood 2006:108 (11): Abstract 18

[2] Burnett et al. JCO 2013: 31(27);3360-8