If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Address for correspondence: Dr Riana van der Linde, Department of Haematology, ICPMR, Westmead Hospital, NSW Health Pathology, Westmead, NSW, 2145, Australia.
Department of Laboratory Haematology, ICPMR, Westmead Hospital, NSW Health Pathology, Westmead, NSW, AustraliaSydney Medical School, Faculty of Medicine and Health, University of Sydney, Camperdown, NSW, Australia
Sydney Medical School, Faculty of Medicine and Health, University of Sydney, Camperdown, NSW, AustraliaFlow Cytometry Unit, ICPMR, NSW Health Pathology, Westmead Hospital, Westmead, NSW, AustraliaDepartment of Clinical Immunology, Westmead Hospital, Westmead, NSW, Australia
Sydney Medical School, Faculty of Medicine and Health, University of Sydney, Camperdown, NSW, AustraliaFlow Cytometry Unit, ICPMR, NSW Health Pathology, Westmead Hospital, Westmead, NSW, AustraliaDepartment of Clinical Immunology, Westmead Hospital, Westmead, NSW, Australia
Department of Laboratory Haematology, ICPMR, Westmead Hospital, NSW Health Pathology, Westmead, NSW, AustraliaSydney Medical School, Faculty of Medicine and Health, University of Sydney, Camperdown, NSW, AustraliaFlow Cytometry Unit, ICPMR, NSW Health Pathology, Westmead Hospital, Westmead, NSW, Australia
Measurable residual disease (MRD) monitoring in acute myeloid leukaemia (AML) is becoming increasingly important and is predominantly performed by multiparameter flow cytometry (MFC) or quantitative polymerase chain reactions (RT-qPCR). We investigated the use of multidimensional plots (MD-MFC) for AML MRD monitoring in an adult cohort. AML MRD was determined using a novel MD-MFC method for 115 MRD samples. Results were correlated with traditional two-dimensional MFC (2D-MFC) and molecular methods. Using the standard cut-off of 0.1% CD45+ cells, concordance was 99/115 (p=0.332). Eighty-four of 115 were concordant using a very low reporting limit of 0.01% (p=0.216). MRD <0.1% by either method was present in 40 of 115 samples. Fifteen of 40 were MD-MFC positive and 2D-MFC negative. Of these two of 15 had a molecular MRD marker and both were positive. Molecular MRD markers were available in 36 of 115 cases. Twenty-one of 36 (58%) were concordant with MD-MFC. Eight of 36 had detectable molecular MRD only and eight of 36 had positive MD-MFC only. There was no correlation between either the MFC method and the molecular results. In summary, there is good correlation between MD- and 2D-MFC-MRD and no correlation between the MFC and molecular methods.
The European Leukaemia Network (ELN) groups AML into three different risk categories based on recurrent genetic and molecular results: favourable, intermediate, and adverse risk.
Approximately 20% of adult and 35% of paediatric AML cases carry recurrent gene fusions, including the RUNX1-RUNX1T1, CBFB-MYH11, and KMT2A fusion transcripts. NPM1 mutations occur in around 30% of adult cases.
The outcome for many AML patients remains poor despite recent advances in management. Cure rates differ between risk groups and range from 60% in the favourable risk group to 10–15% in the intermediated risk group. The outcomes are poorer for the adverse risk groups and for patients older than 60 years.
Minimal measurable residual disease (MRD) is defined as the presence of a small population of malignant cells below the sensitivity of morphology which is 5%.
The ideal MRD test should be able to detect and quantify residual leukaemic cells with high sensitivity and specificity and also allow for implementation with inter-laboratory standardisation with reproducibility.
Current minimum time-points for molecular MRD are at diagnosis, after two cycles of induction/consolidation chemotherapy, at the end of treatment followed by every 3 months for 24 months.
AML MRD monitoring is becoming increasingly important, especially for the favourable and intermediate risk groups. It may also be important for adverse risk groups,
Both the Food and Drug Administration (FDA) and ELN have recently compiled guidelines and recommend reporting the limit of detection (LOD) for MRD negative results and selection of markers that detect leukaemia and not clonal haematopoiesis.
Both of these guidelines highlight the importance of clinical validation of MRD results. The current ELN guidelines recommend a cut-off of 0.1% of CD45+ cells, confirmed by published data, but they note that lower levels of MRD, between 0.01–0.1% CD45+ cells, may be clinically significant.
Measurable residual disease detected by multiparameter flow cytometry and sequencing improves prediction of relapse and survival in acute myeloid leukemia.
