VJHemOnc are delighted to bring to you a lecture from one of the pioneers in myelodysplastic syndrome research – Prof. John M. Bennett, MD, from the University of Rochester Medical Center, Rochester, NY.

In this presentation, Dr Bennett takes us through a historical journey from the origins of MDS, current classification, advances in molecular mutations, the prognosis of patients with MDS and what the future holds for patients.

Core MDS characteristics


“So, for definitions, these are malignant neoplastic disorders characterized by ineffective hematopoiesis that involves either the erythroid, granulocytic, monocytic or megakaryocytic cell lines with a variable percentage of leukemic blasts that can range from as low as 1-2% to as high as 19%. The median age is 70. About 30% of patients will progress within a year or longer, upwards of three or four years, to a form of acute myeloid leukemia.”

– John Bennett



Classification and genomics


“So, have changes in the WHO broadened the definition a bit? If you have monosomy 7, we know that’s not good. We strongly recommend the TP53 mutation be included and if the blast percentage gets beyond 5% and/or dysplasia, these patients are not going to behave as well as if you have a deletion 5q. And that’s very important because of the introduction of lenalidomide, the immune modulator which certainly provides an excellent response rate”.

– John Bennett

Prognostic summary


“We will all pay taxes forever. And about every five years, we all have to endure classifications or evolutions. “

– John Bennett

Full Transcript

John Bennett
Hello, everyone. My name is John Bennett, and I’m a professor of Pathology, Laboratory Medicine, and Medicine at the University of Rochester, particularly assigned to the James P. Wilmot Cancer Institute, and currently active in the Department of Pathology, Hematopathology Division where I do consultations and work primarily in the area of leukemia diagnosis.

Today, I want to talk to you about a very interesting myeloid malignancy referred to as myelodysplastic syndromes, go into the diagnostic aspects, clinical staging, and areas of new advances such as any molecular mutations that have recently been discovered.

So, for definitions, these are malignant neoplastic disorders characterized by ineffective hematopoiesis that involves either the erythroid, granulocytic, monocytic or megakaryocytic cell lines with a variable percentage of leukemic blasts that can range from as low as 1-2% to as high as 19%. The median age is 70. About 30% of patients will progress within a year or longer, upwards of three or four years, to a form of acute myeloid leukemia.

In the United States, the incidence is approximately 15,000 cases annually and the US prevalence is anywhere from 35,000 to 55,000 cases. The vast majority of these, obviously, are in adults. MDS in children is a very, very uncommon disease. The majority present with moderate to severe anemia.

Now, this is a diagram that outlines where we’ve been and where we’re headed. The diagnosis originally was developed by the French-American-British Working Group and published in 1976, modified in 1982, and was primarily based on morphology and blast percentage.

In 2001, a first of three editions of the WHO Blue Book was produced, which took the original FAB classification, modified it a bit by focusing a little more on dysplasia, adding cytogenetics, as well as retaining a blast percentage. And this went through a series of changes, some of which we will outline in a few minutes.

Prognosis by the International Prognostic Scoring System incorporated a lot of the FAB and WHO criteria. This was then modified several years ago, referred to as an International Prognostic Scoring System Revised, which we’ll talk about towards the end of this presentation.

Molecular mutations were initially introduced in 1997 and then additional minor changes, such as elevated LDH, additional chromosomes modified here included in the IPSS-R, performance status, ferritin, fibrosis, beta-2M. But the basic hallmarks of the diagnosis and prognosis remain carefully, accurately, calculating the percentage of blasts, looking at chromosomes and looking at the number of cytopenias.

For all of these conditions, myeloid malignancies, lymphocytic leukemias, chronic myeloid malignancies, et cetera, we are dependent upon looking at a bone marrow aspirate, as well as a peripheral blood smear. We like to have good spicules that’s cellular enough to assess at least 500 cells. Sometimes, we can’t do this because of fibrosis or because of hypocellularity.

We also use flow cytometry in virtually every instance to help us assign a lineage of the blast populations. And clearly, FISH techniques, G-banding, and cytogenetics is very important, as well as a series of mutations that we’ll describe later on.

