Educational content on VJHemOnc is intended for healthcare professionals only. By visiting this website and accessing this information you confirm that you are a healthcare professional.

 A session between leading experts Amer Zeidan, Sanam Loghavi, Rami Komrokji, and Somedeb Ball, who discuss the latest updates in the classification of MDS & their impact on clinical practice.


 

Read the full AI-generated transcript

Hi, everyone. Welcome to the MDS Sessions of VJHemOnc. It’s a pleasure to be with you today. My name is Amer Zeidan from Yale university, and it’s really a pleasure to have three of my esteemed colleagues and friends. All of them are experts in myelodysplastic syndromes. And today we will be talking about classification updates, in particular in myelodysplastic syndromes ,and how that is going to impact practice day to day. So it’s a pleasure to be joined by Doctor Sanam Loghavi from the MD Anderson Group, as well as Doctor Rami Komrokji from Moffitt Cancer Center and Doctor Somedab Ball from the Moffitt Cancer Center. Thank you so much for joining me. Sanam, maybe we can start by talking about the evolution of, uh, the classification systems from a heme-path perspective in myeloid malignancies in general, but in particular MDS just to give a perspective to people. Uh, the situation of having two different classifications is somewhat novel because, compared to lymphoma where I think for many years they used to have three different classifications and, um, somewhat used to it in in the myeloid community. For a long time, we only had the FAB and then the WHO. So we are somewhat in a in a new era for us. So maybe you can walk us through the evolution of, uh, those classifications and then maybe discuss, um, the major changes in the 2022 WHO and ICC classifications,

Of course. So, hi, everyone. It’s a pleasure to be here with all of you. Thank you very much for having me. Um, yeah, this is, you know, without without dating myself and maybe aging myself. I would say that like you said, you know, the really the French American British or the FAB system was for the longest time, what was used, Um, essentially based on, you know, with different blast percentages, but essentially based on blast percentages for the distinction of refractory anemia, which was, which was you know what we used to call MDS, uh, from refractory anemia with excess blasts from refractory anemia with excess blast and transformation and then acute myeloid leukemia. And of course, you know, a lot of FAB. While it was morphologically defined right, we had the it was defined by different, aml was defined by differentiation. But if you think about it, there is a you know, phenotype or the morphology was a manifestation of the genotypic underpinnings of disease. Right? So if you go back and think about it now, the monoblastic leukemias, the M5 leukemias were probably enriched for the MLL or KMT2A rearranged cases. The acute myelomonocytic leukemias were probably enrich for NPM1-mutated cases. Of course, M3, everyone knows was, you know, um, the APL or acute promyelocytic leukemia and then inversion 16 was the AML M4 with eo. So a lot of it was genetically linked or defined, but it wasn’t per se a genetic classification. Um, you know how we recognized genetic classification now? So in 2001, the WHO, um, came out with its first classification. Uh, so we the the initial publication, the hematolymphoid Book or the Blue Book was published. Uh, there was a second iteration. Um, I believe that was published in … I don’t know which one. The when the second one came out, actually. But there was a second iteration that came out in 2016, we had an update to the 2008 version of the WHO And in 2016, you’ll notice that the WHO starts becoming much more genetically defined. This is now where the mutations make it into the classification. Let’s say AML with NPM1 becomes a provisional entity, right? AML with mutated FLT3. Actually becomes an entity. I think there is a lot of overlap between the two systems because, you know, essentially the two systems are based on the same data, right? You would hope that you know that. You know, people aren’t making up entities, really. They’re using data to back up the classification, so there is a lot of overlap. But I think there are subtle, um, yet clinically important differences in the to classification systems in terms of how clinicians utilize these classification systems to treat patients. And, you know, I think for for us as hematopathologists, it’s important to recognize that our care of the patient doesn’t really end with our report. Right. Your your report is going to be what dictates, really eligibility for a clinical trial. Your report is going to be what dictates whether or not a patient can get first for a particular drug. So I think there are several connotations and several downstream effects, uh, to having two classifications. That is definitely not ideal. So I hope that at some point we can unify this classification and have one classification that benefits the entire patient population. But for now, we do have two classifications and maybe, you know, we can review them in details and see what we can learn today.

