Detection of Immunophenotypic Residual Disease After Induction Therapy Is An Independent Prognostic Factor for Duration of Remission In Older AML Patients Treated Intensively

The majority of acute myeloid leukaemia (AML) patients are older than 60 years but little progress has been made in improving their poor outcome. Recent studies have used analysis of pre-treatment risk factors to identify patient subgroups most likely to tolerate and respond to more intensive treatment approaches. There are as yet no data on whether treatment effectiveness measured by the level of residual disease has independent prognostic value in determining the duration of response for older patients achieving a remission. Here we report the largest prospective series of immunophenotypic residual disease monitoring in older adult AML.

We prospectively and sequentially assessed the utility of residual disease (MRD) monitoring by multiparameter flow cytometry in older AML or high risk myelodysplastic (MDS) patients (excluding APL) receiving intensive treatment as part of the ongoing MRC AML16 trial for adults over 60 years. Patients were randomised to 2 or 3 courses of either daunorubicin/cytarabine or daunorubicin/clofarabine with or without mylotarg for the first course. Samples from consented patients (treated from 2006 to 2009) were processed by reference laboratories using an identical 4 colour antibody panel and standard operating procedure. Leukemic aberrant phenotypes (LAPs) were defined in pretreatment bone marrow and/or blood samples and then quantified using selected panel antibody combinations in post chemotherapy cycle bone marrow samples. MRD was reported as positive if detected above the LAP sensitivity threshold.

As previously shown for younger AML patients, immunophenotypic MRD monitoring was applicable to the majority of AML/high risk MDS patients over 60 years treated intensively. At least one LAP was identified in >90% of 224 analysed patients (median age 66, range 57–77) in whom an adequate pretreatment sample was received. Selected LAP sensitivity thresholds ranged from 0.05% to 0.2% as determined by analysis of control bone marrows. Most of patients in whom suitable LAPs could not be identified had AML M5. Aberrant CD7 expression in patients pretreatment (n=49) was associated with a poorer outcome (2 year OS of 11% (CD7+) v 40% (CD7-), HR adjusted for baseline characteristics 2.22 (1.33-3.70), p=0.002).

170 patients in remission were evaluable for MRD after cycle 1 (C1). 90/170 patients (58%) were MRD positive (MRD+) of whom 12% were high risk MDS (10-20% blasts) pretreatment v 11% of MRD negative (MRD-) patients. 2 year cumulative incidence of relapse was 88% (median RFS 8 months) for C1 MRD+ patients v 45% (18 months) for C1 MRD- patients. Adjusted analysis (allowing for age, sex, WBC, cytogenetics, secondary disease, performance status) shows MRD+ after course 1 is highly prognostic for relapse; HR 4.08 (2.05-8.10) p<0.0001.

Post cycle 2 (C2) 57/147 patients (39%) were MRD+ with a 2 year cumulative incidence of relapse of 78% (median RFS 8 months) v 66% (13 months) for C2 MRD- (adjusted HR for relapse 2.36 (1.33-4.20) p=0.003). Only 30 patients were evaluable after 3 cycles of chemotherapy. MRD positivity at this time point was again prognostic from adjusted analysis of relapse but with wide confidence intervals due to small numbers: HR 83.58 (1.91-3654.09) p=0.003.

These results suggest that flow cytometric analysis of residual disease in AML patients >60 treated intensively, particularly after course 1, may be useful in predicting the duration of remission, thus informing risk assessment for further treatment. Further data are needed to determine whether available treatment options can improve outcome of older AML patients with MRD positive disease.

This research was funded by a grant from Cancer Research UK.

No relevant conflicts of interest to declare.