In most cases, we do a bone marrow biopsy. We like to have at least a 1 to 2cm sample and this allows us to evaluate cellularity, topography, the presence of abnormal localization of immature precursors which is called ALIP and excluding other bone marrow disorders or bone marrow infiltrations by solid tumors, particularly lung cancer and cellular tumors, and to look and define whether there is fibrosis present.

And clearly, in more recent years, molecular genetics has entered our armamentarium and we almost invariably, at some point in the assessment and diagnosis of patients with myeloid malignancies, obtain a panel between 30 and 50 genes and look to see whether there are specific mutations that can help us not only in diagnosis but also in prognosis.

So, what defines a blast population? This was initially published by us in Leukemia Research in 2012. And as far as a blastic concern, we have a granular blast with a nucleus that has fine, uncondensed chromatin, usually with an obvious nucleolus or we have the same nucleolus with a presence of occasional, anywhere from 1 to 10 or 15, azurophilic granules.

This is clearly separated from promyelocytes, which have a very prominent, easily identifiable Golgi zone. The nucleus has become more eccentrically located. The chromatin is a little more clumpy or coarse, and the granules persist. And then we see abnormal promyelocytes often with the same Golgi zone but innumerable azurophilic granules. So, this defines the blast population. This defines the beginning of the differentiation of the myeloid cell lines.

The other important aspect in MDS is defining dysplasia of the major cell lines. So, here, we’re seeing a variety of different erythroid precursors. Here is one, for example, that shows unequal division. The nuclei, the dotted nucleus here looks different than this. You can see more chromatin appearing here and the parachromatin between it, the Howell-Jolly body. And the cells mature, but they maintain some of the fine chromatin characteristics of the para-cell. And you can see here a myeloblast that has no granules and a promyelocyte and myelocyte.

Often, we will see multinucleated cells. Here’s one with four nucleoli and another blast here. Here’s one with three, vacuolization on the cytoplasm, and note again that the chromatin is very fine and uncondensed. In the lower left, we see a series of erythroid precursors that have not advanced beyond the polychromatophilic stage as you can see here. No orthochromic cells are identified. Here is a giant orthochromic erythroid precursor with what some people refer to as a bag of worms, very megaloblastic in appearance. And in this projection one can see three agranular blasts and then a very abnormal dysplastic erythroid precursor with multiple Howell-Jolly bodies.

Now, I don’t mean to imply that these changes are pathognomonic of MDS because some of these changes can be seen in non-malignant disorders such as congenital dyserythropoietic anemias and even megaloblastic anemias that are responsive to B12 or folic acid. But it would be very uncommon to see such forms.

Now, we also do an iron stain to identify both storage iron as well as the presence of what we call ring sideroblasts. So, here is a granulocyte that contains no iron, red cells only one, had some siderophilic granules that you can see here and here. And then there are rings of small, deeply stained, blue ferritin granules surrounding the nucleus of the erythroid precursors. And we say you can have a ring sideroblast if you have at least five or more such granules, and this defines a ring sideroblast and is clearly an abnormal finding.

For now, I’d like to focus on morphologic dysplasia that is specific for the megakaryocyte cell line and megakaryocytes that we see in the bone marrow. Just for introduction, the megakaryocyte is a very interesting cell as far as bone marrow and hematopoiesis is concerned because it does not undergo major cell division but instead, the nucleus replicates multiple times, it can go as high as 16 and with the normal cell line, of course, of all cells is diploid.

And the result of that is that you see multiple nuclei in the megakaryocyte, but the cytoplasm does not mature very much beyond the basophilic stage and then becomes very granular and stays about the same for the duration of the life of the megakaryocyte. So, the normal megakaryocyte is what we call poly-diploid and can have multiple nuclei, as many as eight, nine or 10, but they’re all connected by a nuclear filament. So, that helps us understand the difference between that and what we see in megakaryocytic dysplasia.

The second point is that the major differential diagnosis in megakaryocytic dysplasia that we see in both MDS and myeloproliferative neoplasms is that in ITP or autoimmune thrombocytopenic purpura, where there is often an increased number of early megakaryocytes with basophilic cytoplasm. So, that, clearly is something you have to be careful about.

