How I Approach MDS Diagnosis and Risk Assessment

Dr. Hetty Carraway
Hetty E. Carraway, MD, MBA, FACP
Director, Leukemia Program
Vice Chair, Strategy and Enterprise Development
Professor of Medicine
Hematologic Oncology and Blood Disorders
Taussig Cancer Institute, Cleveland Clinic
Cleveland, Ohio

Join Dr. Carraway as she outlines her strategy for diagnosing myelodysplastic syndromes (MDS) and assessing disease risk. In this newsletter, she reviews differential diagnoses for cytopenias/dysplasia, critical diagnostic tests to include in the MDS diagnostic pathway, MDS subtyping and classification, important prognostic factors and scoring systems, and the impact of somatic mutations on disease course and treatment selection.

How do you approach the initial evaluation of a patient with myelodysplastic syndromes (MDS) or suspected MDS?

Many of the patients I see will present with a pancytopenia of some description, although for the majority, anemia is the driving issue that brings them to our clinic. For these patients, typically we start with a comprehensive evaluation to look for non-MDS entities that we could treat to reverse the pancytopenia or anemia. There are a host of abnormalities such as nutritional deficiencies, inflammatory states such as lupus or rheumatoid arthritis, and iron, vitamin B12, folate, copper and/or zinc deficiencies or even medication interactions to consider in the evaluation of a new patient with cytopenia(s). Infections are also a consideration, so evaluation for HIV, hepatitis, and other potential infections are a part of the initial work-up.

Obtaining a bone marrow aspirate and biopsy is an essential part of the evaluation in order to assess for the presence of dysplasia, and we also include analysis by flow cytometry, as well as comprehensive cytogenetics and myeloid mutational profile testing by next-generation sequencing. Some clinics or institutions will send off fluorescence in situ hybridization (FISH) testing for alterations in specific chromosomes such as chromosome 5 and chromosome 7. With regard to the next-generation sequencing (NGS) testing, each institution may have their own myeloid panel and the number of genes tested can vary from 15 to over 100 genes, and this information can be helpful to guide diagnosis, classification, prognosis, and even treatment options. For example, the presence of a mutation of SF3B1 can help with both MDS classification and prognosis. As shown by Malcovati and colleagues, the presence of a SF3B1 mutation helps to identify a subset of patients with MDS with ring sideroblasts where the clinical course can be more indolent and is associated with a more favorable clinical outcome.1

There can be co-occurring issues in a patient with an MDS diagnosis. About 10%-15% of our patients with MDS will have a co-occurring monoclonal gammopathy of undetermined significance (MGUS). Then there are other entities such as large granular lymphocyte (LGL) and paroxysmal nocturnal hemoglobinuria (PNH) also that can come in concert with MDS, which can also impact therapeutic decisions for patients that end up having concurrent MDS, or who do not have MDS and just either have an LGL population or a PNH population.2 So altogether, we want to do a comprehensive workup so we have a clear idea of the contributors to the pancytopenia that we're faced with.

Could you summarize the 2016 World Health Organization (WHO) classification and explain how it helps guide evaluation of MDS patients?

As physicians dealing with hematologic malignancies, one of the most important things that we rely on is working together with our pathologists. They are really central to helping us understand exactly what's going on in the bone marrow. The bone marrow aspirate in a patient with MDS is really the keystone as to whether or not there is dysplasia and the additional information of comprehensive cytogenetics and mutational data fully annotate the associated profile.

There are some challenges to this. We have to ask whether the bone marrow aspirate and biopsy represents an adequate sample in order to evaluate for the presence of dysplasia. Sometimes the sample can be very small and/or limited by crush artifact and the conclusions of the slide are truly not conclusive as a result. Then there can be outside influences on the presentation of dysplasia; for example, if patients are on erythropoietin-stimulating agent (ESA)-based therapy or even agents such as eltrombopag or granulocyte colony stimulating factor (G-CSF), those medications can cause the cells to appear dysplastic. The hope is that the patient being evaluated is not on such therapy when they undergo a bone marrow biopsy procedure.