Yeah, that’s that’s a great overview. So maybe Sanam, you can show us some of the main highlights of the 2022 classifications. A lot of our audience might not be as super specialized in, uh like the rest of us. So maybe focusing on the major, um, kind of changes in a big picture?

Okay, so here’s, um, as I said, in 2022 the two classifications were published. The summary of the WHO, the 2022 version, or the fifth edition of the WHO was published in the Journal of Leukemia and the ICC or the International Consensus Classification, which is now a novel classification system. Uh, was published earlier that year in June of 2022 in Blood. So both of these, uh, papers are available. Um, for reference for anyone that is interested. The beta version of the WHO book is online, and I’ve listed here the website for you where you can access this. Um, there is a subscription fee, but I think you get access to all of these these books, including the other systems for anyone that’s interested in the other systems. But essentially, the beta version of the book is still open for comments. So if you should, you know, so feel inclined to make a comment on any of the entities. I think the editors actually do receive these comments. So, um, I really wish, you know, want to provide an overview of, um, the concept of these new classifications. And it’s if you look at the first of all, the precursor lesions have been, um, introduced in the classification system, and these are the clonal hematopoiesis and CHIP and CCUS entities being clonal cytopenia of undetermined significance Um, and essentially, this is now, You know, we had recognized this as a field for a while. That CHIP exists, clonal hematopoiesis exists. And it is a precursor lesion. Two myeloid malignancies, although, you know, very important to recognize that not everybody with clonal hematopoiesis gets a myeloid malignancy. In fact, a very small fraction of people actually get a myeloid malignancy. But again, it’s not formally recognized as a entity, uh, in the WHO book as well as in the International Classification System

07:40
08:06
or consensus classification system. Now, uh, what is very important is that while the border between these clonal cytopenias, uh, and MDS is really, um the distinction is made using dysplasia and that is 10% dysplasia in any cell lineage, including the granulocytic, erythroid or megakaryocytic series

08:06
08:29
Last count that we were using previously really, to distinguish MDS from AML has now become much less relevant because you’ll see that in the AML category. A lot of entities are now genetically defined. So now we have genetically defining AML entities, and this includes you know, before as we said, we had the acute promyelocytic leukemia.

08:29
08:52
We had the core binding factor leukemias, where the blast count was not relevant. You could make a diagnosis of acute myeloid leukemia in the presence of these genetic abnormalities, regardless of the blast count. So now this this broad category of genetically defined AML has kind of expanded, really, making the blast count less relevant. But you’ll see that there’s some subtle differences in the blast count percentage that we use.

08:53
08:59
Um, all right. So I’m going to start with the updated classification, um, for MDS and the WHO

09:00
09:26
So there’s two broad categories. One is MDS with genetically defined abnormalities. And under this category, first of all, this category trumps the other category, which is the MDS that is morphologically defined. And under this category, you have MDS, um, with isolated deletion 5q. The name here is a little bit misleading because despite the isolated,

09:26
09:51
being in the, you know terminology and the name, you actually allow up to one additional cytogenetic abnormality in this category, unless it’s a deletion of chromosome 7. So monosomy 7 or deletion of the long arm of chromosome 7, which is del(7q). Because obviously those cases have very bad prognosis, and so you don’t want to lump those with isolated del(5q). Which is supposed to be a better prognostic category,

09:52
10:21
and then you have MDS with low blast count and SF3B1 mutation again recognizing that this is a distinct group of myelodysplastic syndromes with SF3B1 mutations and ring sideroblasts. Now, one of the subtle differences between ICC and as a separate entity, as a sub entity of this and that is MDS cases with increased ring sideroblasts meaning more than 15% ring sideroblasts even if they don’t have an SF3B1 mutation.

10:21
10:51
And the reason for that is, you know, we had known, uh, from the results of the luspatercept trials. That patients with ring sideroblasts, even if they don’t have an SF3B1 mutation, they respond to luspatercept. So I think this category was added here to allow for these patients to get treated because you don’t want to bury the ring sideroblast part in, you know, somewhere in the text of your report, where nobody sees it right. You want to put that in the diagnosis so that the clinician knows that this patient may actually respond to luspatercept

10:52
11:01
And then, um, obviously, here you cannot have deletion 5q, chromosome 7 abnormalities or a complex stereotype because those categories will trump SF3B1

11:01
11:27
and then you have the novel entity of MDS with biallelic inactivation of TP53. And I’ll talk to this a little bit more in detail. But essentially, here you have a myelodysplastic syndrome that has loss of both functional copies of TP53 either through multiple mutations or a mutation and a deletion of the wild type copy or a mutation and copy neutral loss of heterozygosity of the other copy.