MRD monitoring is predominantly based on two methods: molecular and immunophenotyping by multiparameter flow cytometry (MFC). Molecular MRD involves identification of a specific molecular abnormality related to the leukaemic clone at diagnosis. The most common form of testing is real-time polymerase chain reactions (RT-qPCR). More recent guidelines include next generation sequencing (NGS) as a potential frontier method for tracking MRD.
The caveat is that only approximately 40–50% of AML cases have a molecular marker that can be used for MRD, and even then the molecular marker may be lost due to the inherent disease heterogeneity related to clonal evolution.
Additionally, most molecular MRD markers are associated with favourable risk AML, and therefore poor risk patients are often monitored using flow cytometry.
NGS is a promising emerging method, with the 2021 ELN guidelines recommending that all mutations can be included on NGS platforms, with the exclusion of known germline mutations, mutations associated with signalling pathways or age-related clonal haematopoiesis.
Flow cytometry is currently a cornerstone of MRD monitoring. Normal cells show consistent patterns of maturation and an expert understanding of normal maturation, including changes that occur after treatment, is essential.
Flow cytometry has several advantages over molecular MRD, including applicability to most AML cases and a shorter turn-around time. MFC MRD can be as sensitive as RT-qPCR, but in general practice the lower limit of quantitation is often between 0.01% and 0.1%.
The principle of MFC MRD monitoring is based on the evaluation of the leukaemia associated immunophenotype (LAIP) and determining the differences from normal (DfN) immature haematopoietic cells.
Limitations of flow cytometry can be grouped into technical and analytical aspects. Technical aspects include sample quality, which is influenced by the cell viability, timing of sample, haemodilution and hypoplasia.
Data analysis has become more complicated due to the increasing number of fluorochromes used in single tubes, leading to greater data complexity. Jafari et al. have shown that the use of multidimensional radar plots can improve investigation of large data sets, with the added benefit of seeing the different cell clusters in relationship to each other.
Multidimensional plots allow for incorporation of multiple parameters in the same graph which is then displayed in a two-dimensional diagram. The advantage is that multiple populations can be viewed simultaneously, while also examining their relationships to each other.
Several flow cytometric analysis software packages have this capability including the radar plot function in Kaluza software (Beckman Coulter Life Sciences, Australia). Radar plots allow for the addition of multiple parameters. Populations can then be identified based on multiple characteristics by adjusting the degree and length of the different parameters. These features can potentially be incorporated into MRD analysis to identify abnormal populations that differ from normal haematopoiesis.
We developed a multidimensional (MD)-MRD pipeline and applied it to >100 consecutive adult AML samples undergoing flow cytometry (FC)-MRD analysis. Here we describe a method using multiparameter plots and incorporation of both LAIP and DfN methods that is comparable to traditional two-dimensional flow cytometry. We also compare both flow cytometry methods to molecular tests where available.
Method
Study design
A total of 115 consecutive bone marrow aspirates sent for routine MRD analysis at NSW Health Pathology (NSWHP), Institute of Clinical Pathology and Medical Research (ICPMR), Westmead Hospital, were studied between August 2020 and March 2021. Our centre is a major quaternary and bone marrow transplant centre that serves Western Sydney and regional and rural hospitals in New South Wales. Demographic and diagnostic information collected included age, sex, diagnosis, bone marrow morphology reports, cytogenetic and molecular results. The data were stored in a soft-copy format on a secure password-protected computer server with limited access according to NSW Health policies. Normal maturation patterns were generated by collation marrows from previous samples taken between 2017 and 2020. Bone marrows were deemed normal if there was no history of benign or malignant bone marrow disorders, with normal morphology, immunophenotyping and cytogenetics. Thirty samples for the myeloid panel and 18 for the monocytic panel were pooled and then analysed to create normal maturation patterns. This study was approved by the Human Research Ethics Committee at Westmead Hospital (2020/ETH03275).
Sample preparation, instrument set-up and cell acquisition
Samples were collected in Roswell Park Memorial Institute media (RPMI) and processed within ∼48 h. Samples were prepared using commercial ammonium chloride 9% lysing solution (Kinetik, Australia). Events were acquired using the Beckman Coulter Gallios instrument (Beckman Coulter Life Sciences, Australia). At least 500,000 CD45+ events were acquired for each MRD assessment and the MRD was expressed as a percentage of CD45+ cells.
