Patients with NPM1-mutated acute myeloid leukemia (AML) who were measurable residual disease (MRD) negative as measured by real-time quantitative PCR (RT-qPCR) after a fourth cycle of venetoclax-based treatment had improved outcomes compared with patients with detectable MRD, according to data published in Blood.1
Specifically, patients achieving MRD negativity had a cumulative incidence of relapse of 10% at two years compared with 72% in patients who remained MRD positive.
This study demonstrated that RT-qPCR MRD can add important prognostic information to the treatment of these patients with venetoclax combinations by providing “an important milestone for identifying patients who are likely to do badly if they discontinue their current treatment,” said study researcher Richard Dillon, MA, PhD, MRCP, of Guy’s and St. Thomas’ NHS Foundation Trust in London.
Although flow cytometry was prognostic in patients treated with venetoclax and azacitidine in the VIALE-A study, flow MRD did not predict outcomes well in patients with NPM1-mutated AML.2
“This is a large subgroup of patients, accounting for about one-third of adult AML, and they frequently lack an informative leukemia-associated immunophenotype, limiting the sensitivity of flow MRD,” Dr. Dillon said.
Patients with NPM1-mutated AML have very good outcomes with venetoclax-based non-intensive treatment, and because of this, there may be a tendency to treat some slightly younger and fitter patients whose AML has an NPM1 mutation with these regimens. This makes it all the more important to be able to accurately identify patients who are not going to do well and consider alternative treatment approaches for them, Dr. Dillon said.
In this study, researchers identified 76 patients from a national real-world cohort in the U.K. and from patients treated in Australia. Patients had received venetoclax with either hypomethylating agents (HMAs) or low-dose cytarabine (LDAC) as firstline therapy for NPM1-mutated AML, achieved complete response (CR) or CR with incomplete hematologic recovery, and had at least one bone marrow MRD assessment in the first four cycles of therapy. RT-qPCR was performed.
A median of 2.5 (range = 1-6) bone marrow MRD assessments were performed during the first six cycles. The best MRD response at any time during therapy was MRD negative in 58% of patients, positive with at least 4 log10 reduction from baseline in 18%, and less than 4 log10 reduction in 24%.
With successive cycles of therapy, MRD response deepened. At the end of two cycles, the cumulative rate of MRD negativity was 25%; by four cycles, 47%; and by six cycles, 50%.
“We showed that kinetics of MRD response with venetoclax are slower than typically seen with intensive chemotherapy protocols. Fewer than 10% of patients achieved MRD negativity after the first cycle, and some patients took more than six cycles to achieve this,” Dr. Dillon said. “Hopefully this will encourage doctors not to panic too much if the early MRD response is not as good as they are used to.”
With a median follow-up of 28 months, the two-year overall survival (OS) was 63% (95% CI 53-76), and the two-year event-free survival was 52% (95% CI 40-66). Among patients who achieved MRD negativity by the end of cycle four, the respective two-year OS was 84% compared with 46% in those who remained MRD positive.
Patients with IDH1 and IDH2 co-mutations had a high rate of early NPM1 BM MRD negativity (83% and 67%, respectively) compared with secondary AML (29%) and FLT3 mutations (28% with FLT3-ITD and 36% with FLT3-TKD, respectively). Multivariable analyses showed achievement of MRD negativity in four cycles (as a time-dependent variable) to be the most important factor associated with survival with a hazard ratio of 0.21 (95% CI 0.08-0.55; p=0.002).
Twenty-two patients elected to stop therapy in MRD-negative remission; they had a two-year treatment-free remission of 88%. One patient was in treatment-free remission for more than eight years, Dr. Dillon said.
“Although flow MRD measurement may be suitable for the majority of patients treated with venetoclax, if you have a patient with AML with NPM1 mutation, molecular MRD measurement using RT-qPCR should be considered and likely has much better prognostic value,” Dr. Dillon said. “This is especially important if your patient may be able to access alternative therapies or could be a transplant candidate. There is a clear need for prospective studies in this population.”
The researchers cited several limitations to the study, including its retrospective nature, which introduced the potential for selection bias, though steps were taken to mitigate this. Additionally, the timing of MRD measurements was heterogeneous, and not all patients underwent a bone marrow assessment after each cycle.
Any conflicts of interest declared by the authors can be found in the original article.
References
- Othman J, Tiong I, O’Nions J, et al. Molecular MRD is strongly prognostic in patients with NPM1-mutated AML receiving venetoclax-based non-intensive therapy [published online ahead of print, 2023 Aug 30]. Blood. doi: doi.org/10.1182/blood.2023021579.
- Pratz KW, Jonas BA, Pullarkat V, et al. Measurable residual disease response and prognosis in treatment-naïve acute myeloid leukemia with venetoclax and azacitidine. J Clin Oncol. 2022;40(8):855-865.
Perspectives
MRD monitoring in AML presents several challenges for broader clinical use and interpretation, not the least of which is whether all leukemic markers are equally good for MRD tracking. Mutations in NPM1 have been proposed as one such marker to track during therapy; an important implication is that if tracking NPM1 mutation levels is truly representative of disease burden, it should be prognostic regardless of the therapy used to get into remission. NPM1 mutation levels have been shown to correlate with outcomes in intensive chemotherapy regimens, but to date it is less clear whether this molecular tracking is also significant with less intensive regimens such as HMAs plus venetoclax or LDAC plus venetoclax.
In contrast to intensive chemotherapy, responses can require more cycles with venetoclax combinations; interestingly, even if it took more than four cycles to achieve an MRD-negative response (threshold of <0.002 copies per 100 ABL1), these patients appeared to have similar survival benefit to those who responded early in treatment.
This study suggests the value of using NPM1 MRD testing serially during treatment of patients with AML harboring an NPM1 mutation, regardless of the therapy used.
The data also support the fact that undetectable NPM1 transcripts are associated with excellent long-term outcomes; what is less clear, however, is whether patients who don’t achieve such deep responses can benefit from alternative therapies, for instance allogeneic transplant or targeted therapies such as menin inhibitors.
This study brings up some exploratory considerations. First, it suggests that peripheral blood MRD testing is not as sensitive as bone marrow testing; this is in contrast to other assays, for instance BCR-ABL monitoring in chronic myeloid leukemia, and should be further evaluated in larger patient cohorts for validation. Moreover, as assays improve over time, it remains unclear what optimal threshold should be used to define residual MRD; what is negative today may not be the same as our technologies improve.
In addition, the idea of treatment cessation in LDAC- or HMA-based regimens for AML remains an emerging concept. This study provides an interesting series of 22 patients who stopped therapy without detectable NPM1 transcripts; at 15 months, only two had relapsed. Although not ready for standard clinical practice, such data add to literature supporting a prospective clinical trial to investigate whether some patients may indeed stop treatment and have durable remissions, or perhaps even cures, using venetoclax-based therapy.
Andrew M. Brunner, MD
Massachusetts General Hospital Cancer Center
Harvard Medical School
Boston, Massachusetts