11:27
11:53
And then, if you don’t fit into any of these categories, you’re left with the morphologically defined entities. And here we have MDS with low blasts and the percentage for low blasts, the cutoff in the bone marrow is 5% and in the peripheral blood, 2% you have hypoplastic MDS, which is defined as a myelodysplastic syndrome that has a age adjusted cellularity of less than 25% of what would be normal. And then MDS

11:53
12:19
with increased blast is broken up into category two categories. One is MDS with increased blasts, and the percentages of the blast are the same as they used to be before and then MDS with increased blasts to again same percentages or if you have Auer rods and again important here to note that the WHO recognizes that these cases may be treated as AML should the treating clinician decide that this patient

12:19
12:44
is better off being treated as AML. And then MDS with fibrosis again. Uh, you know, it’s recognized as a distinct entity, but I do want to highlight that the majority of cases with MDS with fibrosis actually do have TP53 mutations and biallelic loss, so that category would trump this category. So this category only qualifies for the cases that don’t have a, uh, TP53 mutation

12:45
12:57
and then morphologic dysplasia is a qualifier. Remember we used to say, in the previous classification, we used to make a distinction between unilineage dysplasia and multi-lineage dysplasia, and that has been omitted from the WHO classification.

12:58
13:23
The ICC classification is very similar. Um, there are some subtle differences. As I said again, you know, the ICC breaks MGs down to morphologically defined and genetically defined, that genetically defined MDS del(5q) is essentially the same. There’s no difference in classification between the two systems. Uh, the ICC restricts MDS with mutated SF3B1 to cases that have the mutation. So

13:23
13:49
if you have increased ring sideroblasts but you don’t have an SF3B1 mutation, you would be, um, under ICC, you would be subclassified as MDS NOS right. And then, obviously you can’t have an isolated deletion 5q or minus seven. And here they require that SF3B1 mutation have a VAF of 10% or greater, and you cannot have mutations in TP53 or RUNX1

13:49
14:13
MDS with mutated TP53 is essentially the equivalent of the biallelic inactivation. Here you require multi hit TP53 alterations and then if you don’t meet any of these categories, then you’re left with the morphologically defined here. Note that the hypoplastic MDs and the MDS with fibrosis that are defined in the WHO fall under the NOS category.

14:13
14:33
But the ICC does actually make a distinction between single lineage dysplasia and multi lineage dysplasia, which I think some will talk about. And, you know, it is apparent from some studies that the degree of dysplasia may actually still have some prognostic, uh, implications. And then MDS with excess blasts

14:33
15:01
is, um, you know, a little bit different from the WHO. Because you’ll notice the threshold of blast percentages here is different. So you only qualify for MDS up to 9% blasts in peripheral blood and bone marrow. Once you reach or pass the threshold of 10% you’re into this novel category, which is called MDS/AML. Again giving the clinicians some leeway to treat the patient as MDS or AML if they see fit

15:01
15:21
and, you know, to to fit into this category, you cannot have any of the AML-defining genetic abnormalities, which are the translocations, NPM1 mutation, um, and then CEBPA mutation for the WHO and TP53. So any questions about the classification or can I clarify anything?

15:21
15:46
Thank you so much then, Uh, this was very helpful. And, um, I think, um, from a community practice point of view or many of the physicians who don’t see patients in a in a large center with a very specialized impact, Um, it’s not fully clear to us yet how this is gonna be reported. Are people gonna report using the two different, uh, schema or is it gonna fall on the

15:46
16:11
on the physician to kind of decide based on the description how to, uh, to classify the patient? But, you know, I think some of the categories, such as MDS with fibrosis in one classification being reported as MDS with excess blasts in the other. Or um uh, as you said MDS ring sideroblasts. So I think because of all of these, um, reasons, there was an effort initiated by the Moffitt group by Doctor Komrokji

16:11
16:29
um mentoring, uh, Doctor Somedeb Ball. And eventually this expanded under the International Consortium for MDS to try to compare first, validate the two systems in terms of their prognosis, but also to compare the differences. And eventually the goals

16:29
16:47
is to try to come up with a roadmap so that we can go back to having one system, which I think is, uh, something that everybody wants. So maybe, uh, Doctor Ball, you can go over the main highlights of your presentation. That was an oral in in ASH 2022.