Immunophenotypic MRD analysis
MRD analysis was performed using two panels: a general myeloid panel and a monocytic panel. The monoclonal antibodies used in each tube are described in Table 1.
Table 1Monoclonal antibodies used for analysis of minimal residual disease in AML
Flow cytometric data were analysed using Kaluza software (Beckman Coulter Life Sciences, Australia). The leukaemia associated immunophenotype (LAIP) was established at diagnosis using sequential Boolean gating and bivariate plots. MRD samples were subsequently analysed based on the presence of the LAIP. The samples were also reviewed for any abnormal populations that were different from normal (DfN), based on the expression intensity of normal markers and/or expression of aberrant markers.
The MD-MRD method was created using the radar plot function in Kaluza software. To improve the identification of myeloblasts, two initial radar plots were created. The first was gated on all CD45 positive cells (and may include CD45 negative blasts if present) and separated the lymphoid and myeloid populations (Fig. 1B). The second radar plot, gated on the myeloid population, identified the myeloid blasts, and demonstrated maturation into either granulocytic or monocytic populations (Fig. 1C). A third radar plot, gated on lymphocytes, separated haematogones from mature lymphocytes.
Fig. 1Multidimensional measurable residual disease (MD-MRD) analysis template which consists of a series of consecutive radar plots. Radar plot A is the traditional CD45/SSC plot used as a reference of populations. Radar plot B separates the myeloid and lymphoid populations. Radar plot C is gated on the myeloid population and identifies the myeloblasts (orange population). It also shows myeloid/monocytic maturation patterns denoted by the black arrows. Radar plots D, E and F consist of all the parameters in the tube in identical configuration. Plot D represents myeloblasts in the diagnostic sample and contains two gates: the leukaemic population and normal link gate. The normal link gate is linked to the normal gate in radar plot E. Radar plot E represents expected normal maturation patterns with myeloblasts in red, granulocytic maturation in khaki and monocytic maturation in dark and light blue. Radar plot F represents the MRD sample with the RD gate linked to the leukaemic population gate on plot D. Populations: Granulocytes (bright green), monocytes (blue), leukaemic clone (orange), normal myeloblasts (red), promyelocytes (khaki), presumed MRD (purple), lymphocytes (dark green).
The myeloblast population was then graphed on a radar plot that consisted of all the parameters in the tube (Fig. 1D). A second identical plot was graphed next to this radar plot that contained the merged normal samples (30 myeloid samples and 18 monocytic samples) to depict where normal blasts, promyelocyte and promonocyte populations were situated (Fig. 1E). The expected area of normal maturation was represented on the plot containing the leukaemic population by a linked gate (‘normal link’ gate). The parameters were then adjusted to place the leukaemic populations in a non-overlapping space separate from the area where the normal cells were expected (Fig. 1D,E).
Once the settings were established, this formed a unique template for each patient that was applied to subsequent MRD analyses by adding the MRD sample to a composite containing the leukaemic sample, merged normal samples and the MRD sample. Any events located in the previously identified leukaemic space were interrogated. This included comparison to normal maturation patterns and the leukaemic population using 2D and overlay plots (Fig. 2). If no clear MRD population was identified, the whole myeloblast/immature myeloid population was compared to the expected maturation pattern to identify any significant phenotypic shifts and DfN populations.
Fig. 2Measurable residual disease (MRD) analysis using 2D and overlay plots. The first row represents the leukaemic population in orange (row 1). The normal granulocytic maturation pattern is denoted by the solid arrows and monocytic maturation by dashed arrows. The plots in row two is identical to row 1, but represents the measureable residual disease (MRD) sample with MRD in purple (row 2). Rows three and four are overlay plots comparing the MRD population (blue) to the leukaemic population (green) and normal maturation (red) for both granulocytic (row 3) and monocytic (row 4) populations. Row five represents the position of the leukaemic population (orange) an MRD population (purple) on CD45/SSC plots.
MRD was deemed positive if the population fulfilled the following criteria: a clustered population in the blast region consisting of >20 events not associated with normal developmental pathways. An example of MRD positive and MRD negative samples for the same patient is presented in Supplementary Fig. 1 (Appendix A).