So, on our picture here is a series of four different versions of megakaryocytes taken from patients with proven MDS. In the upper left, we have a plasma cell here. Then we have a large mononuclear megakaryocyte with a single nucleus and granular cytoplasm often called monolobated. Over here in the right is a binucleated megakaryocyte and notice that there’s no connection between the nuclei. It’s been often referred to as a pawn-ball nucleus. And this is a dysplastic neutrophil here with a granular cytoplasm. And here on the lower left is the one that is very classically seen in many patients with dysmegakaryopoiesis, and this is almost close to a blast. You can see the cytoplasmic projections, a little bit of granular cytoplasm and a deeply staining nucleus. However, this type of megakaryocyte can be seen in ITP. So, you have to be very careful.

And then finally, this huge cell over in the right is not an osteoclast but is a multinucleated megakaryocyte. And you can see one, two, three, four, 10, almost 18 to 20 different lobes and they’re virtually all connected together. So, for megakaryocytic dysplasia, the minimum numbers that we must see as dysplastic is 10%, but many of my colleagues feel that they should see 20% or two out of five to be sure that you’re defining enough dysplasia to separate it out from non-malignant disorders.

Moving on to the granulocytic line, there are a series of changes that we recognize. We recognize cells that have lost their granules in the cytoplasm and are agranular. There are cells that have intensely clumped chromatin, what we call pseudo-Pelger Huët cell as in this P1 model here. We have cells that have multiple nuclear projections – one, two, three, four, five, six, seven, eight, another dysplastic feature. And a loss of granules, but not nearly a loss of granules as prominent as this cell, which is totally agranular, and this has lost about 75-80% of the granules. If there’s a loss of one-third or more of the granules in the cell, we call that granulocytic dysplasia.

So, originally, we had a classification described by the French-American-British Group. In the low-risk patients, we had refractory anemia, less than 5% blasts, no ring sideroblasts. We had refractory anemia with equal or greater than 15% ring sideroblasts. In the WHO classification, we added the concept of having multinucleated dysplasia, which was mentioned by the French-American-British Group but not really emphasized. We also defined for the first time an abnormality associated with MDS, which is the 5q- syndrome and we’ll come back to that subsequently. And then we always included a group of patients that we felt uncomfortable assigning a specific classification, but they were less than 5% blasts. Some of these patients had myelofibrosis in which we couldn’t actually define precisely a percentage of blasts.

And the same terminology was applied to those patients that had ring sideroblasts. They can have it without dysplasia of the granulocytic or megakaryocytic line and we call those patients refractory cytopenia, multilineage dysplasia with ringed sideroblasts. So, an expansion of these groups in contrast to the original FAB classification.

In the high-risk patients, the French-American-British Group had only one type, 5-19%. And by looking at outcomes with lower percentages of blasts, we were able to define two groups of patients with 5-9% blasts, 10-19% blasts that were called RAEB-I or II or excess blasts-1 or excess blasts-2.

The FAB Group had a small group of patients, around 5%, that we call refractory anemia with excess blasts in transformation. They had 20-29% blasts and these patients often behaved more like AML than they did like MDS. There was a subset that did not, but the WHO committee felt that they could be managed. And so, they had AML, particularly if they were young enough to receive combination chemotherapy. So, this group got removed.

We included in the original FAB classification a group of patients that we call chronic myelomonocytic leukemia. These were patients that had normal white counts, but they had a higher percentage of monocytes, numbering greater than 20%, but they had a blast percentage less than 20%. So, they behaved in many ways identical to patients that had MDS and they indeed did have a dysplasia.

These were removed by the WHO committee because many of these patients actually presented with elevated counts or progressed to a count above, say 11,000. So, they ended up in a separate category, which we call MDS or multiple myeloproliferative disorder or myeloproliferative neoplasms, sort of a combined category. They had features of myeloproliferative neoplasms like an elevated white count with monocytes that had the morphologic features of patients with MDS.

So, if you look at a pie diagram, what happened with the FAB categories, you can see here that the patients that had high-risk MDS that had a very short survival, basically very similar to AML untreated. There were patients that had less than 5% blasts that had a reasonably good outcome, much shorter than controls for that age group but could live on the average of three years. And in the FAB experience, the median survival here was about the same, and you’ll see subsequently the revised discussions and publications shows that this group actually does even better. The RAEB group was, I’d say, 5 to, actually 30%, 20% compared here and you can see that their survival was 18 months.