16:47
16:55
Yeah. Glad to be part of the panel today. I’m happy to discuss our findings, which, as Doctor Zeidan mentioned we presented in ASH, recently concluded.

16:55
17:24
So what we tried to do is we looked into our database institutional database and Moffitt Cancer Center. Um and we, uh, we took all the cases which confirmed diagnosis of MDS, uh, by wh 2016 criteria, which Doctor Loghavi went over with available MDS results at diagnosis. And then we reclassified those cases by WHO 2022 and ICC 2022 proposed criteria, which we just went over. We, uh

17:24
17:54
we defined a TP53 mutated case to be multi hit if they had a variant allele frequency of 50% or more. If they had two different the TP53 mutation or they had one mutation, but they had additional 17p deletion, which we check usual karyotype and FISH analysis. We had a little over 2000 patients who were molecularly annotated in the study with a median follow up duration About five years

17:55
18:25
here you can see in the table we we have summarised median estimate of leukemia-free survival and overall survival for all the subgroups, MDS subgroups by WHO 2022. The thing I would like to highlight is as we reviewed, uh, and consistent with the previous literature MDS with SF3B1 have the best outcome of all subgroups with the median overall survival of a little over 100 months, whereas MDS with biallelic TP53 inactivation

18:26
18:35
by WHO had the worst outcome of all as designated by the red curve here with the median overall survival of only 13 months.

18:36
18:58
Then when we looked into the ICC 2022 classification subgroups, uh, there are some subtle differences as we reviewed between WHO and ICC, So ICC’s SF3B1 subgroup had a little better overall survival, but overall still a lot improved compared to other subgroups similar to WHO

18:58
19:20
so um TP53 mutated subgroups be it MDS with TP53 or MDS AML with TP53 in ICC, which required a blast count of 10% or more had the worst outcome of all with MDS/AML with TP53 having the median overall survival of less than a year. As you can see here,

19:21
19:51
then we reviewed that WHO had had provisions to call someone even with the increased ring sideroblast without an SF3B1 mutation as a, you know, as a designated entity. But ICC did not So in our cohort, at least when we compared this subgroup who did not have an SF3B1 mutation and had increased sideroblast ring sideroblast you know they had similar outcomes compared to other low blast subgroups. As you can see here,

19:52
20:09
MDS with multi lineage dysplasia had worse outcomes. Uh, leukemia-free survival and overall survival compared to single lineage dysplasia. As you saw that ICC recognized it as separate entities. But WHO did not. But it seemed to matter in terms of outcomes.

20:11
20:40
Then we tried to look at the blast cut-off. We know you know we have some kind of a 5%, 5 to 9% in wh is called MDS with increased blast 1, uh So when we compared them to non molecularly defined low blast patients like patients who did not fall into SF3B1 or difficulty mutated categories? Uh, when we compared MDS with IB1 one had a worse, uh, leukemia-free survival and overall survival compared to low-blast subgroups.

20:42
21:03
Then we looked at last count, count about 10%. So 10 to 19% patient’s, uh, patients with blast count of 10 to 19%. They had similar overall survival. All the worse leukemia-free survival, but similar overall survival. As you can see, the curves overlapping here compared to, uh, those with 5 to 9% blasts in our patients

21:04
21:22
WHO had a designated category with MDS with fibrosis, increased blasts patients in our dataset. MDS with fibrosis had a worse leukemia-free survival and overall survival compared to MDS with increased blasts.

21:24
21:52
Uh, we talked about this multi-hit 53. Usually when you have two different hits, meaning 12 mutation on one mutation with another seven p deletion. Usually that carries worse prognosis. You try to make a sense of how it, uh you know how it comes across when we look at it across the different blast subgroups you can clearly see when the blast cut off. Blast count crosses 5%. The outcomes are generally lot worse.

21:53
22:20
So to conclude, both WHO and ICC classification systems, we conclude that they have room for improvement. Molecularly defined entities like SF3B1, del(5q) and TP53, they are clearly unique and TP53 mutation in both classification predicted the most dismal leukemia-free survival and overall survival. Multi-hit TP53 status of course carried, uh, it was an independent predictor of survival.

