Peripheral blood minimal residual disease may replace bone marrow minimal residual disease as an immunophenotypic biomarker for impending relapse in acute myeloid leukemia.

As relapses are common in acute myeloid leukemia (AML), early relapse prediction is of high importance. Although conventional minimal residual disease (MRD) measurement is carried out in bone marrow (BM), peripheral blood (PB) would be an advantageous alternative source. This study aims to investigate the specificity of leukemia-associated immunophenotypes used for MRD detection in blood samples. Consistency of PB MRD as compared with BM MRD was determined in flow cytometric data of 205 paired BM and PB samples of 114 AML patients. A significant correlation was found between PB and BM MRD (r=0.67, P<0.001), while median PB MRD percentage was factor 4-5 lower compared with BM MRD. Primitive blast (CD34+/CD117+/CD133+) frequency was significantly lower in PB (median factor 23.7), indicating that PB MRD detection is more specific than BM. Cumulative incidence of relapse 1 year after induction therapy was 29% for PB MRD-negative and 89% for PB MRD-positive patients (P<0.001). Three-year OS was 52% for MRD-negative and 15% for MRD-positive patients (P=0.034). Similar differences were found after consolidation therapy. As PB MRD appeared to be an independent predictor for response duration, the highly specific PB MRD assay may have a prominent role in future MRD assessment in AML.