We included patients with CMML who had known proliferative and it’s interesting that their median survival was about the same as the median survival of all the other categories. And that was because we did not define a percentage of blasts in our original FAB classifications, and these patients could have either a low blast percentage or a blast percentage that ranged as high as 20%.

And if I looked at a survival curve that we published subsequently in 1997, you can see that the MDS classification, as simple as it was, allowed one to define various risk groups. Refractory anemia with ringed sideroblasts having the better survival than RA patients, again less than 5% blasts. RAEB and CMML right about located here. And you can see, of interest, in green, the CMML group sort of puts all of these other three groups together and this actually turns out to be about the median survival of patients who have MDS without CMML. And that is dictated very much by the percentage of blasts that we report as hematopathologists.

So, moving forward into the WHO classification. And this is a pie diagram that I haven’t seen published anywhere but I thought it’d be interesting to show it. It shows you what happens when you make moves that separate out subgroups. So, the WHO category took the RAEB-T group and took it away. This is now AML. They took out CMML and put it into an overlapped category of MDS/MPN. This represented 25% of the patients that were in the original FAB classification.

So, what happens when you take this out is, you’re left now with just three categories. RAEB, RARS now account for about 50% of the patients you see. Whereas RA, refractory anemia, by itself accounts for 50% of the patients. So, you get sort of a totally different appearance of your patient population that you’re seeing in the clinic on a daily or weekly basis because you removed two categories.

And now, what happens is when you look at the WHO classification, you end up with several different groups. The AML group in blue are the patients that had RAEB-T in the FAB classification and a median survival of less than a year. The RA/RARS group combined is up here. If it’s a refractory cytopenia, refractory cytopenia with multilineage dysplasia and/or RARS, you’re here and you can see that the addition of dysplasia outside of the erythroid line. So, these patients have less than 5% blasts, but they have either dysplasia of the erythroid and granulocytic or dysplasia of the erythroid and megakaryocytic lines or all three. But as soon as you cross the threshold of 5%, down goes the median survival and when you break across 9%, it falls down here, very close to where you are with some patients with AML. And that’s why many investigators sort of look at this RAEB-II group patients if you’re young enough to consider treating patients with chemotherapy that you use for AML. Whereas some of these other groups, you would be more likely to use hypomethylating agents.

So, a final update occurs in 2016 actually the paper appeared only two years ago. It took a while to get published. Refinement of cytogenetic risk groups, more quantitative use of cytopenias, defining interest in maybe less than 5% blast group in the two groups, less than 2% and 2-5% which you can quantify in expansion of the group of patients with the IPSS-R from 800 to 6,000 patients in the patient population. And for the first time, identifying point mutations basically that appear exclusively in patients with ringed sideroblastic anemia with a SF3B1 mutation.

Now, just to show you how sensitive the percentage of blasts can be, this is a database in Dusseldorf, Germany that I reviewed on a mini sabbatical several years ago. And you could see that when you make these various cuts based on the percentage of blasts alone, the median survival drops precipitously. And even the separation of 1% in blast count takes a median survival and drops it almost 50% to 37%. So, 0-2 is a meaningful way to report blast percentages and we ask our hematopathology colleagues to not just say the blast percentage is less than 5%, to really define it precisely.

So now what happens with the new classification, we have this very large group that I indicated before. It stays about the same. Multilineage dysplasia without ring sideroblasts close to 50%. Ring sideroblasts, 11%. Single-lineage dysplasia, so this is this group but no ring sideroblasts. RCMD, this had multilineage dysplasia with ring sideroblasts and then deletion 5q, a separate category. And again, we always have one category that we include for patients that we feel uncomfortable assigning to one of these other categories.

And not surprisingly, we end up with a variety of different survival curves. The AML group is left out of this. RAEB-II here and you can see there’s a separation between RAEB-I and RAEB-II. Refractory cytopenia with multilineage dysplasia has a median survival of about 40 months. As soon as you drop down to above 5% blasts, the median survival, as you can see, drops in half. Patients with deletion 5q or unilateral dysplasia or RARS, all sort of lumped here together. It’s hard to vary, to separate them out. But as soon as one hits a group of patients that have neither 5q- or ring sideroblasts, down goes the survival.