The pathologist is really the person that helps us to best classify the particular entity under the microscope, and the way in which our pathologists do that is following the 2016 WHO classification of MDS.3 There are a number of different MDS subtypes:

  • There are specific criteria by which we're able to follow in order to designate whether a patient has the diagnosis of MDS with a single-lineage dysplasia (MDS-SLD), ie, either a pure anemia alone or a thrombocytopenia alone. In any patient, there has to be the presence of dysplasia in ≥10% of one cell line. Those patients also have a low blast percentage (≤5%).
  • There are other entities such as MDS with ring sideroblasts (MDS-RS), which requires a certain threshold of erythroid precursors that have ring sideroblasts; if it’s 15% or above, then the patient meets the criteria of that particular WHO classification. If they don’t meet that cutoff, but they do meet a 5% cutoff and the SF3B1 mutation is present, then those patients also meet the criteria. This is an important subtype to identify because there are now directed therapies that we may be able to offer that patient (ie, luspatercept).
  • Another classification, MDS with multi-lineage dysplasia (MDS-MLD), highlighting that the dysplasia is seen in more than one cell line, so not just the red blood cell (RBC) dysplasia, but also the thrombocyte or neutrophil cell line. This classification is very descriptive in that it helps you to understand the degree of dysplasia and how many cell lines are affected.
  • We also worry about MDS patients with regard to progression to an acute myeloid leukemia. Some of that has to do with cytogenetic abnormalities, but we also look at the blast population. If the blast percentage is between 5% and 9%, then that is referred to as MDS with excess blasts-1 (MDS-EB-1), and when the percentage is 10% to 19%, then the MDS is classified as an MDS with excess blasts-2 (MDS-EB-2).
  • There are cytogenetic abnormalities that help us know whether a patient may do fairly well with lenalidomide-based therapies. There is an entity called MDS with isolated del(5q). Those patients often present with anemia, though their platelet count can sometimes be elevated. That cytogenetic abnormality can be really important in knowing how best to help that patient in the face of anemia and transfusion burden.
  • Patients that don't fit into any of the categories that I just mentioned are referred to as MDS, unclassifiable (MDS-U).

What prognostic factors need to be evaluated in patients with MDS, and why is this so important?

Once a diagnosis of MDS is confirmed, then there is a normal query with regard to clinical prognosis for that individual patient. MDS is a cancer diagnosis, and for any patient, that is often overwhelming and life-changing. Having a conversation about goals of therapy is paramount with regard to next steps. When we meet together to have that discussion, there are certain features of the MDS that help to weigh in on what is the prognosis for that specific case.

One factor is dysplasia in the bone marrow biopsy, as I discussed earlier. When the blast percentage is elevated, we worry about a higher likelihood of disease progression.4 We also worry about complex cytogenetics (or three or more standard cytogenetic abnormalities) which is associated with a worse prognosis.5 There can be specific molecular mutations that also make us concerned, such as the presence of a TP53 mutation, which tends to indicate that patients are going to be less likely to respond to therapies, even to curative therapies such as a bone marrow transplantation.6 In contrast to that, the presence of an SF3B1 mutation indicates that the prognosis is actually better, in that the progression of the MDS is often slower.1 This type of information, when you have it, can help patients make decisions that are aligned with their goals of care.

How are prognostic scoring systems evolving in MDS?

The International Prognostic Scoring System (IPSS) was established decades ago and is still a useful and important standard for assessing prognosis in MDS.7 It’s easy to follow and easy to remember, so for that reason, some people default to using it. However, I would say the revised IPSS (IPSS-R)4 is favored now because it gives you a little more specific information with regard to prognosis.

In general, the IPSS (and the IPSS-R) are based on three things: (1) the chromosome changes that are appreciated in the bone marrow, (2) the number of cell lines that are affected, and (3) the percentage of blasts. The IPSS-R is based on those same three points but includes several refinements in the specific enumeration of the blast categories, more details on exact cytogenetic abnormalities and associated risk groups, and also includes details on the depth of cytopenias (with annotation of specific parameters of the cell counts), and inclusion of differentiating features.4

As a result, we have learned to rely on the IPSS-R and find that it's more reflective of an accurate prognosis. Nowadays, because of the internet, we can very quickly find an IPSS or IPSS-R calculator, and you can even download an app to your phone to have access to these prognostic scoring systems very easily, irrespective of wherever you are.

