Immunophenotypical and/or molecular aberrancies have been known for decades to mark either the majority of or particular subsets of acute lymphocytic leukemia (ALL), acute myeloid leukemia (AML) and chronic myeloid leukemia (CML). In ALL, the relatively homogeneous aberrant immunophenotype as well as the molecular aberrancies covering most cases (reviewed in Citation[1]) have already enabled the early assessment of the value of these aberrancies for the detection of residual disease after therapy and to define its prognostic impact. Moreover, for childhood ALL, minimal residual disease (MRD) assessment has been incorporated in clinical studies in risk assessment Citation[2,3]. With the molecular aberrancy t(9;22) present in all cases of CML, a highly sensitive quantitative PCR has been developed to monitor the effectiveness of therapy.
How different is the situation for AML? Both for the immunophenotypical and molecular make-up, there is huge heterogeneity from patient to patient and, at least with regards to the immunophenotype, usually also within the bone marrow (BM) blast cells of individual cases of AML. There have been no particular single molecular aberrancies identified applicable for MRD detection in the vast majority of patients. Depending on the ethnographical situation, eligibility added up to only a few tens of percent of patients. This has led to the assumption that an immunophenotypical approach might be instrumental to cover a much higher percentage of patients and, for AML, this approach might be the preferable, if not the only, adequate MRD approach.
Over the next few paragraphs, current ideas from BM immunophenotypic and molecular studies as well as new perspectives emerging from different approaches will be dealt with.
Current status of immunophenotypical MRD in AML
An increasing number of studies have shown that MRD cell frequency after different cycles of therapy offers a highly independent prognostic factor, both in adult and childhood AML Citation[4–9]. This can, at least retrospectively, be understood, since MRD represents the sum of the effect of all relevant cellular resistance mechanisms, pharmacokinetic resistance and other known and unknown factors affecting the effectiveness of treatment. Except for one childhood AML study Citation[10], these data were all derived from retrospective studies. Moreover, the studies were almost exclusively performed in a single-institute setting, all resulting in well-known potential pitfalls, such as bias in patient groups and subjective judgment. The next logical and obvious step would be to perform the studies in a prospective and preferably multicenter way.
All studies identified cutoff levels that allowed the definition of two or more patient groups with different prognoses, thereby opening the way for risk stratification. Cutoff values in the order of 0.01–0.1% of white blood cells are emerging from these studies Citation[4–10]. Despite these seemingly well-defined cutoffs, good-prognosis patient groups nevertheless also harbor poorly performing patients and, vice versa, poor-prognosis patient groups may contain good-performing patients. Such ‘unwanted’ contributions are logically dependent on the particular cutoff levels defined. At present, such studies can not yet be used to establish prognosis for individual patients at an early time point, except for the very poor-performing patients Citation[4] and the extremely good-performing patients Citation[8]. Improving sensitivity and specificity of immunophenotypical MRD detection may contribute to this highly valuable goal, but it is likely that the ultimate goal of 100% specificity and 100% sensitivity will not be reached.
Minimal residual disease assessment can also be instrumental in other settings first, to monitor the graft-versus-leukemia effects after allogeneic stem cell transplantation Citation[11] to fine-tune immune suppression and, second, to sequentially monitor for increases after the end of therapies in individual patients in order to be able to predict relapses early. It is well known, for example for CML, that it may not just be the absolute level of MRD but also a rise in levels of MRD during sequential sampling instead that accurately predicts disease progression and forthcoming relapses. For AML, with its relatively fast relapses, it has been calculated that such sequential BM sampling should take place with 3-month time intervals at most Citation[4]. Schittger and colleagues, based on their molecular studies, calculated that 75% of relapses can be anticipated using 42 days sampling intervals Citation[12]. Such an approach at first sight may not be realistic to perform in the majority of patients. High frequency of BM sampling would be a physical burden for both the patient and the physician/nurse, and, moreover, a psychological burden for the patient, especially given the fact that a considerable part of the good-prognosis patient group would have been cured anyway. Last, the exercise would be costly. Possible ways to circumvent or minimize part of these problems will be discussed in the ‘Perspectives’ section.
