Minimal residual disease assessment in B-cell precursor acute lymphoblastic leukemia
Assessment of minimal residual disease during B-cell precursor acute lymphoblastic leukemia provides a good indication of therapy response and prognosis. This investigation can be done in several ways, including flow cytometry. Techniques in this field have developed in recent years and CLI chatted to Dr van der Velden (Department of Immunology, Erasmus University Medical Center, Rotterdam, the Netherlands) to discover more about their findings.
What is B-cell precursor acute lymphoblastic leukemia?
B-cell precursor acute lymphoblastic leukemia (BCP-ALL) is a form of acute leukemia involving the B cells. B cells normally mature in the bone marrow, but in this disease malignant transformation results in an overproduction of immature B cells and normal hematopoiesis is blocked, resulting in complains such as anemia, bleeding, and all kind of infections. It’s the most common form of cancer in childhood, but it also occurs in in adults. It’s a relatively rare
disease; in the Netherlands, we have about 200 new cases per year from a total population of approximately 20 million. It’s an acute leukemia because the complaints are really quite severe in the short term, and so patients need intensive treatment at an early stage.
How is it normally treated and what are the survival rates?
The classic therapy consists of chemotherapy, different types of drugs supplemented in high-risk patients with an allogenic stem cell transplantation. The outcome for children is relatively good – nowadays more than 90% of children will have an overall survival over 5 years, which has really improved over the last decades. For adults, it’s still less good, and depends on the age of the patient. The older the patient, the worse the outcome will be. The survival varies between approximately 20 and 50%.
Why is it important to assess minimal/measurable residual disease?
Minimal or measurable residual disease (MRD) is important because it detects the number of remaining leukemic cells present – mainly in the bone marrow, which is the preferred site for analysis although blood samples could also be analysed. What has been shown, especially initially in childhood ALL, is that the kinetics of the MRD is of highly prognostic significance for the outcome of the patients. It is important to say that assessment of MRD is not mainly about detecting upcoming relapses but it’s really identifying the initial effect of therapy on the leukemic cells and the disappearance hopefully of the leukemic cells. It has been clearly shown that patients that have no detectable residual disease already at 4 weeks after start of treatment really have an excellent outcome. However, patients that were relatively strongly positive at that early time point but also after 3 months of therapy – although being judged negative by the morphological techniques – still had a really very poor outcome and most of those patients would have a relapse. Hence, assessment of MRD after 3 months of a total therapy time of at least 2 years in children already provides a very good prediction of their outcome. This means that on the one hand, the patients that have a high risk of disease recurrence clearly need more-intensive types of treatment. On the other hand, the patients that have a very low risk of disease recurrence now become candidates for therapy reduction. We know that this is possible because in the early days of treating ALL (let’s say in the 1970s or 80s) there was already a group of pediatric patients that survived – even though the therapy at that time was relatively mild. In the meantime, however, the therapy has been intensified in order to reduce the risk of relapse, but this also means that some patients have been overtreated. So assessment of MRD is also valuable for identifying patients whose treatment can be reduced without having a major impact on their survival rates. That is how assessment of MRD is mainly used in pediatric patients. In adult patients, it’s more or less similar but in this group the more continuous monitoring of MRD is important as well.
How is MRD usually assessed?