Molecular MRD analysis
Quantitative RUNX1-RUNX1T1 [t(8; 21)] analysis was performed in the NSW Health ICPMR Westmead Hospital Diagnostic Molecular Laboratory using an intercalating dye, Syto9, to detect and quantitate the amount of transcript during reverse transcriptase PCR. The final results are presented as a ratio of AML/ETO transcript divided by ABL1 transcript. The quantitative molecular tests for both NPM1 and CBFB-MYH11 (inv16) were performed at separate external diagnostic laboratories. The quantitative NPM1 mutational analysis was performed at Royal Prince Alfred Hospital, NSW Health Pathology, Sydney, using the Ipsogen NPM1 Mut A MutaQuant Kit (Qiagen, Germany). The CBFB-MYH11 A [inv(16/t(16; 16)] mutational analysis was performed at Peter Mac Pathology, Melbourne, using the Ipsogen CBFB-MYH11 A kit (Qiagen).
Data analysis
The results from MD-MRD were compared to 2D-MRD analysis, as well as molecular MRD results, where available. Correlation was calculated using Pearson product-moment correlation coefficient and diagnostic capability by McNemar methods using Microsoft Excel (Microsoft Office 365; Microsoft, USA) and SPSS Statistics for Windows, version 27 (IBM, USA). A p value <0.05 was considered statistically significant.
Results
Patient demographics and treatment
There were 115 single-patient samples in this cohort. The median age of the study population was 53 years, less than the reported median age of 65.
Distribution according to ELN risk stratification was as follows: 38/115 (33%) favourable risk, 20/115 (17%) intermediate risk and 53/115 (46%) adverse risk. Three of 115 (2%) were referred from a distant site without adequate information for risk stratification and 1/115 (1%) was an AML transformed from primary myelofibrosis. Treatment modalities included: 53/115 (46%) receiving either intensive chemotherapy, hypomethylating agents or low dose cytarabine, 53/115 (46%) being post-transplant and 7/115 (6%) having surveillance marrows. Thirty-six of 115 (31%) cases had a measurable molecular MRD marker: RUNX1-RUNX1T1 [t(8; 21)] in 14/36 (39%), CBFB-MYH11 [inv(16)] in 3/36 (8%), and NPM1 mutations in 19/36 (53%) (Table 2).
Induction regimens: idarubicin 12 mg/m2 IV day 1–3 and cytarabine 100 mg/m2 IV day or fludarabine 30 mg/m2, cytarabine 2 g/m2 IV and idarubicin 10–12 mg/m2 IV. Consolidation regimens: cytarabine 1.5 g/m2 IV day or cytarabine 1,000 mg/m2 IV.
a Induction regimens: idarubicin 12 mg/m2 IV day 1–3 and cytarabine 100 mg/m2 IV day or fludarabine 30 mg/m2, cytarabine 2 g/m2 IV and idarubicin 10–12 mg/m2 IV. Consolidation regimens: cytarabine 1.5 g/m2 IV day or cytarabine 1,000 mg/m2 IV.
Comparison of multidimensional to two-dimensional flow cytometric MRD analysis methods
The 2D-MFC results ranged between 0% and 3.5% (median 0.01%) and MD-MFC between 0% and 3.5% (median 0.02%). There was a positive correlation between the two flow cytometric methods, r=0.68 (p<0.001; Fig. 3). Comparison of 2D-MFC results to the MD-MFC method using 0.01% of CD45+ cells as the lower limit showed 84/115 (72%) results were concordant. Twelve of 115 (10%) were positive using the 2D-MFC method but negative using the MD-MFC method, while 20/115 (17%) were only positive using MD-MFC (Table 3). There was no statistically significant difference between the methods (p=0.216). Adjusting the reporting limit to 0.1% of CD45+ cells resulted in 99/115 (88%) being concordant. Six of 115 (5%) were 2D-MFC MRD positive and MD-MFC MRD negative, and 11/115 (9%) were 2D-MFC MRD negative and MD-MFC MRD positive (Table 4). Again, there was no statistically significant difference between the methods at a 0.1% cut-off (p=0.332).
Fig. 3Correlation plot between multidimensional measurable residual disease (MD-MRD) and 2D-MRD. A strong correlation was observed between MD-MRD and 2D-MRD (Pearson correlation coefficient = 0.68; p<0.001).
MRD results between 0.01–0.1% by either test were present in 40/115. Sixteen of 40 (40%) results were concordant. Fifteen of 40 (38%) results were MD-MFC positive and 2D-MFC negative. Of these, 2/15 had a molecular marker [one NPM1 and t(8; 21)], and both molecular MRD results were positive at 0.025% and 0.009%, respectively. A third case had one residual cell (1/40) showing t(3/5) (q21; q31) on conventional karyotyping. In 9/40 (23%) cases MD-MFC was negative but 2D-MFC was positive. A measurable molecular marker was available for 2/9 of these cases. The first case was molecular MRD negative for t(8; 21) and the second molecular MRD positive for the NPM1 mutation, at 0.024%.