There are other models worth mentioning. We know that patients with lower-risk MDS who are highly transfusion dependent over time may have a worse outcome, so there are models that are more reflective of the lower-risk MDS population as compared to the higher-risk MDS population.

There are also several prognostic systems for chronic myelomonocytic leukemia (CMML), a syndrome with overlapping features of MDS and myeloproliferative neoplasms.8 If a community provider has a patient with CMML, it may be worthwhile re-evaluating the patient using a CMML-type model, though this entity is very rare and it may be beneficial to reach out to a specialist, should that be the case.

I predict that our prognostic scoring systems will still even get better with time. We're always trying to refine our ability to quantify risk of progression to leukemia, and determining the ideal time for an intervention such as medical therapy or transplantation that are going to potentially change and impact a patient’s quality of life. And that may mean that it's changing quality of life for the worse—it’s not always the case, but one concern of patients is that an intervention would require more visits to the doctor’s office, and consequently decrease their independence from the medical system and its demands.

The way in which we're trying to move the needle in prognostic scoring systems is by adding information from molecular mutations. This is an active area of research. As I discussed to earlier, the presence of an SF3B1 mutation helps us know that the patient may have a slower progression of disease, while in contrast, the presence of a TP53 mutation would be a concerning finding and worrisome for a less beneficial clinical outcome comparatively given a concern for quicker progression and/or relapse of MDS despite therapy. The mutation profiles are entities that we're trying to incorporate into prognostic scoring systems but are not formally incorporated as of yet. It is likely that, in the near term, we will have a model that accomplishes that.

What somatic gene mutations are most useful in the context of MDS?

More than 95% of our patients with MDS do have the presence of a mutation in the peripheral blood or in the bone marrow sample. These mutations can be present even before a patient has dysplasia present in their marrow, and even without cytopenias. The presence of these mutations help us in various ways: not only can they help us make the diagnosis of MDS (classification), as I talked about earlier, but they can also help with prognosis.

We know from work by Bejar and colleagues that mutations not only in TP53, but also in EZH2, ETV6, RUNX1, and ASXL1 independently predict for a poor clinical survival in patients diagnosed with MDS.9 If one of those genes is mutated, that will up-score a patient from a lower-risk to an intermediate-risk, or from an intermediate-risk to a higher-risk IPSS score for MDS.

The presence of mutations can also be helpful in terms of being predictive. We know that if a patient has a TP53 mutated MDS prior to a hematopoietic stem cell transplant (HSCT), we should be more concerned about a risk of relapse in the post-transplant setting (compared to a non-TP53 mutated MDS). There is a lot of effort focused on whether we can use targeted therapies directly to MDS cells that harbor specific mutations. For example, there are agents targeting the TP53 mutation as well as others. The ultimate goal is to optimize the treatment of MDS (obtain a deep remission) such that when/if a patient is able to move to curative therapy with a HSCT, perhaps we will see better long-term patient outcomes for all.

These molecular mutations give us information that we can use in clinical decision making. For example, a patient with an SF3B1-mutated MDS may be less likely to have progression of the disease, so it allows for monitoring and watchful waiting in some situations and can spare them from intrusive therapy if they don't otherwise need it. In contrast, in the setting of high-risk mutation, there is a concern regarding the risk of disease progression and associated poor prognosis, even despite a low blast percentage; and now may be the best time to move forward with a curative therapy approach.

We also have noted specific patterns in relation to these genomic mutations. For example, the presence of an SF3B1 mutation is often associated with the presence of ring sideroblasts in the bone marrow aspirate. Another entity that we didn't talk about yet is refractory anemia with ring sideroblasts (RARS) and RARS with thrombocytosis; these patients tend to have a pattern of mutations that is inclusive of an SF3B1, in addition to a JAK2 mutation.10 By contrast, SRSF2 mutation is most commonly observed in CMML and in the right context can be considered a diagnostic marker of the disease.11

How can prognostic data be used to help counsel patients or caregivers and set expectations for outcomes?