22:20
22:31
Survival of MDS with ring sideroblasts, SF3B1 wild-type was similar to low blast subgroups, and MLD had worse outcome than single lineage dysplasia.

22:31
22:58
A blast count of 5% cut-off correlated better with overall survival in, uh than 10% in our cohort. However, we’re trying to still decide on precise blast cut-off, and there are some collaborations we’re trying to put together to answer that question. Uh, grade two or three fibrosis in increased blast subgroup carry the worst prognosis. And as I said, validation in a multi center data service ongoing to further support our findings.

22:58
23:15
So with all this knowledge generated from this data set and also from existing literature, we propose this unified approach to classification in MDS. Um, considering incorporating the prognostically significant subgroups from both WHO and ICC

23:15
23:31
uh, so in a patient with the diagnosis of MDS, I think we should look at the presence of TP53 mutation, whether this is multi hit or single hit and accordingly classify them. If that is not present, then we should look at the blast count. In our data set, 5% made more sense.

23:31
23:47
So as prognostically So 5% or more, we look at fibrosis grade. If that is two or more, we call it MDS with fibrosis. If that is not present, then we call them MDS with excess blasts which can be further subdivided in EB1 and EB2

23:47
24:16
Less than 5% blast count patients can be further looked at for presence of deletion 5q, SF3B1 mutation. Hyperplastic have a plastic marrow. Um, if none of them are present, then they should be called MDS with low blast, which can be further subdivided based on number of dysplastic lineages in bone marrow to MDS single lineage dysplasia, splash and multi-lineage dysplasia. Thank you so much and have to take any questions you have.

24:17
24:44
Yeah, thank thank you. Excellent presentation. And I do think this hopefully, um, this and similar efforts and hopefully larger multi center efforts will provide a roadmap to try to harmonize things on an evidence based fashion rather than just expert opinion. But of course, that’s going to take some time to happen. And until it happens, we have to deal with two different classification systems. And, um uh,

24:44
25:11
Rami, I want to take your insights into two specific aspects. One is what are the downstream implications both in terms of the clinical care but also from clinical trial perspective, registering other aspects that could be impacted when you have two different classifications in which there is not one that is uniformly adopted. And the second aspect since we are going to have to deal with this day to day in the clinic.

25:12
25:23
Um, how are you going to navigate discussions with the patients and, uh, kind of making sure that there is no confusion and aligning your clinical practice with treatment

25:23
25:23
visits.

25:24
25:45
Absolutely. Thank you Amer so I think first, thank you, Sanam and Som, they set the stage perfect. I think my part is to think philosophically how to take this to practice. And, uh, I think the first point to make is like, we really mix a lot between classification and prognostication tools, and that’s even mixed. And even drug approvals, like you’ve

25:45
26:06
find drugs approved for higher risk MDS based on the IPSS historically – we even don’t have the revised IPSS maybe recently. So a classification is really reflection to the biology. And as Sanam mentioned, like even 30 or 40 years ago, those entities that we saw under the microscope correlated with some biology or like mutations that we know about,

26:06
26:27
So the classification should be really unique based on biology. The first important thing of the classification is making the diagnosis because obviously that’s very crucial for, you know, treating the patient’s the reimbursement, all that stuff, and I think you know That’s one important thing. Then the second thing is putting the patients into groups that are homogeneous

26:28
26:53
now, not all the time. A classification system have to be the perfect prognostic system because obviously there are other variables you know, you’re looking at, you know, the cytopenias, the depths. Probably the blast percentage does matter as a continuous variable. It’s not like, you know, groups, but we try to make it more homogeneous groups, uh, and so forth, I think. You know, obviously

26:54
27:14
the, uh the third thing to think of a classification is how generalizable this classification. So, like, we are here making a leap of faith that everybody has access to genetic data and information, and probably that’s probably less than 10% of the places in the whole world. Uh, so, like to the point that Sanam made, like, you know, if

27:14
27:35
you have MDS with ring sideroblasts like, yes, maybe the ring sideroblasts without SF3B1 have less better outcome or inferior outcome to SF3B1. But the reality is, if you see ring sideroblasts on a bone marrow, you probably have 80-85% chance of having SF3B1. So in a place where there is no access to genetic data, I think

27:35
28:04
I think you have to make that assumption of treating the patients based on that because that’s important. Then obviously think that we, you and I discussed, like, obviously ring sideroblasts. If you have molecular access and you don’t have that SF3B1, make sure it’s, for example, not a nutritional deficiency like copper deficiency that’s mimicking MDS. That’s what we see on our side. Same thing, even when it comes to P53. Sometimes you could do immunohistochemistry, and there is some good correlation there. Or if you have a complex karyotype, you can make that assumption.