Comparison of MRD by MFC and molecular methods
Molecular markers were available in 36/115 cases. Fourteen of 36 (39%) had t(8; 21), 19/36 (53%) had NPM1, and 3/36 (8%) had inv16 (Table 1). The molecular MRD values ranged between 0.0% and 0.41% (median 0.0002%), the 2D-MRD cohort ranged between 0% and 0.38% (median 0.01%) and the MD-MRD cohort ranged between 0% and 1.03% (median 0.02%). Comparison of MD-MRD with molecular MRD demonstrated that 20/36 (56%) were concordant. Eight of 36 (22%) were positive by molecular method and negative by MFC (ranging between 0.0008 and 0.024%). These comprised four NPM1, three t(8; 21) and one inv(16) cases. Eight of 36 (22%) were positive by MD-MFC but negative with molecular methods, ranging between 0.03% and 0.39%. These cases consisted of three NPM1 and five t(8; 21) cases. The correlation between the MD-MRD and molecular MRD was poor (r=0.011; p=0.949; Fig. 4). Comparison of 2D-MRD and molecular MRD showed concordance in 15/36 (42%) with poor correlation between the two groups (r = –0.13; p=0.456). Eleven of 36 (28%) were positive by molecular MRD only (range 0.0008–0.41%), consisting of six NPM1, three t(8; 21) and one inv16 cases. Ten of 36 (31%) were positive by 2D-MRD only (range 0.03–0.48%) consisting of six NPM1 and four t(8; 21) cases. Four of 36 cases were positive for both MD-MRD and molecular MRD but negative for 2D-MRD. There were no cases positive for 2D-MRD and molecular MRD but not MD-MRD.
Fig. 4Correlation plots between molecular MRD, 2D-MRD or multidimensional (MD)-MRD. There is no correlation between measureable residual disease (MRD) as measured by (a) mulitparameter flow cytometry by multidimensional radar plots (MFC-MD; Pearson correlation coefficient = 0.01, p=0.949) or (b) multiparameter flow cytometry by traditional 2D analysis (MFC-2D; Pearson correlation coefficient = –0.13; p=0.456) and molecular measures.
Flow cytometry remains a major AML MRD testing modality, due to the lack of established molecular markers in ∼50–60% of cases. It is a cost-effective test with a short turnaround time. Limitations of flow cytometry are often related to the heterogeneity of leukaemic cells, treatment induced aberrancies in background normal cells and an unstable immunophenotype.
The ability of MD-MFC to view several populations and maturation pathways simultaneously makes it potentially valuable for MRD analysis, while reducing analysis time and subjective interpretations.
Here, for the first time, we validated the use of MD-MFC for AML MRD in the diagnostic setting.
By creating unique templates for each patient, every subsequent analysis is also personalised. Another advantage is that this method incorporates both the LAIP and DfN approach, while reducing the effect phenotypic shifts have on analysis, as its position in the plot is dependent on multiple markers, rather than two markers, diminishing the effect of a single change in one parameter. There is a risk that the personalised template may mask a change in phenotype, and therefore it is important to also compare the results with the expected maturation pattern.
Comparison of our MD-MFC method to the current method shows good correlation between the two groups and no statistically significant difference. Potential advantages of the radar plot method include a reduced analysis time and reduced subjectivity when reviewing the sample for aberrant populations. Phenotypic shifts, for which AML is well known, may lead to a change in the MRD population position. We observed that this effect is reduced, but not entirely eliminated, when using multidimensional plots. Additionally, phenotypic overlap with normal can still occur in MD-MRD during disease progression.
Limitations of this method include the initial time investment on creating the initial analysis template and additional training required, but this is largely mitigated by the improved turn-around times for subsequent MRD analysis. Lastly, regenerating marrow in the context of treatment may produce abnormal maturation patterns and populations in the absence of leukaemia.
Time point-dependent concordance and prognostic significance of flow cytometry and real time quantitative PCR for measurable/minimal residual disease detection in acute myeloid leukemia with t(8;21)(q22;q22.1).
compared MFC MRD to PCR for t(8; 21) and found a poor correlation of <50% at three timepoints: post-induction, post-consolidation 1 (PC1) and PC2. Most of the discordant results were PCR positive and MFC negative, suggesting the potential increased sensitivity of PCR. The poor correlation between molecular and MFC MRD has also been noted in other studies where the concordance ranged between 15–67% depending on treatment, with the post-induction timepoint usually displaying the poorest correlation.
