Sensitivity of minimal residual disease in acute myeloid leukaemia in first remission – methodologies in relation to their clinical situation Peter Hokland, Hans Beier Ommen, Charlotte Guldborg Nyvold and Anne Stidsholt Roug Department of Haematology, Aarhus University Hospital, Aarhus, Denmark Summary The concept of minimal residual disease in acute myeloid leukaemia has been steadily developed pre-clinically, with quantitative polymerase chain reaction (qPCR) leading the way with highly validated assays for patient-based risk strati- fication at post-treatment time points, which are being integrated in clinical trials both at evaluation of first com- plete remission (CR1) and after attaining CR1. Moreover, multicolour flow cytometry (MFC) has been increasingly employed in identifying leukaemia-associated immunopheno- types (LAIPs) with significant progress being made in standardization. In translating these widely varying method- ologies to parameters useful for individualized patient deci- sion-making, one of the obstacles has been that the assays entail varying sensitivities dependent on a number of vari- ables. For qPCR, sensitivity depends on target type (i.e. fusion transcript, mutated gene or even overexpressed gene) and – in the case of overexpressed genes – on expression in healthy haematopoiesis. For MFC, sensitivity is likewise lar- gely a function on whether the same phenotype is seen in normal immature cells and, in addition, antigen drift/shift with LAIPs changing at relapse is a well-known problem. In considering which sensitivity to opt for, a further variable is the situation of patient, most importantly the level of cytore- duction intended. Here we will attempt to give an overview of these pertinent questions intended for the practicing hae- matologist, focusing on where the field is heading at the clin- ical level. Keywords: Acute myeloid leukaemia, minimal residual disease, quantitative PCR, multicolour flow cytometry, sensitivity levels. Introduction In acute myeloid leukaemia (AML), it is now well established that prognostic biomarkers derived from karyotyping or molecular biology are instrumental in stratifying patients to different levels of cytoreduction and in deciding whether an allogeneic stem cell transplantation is needed (Grimwade et al, 2010a). On the other hand, prognostic markers are not absolute and, in fact, fail to predict clinical outcome in a sizeable fraction of patients attaining a complete remission (CR). Thus, as recently discussed (Dinardo & Luger, 2012), additional variables assessed during CR are needed to better individualize decision-making in CR. During the most recent decade, significant technological advances have been made in targeting the malignant cells in AML with respect to their identification in minute amounts in phases, which would be deemed as being devoid of malig- nancy by morphological criteria. This state of minimal resid- ual disease (MRD) is steadily gaining acceptance as a new candidate for clinical decision-making in the evaluation of response to therapy, with regard to guiding the type and intensity of cytoreduction or allocation for allogeneic stem cell transplantation for individual patients (Cheson et al, 2003; Hokland & Ommen, 2011) and as a powerful and vali- dated tool for adjusting dosages and types of tyrosine kinase inhibitors (TKIs) in chronic myeloid leukaemia (CML) (White et al, 2010). The detection of MRD with a sensitivity reaching the 10 À5 level for certain quantitative polymerase chain reaction (qPCR) assays and 10 À4 for optimized multicolour flow cytometry (MFC) assays in conjunction with their pre-clini- cal validation represents a considerable feat in terms of cross-platform approaches and heterogeneity of targets. Unfortunately, both approaches are limited in systematic integrated clinical use due to lack of standardized assays and cut-offs. On the other hand, formal proof as to the clinical value of MRD determinations is still lacking. Moreover, even though much knowledge has recently been gained concerning the kinetics of relapse in both AML subtypes as well as in CML, the level of sensitivity of MRD needed for effective patient management remains to be determined. In other Correspondence: Peter Hokland, Department of Haematology, Aarhus University Hospital, Tage-Hansens Gade 2, DK-8000, Aarhus C, Denmark. E-mail: phokland@ki.au.dk ª 2012 Blackwell Publishing Ltd First published online 28 June 2012 British Journal of Haematology, 2012, 158, 569–580 doi:10.1111/j.1365-2141.2012.09203.x review