28:04
28:12
So So we have to think of, like, how generalizable. How can we get by a little bit without having access to complete genetic information?

28:13
28:35
Now, how to translate this in practice, I think from clinical trials down the road, we really have to focus a little bit on more homogeneous group, part of unfortunately, our failure developing drugs is we just took MDS in general and sometimes are saying higher risk or lower risk and we’ve not been able to really, you know, show benefit. So I think,

28:35
28:58
for example, the molecularly defined groups, Probably those should be in different trials alone. Like, if you have MDS with SF3B1. Uh, you know, if you have MDS with del(5q) or biallelic p53 I think those should be in separate trials. Um uh, and in terms of how to use this in practice, that’s also like a dilemma. So I think

28:58
29:20
still, obviously the prognostication using tools like the IPSS, IPSS-R, IPSS-M are going to be important on deciding, you know, the only curative option for patient’s, which is allogeneic stem cell transplant. So if patients have high risk of disease or in another word that you know, a mortality expected from the disease in the coming

29:20
29:35
a couple of years, then you’re gonna think of transplant upfront for those patients’ Now, you know, in terms of management, like I think the cut off of the blast is a little bit tricky, and in a way it could be helpful

29:36
29:55
that you start, including those patients in AML trials. You start having access of drugs approved for AML in somebody that has, like, 19% blasts that often the insurance will deny it because of not meeting the AML well, probably if you count twice, you could get to 21 and 19. Uh, so I

29:55
30:14
so it can have that. But it also will create the anxiety for some patients because, like obviously AML always have this, you know, more serious disease. MDS sometimes it’s not even presented to patient’s as a type of a blood cancer, so it creates some dilemma. But it may allow some access. And how to include those on trials.

30:14
30:41
I think it’s it’s very hard. I try to simplify things in my mind. I try to think of like I think the MPN people have it nicely done, like you have chronic MPN. So you have, like almost chronic MDS. The blasts are not increased. You have an accelerated MDS. I don’t know what exactly the cut off of the blast are, but let’s say 5 to 19 and then you have a blast phase MDS that basically a secondary AML or or gone into AML

30:41
31:06
So I think that’s how I think of it. And it’s definitely gonna evolve. You alluded a little bit also to the difficulty, and we just had those discussions even internally. Like, how is the hematopathologist gonna sign this without access to the NGS. Most of the places in the country don’t get access or information from the NGS before 10 days, 14 days, the best case scenario, and our hemepath people sign, you know,

31:06
31:32
40 50 reports of those per day. If they are going to come back and amend the information when the data comes, it’s impossible to do. So we’re having having those discussions to your point. It’s just like a practical point. And you have to think of this in the community practice how this is going to happen. So we have to think of also some of those practicality, like you need to get the diagnosis out, you know, and the physicians need to have,

31:32
31:57
uh and then act on it. So I I think in summary, it’s gonna evolve. I think definitely we need a unified criteria for, you know, classification classification should be based on biology and identifying groups. I think in clinical trials we start separating some of those clearly very unique groups out and then in in, you know, practice. I I think you know the line between

31:57
32:22
MDS and AML and access to drugs we should think of. The last thing is we didn’t talk much about is really introducing the three MDS conditions, which I think is important because those are existing. It’s challenging. So, like, clinical cytopenia of unknown significance that our patients that need treatment recognizing the CHIP and starting to think of prevention strategies for the patients that are higher risk are all important aspects.