Traditionally, there are three methods: one is flow-cytometric immunophenotyping; the second is PCR analysis of rearranged immunoglobin and T-cell receptor genes; and the third is RT-PCR analysis of fusion gene transcripts. The latter has been of limited use in the last couple of decades because it is only applicable in small subsets of patients and using RNA has some limitations over the other two methods. Therefore, flow cytometry and PCR of immunoglobin receptor gene rearrangements are the more preferred methodologies. Classically, in Europe, most initial studies have focused on the molecular (PCR-based) analysis of MRD. There’s a European network called EuroMRD (https://euromrd.org/) that has already been organizing education and quality assessment rounds for over 25 years. Importantly, this means that he the different labs in Europe have aimed to standardize the methods used as well as the means of data interpretation. In Europe, that’s more or less the gold standard for MRD assessment. However, flow-cytometric immunophenotyping is also possible. This technique is applicable in the vast majority of patients but its use has been limited by its sensitivity compared to PCR until a couple of years ago. At that time flow cytometrists were relatively limited in the number of colours – and hence parameters – that they could detect but that has been increased over recent decades from two to three to four to six to 8 to 12, now to possibly even 25, which is an amazing increase. So flow-cytometric immunophenotyping is used more and more in MRD protocols. I should say in the United States of America, it’s a little bit different because, at least initially, people there mainly focused on flow cytometry, but now with the introduction of next-generation sequencing, people in the USA are making much more use of the molecular sequencing methodologies. In fact, those two methods are the main methods that can be used, and, although in principle they’re not identical, they can provide generally similar results. For assessment of MRD by flow cytometry, we use an 8-colour panel that was in fact developed within another European network, EuroFlow (https://euroflow.org/), and the markers that are in that panel were carefully selected based on all the different experiences and experiments that have been performed. This panel includes CD19 (which is an important marker frequently used for identification of B cells) and CD10, CD20, CD34, CD45, CD81 and CD38, which are also expressed on normal B-cell precursors during different stages of development. We also include some markers that are normally not present on B-cell precursors but that can be abnormally expressed on the leukemic cells of a meaningful number of patients and these, for example, are CD66c, CD123, CD73 CD304 . We need all of the eight markers to have an optimal separation of the leukemic cells and the normal cells that can also be present.
Are there any limitations to the current flow cytometry SOP?
I think the main limitation is related to the fact that more and more patients now are not only treated with chemotherapy but also with immunotherapies. Some of those immunotherapies (which can be antibodies or CarT cells) target CD19 and that has consequences for the flow-cytometric assay for multiple reasons. One technical reason may be that use of, for example, a therapeutic anti-CD19 antibody may interfere with the binding of the diagnostic anti-CD19 antibody to the leukemic cells, so that they seem to be CD19-negative. However, the therapeutic anti-CD19 antibody can also drive therapy-resistance or evasion strategies, where the leukemic cells can become truly CD19-
negative so that the therapy will no longer be capable of targeting those cells. This loss of CD19 can happen in two ways. Firstly, because of mutations in the CD19 gene or possibly mutations in other genes that are important for CD19 expression. Secondly, the leukemic cells – originally B-cell precursor cells – can undergo a so-called lineage switch and so transform from a B cell to being much more like a myeloid cell, which also lacks CD19 expression. In all of those situations you can no longer use CD19 as the B-cell gating or identification marker. Hence, the reliance of the 8-colour panel on CD19 is a possible limitation. I say ‘possible limitation’, because when we in the EuroFlow consortium evaluated the performance of the 8-colour panel without using CD19 but using the other markers present, we found that in most cases it is still possible to identify the leukemic cells. However, that situation requires a focus on the inclusion of CD10 and CD34, but there are also leukemias that do not have CD34 and also do not have CD10 and then the efficacy of the 8-colour panel starts to fade.
Additionally, manual gating is a challenge and a limitation. When we analyse the flow cytometry data, we really need to decide per population – or per cell maybe even – what type of cell is it. This means that the people who do that analysis really need a lot of expertise and experience to look at the different cells and to judge what type of cells those are – whether they are normal B cells, whether they are abnormal B cells, or whether they are other types of cells, such as T-cells or myeloid cells, and so on. So this can be quite subjective and that also means that it can result in discrepancies between different labs or between different people. Therefore, within the EuroFlow network we also tried to develop a tool that could at least assist in the analysis. It’s not a fully automated analysis, but we made a database of normal bone marrow samples that were also stained with the 8-colour or 12-colour marker panel and we analysed them manually so that we could assign all of the cells to one of the normal populations. In that way we were able to build a database of all the different, normal populations that can be identified from the bone marrow using those tubes. And what we then can do is that if we now have a patient with a BCP-ALL, either at diagnosis or especially during follow-up, we can we can ask the computer to make clusters of cells so that the computer identifies in a multi-dimensional space which cells form a nice cluster. Each cluster, then, is compared with the database, and if the cluster is highly comparable to a population present in the database, the computer can identify this cluster as that population of cells. In that way, all the normal populations can be identified and what is left is either a leukemic cell or abnormal cell, or some kind of technical flaw (something that is not optimally stained or not optimally measured or whatever). So the computer can already ‘get rid’ of most of the cells that are normal cells and then the remaining cells are checked and identified by an expert. Again, even at this stage the computer can give some suggestions about assigning the cells as normal or abnormal. This helps to have a more standardized analysis, but it’s still not a fully automated analysis – someone with expertise, skill and experience is still needed for the final checks.
