SOHO 2019 | MRD in AML

During the annual meeting of the Society of Hematologic Oncology (SOHO), held in Houston, TX, US, Pau Montesinos, Hospital La Fe, Valencia, ES, discussed the latest data regarding the use of minimal residual disease (MRD) in acute myeloid leukemia (AML).¹

Dr Montesinos began by highlighting the unanswered questions about MRD in AML:¹

1. What method should be used to measure MRD, and at what time point should the prognostic impact of MRD be assessed?
2. Is MRD an acceptable endpoint?
3. What is the prognostic significance of MRD in young, intensively-treated patients with AML, and older, less intensively-treated patients with AML?
4. How can we use MRD in clinical practice?

Evolving the use of MRD in AML²

The 2017 European LeukemiaNet (ELN) guidelines categorize patients with AML who are both, in complete remission (CR) and MRD negative (CR_MRD-), as a separate response group to those in CR, CR with incomplete hematologic recovery (CR_i), morphologic leukemia-free state (MLFS) or partial remission (PR). This distinction has been made because relapse is more likely in patients in CR or CRi with MRD.² The guidelines recommend MRD to be assessed by the presence of a genetic marker by real time quantitative polymerase chain reaction (RT-qPCR) or by multi-parameter flow cytometry (MFC). The assessment should be done at early time points, such as post-induction and consolidation so that remission status can be determined, and kinetics of disease response can be evaluated. Assessments may also be done beyond consolidation therapy to identify morphologic relapse that may occur.²

MRD- versus morphologic-based definitions of remission³

In AML, morphologic CR is distinct from active disease (≥5% marrow blasts by morphology). Patients undergoing transplant who are in morphologic CR but who have MRD (CR_MRD+) have poorer outcomes compared to those who are CR_MRD-. Araki D et al., retrospectively compared the outcomes of patients with AML (N= 359) who underwent allogeneic transplant (allo-transplant) with active disease, to those in CR_MRD+ and CR_MRD-. In these cases, transplant source was peripheral blood stem cells (PBSCs) or bone marrow (BM) stem cells and MRD was analyzed by 10-color MFC on BM aspirates prior to transplantation. Survival outcomes by MRD status are shown in Table 1.³ This study found:

• CR_MRD- status pre-transplant was significantly associated with longer OS and progression-free survival (PFS) compared to patients with active disease or CR_MRD+
• Patients with CR_MRD+ status pre-transplant had similar outcomes to those with active disease
• The authors supported the use of MRD-based treatment algorithms over morphological assessments³

Changing paradigm of treating AML

Dr Montesinos explained how the treatment of AML has evolved over time, beginning with a one-size-fits-all chemotherapy approach when AML was believed to be a single entity. Since then, tailored, risk-adapted therapies also using allo-transplant have been developed based on risk stratification of distinct molecular and chromosomal AML subtypes. The future direction of treatment likely lies in the continued investigation of molecular targets, as markers for MRD detection and for the development of targeted therapies.

MRD to guide treatment decisions⁴

In a review article by Buccisano et al., the potential prognostic and therapeutic indications of MRD in AML were summarized. The authors explained that MRD assessment is a key measure of response to chemotherapy and it enables clinicians to plan post-remission strategies that are driven by individual risk of relapse. They noted that MRD assessment at the end of consolidation, in combination with pre-treatment cytogenetic information, will assist in determining which patients are suitable for autologous stem cell transplant (ASCT), based on specific risk of relapse rather than availability of an appropriate donor.⁴

Buccisano et al., also evaluated the outcomes of an MRD-guided prospective high-risk cohort where the decision for ASCT was based on pre-treatment genetic/cytogenetic information and MRD-positivity post-consolidation. Outcomes of the “prospective” cohort were compared to an “historical” high-risk cohort where treatment decisions were not based on risk stratification (Table 2). They showed at a median follow-up of 18 months that survival estimates were superior for the prospective cohort compared to the historical cohort, supporting MRD-guided decision making.⁴

Results of MRD studies in AML

HOVON/SAKK AML 42A

The HOVON/SAKK AML 42A study evaluated MRD status in BM samples of adult patients with AML (<60 years old) in CR.⁵ MRD was assessed by flow cytometry, with the absolute numbers of MRD cells/ml determined by multiplying the MRD frequency with white blood cell count (WBC)/ml in the BM.⁶ To compare flow cytometry MRD assessment with molecular MRD assessment, the authors also assessed MRD by qRT-PCR for the NPM1 mutation, AML1-ETO, and CBFB-MYH1.⁵

