Early peripheral blast cell clearance predicts minimal residual disease status and refines disease prognosis in acute myeloid leukemia.

ABSTRACT

Minimal residual disease (MRD) assessment in acute myeloid leukemia (AML) is increasingly used in risk stratification. However, several issues around this use are unresolved, including, among others, the most suitable time-point(s) for its application. Overall, late assessments appear more effective at distinguishing outcome but, in some studies, the early evaluations were already highly informative, anticipating the value of later ones. Our work integrated MRD with peripheral blast clearance (PBC), a treatment-related biomarker previously demonstrated to be a powerful predictor of response. From 2007 to 2014, we have studied 120 patients treated according to the NILG 02-06 trial and who achieved CR after induction. Patients in PBC-defined categories (separated by a 1.5-log threshold) showed significantly different probabilities of attaining MRD negativity, after either induction (MRD1) or consolidation (MRD2). Peripheral blast clearance combined with MRD1 largely anticipated MRD2-related information when both biomarkers predicted chemosensitive disease (PBChigh /MRD1neg ), the rate of MRD2-negativity was 90%, and DFS and OS estimates were 68% and 76% at 3 years, respectively. When both markers were unfavorable (PBClow /MRD1pos ), rates of MRD2 negativity, DFS, and OS were 20%, 34%, and 24%, respectively, at 3 years. In fact, MRD2 added prognostic value only in cases with discordant PBC/MRD1 data. Our data support a reasoned timing for MRD-based therapeutic decisions, modulated on individual chemosensitivity, an approach we have implemented in a forthcoming prospective multi-center trial by Gruppo Italiano Malattie EMatologiche dell'Adulto (GIMEMA).