
Acute lymphoblastic leukemia (ALL) is a neoplasm of lymphoid progenitor cells committed to the B- (80-85% of diagnoses), T- (20-25%), and natural killer cell lineages (1%) [1]. As it represents one-quarter of diagnosed childhood cancers, it is the most common malignancy in minors [2,3]. Risk-adapted therapy and dose intensification strategies have greatly improved the outcomes of children with ALL, with an overall complete remission (CR) rate of >95% for B-ALL and 70% for T-ALL [4,5]. However, relapses still occur in 20% of children with this disease, implying the existence and proliferation of residual ALL cells that are not eradicated during therapy but remain undetected during conventional cytomorphologic assessment in which CR is defined as <5% visible leukemic blasts in the bone marrow and blood [4,6-8].
Minimal residual disease (MRD) is a consequence of the abovementioned undetected cells that are responsible for relapse [4,6-8]. It can only be detected and quantified using techniques with higher sensitivity and specificity for leukemic cells and is therefore also referred to as “measurable residual disease” [9-12]. The measurement of MRD is a response-based assessment that cannot be predicted via conventional pretreatment covariates such as age, white blood cell count, and cytogenetics, indicating that responses to therapy are patient-specific and reflect the
Classical MRD assays widely used for pediatric ALL include multiparameter flow cytometry (MFC) and real-time quantitative polymerase chain reaction (RQ-PCR) [9-11,30,31]. Newer MRD assaying technologies such as next-generation flow cytometry (NGF), next-generation sequencing (NGS), and droplet digital PCR (ddPCR) have also been developed [9-11,24,30-33]. Frequently used targets for MRD detection in pediatric ALL include aberrant immunophenotypes, fusion gene transcripts, and clonal immunoglobulin (
MFC identifies the aberrant immunophenotypes in ALL cells. RQ-PCR is used to detect fusion transcripts and clonal
Table 1 . Characteristics of classical minimal residual disease assays used in pediatric acute lymphoblastic leukemia.
Assay technique and targets | Applicability and sensitivity | Advantages | Disadvantages |
---|---|---|---|
Multiparameter flow cytometry for LAIPs | >90% | Wide applicability | Lower sensitivity than RQ-PCR |
3-4 colors: 10−3-10−4 | Low cost | Requires fresh samples (<24 h old) | |
Short turnaround time | Requires diagnostic sample to identify LAIPs | ||
6-8 colors: 10−4-10−5 | Exclusion of apoptotic cells lacking leukemogenic potential | Immunophenotypic shifts may cause false negatives | |
Requires high level of expertise to interpret data | |||
Analysis at cell population or single cell level | Limited standardization | ||
Real-time quantitative PCR for fusion transcripts (mainly on RNA) | B-ALL: 25-30% | High sensitivity | Limited applicability |
Short turnaround time | RNA instability | ||
T-ALL: 15-25% | Stable target throughout treatment | Risk of contamination | |
Wide availability of primer sets | Requires standard curves | ||
10−4-10−5 | Standardization for recurrent fusion transcripts | Risk of inaccurate quantitation | |
False positive results owing to nonspecific amplification of normal DNA or of cells without leukemogenic potential | |||
Real-time quantitative PCR for | 90-95% | High sensitivity | Long turnaround time |
10−4-10−5 | Wide applicability | Generation of patient-specific allele specific oligonucleotide primer sets is cumbersome | |
Standardized protocol and data interpretation | Requires prior knowledge of | ||
Clonal evolution can lead to false negatives | |||
Relative clone load quantitation is affected by the proportion of B/T lymphoid cells |
B-ALL, B-cell acute lymphoblastic leukemia;
Leukemic cells show expression patterns that differ from those of their normal counterparts and are referred to as leukemia-associated immunophenotypes (LAIPs). These include the asynchronous co-expression of early and late antigens, over- or under-expression of normally expressed antigens, and expression of cross-lineage antigens [34-36]. The principle of the MFC MRD assay is ALL cell recognition through patient-specific LAIP markers combined with backbone markers that enable the gating of lymphoid precursors [34-36]. The patient-specific LAIP should be identified at diagnosis (i.e., before commencing any therapy) by comparing the immuno-phenotypic profile of the ALL cells to reference bone marrow samples using various combinations of monoclonal antibodies.
MFC MRD is applicable for >90% of pediatric ALL [9-11,29,30,37]. Its sensitivity is between 10−3 and 10−5 (3-6 colors), which is mainly determined by the number of cells acquired, number of antigens used, degree of immunophenotypic deviation of the leukemic blasts, and the proportion of normal counterparts [26,28,36,38-40]. For example, 1,000,000 events should be acquired to achieve a sensitivity of 10−4 with a coefficient of variation of 10% [36,41]. The quantity of MRD is usually expressed in percentages; the denominator depends on the protocol used and may be total nucleated cells, total white or nonerythroid cells, or mononuclear cells. Therefore, caution should be taken when comparing MFC MRD results from different laboratories [36].