Evaluation of minimal residual disease by real-time quantitative PCR of Wilms' tumor 1 expression in patients with acute myelogenous leukemia after allogeneic stem cell transplantation: correlation with flow cytometry and chimerism.
Comparison of multiparameter flow cytometry immunophenotypic analysis and quantitative RT-PCR for the detection of minimal residual disease of core binding factor acute myeloid leukemia.
The molecular markers and testing methods were variable among these studies, and most were limited to one marker only. Our study is consistent with no correlation found when comparing either MD-MRD or 2D-MRD with the molecular MRD incorporating all ELN recommended molecular MRD markers: NPM1, t(8; 21) and inv(16).
The potential increased sensitivity of molecular methods is also suggested in our results with very low-level molecular MRD of <0.01% present when MFC-MRD is negative.
It should be noted that low-level molecular MRD does not always translate to an increased risk of clinical relapse. The ELN defines molecular MRD with low copy numbers (MRD-LCN) as a transcript level <1–2% with a <1-log change between any two positive samples at the end of treatment.
Similar studies are not yet available for other molecular markers. Other work has shown that discrepant results, (with either MFC MRD or molecular MRD being positive) had prognostic value in that better outcomes were observed as compared to cases with concordant positive results.
The predictive value of minimal residual disease when facing the inconsistent results detected by real-time quantitative PCR and flow cytometry in NPM1-mutated acute myeloid leukemia.
Multicolor flow cytometry and multigene next-generation sequencing are complementary and highly predictive for relapse in acute myeloid leukemia after allogeneic transplantation.
We postulate that the patient group with positive MFC-MRD but negative molecular MRD, are likely more heterogeneous. Potential explanations include: the propensity of AML to display clonal evolution, bone marrow regeneration following treatment, and/or or pre-leukaemic myeloid populations that confound MRD result interpretation.
Time point-dependent concordance and prognostic significance of flow cytometry and real time quantitative PCR for measurable/minimal residual disease detection in acute myeloid leukemia with t(8;21)(q22;q22.1).
We found substantial agreement between the two flow cytometry analysis methods, demonstrating the radar plot method is fit for purpose. There were notable differences between MFC-MRD and the molecular methods which is in keeping with results from previous studies.
Time point-dependent concordance and prognostic significance of flow cytometry and real time quantitative PCR for measurable/minimal residual disease detection in acute myeloid leukemia with t(8;21)(q22;q22.1).
Comparison of multiparameter flow cytometry immunophenotypic analysis and quantitative RT-PCR for the detection of minimal residual disease of core binding factor acute myeloid leukemia.
We are currently conducting a prospective MRD study involving long term follow-up, which will help us to better understand these findings within the clinical context.
Acknowledgement
The authors would like to thank all the staff from the ICPMR flow cytometry laboratory for their assistance during this project.
Conflicts of interest and sources of funding
This work was supported by the ICPMR ROPP trust fund and a grant from the ICPMR Jerry Koutts Scholarship. The authors state that there are no conflicts of interest to disclose.
Appendix A. Supplementary data
The following is the Supplementary data to this article.
Measurable residual disease detected by multiparameter flow cytometry and sequencing improves prediction of relapse and survival in acute myeloid leukemia.
Time point-dependent concordance and prognostic significance of flow cytometry and real time quantitative PCR for measurable/minimal residual disease detection in acute myeloid leukemia with t(8;21)(q22;q22.1).
Evaluation of minimal residual disease by real-time quantitative PCR of Wilms' tumor 1 expression in patients with acute myelogenous leukemia after allogeneic stem cell transplantation: correlation with flow cytometry and chimerism.
Comparison of multiparameter flow cytometry immunophenotypic analysis and quantitative RT-PCR for the detection of minimal residual disease of core binding factor acute myeloid leukemia.
The predictive value of minimal residual disease when facing the inconsistent results detected by real-time quantitative PCR and flow cytometry in NPM1-mutated acute myeloid leukemia.
Multicolor flow cytometry and multigene next-generation sequencing are complementary and highly predictive for relapse in acute myeloid leukemia after allogeneic transplantation.
00685-7&id=gr1.jpg){kind=link}
00685-7&id=gr2.jpg){kind=link}
00685-7&id=gr3.jpg){kind=link}
00685-7&id=gr4.jpg){kind=link}