The authors demonstrated that MRD by flow cytometry was an independent prognostic factor for OS and PFS. Table 3 shows the median MRD values at each stage of treatment. Multivariate analysis after cycle two showed that MRD-positive patients had a significantly higher risk of relapse across multiple subgroups (late CR, WBC>100 x 10⁹/L, and intermediate risk vs good risk).⁵

The relapse incidence was higher for MRD-positive compared to MRD-negative patients (Table 4). The results confirmed MRD values of >0.1% of WBC after cycle two (when consolidation treatment decisions are made) were associated with a higher risk of relapse. The authors recommended that patients in the intermediate cytogenetic risk group who were MRD-positive with a late CR after cycle two, should be treated as if they were poor risk.⁵ These results suggest MRD status may be more predictive further along the treatment pathway.^1,5

PETHEMA⁷

San Miguel et al. evaluated MRD in BM samples from patients in morphologic CR following induction (n= 126) who had aberrant phenotypes at diagnosis. They categorized patients into four risk groups (very low, low, intermediate, high) based on MRD level of LAP⁺ cells as assessed by flow cytometry and evaluated their outcomes (Table 5). It was shown that MRD level influenced the three-year relapse-free survival (p= 0.0001) and OS (p= 0.003) of patients.⁷

Dr Montesinos then presented currently unpublished data, of 1118 patients in first CR (CR1), who were categorized by local MRD level post-induction (<0.01%, >0.01–0.1%, >0.1%). MRD was detected as per PETHEMA protocols listed above. Both OS and CIR were significantly associated with MRD level (p< 0.001 and p= 0.01 respectively). 

However, there was no significant difference in CIR related to cytogenetic risk (Table 6). Additionally, it appears that after allo-transplant and ASCT, the prognostic value of MRD is lost (p= 0.54 and p= 0.38 respectively), indicating MRD status may only be prognostic in patients receiving chemotherapy as post-CR therapy.¹

FLUGAZA⁸

Little data is available on the impact of MRD in older patients with AML treated with less intensive therapies. The phase III PETHEMA-FLUGAZA trial evaluated the quality of CR by MRD-status after induction and consolidation as a secondary objective in patients with AML (n= 285) aged >65 years. MRD was assessed by MFC.⁸

• Patients were screened and split into two arms:⁸
  • Fludarabine, cytarabine and G-CSF (FLUGA; n= 141)
    • Three induction cycles of FLUGA
    • Six consolidation cycles of reduced-intensity FLUGA
  • Azacytidine (AZA; n= 144)
    • Three induction cycles with AZA
    • Six consolidation cycles with AZA

OS was similar in patients in CR with detectable MRD (12 months) and patients in PR (13 months) compared to patients with undetectable MRD (21 months, p= 0.12). MRD >0.1% conferred a significantly higher risk of relapse and inferior OS in elderly patients with AML. Dr Paiva, who presented this data at the 60^th American Society of Hematology (ASH) annual meeting, stated that the risk of relapse in patients with undetectable MRD is still high and thus innovative approaches are required to maintain MRD-negativity.⁸

Ongoing studies which use MRD for treatment decisions in AML

HOVON/SAKK AML 132 (HO132)⁹

Dr Montesinos then showed the treatment pathway of the ongoing HO132 study, which is a randomized study to assess standard remission-induction chemotherapy and post-remission treatment plus/minus lenalidomide in patients aged 18–65 with previously untreated AML.⁹

Following achievement of a CR/CRi after two cycles of induction therapy, patients in both arms of the study will undergo PBSC mobilization and treatment decisions based on MRD-status, cytogenetics, and patient status:⁹

• Where ASCT is not possible: mitoxantrone + etoposide
• Good risk, or intermediate risk + MRD-negative: busulfan + cyclophosphamide and ASCT
• Poor risk or very poor risk: allo-transplant

MRD will be assessed by flow cytometry (immuno-MRD) and by NPM1 levels (molecular-MRD).

FLOW-2019¹

The design protocol for PETHEMA FLOW-2019 was also presented, which is a study of patients who are NPM1/core binding factor (CBF) negative (intermediate risk) and is based on centralized, rather than local, MRD assessment by RT-qPCR. Patients achieving a CR after first induction (Ara-C + IDA) with intermediate cytogenetics will proceed to second induction followed by centralized MRD assessment. Patients with MRD>0.1% will receive allo-transplant, and ≤0.1% will receive further chemotherapy and ASCT.¹

Conclusion

Although many trials have confirmed that CR with detectable MRD is not better than partial remission, Dr Montesinos concluded that there is currently no gold standard for the use of MRD in the treatment of AML. To date, MRD has only formed an exploratory endpoint for clinical trials. Currently, Dr Montesinos’s recommendation is to use MRD in well-defined interventional trials and for research purposes.