Measuring minimal residual AML in remission using a novel multi DNA mutation based next generation sequencing assay

Abstract Introduction: Despite achievement of complete remission (CR) following chemotherapy, Acute Myelogenous Leukemia (AML) relapses in the majority of adult patients. The identification of cellular and molecular mechanisms that underlie resistance, persistence, and competitive regrowth of leukemic cells is of high interest to improve the outcome of AML patients. In this study, we have investigated a newly developed assay (v96) as a tool to measure and characterize residual disease burden and its subclonal structure in NPM1 mutated AML. Methods: We have developed a refined version of a next-generation sequencing-based MRD assay named “v96” and have applied this assay to diagnosis, remission and relapse DNA of 57 AML cases carrying NPM1 mutations, of which 46 received intensive induction chemotherapy and 11 were treated with Venetoclax/Decitabine. This assay tracks up to 96 leukemia- and patient-specific mutations detected via whole genome sequencing resulting in very high sensitivity for detection of tumor DNA in blood. The accuracy of the assay derives from the use of duplex sequencing and other strategies that dramatically reduce the error rates associated with amplification and sequencing, which has hindered the application of most MRD assays. We applied v96 to paired diagnosis and remission marrow or blood derived DNA samples and to nine paired relapse samples. We correlated various measures of residual DNA mutation burden, including the mean allele frequency (MAF), and total # of mutated amplicons/total # of amplicons with survival. Results: All patients had molecular evidence of residual leukemia in either the blood and/or bone marrow at the time of clinical remission. In addition, the large number of tracked mutations allowed us to define the subclonal structure at remission using clustering via MAF values. We found that remission AML samples harbored between one and eight distinct subclones and that relapsed AML comprised NPM1 WT and NPM1 mutated subclones. Mutations in DNMT3A and TET2 were common in clonal hematopoiesis and pre-leukemia and were carried forward into frank leukemia as part of stepwise transformation. We delineated the DNMT3A or TET2 containing subclones in the remission marrows and facultatively included or excluded these subclones in outcome calculation. We found that overall survival correlated with the tumor mutation burden at remission and that patients with high amounts of circulating tumor DNA had poor survival. This was independent of usage of BMT as consolidation. Conclusions: We applied a novel DNA based MRD test to NPM1 mutated AML as a proof-of-principle study for wider application to all AML patients, including those who do not carry traditional driver mutations used for MRD. We found that every patient had molecular evidence of residual disease at the time of clinical remission, and that the induction therapy does not eradicate all leukemia cells. Nonetheless, there were marked differences in residual mutations burden, which were prognostic. In particular, we identified a portion of AML patients with high residual mutation load that displayed poor long term survival. These results show that v96 assay can be applied to all AML patients, independent of preexisting traditional MRD driver mutation markers, to detect minimal residual disease during treatment.