Minimal Residual Disease in AML: Worth Looking?

Get Permission

Selina M. Luger, MD

Marlise R. Luskin, MD, MSCE

Further study must focus on extending applicability across clinical approaches and AML subgroups, as well as proving that a minimal residual disease–based therapeutic intervention can improve outcomes.

Selina M. Luger, MD, and Marlise R. Luskin, MD, MSCE

As summarized in this issue of The ASCO Post, Ivey and colleagues demonstrated that assessing for NPM1-mutated gene transcripts by reverse-transcriptase quantitative polymerase chain reaction assay is a feasible approach for measuring minimal residual disease after acute myeloid leukemia (AML) induction therapy.1 In addition, the researchers showed that the presence of NPM1-based minimal residual disease in this setting is strongly associated with inferior outcome.1

Clinical adoption of minimal residual disease assessment in AML has been hampered by the difficulty of standardizing flow-based minimal residual disease assays for AML, as well as lack of clarity regarding appropriate time points for assessment and uncertainty surrounding the clinical utility of the information obtained.2 This study demonstrates the value of assessing minimal residual disease in AML by rigorously defining and validating the prognostic performance of an AML minimal residual disease assay in a large cohort of patients uniformly treated on the UK National Cancer Research Institute AML17 trial. However, clinical benefit remains to be demonstrated.

We have had few tools to accurately predict long-term outcome in patients with intermediate-risk AML. Recently, the identification of recurrent gene mutations in AML has helped refine AML prognostication, especially for patients with intermediate-risk disease.3-5 In this group, the presence of a DNMT3A and/or an FLT3 internal tandem duplication (ITD) mutation confers a poorer prognosis, whereas the presence of an NPM1 mutation in the absence of the aforementioned mutations is associated with an improved prognosis.6-9

Heterogeneity of Outcomes

Despite the prognostic contribution of molecular profiling, the heterogeneity of long-term outcomes among AML subgroups remains profound, and there is a pressing need for the development of new prognostic tools. Minimal residual disease–based assays are prime candidates. The theoretical advantage of a minimal residual disease–based prognostic marker relies on the assumption that relapses arise from submicroscopic disease reservoirs remaining after initial treatment and that minimal residual disease is the result of a combination of factors that underlie an individual’s response to therapy (including leukemic stem cell biology, mechanisms of leukemogenesis, and drug-resistance mechanisms). Minimal residual disease reflects an individual’s actual response to therapy as opposed to his or her expected prognosis based on genetic group.

Ivey et al demonstrated that the persistence of detectable mutated NPM1 transcript in the peripheral blood after two cycles of induction chemotherapy (standard therapy in UK AML induction protocols) reliably predicts risk of relapse. In these patients, the risk of relapse was 82% at 3 years vs 30% in the absence of such transcripts (hazard ratio [HR] = 4.8, P < .001), with a 3-year overall survival of 24% vs 75% (HR = 4.38, P < .001).

The authors compared the prognostic prowess of minimal residual disease to that of diagnostic features, including extended molecular profiling (using a 51-gene panel). Minimal residual disease–positive status had by far the strongest association with risk of relapse and death, and multivariable analysis showed minimal residual disease to be the only significant prognostic factor for both measures (HR for relapse = 5.01, P < .001; HR for death = 4.84, P < .001).

Practical Implications

What does this mean for clinical practice? Genetic characteristics are merely one facet of complex disease biology that is likely better assessed by evaluation of disease response via a reliable and validated minimal residual disease tool. For example, Ivey et al showed that within a high-risk molecular subgroup (DNMT3A-mutated and FLT3-ITD–positive), those achieving minimal residual disease negativity (for NPM1) have excellent outcomes (86% survival at 2 years), while the converse was demonstrated for the favorable molecular NPM1-mutated, DNMT3A–wild type, FLT3-ITD–negative subgroup, in which patients with minimal residual disease had poor survival.1

The value of extended molecular profiling at diagnosis still has relevance for defining the presence of therapeutic targets, determining AML ontogeny,10 and—as shown here—identifying the appropriate target for minimal residual disease assessment. However, the value of molecular profiling for AML prognostication has its limitations. Genetic complexity at diagnosis is staggering (in this study, 150 subgroups were identified), making subgroup study difficult; this does not even account for differences in clonal diversity or sequence of mutation acquisition, additional genetic factors that may affect prognosis.