Prospective clinical MRD studies
By definition, prospective MRD studies performed in a blinded fashion would validate the cutoff frequency points established in retrospective analyses. Owing to the complexity of AML BM MRD analyses, in 2004 a few specialized flowcytometry laboratories affiliated with institutes that participate in clinical AML studies of the cooperative clinical study group for hematooncology diseases in Netherlands and Belgium (HOVON)/cooperative clinical study group for hematooncology diseases in Switzerland (SAKK) commenced a study exploring the feasibility of immunophenotypical MRD analyses in a multicenter setting. A similar exercise has been performed for MRD in ALL Citation[13]. The results of the study in AML [Feller et al., Unpublished Data] implied that immunophenotypical MRD studies should be performed in a limited number of specialized and experienced centers. With that in mind, two prospective MRD studies were started in parallel with two new HOVON/SAKK clinical studies: one for younger patients (<60 years) and one for elderly patients (>60 years). Studies have recently been closed after reaching the target number of patients of 518 and 220 patients, respectively. MRD determination was performed in four laboratories and MRD correlates with clinical outcome are expected soon.
Current status of molecular biological MRD
Minimal residual disease studies have, for many years, been restricted to PCR-able chromosomal aberrancies, such as t(15;17), t(8;21) and inv16 (reviewed in Citation[1]), together representing only a minority of patients of whom many have good prognosis. Even in these cases the sensitivity is sometimes questionable. Initially, the discovery that the FLT3 mutations occurred in over 25% of AML cases evoked enthusiasm regarding its suitability for molecular MRD. However, different reports on the stability of FLT3 internal tandem duplications (ITDs) show that mutations present at diagnosis may have disappeared at relapse Citation[12,14,15], while new mutations may arise Citation[12,14]. Also, the need for patient-specific PCRs makes the approach labor intensive. Since the percentage of disappearing mutations may not be that high, FLT3 ITDs may have a role in molecular MRD, probably in concert with other targets Citation[12]. Similarly, other mutations, such as Ras and WT1, are quite unstable, with significant losses at relapse Citation[16]. More recently, other PCR-able chromosomal abnormalities have changed the perspective for molecular MRD detection: the possibility to use nucleophosmin (NPM)-1 mutations Citation[12] on the DNA level Citation[17] and on the gene-expression level Citation[18], together with its high sensitivity (better than 1:105 cells), its stability during disease/treatment Citation[19] and its high frequency (∼30%, mainly in normal cytogenetics group) shows that molecular MRD may have a more widespread future in MRD monitoring. Other less frequent but stable mutations, such as those in CEBP-a, may also serve as suitable targets Citation[1]. It has to be noticed, however, that the best molecular targets, t(8;21), inv16 and NPM-1, are in the group of good-prognosis patients, emphasizing the need for additional molecular markers for the poor-prognosis patients.
The use of gene-expression analysis for MRD detection has been controversial. Wilm’s tumor (WT)1 for a long time has been known to be overexpressed in approximately 80% of AML cases and is, thereby, potentially usable for MRD detection. Relatively high expression in normal BM may hamper the applicability. Recent studies on WT1 expression, including a validation study in the European LeukemiaNet Citation[20], nevertheless suggest its potential applicability. Similarly, other genes such as PRAME, overexpressed in 30–40% of AML cases Citation[1], may be considered for gene expression MRD studies Citation[21]. An interesting approach has been adopted by Steinbach and colleagues for childhood AML in which a multiplex approach expression of a set of seven genes, including WT1 and PRAME, overexpressed in the majority of childhood AML, was used for MRD detection Citation[22].
Perspectives
The aforementioned developments for both prospective immunophenotypical MRD studies and growing opportunities for molecular (DNA- and RNA-based) MRD, indicates that the time has come to initiate comparative studies that, for obvious reasons, have not yet been performed on a large scale. Even in the absence of a well-defined final choice for particular molecular targets, it thus seems sensible to let the sampling for immunophenotypical MRD be paralleled by sampling of DNA and RNA at the same follow-up time points. We did so in the patient studies mentioned earlier in the ‘Prospective clinical MRD studies’ section, now enabling us to perform the comparisons with already chosen targets.
