MRD-guided timing of allogenic transplantation for AML

Allogeneic hematopoietic stem cell transplantation (allo-HSCT) is still considered as the only curative treatment of acute myeloid leukemia (AML). However, approximately 40% of patients experience a post-transplant relapse.^1,2, Even small amounts of residual tumor cells after chemotherapy induction predispose to relapse. Measurable (minimal) residual disease (MRD), referred to as sub-microscopic disease during or after therapy, has recently received lots of attention.  However, the sensitivity of methods used to detect these remaining leukemic cells, which can help to predict the risk of relapse, varies between technologies.³

A review of the latest developments in the technologies used for MRD detection, as well as its value and impact on the timing of allo-HSCT, by Omatayo Fasan from Novant Health Hematology, has been recently published in Current Opinion in Hematology.⁴ A summary of this article can be found below.

Techniques for MRD assessment

A range of techniques can be used for the assessment of MRD. The European LeukemiaNet (ELN) MRD Working Party has recently issued a report to guide practice and standardize the evaluation of MRD.⁵

Multiparameter flow cytometry (MFC)

MFC detects abnormal antigens on leukemic cells but requires high-quality sample material. Ideally, a fresh (<3 days old) bone marrow aspirate should be used as it contains higher diagnostic material and is considered to be more representative. However, a peripheral blood sample can also be used. Analysis of at least 500,000–1 million white blood cells (WBC), performed on equipment suitable for a minimum of 8-color MFC, is recommended.

WBC cells are usually assessed for remaining leukemic cells based on leukemia-associated immunophenotype (LAIP) and a ‘different from normal’ phenotype if pre-treatment samples are not available. A comprehensive antigen panel should be used, containing immature, differentiation and aberrant antigens including CD7, CD11b, CD13, CD15, CD19, CD33, CD34, CD45, CD56, CD117, and HLADR. Additionally, in cases of suspected monocytic leukemia, CD11b, CD14, and CD64 markers should be added to the panel. The ELN recommends using 0.1% as the threshold for MRD- status but advises adding a comment on the lack of clinical validation of this cut-off.

Molecular techniques

Molecular detection methods, such as quantitative polymerase chain reaction (qPCR), digital droplet PCR, and next-generation sequencing (NGS) have high sensitivity and are able to detect one abnormal cell among 10³–10⁶. Patient samples are screened for the presence of somatic mutations in relevant genes belonging to several important groups⁶:

• Signaling and kinase pathways (FLT3, KRAS, NRAS, KIT, PTPN11, NF1)
• Epigenetic modifiers (DNMT3A, IDH1/2, TET2)
• Transcription factors (CEBPA, RUNX1, GATA2) 
• Tumor suppressors (TP53)
• Spliceosome complex (SRSF2, U2AF1 SF3B1, ZRS2) 
• Cohesion complex (RAD21, STAG1/2, SMC1A/3)
• Nucleophosmin (NPM1)

Detection and monitoring of fusion genes, associated with the pathogenesis of AML and known to be strong predictors of relapse, such as PML-RARA, RUNX1-RUNX1T1, or CBFB-MYH11, using qPCR can also be useful to monitor MRD status.^7,8 However, the presence of a somatic mutation does not necessarily mean it can be used for the detection and monitoring of MRD. For example, WT1 mutations are only recommended to be used if no other MRD markers are available as they poorly correlated with the risk of relapse. Moreover, the ELN advises against the use of NPM1, FLT3-ITD, FLT3-TKD, NRAS, KRAS, IDH1/2, and EVI1 as single markers of MRD. Instead, they recommend a combination with a second MRD marker.

The ELN defines molecular progression as an increase of MRD copy numbers ≥1 log10 between two positive samples and encourages reporting absolute copy numbers in addition to the fold increase.

Impact of MRD and 0.1% positivity threshold cut off on patient outcome after transplantation

Several studies have recently examined the value of MRD in predicting the outcome for patients undergoing transplantation.

One of them analyzed clinical data from 185 high-risk patients with AML (>65 years old) undergoing allo-HSCT in order to design an MRD status-based risk score.⁹ Bone marrow aspirates taken before transplantation were analyzed by MFC to establish the MRD status. Based on the results patients could be stratified into three risk groups:

• High-risk: MRD+ with ≥1 additional risk factor or a major medical complication (MMC)
• Intermediate-risk: patients who did not fit into high-risk or low-risk criteria
• Low-risk: MRD− with ≤3 additional risk factors but no MMC

A high-risk score was associated with a lower 2-year overall survival (OS) of 7.7%, compared to 32.2% in the intermediate group and 76.2% in the low-risk group.