Figure 1. Identification of different cell types using the 12-marker panel
The leukemic cells are shown in red, whereas all other leukocytes are coloured grey. The level of MRD was 0.005%. The BCP-ALL cells express CD19, CD22 (weak), CD24 and HLA-DR and are positive for CD10 and CD34, whereas CD38 and CD20 are negative.
Are there any developments that would improve BCP-ALL MRD assessment?
Within the EuroFlow network, we were relying on CD19 as a B-cell identification marker with the 8-colour tubes (Table 1). Therefore, with the 12-colour panel, we wanted to add some additional markers that would allow us to identify the B cells – both the normal cells as well as the leukemic cells – in an alternative way (Table 1, Fig. 1). So we looked again at a large number of diagnostic samples to see which markers are present on virtually all the leukemic cells and the normal B cells. These were particularly CD22 and CD24. I should say that both of these markers are expressed on B cells but they can also be expressed on some other cells, so they are not as clear cut as CD19 but those markers were definitely assumed to be of help for identification of B cells. We also included two markers CD3 and CD7 that are not expressed on BCP-ALL cells or on normal B cells but are expressed on other lymphocytes, especially T-cells and NK cells, so those markers could be used as exclusion markers. We also added a marker called HLA-DR, which is a marker that is virtually always expressed on B cells as well as on other antigen presenting cells. So it’s not a specific marker for B cells alone, but at least it could further support the identification of B cells. We tested that 12-colour combination in different patients – with the limitations that the number of patients that have CD19-negative BCP-ALL MRD cells is small and that these samples generally have very low number of cells, which makes comparisons very difficult. We also included other types of samples and ultimately we found that the 12-colour panel did not perform ‘better’ than the 8-colour panel but that the people doing the analysis felt an increased confidence in their identification of cell type when they used the 12-colour tube.
Table 1. The different cell markers identified by the 8- and 12-colour tubes
The colours refer to the lasers required to excite the fluorochromes (violet, blue and red).
Another benefit of the 12-colour tube is that it includes antibodies for CD22 which is also marker that can be targeted by certain therapies, for example the CD22 immune conjugate Inotuzumab. So that’s also important information for the for the clinician. Also, the results from the 12-colour tube could also be analysed using the database that we developed for the 8-colour panel. The 12-colour panel results therefore in an improvement in the reliability and hopefully then reproducibility of the data.
However, as I said, the data interpretation is not a simple process and a lot of training is needed. I hope that in the near future AI or other kinds of tools will provide further software support for this type of analysis. I will not say that it should be completely fully automated but it would definitely be helpful if people have more support in this highly complex decision-making process.
The advantage for the patient at this moment is mainly that the patient will get a more reliable result for the level of MRD. In general, the advantage of the flowcytometric immunophenotyping is that you get information about the number of residual leukemic cells but you get also information about the proteins that are expressed on the cells, which is relevant for the targeted therapies. That kind of information is really unique to flow cytometry compared to the other molecular assays. That information may in the end also help clinicians to identify which patients may profit from other relatively expensive types of therapies with a high chance of success.
Do you have any final comments?
The system that we have now uses 12 colours, which is already quite a lot – certainly for clinical diagnostic labs. However, in the last couple of years we have had the introduction of spectral flow cytometers which can acquire and measure from 25 to over 50 different types of colours. At this time point these instruments are mainly being used for research but they open up new possibilities to further extend the choice of markers analysed. Then the question becomes about what do you really need? Because of course you can use 25 or 50 markers, but in the end if it’s the same result it’s just more expensive. So I think the next challenge will be to determine what is really needed and what can then be included. For example, it also opens up the possibilities not only to look at the leukemic cells by all the different markers but maybe – in the same tube of the same limited amount of material –you may also investigate if CarT cells are still present and, if so, what type of CarT cells are there. Or, if you think about giving a patient targeted therapy that needs T cells you may also identify the different T cells that are present in the patient and check whether those T cells are still functional or whether they are already in a senescent state. So in that way spectral flow cytometry creates a lot of possibilities for combining different questions into a single tube. So, lots of exciting options!
