MDS: When the Prognostic Models Fall Short

Transcript:Jamile Shammo, MD: The other unfortunate part is that even though the prognostic systems are a good way to start, like in terms of risk assessment, etc, we really don’t have a good prognostic system that’s dynamic, in essence, that you can utilize there. And also, I can’t overemphasize the point about making the correct diagnosis, because in the days of next-generation sequencing, a lot of times you don’t have a good diagnosis for MDS, but you have the next-generation sequencing data. Then, what do you do? I just want to emphasize that in the absence of dysplasia, the presence of any kind of a mutational abnormality does not constitute a diagnosis of MDS. That’s very important.

Thomas Prebet, MD, PhD: We’re still in the process of defining what we need to do for these patients with mutations without any real signs of MDS, and probably most of them will need to have some real follow-up that needs to be repeated.

Rami S. Komrokji, MD: Also to add to the point of dynamic models, which we don’t incorporate and maybe Thomas can comment on that, is the effect of prior therapies. So, if somebody had ESA (erythropoietic stimulating agent) failure, if somebody had HMA (hypomethylating agent) failure, those are not incorporated in the models, and I know that Thomas had some experience with that.

Thomas Prebet, MD, PhD: I think it’s a really good point. We know that the IPSS, for example, has been developed on untreated patients at diagnosis, same thing for the revised IPSS. It has been developed with a cohort of untreated patients at diagnosis. We know that several groups now have demonstrated that the revised IPSS can be used for treatment with a hypomethylating agent or immunomodulatory drugs. But, you’re completely right, we also need to integrate the prior therapies, the need for transfusion, in the model. And that’s something that will lead us to have more complex models, new risk system to really assess the current risk of the patient. There’s clearly some indication that having been failed by treatments, such as hypomethylating agents, lenalidomide, or ESA, are clearly increasing the risk of a bad outcome in the long-term. We can discuss if the importance of this risk increase is the same for lower-risk myelodysplastic syndrome where survival is pretty prolonged, as in high-risk patients. But, basically, the bottom line is we know that failing therapies will have an impact on the outcome.

Vinod Pullarkat, MD: I think that’s an important point, that there are certain patients who are low-risk by IPSS or IPSS-R (Revised International Prognostic Scoring System) who may have adverse factors, for example, patient-related factors like frailty and genomics, which you touched upon. They may have an adverse genomic feature, which may make them higher-risk than what’s captured by the IPSSR. What about transfusion dependence?

Rami S. Komrokji, MD: Right. I think we know that transfusion dependency is not a good player in MDS. If you look at all MDS patients, I think 90% or more, they will have anemia, and maybe 60% to 70% at one point will become transfusion dependent. This had been looked at, published in New England Journal of Medicine, by the Italian group that transfusion dependent patients do worse. Actually, there is correlation with the degree of transfusion dependence almost by unit. It could reflect partly the biology of the disease, so who is transfusion dependent has more ineffective hematopoiesis. But, it also could be telling you some of the complications of the transfusion itself like iron overload, volume overload, and many other things.

And that brings us to other things like, as you mentioned, those models that we are talking about that don’t capture the whole story. You have the patient-related factors. You have the comorbidity-related factors. You have things that are modifying. We looked, in our MDS Consortium, at the risk models, if one is better than the other. The bottom line is that those models will upstage 25%, as I mentioned. But, if you look at what we call lower-risk MDS, unfortunately, around 25% of the patients still will die within 2 years. And even those new models capture only 50% of those patients, so there are other things we start looking at, as you mentioned, like blood transfusions. The NCCN (National Comprehensive Cancer Network) guidelines and revised IPSS suggest looking at things such as age-modified models, the incorporation of factors like the LDH (lactate dehydrogenase) if the disease is proliferative, and maybe characteristics of MDS MPN. Fibrosis is a very important factor, as are serum ferritin levels. There is some good correlation there. You start putting in all those factors and that could be sometimes now important because we have a category in the IPSS revised called “intermediate,” that sometimes we are not sure are we going to treat those as higher-risk or lower-risk. So, if you start seeing that those patients are more transfusion-dependent, have fibrosis, they are younger age, I think we’re all automatically thinking of this patient as a higher-risk. It’s a very important point.

Vinod Pullarkat, MD: What about therapy related, do you think they’re higher-risk?

Rami S. Komrokji, MD: Right. So, that’s a very good question as well. We are seeing more and more therapy-related MDS. They are like 10% or 20%. At this meeting, we heard about the challenge and whether to call that therapy-related MDS or not. There is also now interesting data coming that if patients have certain clonal hematopoietic events, like mutations detected at the time of solid tumors, and they get chemotherapy, they’re at a higher risk of developing therapy-related MDS. The general principle is that therapy-related MDS patients don’t do as well, and we looked at that, and that’s true. So, if you look at a MDS de novo versus therapy-related stage-by-stage, they don’t do as well. However, the caution I always say, I was trained originally that any therapy-related MDS is a bad disease. You start thinking of transplant. But, over time, when you look at those patients, what drives the outcome in those patients are really the cytogenetics and mutations. So, you have a small subset of patients that will have no bad features at the beginning, and I think those are okay to go stepwise and managing them. But, a majority of the patients actually will have the bad cytogenetics that we are used to seeing nowadays, the p53 mutation detected. But, in general, I definitely think we should step back and always think of therapy-related MDS as a separate entity.

Jamile Shammo, MD: Obviously, everybody is aware that IPSS and IPSS revised excluded people who had therapy-related disease, so we don’t have a way, with the exception of the MD Anderson score, that would include secondary MDS or therapy-related.

Thomas Prebet, MD, PhD: And I think to get back on the question of patient-related factors, we also have these 2 really challenging situations where have some patients with a new diagnosis of therapy-related MDS several years after their treatment, for example, of breast cancer. You also have these patients now with lung cancer sarcomas that benefit from the advances in treatment of this disease, but that begin developing therapy-related MDS at the same time for potentially ongoing disease. And that’s another challenge we have to face now.

Rami S. Komrokji, MD: Absolutely.

Ellen K. Ritchie, MD: Well, all of those patients, we can’t put on clinical trials that have co-existing cancers of one kind or another. And as the population ages, I think we’re going to see more of that.

Transcript Edited for Clarity