New developments in flowcytometry, such as new instruments and new fluorochrome development, now enable the expansion of the limited number of antibody–fluorochrome combinations per tube (usually 4–5) to an ever increasing number. Obviously, if one aims at performing multicenter flowcytometry studies, extensive validation and quality controls are required. In Europe, new eight-color protocols for diagnosis of all hematological malignancies have been developed in a multicenter fashion Citation[23]. The validation of application for MRD detection in AML, however, has not yet started. Despite this, some aspects of the already performed program are of particular interest in that respect: revolutionary new software enables the production of an artificial file in which expression of an unlimited number of antigens, detected using many separate tubes, can be directly compared with each other. In addition to this, while using the same software, a backbone of cumulative normal BM antigen-expression profiles can be produced. This probably results in a more objective assessment of the aberrant profiles used for MRD detection, with less need to be highly experienced in recognizing normal BM differentiation patterns.
It has been known for years that for molecular biological MRD, BM can be replaced by peripheral blood MRD in T-ALL, although not in B-ALL Citation[24]. The applicability for immunophenotypical MRD measurements in AML has only been tested recently and revealed a remarkably good correlation with BM-MRD, and it offered an independent prognostic factor for clinical outcome Citation[25]. In our own institution, we were able to confirm these data. One disadvantage is the lower sensitivity, approximately one log. However, this may be compensated by its higher specificity due to the absence of normal cells that in the BM are characterized by a number of differentiation stages that interfere with the correct recognition of aberrancies. We are currently discussing how PB MRD ultimately may be used in the clinic: changes in peripheral blood MRD to be seen as a guide to perform BM MRD or as a single prospective factor. Last, one study has shown that, similar to childhood and adult ALL Citation[1], blast clearance has a remarkably good prognostic impact Citation[26].
It is now generally thought that MRD is the result of limited outgrowth of therapy-resistant cells with leukemia initiating/stem cell ability. Information on the fate of the leukemic stem cells during therapy/disease might, thus, not only contribute to our understanding of the role of such cells in initiation, persistence and outgrowth of leukemia but also contribute to a correct prognosis for the patient. It turned out that the size of the stem cell compartment correlated with clinical outcome, not only at diagnosis Citation[27], but after therapy as well Citation[28,29]. The size of the stem cell compartment is approximately, however, at least two logs smaller than the MRD compartment, excluding some AML cases from such analyses. Moreover, the analyses have been used for CD34+CD38- stem cells, thereby limiting the analyses to CD34+ cases. There are also two advantages: first, the finding that AML stem cells can be recognized in a highly specific way using AML-specific markers. Similar to PB, this results in a more objective assessment of disease without the need for extensive experience in recognizing normal bone marrow differentiation patterns. Second, specific markers, such as CLL-1, CD123, CD96, CD44 (reviewed in Citation[30]) and CD47 Citation[31], may be used in those cases where no solid aberrant immunophenotypes usable for MRD detection can be formulated (typically 10% of the cases). On the other hand, the stem cell markers cannot replace MRD detection since, although absent on normal CD34+CD38- stem cells, they are clearly expressed on the more differentiated CD34+CD38+ progenitors, which are part of MRD analysis.
Conclusion
Owing to the instability of mutations (appearing and/or disappearing) it is likely that during and after treatment the disease can become more aggressive. Early signals for forthcoming relapses are thus eagerly awaited to ultimately guide early clinical interventions.
