It is widely accepted that patients with acute myeloid leukemia (AML) who test positive for measurable residual disease (MRD) prior to allogeneic stem cell transplantation (alloHCT) have higher rates of post-transplant relapse compared to those who are MRD negative; however, there are several nuances to consider regarding MRD in AML. Outcomes are also influenced by baseline disease characteristics and available MRD tests in different AML subsets. The inherent molecular and immunophenotypic heterogeneity of AML has limited the development of a single ultra-sensitive universal MRD test for all patients with AML. Various methods of MRD testing continue to evolve as more sensitive tests are becoming available for certain AML subtypes. The 2021 ELN MRD guidelines recommend MRD testing using quantitative polymerase chain reaction (qPCR) in those with validated molecular ‘MRD markers’ and multiparameter flow cytometry (MFC) in all other patients [1]. The current validated molecular MRD markers are primarily available for those with favorable risk disease by ELN 2022 guidelines including NPM1, t(8;21) RUNX1-RUNX1T1, inv(16) CBFB/MYH11, and t(15;17) PML/RARA [2]. The qPCR assay for MRD is highly sensitive and able to detect ~1/100,000 events compared to MFC with a sensitivity of only ~1/1000 with greater variability of test results at different centers. Other novel molecular MRD tests are being developed, such as the FLT3-ITD MRD Invivoscribe assay used in the MORPHO study, which may be recommended in subsequent MRD testing guidelines [3]. Patients with ELN-favorable risk disease have the benefit of higher response rates to current therapies and more sensitive MRD detection methods to guide therapeutic management. Ongoing efforts are focused on the development of ultra-sensitive molecular MRD testing for improved response assessments and prediction of those at highest risk for relapse, as well as to guide therapeutic interventions and perhaps de-escalation in some cases. With better disease prognostication, strategies to prevent relapse might be employed, including novel conditioning approaches or escalating conditioning intensity, graft manipulation, immunotherapy, maintenance therapy, and early tapering of immunosuppressive agents. After the development and validation of new MRD tests, well-designed clinical trials will be imperative to inform the optimal use of molecular MRD to guide decision-making.

The pre-MEASURE study generated proof of concept that molecular MRD testing in AML by error corrected next generation sequencing (NGS) prior to allogeneic transplant is better able to predict post-transplant relapse and survival compared to flow cytometry – particularly in those with FLT3-ITD or NPM1 mutations [4]. This was a retrospective evaluation of 1075 patients age 18 and over with AML who underwent HCT in first complete remission, with DNA sequencing performed on pretransplant blood samples done in a centralized setting, detectable at an allele fraction of 0.01% or above in FLT3-ITD and NPM1-mutated AML. Clinical data, along with MFC-MRD data performed locally, were collected by the Center for International Blood and Marrow Transplant Research (CIBMTR). Flow cytometry did not add additional prognostic value to the results of molecular MRD testing, though MFC data reported to the CIBMTR were not centralized. As seen in other studies, myeloablative conditioning was associated with improved survival and lower relapse rates compared to reduced intensity conditioning (RIC) in young patients with NGS-MRD prior to transplant [4,5,6]. Among patients with persistent FLT3-ITD or NPM1 mutations who received RIC, there was improved survival and lower relapse in those patients who received melphalan-based conditioning compared to other RIC regimens. As a complement to the already published data generated from pre-MEASURE, several commonly mutated genes and their significance as ‘MRD markers’ are discussed in this special edition focused on the evolution of MRD in AML.

As we move toward MRD as a surrogate endpoint to direct management, possible scenarios and outcomes must be considered (Fig. 1) as not all detectable markers may reliably predict outcomes [7]. One ideal scenario is that of blinatumomab in acute lymphoblastic leukemia, where the therapy acts as an “MRD eraser” frequently resulting in MFC-MRD negativity, which in turn correlates with improved post-transplant outcomes [8]. Blinatumomab has now shown a survival benefit as consolidation therapy to those testing MRD-negative as well, suggesting effective therapies may have benefit beyond current MRD measurements [9]. In AML, one goal of MRD development is to improve response criteria for specific subsets of disease and subsequently to pair disease-specific MRD tests with well-designed prospective studies evaluating MRD clearance and outcomes. To validate MRD markers, the test result must correlate with response and outcomes in large-scale prospective trials.

Fig. 1: Possible outcomes in attempts to eradicate MRD, reproduced [7].
figure 1

The final, published version of this article is available at https://www.karger.com/10.1159/000535463.

Full size image

After the development of a new potential MRD test, clinical trials will evaluate MRD clearance at different time points (post-induction, post-consolidation, pre-transplantation, post-transplantation) to guide larger scale treatment recommendations, and potentially the decisions about whether and how to pursue allogeneic transplant (Fig. 2). Retrospective data evaluating molecular predictors of MRD clearance may help guide which therapies have the best chance of achieving MRD clearance; however, strategies to eliminate MRD need to be evaluated in a prospective manner [10, 11]. Biomarker-directed clinical trials may also be informative as being done in the myeloMATCH umbrella trial for AML and myelodysplastic syndrome [12]. It is important to distinguish a biomarker predictive of response from MRD testing representative of response.

Fig. 2
figure 2

MRD timepoints and major management questions throughout the treatment course of AML.

Full size image

Baseline disease risk impacts the likelihood of clearing MRD with currently available treatments, and transplant may be the best approach for MRD clearance particularly in ELN adverse risk patients [13,14,15]. These high-risk patients should ideally be enrolled on clinical trials to improve current treatments. As MRD testing improves, an updated disease risk index may be appropriate to guide transplant strategies perhaps using CIBMTR data combined with new molecular MRD information [16].

This issue of BMT is focused on the impact of pre-MEASURE—including additional original data reporting the significance of FLT3-TKD, IDH1, and IDH2 persistence prior to transplantation. Review articles cover how pre-MEASURE has impacted clinical decision-making and the interpretation of NPM1 MRD throughout the treatment course of AML. Other manuscripts address the limitations of MFC-MRD in informing conditioning regimens, the role of the CIBMTR in precision transplantation research, and the importance of representation in clinical research.

Future challenges in this arena may include deciphering the best MRD test(s) in patients with multiple mutations or potential ‘MRD markers’, incorporating patient demographics and biological differences, and incorporation of the depth of MRD detection. Discussion of current methods of data collection for the CIBMTR to analyze these questions on a larger scale are also discussed in this issue. The upcoming prospective MEASURE study evaluating molecular data in the peri-transplant setting will hopefully generate centralized data to answer many of these questions.