Molecular Minimal Residual Disease Detection in Acute Myeloid Leukemia

Minimal residual disease (MRD) detection based on the standardized molecular monitoring of the t(9;22)-related BCR-ABL1 fusion transcript is well established for patients with chronic myeloid leukemia (CML). The levels of BCR-ABL1 serve as a guide to tailor treatment of the CML patient. In acute myeloid leukemia (AML) MRD detection based on polymerase chain reaction (PCR) approaches targeted towards the acquired molecular abnormalities is less well established. MRD measurement of the CBFB-MYH11 and RUNX1-RUNX1T1 fusion transcripts after induction therapy has been shown to be of some clinical importance. However, these transcripts can persist during long term complete remission, without having an effect on treatment outcome. In contrast, sequential MRD monitoring of the PML-RARA fusion transcript in acute promyelocytic leukemia (APL) is a strong predictor of relapse. Initial molecular MRD studies were limited to these favorable AML subtypes. Due to the discovery of novel recurrent abnormalities in AML the potential of molecular MRD detection has increased substantially. Although, certain acquired mutations, such as those in NPM1, are known for a number of years, only recently the application of these molecular abnormalities for MRD detection has been investigated in larger clinical trials. By NPM1 mutant MRD detection we can now recognize patients with higher risk of relapse. Highly sensitive targeted detection of the hotspot mutations in AML subsets is feasible by means of real-time PCR, but detection of patient specific mutations with this technology is still challenging. Next generation sequencing (NGS) revealed that AML is an extremely heterogeneous disease, as illustrated by the multitude of acquired mutations, but this technology has also opened possibilities for detection of MRD in virtually every patient. With NGS there is no need for patient specific assays since practically all mutations are detected. These molecular abnormalities, as single marker or in combination, will most certainly improve MRD monitoring of AML. However, it remains yet to be determined how MRD levels are assessed and which combination of markers in a MRD detection result in clinically relevant information, requiring extensive validation in large clinical AML trials. Smaller studies already demonstrated the variable dynamics of MRD during treatment and associations between somatic mutations persistence and risk of relapse. However, clonal hematopoiesis of undetermined potential, i.e., preleukemic mutations that may persist after treatment, provides an extra layer of complexity to the applicability of MRD detection. For example, the clinical applicability of MRD detection in the setting of mutant DNMT3A and IDH mutations is likely less effective due to the persistent DNMT3A and IDH mutant preleukemic cells following treatment. However, should all mutations be cleared after treatment or can preleukemic mutations in otherwise normal hematopoiesis persist without resulting in relapse? Taken together, there is need for molecular approaches to understand the dynamics of residual disease in AML during treatment.