Detection of minimal residual disease in circulating cell-free DNA in acute myeloid leukemia

Minimal/measurable residual disease (MRD) in Acute Myeloid Leukemia (AML) is defined as persistent leukemic cells below cytomorphological detection threshold. Next generation sequencing (NGS) of circulating cell-free DNA (cfDNA) to profile cancer-associated mutations has been shown to allow for quantification of disease burden in solid tumors and has also been suggested to enable minimally invasive follow-up of AML patients. In this pilot study we investigated the technical sensitivity and potential prognostic implications of cfDNA-based MRD monitoring in AML after allogeneic stem cell transplantation in comparison to donor chimerism analysis or, respectively, after consolidation chemotherapy. 75 cfDNA samples from 29 patients were analyzed by targeted NGS using a commercially available 10- or 37-gene hotspot panel (VariantPlex Core AML or Core Myeloid panel, ArcherDx). Patients' leukemias exhibited 1-7 mutations as determined by routine diagnostics. Only previously identified mutations were considered for MRD evaluation. cfDNA was isolated in sufficient amounts for NGS from all samples (total yield 24 ng-5.2 µg). The sensitivity of variant detection increased with higher overall read count and higher mutation-specific coverage (variant allele frequency [VAF] range 0.08-100%). At least one previously known mutation was identified in 32/55 samples (58%, VAF 0.08-78.04%) which were taken during hematological complete remission (CR) in both patients after allogeneic stem cell transplantation (aHSCT) and patients after consolidation chemotherapy. In patients after aHSCT (n = 25), at least one previously known mutation was detected in 16/29 cfDNA samples (55.1%, VAF 0.08-6.7%) obtained when donor chimerism was ≥ 90% and in 6/6 samples (100%, VAF: 0.88-63.77%) with reduced donor chimerism. Probability of progression-free survival 17 months after aHSCT in patients with donor chimerism ≥ 90% but mutation-positive cfDNA was 64% compared to 100% in patients with undetectable MRD. In patients after consolidation chemotherapy, cfDNA was positive in all samples taken during CR (n = 4; VAF 0.26-29.84%) and non-CR (n = 4; VAF 8.46-100%). Our results indicate that NGS of cfDNA is suitable for MRD monitoring in AML and offers higher sensitivity for detecting residual leukemic cells than chimerism analysis in patients after aHSCT. Further studies are needed to evaluate clinical relevance of MRD status as determined in cfDNA.