MRD in AML: who, what, when, where, and how?

In this issue of Blood, Othman et al use retrospective data to show the utility of measurable residual disease (MRD) monitoring of patients with NPM1 acute myeloid leukemia (AML) in de novo treatment using venetoclax-based regimens.¹ 

MRD assessments have the potential to improve outcomes for patients with AML. However, there are still many open questions. For example, the 2021 European LeukemiaNet (ELN) updated MRD guidelines pertaining primarily to patients receiving intensive chemotherapy.² These guidelines recommend using quantitative or digital polymerase chain reaction (qPCR or dPCR) for MRD monitoring in molecularly defined subgroups, that is, NPM1-mutated and core binding factor AML (containing translocations involving RUNX1-RUNX1T1 or CBFB-MYH11). MRD monitoring by mutiparameter flow cytometry (MFC) is recommended for all other patients. Thus, it is unclear from these guidelines how to incorporate NGS-based monitoring, what depth of detection is needed for each of the modalities, what time points are optimal for MRD assessments, and the best source of assessment (bone marrow [BM] vs peripheral blood [PB]).

MRD monitoring for patients with NPM1-mutated AML undergoing intensive chemotherapy is based on work by Ivey et al.³ Patients with NPM1-mutated AML who achieved complete remission (CR) with 2 cycles of intensive induction were studied. Persistence of NPM1 PCR transcripts in PB after 2 cycles of intensive induction chemotherapy was associated with greater risk of relapse at 3 years. Relapse was also reliably predicted by a rising level of NPM1 transcripts with sequential monitoring. The ELN guidelines recommend assessing MRD in NPM1-mutated AML with NPM1 PCR in PB following 2 cycles of intensive chemotherapy, in the BM at the end of treatment, and in either BM every 3 months or PB every 4 to 6 weeks for 24 months after therapy completion (see figure panel A). These guidelines also define an entity of MRD at low level (MRD-LL) in NPM1-mutated AML as <2% but above the limit of assay detection. MRD-LL is associated with a low risk of relapse when measured at the completion of consolidation. These guidelines also recommend individualized treatment strategies to reduce the risk of relapse if specific MRD trends are noted. The MRD trend groups include MRD positivity defined as ≥2% in the BM at completion of consolidation, failure to achieve a 3 to 4 log reduction in either the BM or PB at completion of consolidation, or MRD relapse defined as conversion from MRD negativity to MRD positivity or increase of MRD ≥1 log₁₀ between any 2 positive samples for patients with MRD-LL. Allogeneic stem cell transplantation may improve patient outcomes with NPM1-mutated AML who have suboptimal molecular responses after induction therapy.⁴ Preemptive therapy may benefit patients in these settings as persistent NPM1 MRD pretransplant has been associated with worse posttransplant outcomes.^5,6 The ELN MRD guidelines recommend treatment on clinical trial whenever possible to establish an evidence-based approach for patients with MRD persistence or relapse.

There are currently no established guidelines related to MRD monitoring for patients receiving less intensive induction regimens, in which venetoclax-based regimens are the new standard of care.^7,8 The VIALE-A study led to approval of venetoclax combined with azacitidine for patients with newly diagnosed AML who are ineligible for intensive chemotherapy. A follow-up study reported on MFC for MRD assessments on VIALE-A and found that 41% of patients achieved MRD negativity (defined as <10⁻³) during treatment. Of these patients, MRD negativity was achieved in 52% by the end of cycle 4 with the rest occurring later in treatment. Patients who achieved MRD negativity had superior overall survival (OS) compared with patients who did not.⁹ Early retrospective data also suggest that there may be patient populations able to stop venetoclax-based therapy without relapse with MRD negativity by MFC being a major predictive factor.¹⁰ 

The study by Othman et al is a retrospective review of patients with newly diagnosed AML with NPM1 mutations who achieved CR using regimens containing venetoclax with low-dose cytarabine or hypomethylating agents. This study assessed the impact of NPM1 MRD by RT-qPCR and found the deepest MRD responses (≥4 log₁₀ reduction from baseline) were predictive of better 2-year OS, with 44 patients (58%) achieving BM MRD negativity and a further 14 (18%) achieving a reduction of ≥4 log₁₀ from baseline as their best response. Achievement of BM MRD negativity at the end of 4 cycles of treatment was the factor associated with greatest improvement in OS on multivariable analysis. A subset of the patients with MRD negativity also stopped further treatment with a 2-year treatment-free remission rate of 88%. The authors also showed that event-free survival was worse for patients who were MRD⁻ in the PB but MRD⁺ in the BM compared with patients with MRD negativity in both sources.

This study provides valuable insight into the kinetics and prognostic significance of molecular MRD by NPM1 qPCR for patients with NPM1-mutated AML treated with less intensive venetoclax-containing regimens. It raises several important questions for additional investigation (see figure panel B). Some patients who achieved MRD negativity were able to stop therapy without experiencing relapse, raising the question of whether certain patients may be cured with venetoclax-based regimens. Future prospective studies of treatment deescalation/cessation for patients achieving MRD⁻ remissions by certain time points would be informative, including how to monitor these patients after treatment cessation. Conversely, could patients who do not achieve MRD negativity by early time points benefit from changing therapies? Another uncertainty is the optimal MRD threshold for clinically actionable decisions. In this study, patients with detectable MRD but ≥4 log₁₀ reduction had survival outcomes that were worse than MRD⁻ patients but better than those who achieved <4 log₁₀ reduction. Thus, it is not clear how patients with detectable, but lower, amounts of MRD should be approached. Patients with MRD-LL (<2% positive) after intensive consolidation are at low risk of relapse, but it is unclear if this is applicable for patients treated with lower-intensity regimens.

This study represents a step forward in continued efforts to understand the significance of MRD and incorporation into clinical decision-making for patients treated with venetoclax-based regimens. It paves the way for further studies exploring how to better define who, what, when, where, and how to use MRD status optimally.

Conflict-of-interest disclosure: A.S.M. has served on advisory boards for Bristol Myers Squibb, Syndax, Servier, Ryvu Therapeutics, Rigel Therapeutics, and AbbVie; has served on Data Safety and Monitoring Committees for Daiichi Sankyo, Jazz Pharmaceuticals, and Foghorn Therapeutics; and also serves as a senior medical monitor for the Leukemia and Lymphoma Society’s Beat AML Study. K.D.S. declares no competing financial interest.