Immunophenotypic MRD studies in AML will become mature if prospective analyses show their prognostic impact. Remaining an as yet unanswered question is how to use the data: for postdiagnosis risk assessment in patients with intermediate cytogenetics; or identification of poor prognostic patients in good cytogenetics patients; and/or in the decision to perform allogeneic transplantation. In addition, MRD assessment may be particularly useful in subgroups of good-prognosis patients and to monitor the effects of allogeneic transplantation. It seems that the definition of subgroups of patients with 100% chance of relapse is possible at an early stage during treatment. It is more difficult to define patient groups with extremely good prognosis and, herein, the new approaches to improve the specificity of the methods (new multicolor approaches, use of peripheral blood and stem cell MRD) may have a prominent role. Moreover, further improvement on this may be expected from the application of the new molecular biological targets, especially if sensitivity is higher than immunophenotyping. Combinations of these approaches may ultimately bring us closer to the ultimate goal of real individualized risk assessment and therapy.
Financial & competing interests disclosure
The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.
No writing assistance was utilized in the production of this manuscript.
References
- Bene MC, Kaeda JS. How and why minimal residual disease studies are necessary in leukemia: a review from WP10 and WP12 of the European LeukaemiaNet. Haematologica94(8), 1135–1150 (2009).
- Schrappe M, Beier R, Burger B. New treatment strategies in childhood acute lymphoblastic leukaemia. Best. Pract. Res. Clin. Haematol.15(4), 729–740 (2002).
- Cazzaniga G, Biondi A. Molecular monitoring of childhood acute lymphoblastic leukemia using antigen receptor gene rearrangements and quantitative polymerase chain reaction technology. Haematologica90(3), 382–390 (2005).
- Feller N, van der Pol MA, van Stijn A et al. MRD parameters using immunophenotypic detection methods are highly reliable in predicting survival in acute myeloid leukaemia. Leukemia18(8), 1380–1390 (2004).
- Kern W, Voskova D, Schoch C, Hiddemann W, Schnittger S, Haferlach T. Determination of relapse risk based on assessment of minimal residual disease during complete remission by multiparameter flow cytometry in unselected patients with acute myeloid leukemia. Blood104(10), 3078–3085 (2004).
- Langebrake C, Creutzig U, Dworzak M et al. Residual disease monitoring in childhood acute myeloid leukemia by multiparameter flow cytometry: the MRD–AML–BFM study group. J. Clin. Oncol.24(22), 3686–3692 (2006).
- San Miguel JF, Martinez A, Macedo A et al. Immunophenotyping investigation of minimal residual disease is a useful approach for predicting relapse in acute myeloid leukemia patients. Blood90(6), 2465–2470 (1997).
- San Miguel JF, Vidriales MB, Lopez-Berges C et al. Early immunophenotypical evaluation of minimal residual disease in acute myeloid leukemia identifies different patient risk groups and may contribute to postinduction treatment stratification. Blood98(6), 1746–1751 (2001).
- Venditti A, Buccisano F, Del Poeta G et al. Level of minimal residual disease after consolidation therapy predicts outcome in acute myeloid leukemia. Blood96(12), 3948–3952 (2000).
- Sievers EL, Lange BJ, Alonzo TA et al. Immunophenotypic evidence of leukemia after induction therapy predicts relapse: results from a prospective Children’s Cancer Group study of 252 patients with acute myeloid leukemia. Blood101(9), 3398–3406 (2003).
- Maurillo L, Buccisano F, Del Principe MI et al. Toward optimization of postremission therapy for residual disease-positive patients with acute myeloid leukemia. J. Clin. Oncol.26(30), 4944–4951 (2008).
- Schnittger S, Kern W, Tschulik C et al. Minimal residual disease levels assessed by NPM1 mutation-specific RQ-PCR provide important prognostic information in AML. Blood114(11), 2220–2231 (2009).
- Dworzak MN, Gaipa G, Ratei R et al. Standardization of flow cytometric minimal residual disease evaluation in acute lymphoblastic leukemia: multicentric assessment is feasible. Cytometry B Clin. Cytom.74(6), 331–340 (2008).
- Cloos J, Goemans BF, Hess CJ et al. Stability and prognostic influence of FLT3 mutations in paired initial and relapsed AML samples. Leukemia20(7), 1217–1220 (2006).