Another study evaluating pre-transplant MRD in 51 patients with NPM1 mutated AML found that patients who were MRD+ before transplant had a higher risk of relapse (64.7% vs 6%; p< 0.001) and significantly lower OS (38.8% vs 71.7%; p=0.014) compared to MRD− patients.¹⁰ Post-transplantation MRD was also predictive of relapse, independent of the detection method, with a hazard ratio (HR) of 4.96 and 4.36 for MFC and NGS of NPM1 mutation, respectively. ¹¹ Interestingly, only patients with pre-transplant MRD+ status detected by MFC, but not NGS, were at risk of relapse (HR of 4.63).¹¹

A different study looked at the implications of having a single cut-off of 0.1% (as recommended by the ELN) for MRD positivity.¹² The authors used MFC to evaluate 292 adult patients with AML and defined the baseline LAIP in normal or regenerating marrow aspirates between 2x10^-5 and 5.71x10^-4. This was used as the threshold ‘individual level for MRD assessment’.  They found that a ELN cut-off of 0.1% was not very informative. The individual MRD− status was associated with significantly improved survival compared with the 0.1% MRD− status (but MRD-individual positivity) and 0.1% MRD+ status. The MRD status after the second consolidation also had prognostic value. Similar doubt on the practicality of the 0.1% cut-off was also cast by another report¹³, where the 5-year cumulative incidence of relapse (CIR) did not differ between patients with MRD+ of <0.1% and those with >0.1% in CBF-AML and NPM1-mutated AML.

Other researchers showed the importance of MRD dynamics. In this study, MRD of 279 patients was evaluated using MFC, before and 21–35 days after transplantation.¹⁴ The worst outcome was observed in patients where pre-transplant MRD− status changed to MRD+ after the transplantation. Patients with negative MRD before and after allo-HSCT had much better outcomes compared to those with positive MRD. The 3-year OS was 76 vs 19%, and relapse-free survival 71% vs 14%, respectively.

MRD status and timing of the allo-HSCT

The author advocates an individualized approach to interpreting the MRD status for transplantation decisions. For each case, clinical background, method of MRD evaluation, and dynamics of the MRD status should be taken into consideration. Hopefully, in the future, the consequences of MRD status in patients in complete remission on the timing of allogeneic transplantation will become clearer. The data currently available indicate similar outcomes for patients with MRD+ disease and those with persistent disease, with a 3-year OS in the range of 20–30% and 3-year CIR of around 65–80%.^14,15

Impact of pre-transplant MRD status on the selection of conditioning regimens and post-transplant maintenance therapy

Currently, there is no data to guide the selection of appropriate conditioning regimens dependent on the MRD status. Although MRD+ status would prompt a more intensive conditioning regimen, the published data suggest that such an approach does not necessarily translate into better outcomes.^9,14,15 However, recent findings of a retrospective assessment of MRD status in patients on the BMTCTN 0901 trial reported reduced OS, disease-free survival, and increased risk of relapse (HR of 6) in MRD+ patients who underwent conditioning with reduced intensity regiments (RIC) compared to those on the myeloablative regimens (MAC). In contrast, for patients who were MRD−, the outcome was not different after RIC or MAC.¹¹

Based on data from the RELAZA and RELAZA2 studies, post-transplantation pre-emptive treatment of MRD+ patients with azacitidine could delay, or even prevent, relapse in MRD+ patients.^13,16

Summary

The author of the review concludes that MRD status is a good prognostic factor but its use should be standardized. Hopefully, the ELN MRD Working Party guidelines on the testing will help to achieve this. The dynamics of MRD add relevant information about the patient's risk of relapse. Therefore, it is important to look at the changes in the MRD status over time, rather than use it as a static value collected at a single time-point.

There is a lot of variability on how the MRD status is incorporated into clinical practice, in particular its impact on treatment decisions and on the timing of allo-HSCT. To address this, prospective randomized trials, with MRD evaluation at defined pre-and post-transplant time-points, are needed.