Myelodysplastic Syndromes: Translating Genetics to Clinical Practice - Episode 4

MDS Risk Stratification and Treatment

Transcript:Mikkael A. Sekeres, MD, MS: Let’s assume that we have a patient, and we’ve made the diagnosis of myelodysplastic syndrome. Rami, I wonder if you could talk a little bit about the Revised International Prognostic Scoring System (IPSS-R) and how you use that in determining prognosis and treatment options.

Rami S. Komrokji, MD: As you mentioned, once we establish the diagnosis, the next step is what I call risk stratification or staging, which is something we apply to all the hematology or medical oncology diseases. The risk stratification provides two things: the information for the patient and the family about the prognosis, and the outcome of the patients. But then, what we’ve used it mostly for in MDS is really to tailor the therapy accordingly because we estimate the risk of the disease to justify a risk of a procedure, such as allogeneic stem cell transplant. So, historically, the IPSS had been the gold standard, which depends on the presence of cytopenia, the blast percentage, and the cytogenetics. And it served pretty well. It can identify patients that are higher risk, but it has some shortages, like addressing lower-risk patients that they could have a little bit of higher-risk features. As we learn more about this, there had been a proposal for the revision, the IPSS-R, which allows a little bit more weight on the depth of the cytopenia. If a patient has platelets of 10, we know clinically they didn’t do as well as patients that have platelets of 90; they were treated equally in the IPSS.

Now, the revised IPSS gives some weight. I think one of the more important changes, also, is we learn more about the depth of the cytogenetics in those diseases. So, the IPSS-R has a little bit more detailed cytogenetic groups. We learned, for example, something like monosomy 7 is worse than deletion 7. We learned about some entities that are rare to occur, but may have favorable prognosis. Then several groups had already validated the IPSS-R. I think that it’s really a better tool at determining prognosis. It now puts patients in one out of five categories, rather than four. So, there’s low or very low, low, intermediate, high, and very high. The question always comes to what to do with this intermediate group. I think there are other things you weigh on, on how to decide on treating that group. But there is no doubt that I think it refines the prognosis for the IPSS-R.

There are several other models that are there. The WPSS (WHO classification-based prognostic scoring system)—which is mostly used in Europe—has not been used much in the United States because it weighs heavy on the WHO classification and giving points on the system, and many times that information is lacking. I think with all the other things, it doesn’t account for many items of the disease. There are the MD Anderson models, which actually are pretty predictive for outcome. As you well know, we looked at the MDS Consortium, at those models, and many of those newer models definitely refine the prognosis value of the IPSS. And it’s interesting, because if we look at what we think is lower-risk disease, still around 25% of those patients will unfortunately die from the disease or its complications within 2 years. It’s not always AML transformation, but also it could be disease-related complications. So, definitely those models can identify those patients that we always thought are at lower risk could have higher-risk features. The question now, how do we translate that to treatment? It’s probably not yet clear when to proceed on to a procedure like allogeneic stem cell transplant if you upstage those patients by the new models.

Mikkael A. Sekeres, MD, MS: So, is the IPSS-R valid in patients who are about to be treated? Because it was developed based on over 8000 patients internationally, but none of those patients received any therapy. Does it work? If I have a patient who’s coming in and I’m not planning any type of therapy and I’m just going to follow that person with watchful waiting, then that’s probably a valid system for prognosis. But what about if I have a patient who comes in and I’m about to start the patient on azacitidine?

Rami S. Komrokji, MD: I think that the original publication and the data set that was looked at were mostly patients untreated. When we looked at it—at least in our experience, whether we validated the IPSS-R—we did it at time of referral, and it still was predictive for the outcome for those patients. The issue of whether they’re dynamic or not is always something that comes up. But I think they do. And at least in our data set, the IPSS-R at time of referral was still predictive for the outcome for those patients. I don’t know the impact of the treatments on changing the prognosis for those patients as well as being just a prognostic marker, rather than predictive markers.

Jamile Shammo, MD: It’s only fair, though, to compare all those patients that had not gotten therapy—to your point—so that way you can estimate survival of leukemia evolution. But when you think about rate of progression, that’s never been addressed in any of the scoring systems, which is why I think it’s probably slightly overemphasized. To me, scoring systems are a great starting point. So, choose one risk score and move on. And, as far as the rate of progression is concerned, I don’t think we know.

Mikkael A. Sekeres, MD, MS: It’s a tough conundrum because Rami referred to these systems not being dynamic, meaning we’re not sure how well they change in accuracy as a patient’s disease evolves. They seem to be pretty good at time point zero. They probably are not as good at time point one-and-a-half years. And we did do a study where we looked at patients who had been treated with hypomethylating agents. Those agents had failed those patients, and then we tried to apply the IPSS or other prognostic scoring systems to those patients at that time. And, in fact, they were not at all predictive of outcome. So, we probably shouldn’t be applying these after patients have been exposed to therapies for long periods of time, and we certainly shouldn’t be using them for clinical trial enrollment at that time.

Jamile Shammo, MD: I agree.

Ellen K. Ritchie, MD: I agree.

Transcript Edited for Clarity