- Kottaridis PD, Gale RE, Langabeer SE, Frew ME, Bowen DT, Linch DC. Studies of FLT3 mutations in paired presentation and relapse samples from patients with acute myeloid leukemia: implications for the role of FLT3 mutations in leukemogenesis, minimal residual disease detection, and possible therapy with FLT3 inhibitors. Blood100(7), 2393–2398 (2002).
- Bachas C, Schuurhuis GJ, Hollink IHIM et al. Frequent instability of type I and II aberrations between diagnosis and relapse observed in pediatric AML; implications for targeted therapy and MRD detection. Blood114, 649 (2009).
- Gorello P, Cazzaniga G, Alberti F et al. Quantitative assessment of minimal residual disease in acute myeloid leukemia carrying nucleophosmin (NPM1) gene mutations. Leukemia20(6), 1103–1108 (2006).
- Chou WC, Tang JL, Wu SJ et al. Clinical implications of minimal residual disease monitoring by quantitative polymerase chain reaction in acute myeloid leukemia patients bearing nucleophosmin (NPM1) mutations. Leukemia21(5), 998–1004 (2007).
- Kern W, Haferlach C, Bacher U, Haferlach T, Schnittger S. Flow cytometric identification of acute myeloid leukemia with limited differentiation and NPM1 type A mutation: a new biologically defined entity. Leukemia23(7), 1361–1364 (2009).
- Cilloni D, Renneville A, Hermitte F et al. Real-time quantitative polymerase chain reaction detection of minimal residual disease by standardized WT1 assay to enhance risk stratification in acute myeloid leukemia: a European LeukemiaNet study. J. Clin. Oncol.27(31), 5195–5201 (2009).
- Tajeddine N, Millard I, Gailly P, Gala JL. Real-time RT-PCR quantification of PRAME gene expression for monitoring minimal residual disease in acute myeloblastic leukaemia. Clin. Chem. Lab. Med.44(5), 548–555 (2006).
- Steinbach D, Schramm A, Eggert A et al. Identification of a set of seven genes for the monitoring of minimal residual disease in pediatric acute myeloid leukemia. Clin. Cancer Res.12(8), 2434–2441 (2006).
- Pedreira CE, Costa ES, Barrena S et al. Generation of flow cytometry data files with a potentially infinite number of dimensions. Cytometry A73(9), 834–846 (2008).
- van der Velden VHJ, Jacobs DCH, Wijkhuijs AJM et al. Minimal residual disease levels in bone marrow and peripheral blood are comparable in children with T cell acute lymphoblastic leukemia (ALL), but not in precursor-B-ALL. Leukemia16(8), 1432–1436 (2002).
- Maurillo L, Buccisano F, Spagnoli A et al. Monitoring of minimal residual disease in adult acute myeloid leukemia using peripheral blood as an alternative source to bone marrow. Haematologica92(5), 605–611 (2007).
- Kern W, Haferlach T, Schoch C et al. Early blast clearance by remission induction therapy is a major independent prognostic factor for both achievement of complete remission and long-term outcome in acute myeloid leukemia: data from the German AML Cooperative group (AMLCG) 1992 Trial. Blood101(1), 64–70 (2003).
- van Rhenen A, Feller N, Kelder A et al. High stem cell frequency in acute myeloid leukemia at diagnosis predicts high minimal residual disease and poor survival. Clin. Cancer Res.11(18), 6520–6527 (2005).
- Terwijn M, Kelder A, Rutten AP, Zweegman S, Ossenkoppele GJ, Schuurhuis GJ. High leukemic cell frequency in remission bone marrow predicts poor outcome in acute myeloid leukemia. Blood112, 881 (2008).
- Terwijn M, Kelder A, Rutten AP et al. Specific flow cytometric detection of leukemic stem cells in acute myeloid leukemia predicts prognosis in remission bone marrow. Blood114, 165 (2009).
- Krause DS, Van Etten RA. Right on target: eradicating leukemic stem cells. Trends Mol. Med.13(11), 470–481 (2007).
- Majeti R, Chao MP, Alizadeh AA et al. CD47 is an adverse prognostic factor and therapeutic antibody target on human acute myeloid leukemia stem cells. Cell138(2), 286–299 (2009).