Compared to those of PCR assays (discussed below), the main advantages of MFC include its shorter turnaround time (TAT; approximately 4 hours), lower cost, and wider applicability. Hence, this assay is the most widely used for MRD detection in pediatric ALL [9-11, 13,30,31,36,37]. It excludes apoptotic cells that could contribute to false positive results in PCR assays by gating out cells with high side scatter [37]. Moreover, MFC analyzes antigen expression at the single-cell level but provides data on all cells in the entire sample simultaneously [9-11,30]. Specific immunophenotypes are indicative of the prognosis and/or particular cytogenetic and molecular genetic abnormalities [32,42].
Nevertheless, the sensitivity of MFC is still lower compared to that of PCR assays, and its requirement of fresh samples (<24 h old) precludes its use with archived specimens [36,43]. Diagnostic samples are required to identify LAIPs, whereas phenotypic shifts are frequent after treatment or elevated on relapse, which can cause false negative interpretations [35,36]. It requires expert knowledge of antigen expression observed during differentiation of normal hematopoietic progenitors, although the assessment could still be subjective [28,36].
A standardized MRD protocol for B-ALL is available [44]. Standardized antibody panels for 8-color flow cytometry were optimized and validated for the diagnosis and subclassification of hematologic malignancies including ALL; these could be used in the MFC MRD assay as well [31,34,37,45].
Leukemia-specific fusion transcripts derived from oncogenic chromosomal rearrangements can be used as MRD targets in pediatric ALL [7,46,47].
The sensitivity of RQ-PCR is comparable to or 1 log higher than that of MFC (10−4-10−5) [14,16,49]. The assay’s TAT is comparable to that of MFC [9-11,13,14,16, 30,47,49]. When compared to MFC or to the RQ-PCR assay in terms of detecting clonal
Nevertheless, the major drawback of RQ-PCR for fusion transcripts is that it is applicable only to patients with particular fusion transcripts. Other disadvantages include the instability of the RNA and the discordance between the number of cells and number of fusion transcript copies, which depends on the cell cycle [14,16, 46-49]. Genomic DNA of rearranged genes can be used instead of mRNA as a starting material owing to its greater stability, easier quantitation (only 1 PCR target is present per cell), and greater availability of information on oligoclonality or clonal evolution [15,16,48,50,51]. However, this is technically cumbersome owing to the high variability of some rearrangement breakpoints between patients thus is not widely used in clinical laboratories [10]. The risk of contamination during the PCR process cannot be avoided regardless of the choice of the starting material.
Standardized RQ-PCR protocols for some recurrent fusion transcripts are available [31,46,47,51,52].
The
The applicability of RQ-ASO-PCR is comparable to that of MFC (>90% of pediatric ALL patients) depending on the primer sets used [8,57,59]. A sensitivity of 10−4-10−5 is achieved through the RQ-PCR process [8,40,59,60]. It detects patient-specific clones and identifies clonal relationships between two lymphoid malignancies in a single patient and can therefore even differentiate between a relapse and second malignancy [47,54,57,58].
Nevertheless, this assay is not widely used in routine clinical practice since the TAT reaches 4-5 weeks given the ASO-PCR setup for individual patients, even though it has been considered the gold standard for MRD detection in pediatric ALL and is extensively used in MRD research studies along with MFC [37,40,59,60]. Moreover, false positive/negative clonal
In practice, the combined application of
Even though this assay is not affected by any treatment-caused immunophenotypic shifts, approximately 20% of minors with ALL lose their original
A standardized protocol for the detection of clonal
Emerging MRD assays aim to achieve higher sensitivities and specificities while overcoming the limitations of classical assays. NGF identifies aberrant phenotypes through a standardized high-throughput process; NGS recognizes clonal
Table 2 . Characteristics of emerging minimal residual disease assays used in pediatric acute lymphoblastic leukemia.