So, let’s move on now to look at what the impact is on patients who are found to have somatic mutations. And in this day and age, I would say that 90-95% of our patients of whom we suspect MDS, or we think have MDS or convinced they have MDS get a myeloid mutation panel. At our institution, that myeloid mutation panel targets 35 genes. In some institutions, it’s even higher and goes as high as 40 genes.

Here, we’ll cover virtually all of the gene mutations and regulators and splicing genes that one can think of, including TP53. And it turns out that the most common mutations that we see are patients who have RNA splicing mutations and then particularly SF3B1. As you can see, the frequency in the patient population and the horizontal curve where they’re commonly seen. So, for example, if you take patients that have an SF3B1 mutation, they virtually all fall into the RAS category or the RCMD categories, this large group and this large group.

And we can go on. So, does this have an impact? Well, it does. And this is a study published by Bejar and colleagues back in 2011 and they’re probably one of the best studies that has been reported. These are mutations that were considered, five altogether. In the upper left is their experience with the International Prognostic Scoring System. Low-risk, intermediate, high-risk, very high-risk. And as you can see, the curves separate out very nicely.

If one introduces the concept of taking all the patients and defining whether they have mutations. So, in this series, mutations were present in about 25% of patients. Some series report more. This is a group that had one or more mutations, and you can see there is a significant decrease in survival, and this almost corresponds to patients that have intermediate-2 risk patients, no mutations or low mutations.

And this is present in patients whether they have low-risk, intermediate-1, or high-risk disease. Notice that the higher the risk, the less separation of the curves, which is really not surprising because once you get to intermediate-2, you really are very high risk and it’s unlikely that mutations are going to make a difference. But they continue to make a difference in intermediate-1 and for low-risk mutation absence, low-risk mutation present. So, what happens is that if you have low-risk absent mutations up here and then you add a mutation to that, anyone of these, you behave as though you have intermediate-1. So, mutations tend to leave you one full risk higher in the sequence.

There is an international study under way which is global and confirmed these findings. We already have published one article which shows that if you have biallelic TP53 mutation, you do indeed behave as though you have very high-risk.

Now, where this is particularly of interest is the presence and absence of ring sideroblasts. So, there’s a very strong correlation between ring sideroblasts as you can see here and the finding of a spliceosome gene mutation SF3B1. Here are patients with the mutation and without. So, there’s always exceptions to the rule. This is the only favorable mutation that we see in patients with MDS. If you have it, you do better. And if you have it, you’re very highly likely to have the SF3 mutation.

So, this is some unpublished data from my good friend, Torsten Haferlach, from the Munich Leukemia Laboratory. This was blinded to the observers. Her is the presence of ring sideroblasts from 0, 1-5, 5-14, way over 15%. And on the vertical curve, the percentage of patients that had mutations. So, greater than 15%, 70% of patients had ring sideroblasts. And the problem is pretty obvious that you can’t rely on either of these modalities to help you out. So, what we do is we do both. We will have patients that have so few ring sideroblasts that it’s hard to quantify them and you have upwards of 5% of these patients will have an SF3B1 mutation. If you don’t see any ring sideroblasts, you’re not going to find a mutation.

Here’s the paper just recently published by an international working group sponsored by the MDS Foundation. It’s entitled SF3B1: the lord of the rings in MDS published in Blood in the year 2020. So, if you based it on the WHO classification, ring sideroblasts can be seen in single lineage or multilineage dysplasia. If they have the wild type, meaning no mutation, there’s a higher risk of dysmegakaryopoiesis, reduced overall survival, and a heterogeneous molecular profile. If you have the mutation, it’s a specific phenotype. It has a very favorable outcome with a restricted spectrum of subclonal mutations. And even in the group in which you don’t see ring sideroblasts, about 5% of these patients will have the mutation and they will do just as well as these patients. So, the international group is now proposing that the next revision of the WHO redefine the ring sideroblast chapter and more precisely define how these patients behave.

