Measurable residual disease assessment by qPCR in peripheral blood is an informative tool for disease surveillance in childhood acute myeloid leukaemia.

Serial assessments of measurable (or minimal) residual disease (MRD) by qPCR may identify nascent relapse in children with acute myeloid leukaemia (AML) and enable pre-emptive therapy. We investigated the kinetics and prognostic impact of recurrent fusion transcripts (RUNX1-RUNX1T1, CBFB-MYH11, KMT2A-MLLT3 or KMT2A-ELL) in 774 post-induction samples from bone marrow (BM, 347) and peripheral blood (PB, 427) from 75 children with AML. BM MRD persistence during consolidation did not increase the risk of relapse, and MRD at therapy completion did not correlate to outcome (HR = 0·64/MRD log reduction (CI: 0·32-1·26), P = 0·19). In contrast, 8/8 patients with detectable MRD in PB after first consolidation relapsed. Persistence (n = 4) and shifting from negative to positive (n = 10) in PB during follow-up predicted relapse in 14/14 patients. All 253 PB samples collected during follow-up from 36 patients in continuous complete remission were MRD negative. In core-binding factor AML, persistent low-level MRD positivity in BM during follow-up was frequent but an increment to above 5 × 10-4 heralded subsequent haematological relapse in 12/12 patients. We demonstrate that MRD monitoring in PB after induction therapy is highly informative and propose an MRD increment above 5 × 10-4 in PB and BM as a definition of molecular relapse since it always leads to haematological relapse.

Patient-tailored analysis of minimal residual disease in acute myeloid leukemia using next-generation sequencing.

Next-generation sequencing techniques have revealed that leukemic cells in acute myeloid leukemia often are characterized by a limited number of somatic mutations. These mutations can be the basis for the detection of leukemic cells in follow-up samples. The aim of this study was to identify leukemia-specific mutations in cells from patients with acute myeloid leukemia and to use these mutations as markers for minimal residual disease. Leukemic cells and normal lymphocytes were simultaneously isolated at diagnosis from 17 patients with acute myeloid leukemia using fluorescence-activated cell sorting. Exome sequencing of these cells identified 240 leukemia-specific single nucleotide variations and 22 small insertions and deletions. Based on estimated allele frequencies and their accuracies, 191 of these mutations qualified as candidates for minimal residual disease analysis. Targeted deep sequencing with a significance threshold of 0.027% for single nucleotide variations and 0.006% for NPM1 type A mutation was developed for quantification of minimal residual disease. When tested on follow-up samples from a patient with acute myeloid leukemia, targeted deep sequencing of single nucleotide variations as well as NPM1 was more sensitive than minimal residual disease quantification with multiparameter flow cytometry. In conclusion, we here describe how exome sequencing can be used for identification of leukemia-specific mutations in samples already at diagnosis of acute myeloid leukemia. We also show that targeted deep sequencing of such mutations, including single nucleotide variations, can be used for high-sensitivity quantification of minimal residual disease in a patient-tailored manner.