Prognostic data is important to share with patients. Patients often ask what will happen to them if they choose to do nothing other than continue to come to clinic and follow their symptoms to direct therapy and/or medical intervention. I think some of the reason for that question is because people are fearful about the word “chemotherapy” and what it means. They may fear the impact not only on themselves, but also the impact on their loved ones. That's what people really care about: “How much are you going to impact my quality of life, and how much are you going to make me depend on my loved ones? Are my kids going to need to leave work to help me? Am I going to lose my independence?” Many patients are fiercely independent and they want to stay that way for as long as possible.

Many of the MDS treatments that are available are well tolerated, and we've given them to patients that are in their 80s and 90s, but often those are also the people who are independent and have a pretty good performance status. For others who are not doing as well, if their disease progresses very quickly and their performance status is rapidly declining, that's a different story. The disease-modifying impact of chemotherapy can be incredibly meaningful to such patients that have had a decline in their performance status that is directly related to their MDS and disease burden. So it’s important to get a good sense of how the patient is doing, and some of that information comes from the patient and also from the person that brought them into the clinic visit.

Ultimately, the prognostic scoring system has its limitations. We wish that those numbers mean that you could predict exactly what would happen to the individual patient in front of you, and that's just not always the case. There are times where patients and caregivers can get very hung-up on the numbers and not realize that there's room for some patients that are outliers to that collective data. I think those are the more challenging issues that we need to navigate as we talk about the data that we have, and what our thoughts are for that individual person.


  1. Malcovati L, Karimi M, Papaemmanuil E, et al. SF3B1 mutation identifies a distinct subset of myelodysplastic syndrome with ring sideroblasts. Blood. 2015;126(2):233-241. doi:10.1182/blood-2015-03-633537
  2. Langerak AW, Assmann JLJC. Large granular lymphocyte cells and immune dysregulation diseases – the chicken or the egg? Haematologica. 2018;103(2):193-194. doi:10.3324/haematol.2017.186338
  3. Arber DA, Orazi A, Hasserjian R, et al. The 2016 revision to the World Health Organization classification of myeloid neoplasms and acute leukemia. Blood. 2016;127(20):2391-2405. doi:10.1182/blood-2016-03-643544
  4. Greenberg PL, Tuechler H, Schanz J, et al. Revised International Prognostic Scoring System for Myelodysplastic Syndromes. Blood. 2012;120(12):2454-2465. doi:10.1182/blood-2012-03-420489
  5. NCCN Clinical Practice Guidelines in Oncology (NCCN Guidelines). Myelodysplastic Syndromes. Version 3.2021 - January 15, 2021. Accessed May 24, 2021.
  6. Lindsley RC, Saber W, Mar BG, et al. Prognostic Mutations in Myelodysplastic Syndrome after Stem-Cell Transplantation. N Engl J Med. 2017;376(6):536-547. doi:10.1056/NEJMoa1611604
  7. Greenberg P, Cox C, LeBeau MM, et al. International scoring system for evaluating prognosis in myelodysplastic syndromes. Blood. 1997;89(6):2079-2088.
  8. Patnaik MM, Lasho T. Myelodysplastic syndrome/myeloproliferative neoplasm overlap syndromes: A focused review. Hematol Am Soc Hematol Educ Program. 2020;2020(1):460-464. doi:10.1182/hematology.2020000163
  9. Bejar R, Stevenson K, Abdel-Wahab O, et al. Clinical effect of point mutations in myelodysplastic syndromes. N Engl J Med. 2011;364(26):2496-2506. doi:10.1056/NEJMoa1013343
  10. Patnaik MM, Tefferi A. Refractory Anemia with Ring Sideroblasts (RARS) and RARS with Thrombocytosis (RARS-T) – “2019 Update on Diagnosis, Risk-stratification, and Management.” Am J Hematol. 2019;94(4):475-488. doi:10.1002/ajh.25397
  11. Meggendorfer M, Roller A, Haferlach T, et al. SRSF2 mutations in 275 cases with chronic myelomonocytic leukemia (CMML). Blood. 2012;120(15):3080-3088. doi:10.1182/blood-2012-01-404863

Provided by MediCom Worldwide, Inc.
This activity is supported by educational grants from Bristol-Myers Squibb and Taiho Oncology, Inc.