Now, moving on to some of the specific types, these are patients with what we call the 5q- syndrome or deletion 5q. They have usually normal cellular nodes, decreased erythroid precursors. The male/female incidence is reverse. So, it’s more common in middle-aged women. And in the bone marrow, we see relatively few erythroid precursors, but you see normal to increased numbers of megakaryocytes and note that they have a single nucleus. You could see here, sort of monolobated.

Sometimes, it can be picked up even easier on a bone marrow biopsy. So, here, you have granulocytes, a few erythroid precursors, and you can pick up very easily this large megakaryocyte with a single nucleus. Another one here and another one here.

So, these patients actually have a pretty good survival depending upon whether they present with just single or one chromosome abnormality as you can see here, once you get one or more, survival drops. Or if you show up initially or develop a mutation involving TP53. So, here, you could see a group of patients who never developed TP53 mutation, and their survival is almost as good as age-corrected normals.

So, have changes in the WHO broadened the definition a bit? If you have monosomy 7, we know that’s not good. We strongly recommend the TP53 mutation be included and if the blast percentage gets beyond 5% and/or dysplasia, these patients are not going to behave as well as if you have a deletion 5q. And that’s very important because of the introduction of lenalidomide, the immune modulator which certainly provides a excellent response rate but primarily in patients that have the isolated syndrome. And if you present with two chromosome abnormalities or more dysplasia and more cytopenias, and 5 or 6 or 7% blasts, you may respond but you won’t respond for very long.

Okay, so this here’s basically a summary that summarizes where we stand. We provide how to approach these patients. Remember now that monocytes are no longer counted because they’re now included in the CMML group which is in a different category. So, we have blood and marrow definitions for each of the subtypes. Presence of dyserythropoiesis, multilineage dysplasia, you must then have dysplasia of erythroid, plus one of the others, which reminds me to indicate that isolated dysplasia of megakaryocytes or granulocytes without dysplasia and erythroid precursors is an extremely rare event. I’ve seen it maybe in 1-2% of all the patients I’ve diagnosed with MDS. And to be sure that you ask your pathologist to define precisely the percentage of blasts so you can then calculate the prognostic scoring system.

For higher-risk patients, life hasn’t changed very much, fortunately. We sort of dropped the term “refractory anemia” since all these patients are anemic and just simplified it by calling these patients EB-1 and EB-2 rather than RAEB-I and RAEB-II and again we exclude an increased number of monocytes. The 1000 level for monocytes, by the way, is two standard deviations above the high normal level in most laboratories.

So, this summarizes again where we are for both the low-risk and the high-risk patients and what we define is necessary to be done and to fine tune the classification a little bit better. And again, we leave this unclassifiable group in. So, these are patients that have, say, a little bit of dysplasia with 1% circulating blasts in the peripheral blood and are present in two consecutive peripheral blood reviews, CBCs, at least two or three months apart.

So, this is all then applied into the IPSS-R, which for sure everyone is using now. Five groups. It incorporates a variety of different chromosomes and if you don’t have cytogenetics, we recommend G-banding rather than FISH. If you can’t do that and you do FISH, you have the results. You can use the classification.

The classification for risk assessment cannot be used without the karyotypes because the karyotypes wait. In a multivariate analysis, it comes very close to waiting, a high degree of waiting for the percentage of blasts. But as you can see, clearly, the median survival goes from five and a half down to less than one year. 25% of patients progress to AML and then you can see the percentage of patients in those categories; very good to good, close to 80% of the patients that you see.

And if you look at the cytogenetic risk groups, you can see what the impact of the cytogenetic risk scores is. It’s very high up on the score criteria for defining risk. As you can see here, poor risk you get a four. There’s no other category where you get a four. Marrow blasts, the highest we give is a three. So, this is extremely important and that’s why we say, “Don’t use either IPSS or IPSS-R unless you have chromosomes that have been done”. And you can see here, again, how the survival is depending on the percentage of blasts in the upper curve and when you look at the whole grouping together, you get a very, very nice separation of the survival curves.