Assay techniques and targets | Applicability and sensitivity | Advantages | Disadvantages |
---|---|---|---|
Next-generation flow cytometry for identifying immunophenot-ypic deviations from normal counterparts | >90% | High sensitivity and wide applicability | Requires fresh samples (<24 h) |
10−5-10−6 | Short turnaround time | ||
Does not require prior information on patient-specific aberrant immunophenotype at diagnosis | Requires 4 million cells for a sensitivity of 10−6 | ||
Lower risk of false negatives caused by immunophenotypic shift during therapy compared to MFC | |||
Excludes apoptotic cells lacking leukemogenic potential | Requires high-level expertise for interpretation | ||
Analysis at cell population or single cell level | |||
Standardized for B-ALL | |||
Next-generation sequencing for | >90% | High sensitivity and wide applicability | Long turnaround time |
10−5-10−6 | Forgoes the need to design patient-specific/allele specific oligonucleotide primer sets | Risk of disproportional target amplification during multiplex PCR | |
Does not require knowledge of | High cost | ||
Can identify oligoclonality and clonal evolution | |||
Provides information on B/T-cell background repertoire | |||
Includes internal quality controls to monitor primer performance, technical variability, and quantitation | |||
Freely available web-based bioinformatics pipeline | |||
Standardized for B-ALL | |||
Potentially useful for other gene mutations | |||
Digital droplet PCR for fusion transcripts or | Applicability varies depending on targets: >90% for | High sensitivity and accuracy | Limited experience for pediatric ALL |
Does not require a standard curve | Not yet standardized | ||
10−5-10−6 | Potentially useful for other gene mutations |
B-ALL, B-cell acute lymphoblastic leukemia;
NGF is a novel high-throughput MRD assay using flow cytometry that was introduced by the EuroFlow Consortium and is based on a multidimensional approach that includes principal component and canonical analyses [31,32,65-68]. The NGF MRD assay differs from classical MFC MRD testing mainly in that its protocol is fully standardized and that it analyzes >4 million cells to achieve a sensitivity of 10−5-10−6 [32,41]. It retains the advantages of MFC including its wide applicability and short TAT [31,32].
Rather than using the LAIP of an individual patient as is performed in the classical MFC MRD assay, the NGF MRD assay recognizes ALL cells using standardized, preset LAIPs that are “different from normal” [32]. This is achieved through comparing the expression of antigenic patterns of ALL cells in numerous patients to that of normal counterpart populations (hematopoietic progenitors of similar lineage and maturational stage) [28,32,38,39]. It does not require information on the immunophenotype of leukemic cells from each patient at the initial diagnosis and is less affected by immunophenotypic shifts caused by treatment or relapse. The standardized 8-color B-ALL panel includes CD38, CD66c/CD123, CD73/CD304, and CD81 owing to their strong ability to discriminate between B-ALL cells and normal B-cell precursors/regenerating B-cells along with the 5 backbone markers CD19, CD45, CD34, CD10, and CD20 for appropriate B- cell precursor gating as well as differentiation between normal B-cell precursors and B-ALL cells [32,37,69-72]. A fully standardized laboratory protocol that includes equipment settings was established [32]. Upon validation, this assay successfully distinguished patients with B-ALL from normal individuals in 99% of the subjects [32].
Using the erythrocyte bulky lysis protocol, the NGF MRD assay counts >4 million cells [as is required for a minimum of 10 clustered events to consider a sample MRD-positive (lower limit of detection) as well as a minimum of 40 clustered events for the accurate quantitation of the MRD level (lower limit of quantitation)], thereby exceeding the sensitivity of RQ-PCR while being comparable to that of NGS [32,73]. This new protocol that involves the resuspension of large amounts of lysed samples was found to increase the number of evaluable leukocytes without significantly altering the cellular composition or increasing the percentage of doublets [32]. By doing so, a concordance of 93% was achieved between the NGF and RQ-PCR assays; most discordances were resolved through NGS for
A standardized NGF MRD assay for T-ALLs is expected to be developed based on the current knowledge of the immunophenotypes of neoplastic T-ALL cells and their normal T-cell counterparts [10,11,34,65]. The high sensitivity of this assay remains be demonstrated in large clinical trials.
NGS for
Like the aforementioned RQ-ASO-PCR method, NGS MRD assays can detect clonal
A standardized NGS MRD assay protocol to detect clonal
As different primers function under the same reaction conditions, NGS is subjected to some variability in the course of library preparation, sequencing, and bioinformatics steps [74]. The ‘central poly-target quality control’ (cPT-QC) is a standardized mixture of different lymphoid samples representing a full repertoire of
The target cells of this assay are lymphocytes, not total leukocytes. Primers for
The correlation between the NGS MRD assay for
ddPCR is a highly sensitive third-generation PCR technology that enables absolute quantification [79]. A DNA-containing sample is compartmentalized into oil droplets, and the fluorescence from each droplet is measured at the endpoint after multiple PCR reactions. The fraction of positive droplets (i.e., droplets containing the target DNA) is fitted to a Poisson algorithm, and the absolute copy number is then derived as copies per 1 mL without the need for a standard curve. The sample partitioning, high ratio of target DNA molecules to PCR reagents, and endpoint measurements contribute to the high sensitivity (10−5-10−6) and accuracy of this assay [79,80]. This method is also able to quantify samples classified as PNQ according to RQ-PCR [24,80].
