Accurate and Sensitive Analysis of Minimal Residual Disease in Acute Myeloid Leukemia Using Deep Sequencing of Single Nucleotide Variations

Abstract

Minimal residual disease (MRD) in acute myeloid leukemia (AML) is of major prognostic importance. The genetic landscape of AML is characterized by numerous somatic mutations, which constitute potential MRD markers. Leukemia-specific mutations can be identified with exome sequencing at diagnosis and assessed during follow-up at low frequencies by using targeted deep sequencing. Our aim was to further validate this patient-tailored assay for substitution mutations. By applying a statistical model, which corrects for position-specific errors, a limit of detection for single nucleotide variations of variant allele frequency (VAF) of 0.02% was achieved. The assay was linear in MRD range (0.03% to 1%) with good precision [CV, 4.1% (2.2% to 5.7%) at VAF 1% and 13.3% (8.8% to 19.4%) at VAF 0.1%], and low relative bias [7.9% (2.5% to 15.3%) at VAF 1%]. When applied to six childhood AML cases and compared with multiparameter flow cytometry for MRD analysis, deep sequencing showed concordance and superior sensitivity. Further high concordance was found with expression of fusion transcripts RUNX1-RUNX1T1 and KMT2A-MLLT10. The deep sequencing assay also detected mutations in blood when VAF in bone marrow exceeded 0.1% (n=19). In conclusion, deep sequencing enables reliable detection of low levels of residual leukemic cells. Introduction of this method in patient care will allow for highly sensitive MRD surveillance in virtually every patient with AML.