32:23
32:48
Yeah, Thank you so much Rami. This is very comprehensive. Um, like, just thinking about this again from a practical point of view. Sanam, as people try to figure out in the community. So I have been talking to colleagues. Um, I have heard three general directions. Some people are talking about just adopting one of the two systems in their institution and putting the reports according to the WHO, the ICC,

32:48
33:17
I had other people talking about having the two different systems on the same report saying that this is what the patient has according to the WHO. And this is what they have according to the ICC. And then the last group, which, uh, Rami also mentioned a little bit before the discussion is just putting it descriptively, you know, excess blasts whatever. And then the physician will have to kind of integrate the molecular data and come up with the diagnosis. Clearly, I think in a community setting, I think this is a

33:17
33:40
a challenging task to ask, uh, you know, kind of positions to to do this integration. And I wonder, I’m sure these discussions are you know, you guys are having, um, as hematopathologists, and I wonder what’s has been like, like, official guidance from the societies. And how do you How do you envision this? Or is this gonna be up to each individual institution to decide?

33:40
34:08
Yeah. I mean, I think it’s this is a very difficult question, right? So I’ll give you my personal philosophy, first. And then I’ll tell you what You know what the guidance is. You know, I think the more I think about this and this may actually apply more to acute myeloid leukemia than it does to myelodysplastic syndromes. But I think you know, my morphologic diagnosis really matters less and less as we learn.

34:08
34:36
more about the genetics. In fact, the clinicians at my institution won’t enroll a patient on the clinical trial without knowing the genetics, right? So they don’t care what I say morphologically. If I call this AML with, you know, however many percent blasts, if it has a TP53 mutation, that’s what they want to know. They want to know if it has an NPM1 mutation. They want to know if it’s a core binding factor leukemia. Really that’s what determines the therapy. So I think I don’t, uh,

34:36
35:04
um really feel the need, uh, to, like shove my diagnosis into one box and then be proven wrong by the by the genetics later. So I really like, you know, the approach that Rami was also, um, I guess advocating for is that I think and I really want to see our classifications go toward this saying that this is a myeloid neoplasm, this is the number of blasts, and these are the genetics right

35:04
35:31
now, in terms of I think, the generating comprehensive reports, I really do think that the, um that burden should be on the hematopathologists. I don’t think that the treating physicians should be looking for, um, you know, the different parts of the ancillary studies and trying to piece meal things together, particularly because, you know, we think about our centers where everything is done at one place. But a lot of the times it will

35:31
35:59
won’t be, you know, done at one place. And it may, you know, your lab, maybe sourcing out it’s NGS to some other place, sourcing out the FISH to some other place. So I really think there should be a comprehensive report that integrates everything together and gives the treating physician the information that they need to treat this patient appropriately. That being said, I also agree that, you know, with the way we work right now and with the number of cases that I’m looking at, every

35:59
36:27
it’s physically impossible for me to go back and make an addendum or an amendment report on everything to do this. But I do think that you know, we have tools and we have resources to do this. Why not use? And, you know, I don’t know how to do this, but I’m sure that there are engineers that can come and do this, you know, write a code and do this an epic where everything gets, you know, together. And it’s, you know, one comprehensive report is generated. So why are we not leveraging AI

36:27
36:54
I to do something as easy as this for us to essentially just get everything and, you know, generate a comprehensive report. I suspect that it’s probably not that hard, but we need to put the resources and, you know, talk to the right administrators in our hospitals to help us do this. Um, and in fact, I know some places actually have full time employees that do this that are essentially, like putting out together, putting, you know, together, all of these reports, and then a physician is signing off on it.

36:55
37:16
No, that’s I I think clearly that the direction in the future is to try to, um, do more of this machine learning and artificial intelligence. And probably I would not be surprised if some people already looking at some of those, uh models, and we’re probably going to see some of them in the next in the next few years. So in the last few minutes we have

37:17
37:35
remaining. I want to see if any of you have any, um, departing thoughts on on my end, kind of to try to finish on a positive note is that I I think both groups have heard it I think very clearly from from the community that everybody wants a harmonized system and I think they are ongoing efforts to try

37:35
37:55
to do that. I I do think the initiative that, uh, Doctor Ball outlined and hopefully other multi center validations could provide some evidence, uh, kind of on how to do this in a way that that makes more sense. And I do think it’s important that we don’t end up in a situation where we have,

37:55
38:23
you know, different, we have already different therapies between kind of the developed world and, uh, countries with limited resources. And, uh, I would be concerned to kind of take that another step with, uh, you know, with the even the classifications where we have classifications that only apply to countries where you can get next gen sequencing or get them within 2-3 days. You know, uh, compared to to the rest of the world because I think that’s going to limit our ability to understand.