Okay, now let’s take a look at some of the nuances of making a diagnosis of MDS. So, this is a cartoon developed by David Steensma who most recently was at the Dana-Farber Cancer Center but has now moved into industry. But he looked at patients who don’t quite have MDS, but may have a mutation, and talked about traditional ICUS. I-C-U-S, clonal abnormalities of unknown significance. Idiopathic because we can’t define what causes it. So, there’s no clonality. There’s no significant dysplasia, which we define as 10% of a dysplastic cell line. There are very few blasts. The overall risk is low.

Now, we’re moving into CHIP, which is really clonal abnormalities picked up in individuals who were screened for cardiovascular risk and turn out to have a clonal abnormality that could be seen in patients with myeloid malignancies like DNMT3A. So, they have clonality but no dysplasia, no cytopenias, and they’re observed.

This group has clonal abnormality, meaning they have a cytogenetic abnormality. Dysplasia, cytopenias. And these patients really behave pretty much like patients that have low-risk MDS even though they don’t have dysplasia. We’ll see a survival curve on this. So, these are the patients that clearly have all three categories. This group has clonality defined by the presence or absence of one of these genes, but as yet, they don’t have cytopenias. They have cytopenias but they remain undefined.

So, a few years ago, the Italian group headed by Luca Malcovati in Pavia published a paper in Blood in which they looked at a group of patients who had normal chromosomes, had mutations, had unexplained cytopenias but could not be defined as they had MDS because they had no dysplasia. And what they found is that the presence of some of these mutations allowed these patients to behave as though they had a high predictive value for myeloid neoplasms and often behaved as though they had MDS, with the exception of SF3B1. And two or more mutations had a very high predictive value for behaving like patients who had MDS.

And you could see that when they compare the patients who had a mutation pattern in blue to patients who had known MDS in red, they looked virtually the same. So, what they’re suggesting is if you have patients with unexplained cytopenias and have a myeloid neoplasm mutation with a high variable allele frequency, you could view these patients as though they have low-grade MDS.
However, 10% of healthy patients, healthy individuals, harbor somatic mutations and the allelic burden can be 10-20%. There is an associated increased risk of hematologic malignancy and death, but currently, just defining the mutations to diagnose MDS in CCUS, you’ve got to be very careful and circumspect about calling them MDS unless you have some evidence to suggest that the cytopenias are such that you want to introduce therapy.

Now, coming up, and this is really now the new world we’re entering into, a recently published paper, multi-authored, from Europe, in JCO, 2,000 patients with de novo MDS, looking at 47 genes, they defined eight subtypes with prognostic and clinical lab differences and no obvious correlation with WHO types except for this RARS/RCMD-RS which has the SF3 mutation and RARS in transformation. 14 genes were unfavorable and only one was favorable.

So, this is why I tried to figure out if I could get a group breakdown. And if you read through paper, they come up with seven different types. An AML pattern, not surprisingly showing some of the mutations we see in AML like NPM1, FLT3-ITD. Another type that has SF3B1 and the JAK mutation pathway, which we now call RARS-TRS, increased platelet counts through the JAK2 mutation as you can see here. And then an SF3B1 mutation, but the blast counts greater than 5%. A zero group with no mutations behaving sort of like hypocellular aplastic anemia, and you go on.

So, here are the various groups they found and pathogenetic lesions with correlations. The fewer you have, the better you do. But note that they weren’t able to break this down by their groupings. This is broken down by the number of mutations. So, if you have greater than five mutations, these patients behaved like they had EB-2 or AML. If you’re a zero on to two, you’re up here.

Now, the best that they could do was look at a post-transplant survival population and then the grouping does show a correlation. But nowhere in the paper could I find this type of analysis from patients who first present, which suggests to me that they weren’t able to carry that out. So, this is a different survival curve of patients who underwent allo-bone marrow transplant and then they looked at their survival. And clearly, if you want to go allo-transplant, they’re a highly selected category. So, you have to take this with a caveat.

So, in conclusion, all of these expressions or statements are true, 100%. No false positives, no false negatives. We will all die someday. We will all pay taxes forever. And about every five years, we all have to endure classifications or evolutions.

Thank you for your attention.

Watch the full session



Explore the latest updates in MDS, including expert interviews from major international congresses, podcasts and e-learning.


John Bennett – No Disclosures

Sign-up for our Newsletter!

Keep up to date with all the latest news with our monthly newsletter