It is worth highlighting that although minimal residual disease was more frequently detected in bone marrow samples, the investigators determined that peripheral blood minimal residual disease measurement was the preferred tissue sample for discriminating clinical outcome. This has the significant advantage of being easier to obtain and not dependent on operator technique, which will facilitate application of these findings.

Note of Caution

Although this study introduces the exciting prospect of a new way to assess prognosis and determine therapeutic approach in AML patients, caution should be exercised before widespread adoption. The conclusions presented are specific to the characteristics and sensitivity of the reverse-transcriptase quantitative polymerase chain reaction assay used, as well as to the timing and therapeutic context of assessment. In the United States, where double induction is not standard, the applicability of assessment after recovery from a second induction cycle is not clear. Would a similar minimal residual disease assessment be useful after a patient achieves remission from a single induction cycle?

Furthermore, this approach only applies to approximately half of the patients with intermediate-risk AML (those with NPM1 mutations) and cannot be applied to those with favorable or adverse cytogenetic risk, which significantly limits its applicability. Whether AML community efforts should focus on developing minimal residual disease assays for specific subgroups or a more universally applicable assay is unclear. Finally, it is extremely important to recognize that this study does not demonstrate that any particular postremission approach assigned based on minimal residual disease results can alter clinical outcomes. This is an absolutely essential next step.

Ivey and colleagues have brought the field of AML prognostication a step forward by successfully demonstrating that a molecular-based minimal residual disease assay performs impressively as a prognostic tool for AML patients receiving induction chemotherapy, with the tantalizing possibility of being able to guide advances in AML therapy. Further study must focus on extending its applicability across clinical approaches and AML subgroups, as well as proving that a minimal residual disease–based therapeutic intervention can improve outcomes. ■

Disclosure: Drs. Luger and Luskin reported no potential conflicts of interest.


1. Ivey A, Hills RK, Simpson JV, et al: Assessment of minimal residual disease in standard-risk AML. N Engl J Med 374:422-433, 2016.

2. Paietta E: When it comes to MRD, AML ≠ ALL. Blood 120:1536-1537, 2012.

3. Patel JP, Gonen M, Figueroa ME, et al: Prognostic relevance of integrated genetic profiling in acute myeloid leukemia. N Engl J Med 366:1079-1089, 2012.

4. Cancer Genome Atlas Research Network: Genomic and epigenomic landscapes of adult de novo acute myeloid leukemia. N Engl J Med 368:2059-2074, 2013.

5. Grimwade D, Ivey A, Huntly BJ: Molecular landscape of acute myeloid leukemia in younger adults and its clinical relevance. Blood 127:29-41, 2016.

6. Ley TJ, Ding L, Walter MJ, et al: ­DNMT3A mutations in acute myeloid leukemia. N Engl J Med 363:2424-2433, 2010.

7. Gale RE, Lamb K, Allen C, et al: Simpson’s paradox and the impact of different DNMT3A mutations on outcome in younger adults with acute myeloid leukemia. J Clin Oncol 33:2072-2083, 2015.

8. Schlenk RF, Dohner K, Krauter J, et al: Mutations and treatment outcome in cytogenetically normal acute myeloid leukemia. N Engl J Med 358:1909-1918, 2008.

9. Peterlin P, Renneville A, Ben Abdelali R, et al: Impact of additional genetic alterations on the outcome of patients with NPM1-mutated cytogenetically normal acute myeloid leukemia. Haematologica 100:e196-e199, 2015.

10. Lindsley RC, Mar BG, Mazzola E, et al: Acute myeloid leukemia ontogeny is defined by distinct somatic mutations. Blood 125:1367-1376, 2015.

Related Articles

Minimal Residual Disease Identified by NPM1 Mutation Is a Powerful Marker for Poorer Outcome in Standard-Risk AML

In a study reported in The New England Journal of Medicine, Adam Ivey, MSc, from Guy’s Hospital, London, and colleagues found that a leukemia-specific marker consisting of mutation in the gene encoding nucleophosmin (NPM1) can be used to identify minimal residual disease in peripheral blood in...