Clonal hematopoiesis and measurable residual disease assessment in acute myeloid leukemia

Abstract

Current objectives regarding treatment of acute myeloid leukemia (AML) include achieving complete remission (CR) by clinicopathological criteria followed by interrogation for the presence of minimal/measurable residual disease (MRD) by molecular genetic and/or flow cytometric techniques. Although advances in molecular genetic technologies have enabled highly sensitive detection of AML-associated mutations and translocations, determination of MRD is complicated by the fact that many treated patients have persistent clonal hematopoiesis (CH) that may not reflect residual AML. CH detected in AML patients in CR includes true residual or early recurrent AML, myelodysplastic syndrome or CH that is ancestral to the AML, and independent or newly emerging clones of uncertain leukemogenic potential. Although the presence of AML-related mutations has been shown to be a harbinger of relapse in multiple studies, the significance of other types of CH is less well understood. In patients who undergo allogeneic hematopoietic cell transplantation (HCT), post-HCT clones can be donor-derived and in some cases engender a new myeloid neoplasm that is clonally unrelated to the recipient's original AML. In this article, we discuss the spectrum of CH that can be detected in treated AML patients, propose terminology to standardize nomenclature in this setting, and review clinical data and areas of uncertainty among the various types of posttreatment hematopoietic clones.

Figure 1.

Illustrations of dynamic patterns of the mutational landscape after therapy for AML. (A) gMRD. At diagnosis (Dx), FLT3-ITD and NPM1 mutations were present. In CR, there is detectable NPM1 mutation, indicating the presence of gMRD despite morphologic CR. (B) CH that was detectable in the original AML. At diagnosis, DNMT3A and NPM1 mutations were present, both at high VAFs. In CR, there is no detectable NPM1 mutation, but a persistent DNMT3A mutation, indicating the presence of CH shared with the AML clone. If absence of NPM1 mutation is confirmed by a sensitive method, this would represent CH without gMRD. (C) CH that was not detected in the original AML. At diagnosis, CBFB-MYH11 rearrangement and KIT mutations were present, which disappeared in the initial CR time points. Subsequently, a TET2 mutation developed, indicating CH that was either present at a very low level prior to treatment (presumably not in the AML clone) or emerged after therapy. (D) RMN. At diagnosis, JAK2, TET2, and GATA2 mutations were present. The JAK2 and TET2 mutations persist in CR and BM shows morphologic features of primary myelofibrosis, indicating RMN despite morphologic CR and disappearance of the GATA2 mutation (which is likely AML-related in this case). (E) Persistent germline mutation. At diagnosis, 2 DDX41 mutations were present, 1 of which was shown to be of germline origin. After treatment in CR, the germline DDX41 mutation persists, with a similar VAF at all time points. (F) Recurrent AML. At diagnosis, NPM1 and FLT3-ITD mutations were present, which disappeared in CR. The same NPM1 and FLT3-ITD mutations are present at the time of relapse. (G) Recurrent AML with AML-related CH-type mutations, but new AML-related mutation. At diagnosis, DNMT3A, IDH1, and FLT3-ITD mutations were present, with persistence of the DNMT3A and IDH1 mutations during CR. At relapse, a new FLT3-TKD mutation has been acquired; nevertheless, the shared unique DNMT3A and IDH1 mutations indicate recurrence of the original clone rather than a new, unrelated clone. (H) New clonally unrelated AML. At diagnosis, RUNX1-RUNX1T1 rearrangement was present, which disappeared in CR. At relapse, there is a new TP53 mutation and a complex karyotype, but no RUNX1-RUNX1T1 rearrangement, indicating a new genetically distinct AML that is considered to be therapy-related. (I) Donor-derived CHIP. At diagnosis, TET2, ASXL1, and IDH2 mutation were present. Following HCT from an unrelated donor, the TET2, ASXL1, and IDH2 mutations disappear and the patient remains in CR, but a new DNMT3A mutation has appeared that was absent in the original leukemia. Chimerism studies show 100% donor chimerism, indicating DNMT3A-mutated CHIP that is donor derived (rather than originating from a low-level undetectable mutation that was present in the original recipient AML).