Bibliography
1. Theunissen P, Mejstrikova E, Sedek L et al; EuroFlow Consortium. Standardized flow cytometry for highly sensitive MRD measurements in B-cell acute lymphoblastic leukemia. Blood 2017;129(3):347–357 (https://doi.org/10.1182/blood-2016-07-726307).
2. Verbeek MWC, Rodríguez BS, Sedek L et al. Minimal residual disease assessment in B-cell precursor acute lymphoblastic leukemia by semi-automated identification of normal hematopoietic cells: A EuroFlow study. Cytometry B Clin Cytom 2024;106(4):252–263 (https://doi.org/10.1002/cyto.b.22143).
3. Verbeek MWC, van der Velden VHJ. The evolving landscape of flowcytometric minimal residual disease monitoring in B-cell precursor acute lymphoblastic leukemia. Int J Mol Sci 2024;25(9):4881 (https://doi.org/10.3390/ijms25094881).
4. Lebecque B, Besombes J, Dannus LT et al. Faster clinical decisions in B-cell acute lymphoblastic leukaemia: A single flow cytometric 12-colour tube improves diagnosis and minimal residual disease follow-up. Br J Haematol 2024;204(5):1872–1881 (https://doi.org/10.1111/bjh.19390).
5. Buldini B, Varotto E, Maurer-Granofszky M et al. CD371-positive pediatric B-cell acute lymphoblastic leukemia: propensity to lineage switch and slow early response to treatment. Blood 2024;143(17):1738–1751
(https://doi.org/10.1182/blood.2023021952).
6. Mikhailova E, Itov A, Zerkalenkova E et al. B-lineage antigens that are useful to substitute CD19 for minimal residual disease monitoring in B cell precursor acute lymphoblastic leukemia after CD19 targeting. Cytometry B Clin Cytom 2022;102(5):353–359 (https://doi.org/10.1002/cyto.b.22088).
7. Mikhailova E, Semchenkova A, Illarionova O et al. Relative expansion of CD19-negative very-early normal B-cell precursors in children with acute lymphoblastic leukaemia after CD19 targeting by blinatumomab and CAR-T cell therapy: implications for flow cytometric detection of minimal residual disease. Br J Haematol 2021;193(3):602–612 (https://doi.org/10.1111/bjh.17382).
8. Dourthe ME, Rabian F, Yakouben K et al. Determinants of CD19-positive vs CD19-negative relapse after tisagenlecleucel for B-cell acute lymphoblastic leukemia. Leukemia 2021;35(12):3383–3393 (https://doi.org/10.1038/s41375-021-01281-7).
9. Aldoss I, Li S, Zhang J et al. TP53 mutations are associated with CD19- relapse and inferior outcomes after blinatumomab in adults with ALL. Blood Adv 2025;9(9):2159–2172 (https://doi.org/10.1182/bloodadvances.2024014986).
The interviewee
Vincent van der Velden, PhD; Laboratory Specialist Medical Immunology; Group leader Leukemia and Lymphoma Diagnostics
Laboratory for Medical Immunology, Department of Immunology, Erasmus University Medical Center, Rotterdam, the Netherlands
For further information see:
Verbeek MWC, Reiterová M, Laqua A, Rodríguez BS, Sedek L, Buracchi C, Buysse M, Oliveira E, Engelmann R, Desterro J, De Jong AX, Boettcher S, Jugooa R, Barrena S, Kohlscheen S, Nierkens S, Rodriques JG, Hofmans M, Gaipa G, Sobral de Costa E, Mejstrikova E, Szczepanski T, Brüggemann M, van Dongen JJM, Orfao A, van der Velden VHJ. Minimal residual disease assessment following CD19-targeted therapy in B-cell precursor acute lymphoblastic leukemia using standardized 12-color flow cytometry: A EuroFlow study. Hemasphere 2025;9(4):e70125 (https://doi.org/10.1002/hem3.70125).