Additionally, ddPCR MRD assays measuring the aforementioned leukemia-specific fusion transcripts or
It was in pediatric ALL that the prognostic significance of MRD quantification was demonstrated for the first time [6]. The prognostic significance of MRD has been extensively investigated in various pediatric ALL settings using classical MRD assays. MRD was strongly associated with early remission, CR, relapse after the first CR or allogeneic stem cell transplantation (allo-SCT), event-free survival, and overall survival [6-8,14-18]. Refining risk groups according to MRD level improved patients’ outcomes by providing better guidance in terms of treatment reduction or intensification, including whether to pursue allo-SCT [4-6,15,19]. Currently, MRD detection is a component of standard pediatric ALL clinical practice; thresholds of 1%, 0.1%, and 0.01% are used for risk stratification and MRD response assessment [20,31,37,59,85].
In terms of compatibility between the classical MRD assays, MFC and RQ-PCR for
The quantitative correlation between RQ-PCR for fusion transcript versus that for IG/TCR gene rearrangements has been investigated less frequently [15,16,77]. RQ-PCR for
Overall, most studies concluded that all the classical assays are efficient tools for monitoring MRD in pediatric ALL [37,86-88] and recommended the combined use of different MRD assays to prevent false positive/negative results and to better refine risk stratification [15,29,85, 86].
To date, the clinical significance of emerging MRD assays, as well as their inter-assay correlations and concordance with classical MRD assays, have been investigated in only a few pediatric ALL studies [12,21-25]. NGF versus NGS MRD assays for
MRD is comparable in the peripheral blood (PB) and bone marrow (BM) of patients with T-ALL; however, those with B-ALL have 1-3 logs lower MRD in PB than in BM. Therefore, BM is the preferred tissue for MRD testing in B-ALL and is also used in T-ALL testing [10,11, 89-91]. The possibility of using PB is still being investigated for both classical and emerging MRD assays given its convenience, but supporting evidence is insufficient to date [86,89-91].
Because MRD testing is quantitative, the representativeness of the sample is critical. BM aspirates submitted for MRD analysis invariably contain some PB [36,92]. The evaluation of PB contamination has been proposed using flow cytometry to detect CD117+ mast cells, B-cell precursors and nucleated red cells [73], the intensity of CD16 on maturing neutrophils [36], as well as plasma cells, CD34+ cells, and CD10+ granulocytes [92]. However, these techniques are not widely used, and practical methods should be developed and standardized to better evaluate the representativeness of samples for MRD assays.
The discovery and evaluation of new targets would contribute to improving MRD monitoring in pediatric ALL.
Early T-cell precursor ALL is a high-risk T-ALL characterized by absent (i.e., not yet occurred owing to its immaturity) or oligoclonal
Immunotherapeutic strategies have moved to the forefront of ALL treatments aimed at reducing MRD levels and/or decreasing conventional chemotherapy-related toxicities [97]. The three representative approaches are the (i) CD3/CD19 bispecific T-cell binder blinatumomab, (ii) CD22-directed antibody drug conjugate inotuzumab ozogamicin, and (iii) CD19-directed chimeric antigen receptor T-cell (the so-called CAR-T therapy) [97]. MRD evaluation can assess the efficacy of these novel treatments and serve as a surrogate marker for the endpoint [30]. However, flow cytometry-based MRD assays could be problematic particularly in B-ALL patients receiving immunotherapy since, in principle, these therapies target particular B-cell markers (CD19 and CD22) to identify B-ALL cells, which are also the markers used to gate B-cell precursors in flow cytometry [32]. Alternative strategies for detecting residual CD19-negative B-ALL cells using other B-cell markers such as CD22 or CD24 could be employed [98]. Overall, a standardized protocol for MRD evaluation in pediatric patients with B-ALL who are receiving immunotherapy remains to be developed.
MRD has been proven to be the strongest prognostic factor in pediatric ALL. MRD evaluation expedites personalized medicine in pediatric ALL by enabling accurate risk group assignment and risk-adapted treatment. For routine use, MRD assays should have clinically relevant sensitivity and specificity, reproducibility, applicability, appropriate TAT, and technical feasibility. MFC- or RQ-PCR-based classical MRD assays show sensitivities of 10−3 to 10−5 and have mainly been standardized by Euro-pean working groups. Novel techniques such as NGF, NGS, and ddPCR are promising alternatives given their improved sensitivities (10−5 to 10−6), specificity for ALL clones, applicability, and technical feasibility compared to classical MRD assays. These methods aim to overcome the drawbacks of classical assays and improved prognostic stratification. Prospective clinical trials ought to clarify the clinical benefit of the high sensitivities of these emerging assays in pediatric ALL. Standardization efforts and quality assurance programs are expected to be pursued through international collaborations to allow for the actual implementation of the emerging assays in clinical laboratories. In the meantime, efforts to unveil new targets and improve existing methods continue.
The authors have no conflict of interest to declare.