38:23
38:41
Uh, you know, the differences between the epidemiology, the impact of different treatments, etcetera. So I do think morphology hopefully will continue to be a big part of how we define things. But maybe I can go around, uh, start with you Rami for any departing final thoughts.

38:43
39:06
I think I agree. Like, I think you summed it nicely. And, you know, I think so now. And like Som provided good, you know, highlights of those I I think the message again, like we need uniform classification. And in reality, also at the end, we need better treatments for our patients. So I think when we get treatments that are effective, sometimes they will change how we

39:06
39:20
prognosticate things. How do we classify things? Uh, but I think the most important is really to not lump all MDS as one disease. This is a spectrum clearly. And think of designing clinical trial and is in that way.

39:22
39:43
Thank you. Sanam I think your presentation was like the start of a nice effort. And I know there are multiple ongoing analyses that are happening, hopefully to kind of inform those, uh, discussion’s that we just mentioned. Maybe we wanna in your departing thoughts tell us a little bit about the subsequent analyses that are being hopefully planned out of this effort.

39:43
40:08
Yeah, sure. So we we, uh, you know, formed a collaboration with an Italian group to try and see, and they have a large number of patients. So we are kind of combining our databases, uh, to try and answer some of those questions related to subtle differences that Doctor Loghavi reviewed between the two classifications, try to see if we really need those complexities. Uh, can we get rid of

40:08
40:34
unnecessary complexities, which would hopefully serve the purpose that you outlined you and Dr Komrokji that you know, more resource limited settings or even quicker assessment of quicker classification? Uh, you know, do we really need to do those additional co-mutations? SF3B1 mutation – how much does that matter? So some some questions of that kind

40:34
41:00
we’re trying to answer. And also within the TP53 mutated group, at least clearly in our data, said we saw more than 5%. It did not really matter much whether you have one TP53 versus two. Is that true in a multi center dataset or not? I think that’s an important thing because some of those additional informations are not readily available in some centers. So definitely looking forward to those data. And also as a general, I think you know,

41:00
41:27
you mentioned it at the beginning. Our our biggest responsibility are liabilities to our patients. So I think patient education resources across platforms used to be updated to reflect some of these changes happening in the field as well. Because a lot of the patients that we treat, they read up and they come and, uh, you know, uh, in clinic asking questions – so that they can get their questions answered when they’re meeting their clinician in the clinic as well.

41:28
41:31
Yeah, thank you. That’s very well said. Sanam, any final thoughts?

41:32
41:53
No, I mean same. Thank you all for your for your thoughts. I think this has been a greatly helpful session for me as well. I do want to emphasize one point for maybe the hematopathologists listening to this and the trainees. I think, you know, even though our classification systems are becoming more and more complex and sophisticated and you know you need all of these sophisticated tools, I

41:53
42:13
I think it’s really important to still recognize that, you know, again going back to the FAB comment that I made in the beginning. A lot of these genetic abnormalities we can actually recognize morphologically. And there are surrogate markers that we can use to recognize some of these entities, including IHC stains, including the morphology of the blast,

42:13
42:34
including the presence of fibrosis. So I really want to emphasize and say, you know, don’t discount the morphology. Don’t discount the traditional teaching because there’s still a lot of value in that. And I think we can really contribute to patient care as good morphology just as good pathologists. And not only leave everything to sequencing Um, so that would be my message for everyone?

42:34
43:00
No. And I fully agree with that and actually kind of just supporting what you just mentioned. I I kept highlighting this in my presentations on the validation of the molecular IPSS as you know, when they compared to the revised IPSS. And the C index went from 0.742 point 81, which highlights what you and Rami mentioned is that you know the morphologic data. The you know, the cytogenetics. Uh, the blood counts the blasts.

43:00
43:29
And all of these, I think convey very important information. And while molecular integration certainly can improve or does improve the prognosis and, uh, classification schema, I still think the bulk of the data is already there. So I think people should feel empowered to act a lot based on on what we have. So thank you so much. This was a fantastic discussion, and I look forward to having you on some subsequent editions of, uh, VJHemOnc MDS sessions. Thank you so much.

43:29
43:30
Thank you for having

EDIT
us.


The MDS Channel on VJHemOnc is supported by

These supporters have no influence over the production of the content.