Updates in Myelodysplastic Syndrome - Episode 12

Phase III MEDALIST Trial of Luspatercept for MDS


Mikkael Sekeres, MD, MS: There is a drug that’s been very heavily studied, and we’re eagerly awaiting news at the time of this recording about whether it will be FDA approved in MDS [myelodysplastic syndrome], and that’s luspatercept. Rami, you were instrumental in the development of this class of drugs here in the United States. Do you want to comment on the MEDALIST study and its results, and where you feel like it fits into the treatments we’ve discussed?

Rami Komrokji, MD: Sure, absolutely. As you mentioned, there is excitement that hopefully we’ll have the next drug in MDS approved for our patients. We didn’t have any in the last 10 years, since the approval of LEN [lenalidomide], probably. It’s hopefully something that’s going to be helpful for the patients. You were part of the studies as well. Those agents, luspatercept, they are termed now as erythroid maturation agents. They work on terminal erythroid differentiation. They are trap fusion proteins that will neutralize the ligands of the TGF [transforming growth factor]-beta pathway that ended up being important in MDS being overactivated, contributing to cytopenia. Some of those ligands, mainly GDF11, seem to be negative regulators of terminal erythroid differentiation. So those drugs will neutralize that.

Originally, they were tested in healthy volunteers. The toxicity was erythrocytosis, which is very welcomed in MDS. Our colleagues in Germany conducted the original phase I/phase II study. In those, there was a signal that the responses may be higher in patients with lower-risk MDS and the ring sideroblast subtype. The MEDALIST study was designed to include that population. Those were lower-risk patients with MDS who had ring sideroblast that were transfusion-dependent. They randomized in a 2:1 fashion to get either luspatercept, which is a subcutaneous injection every 3 weeks, versus placebo.

The primary end point was transfusion independence at 24 weeks, but they also looked at 48 weeks. The results were positive. The transfusion-independence rate was in the range of 30%—around 36%, 37% with the luspatercept—versus 13% with the placebo. This goes back to the point that you mentioned earlier, that sometimes our thresholds for transfusion change once we put patients on trials, or there could be other reasons for bleeding. But nonetheless, the study was positive in terms of transfusion independency, and at 48 weeks it was the same.

Interestingly, when you look at the subset of patients who were not heavily transfusion dependent, less than 4 units every 8 weeks—or those patients where we looked at objective response of hemoglobin increase of 1-1/2 gram per dL or more—more than 50%, originally 69%, of the patients in that population achieved a hemoglobin increase of more than 1.5 g/dL. For me, that is always an objective sign that it’s not just transfusion reduction and that we are seeing some effect on the hematopoiesis.

The treatment is relatively well tolerated with fatigue and some general adverse effects, but really no major adverse effects with it. At least so far, no signal of increased progression of MDS to high-risk MDS, AML [acute myeloid leukemia], or fibrosis on the bone marrow. This year I believe we are going to hear the update on the results, what’s available in the abstract data that basically, when they looked at all through the study, the responses were higher obviously, almost 48% of the patients achieved transfusion independence with luspatercept.

They use this principle of looking at different time periods of 8 weeks or more because of some of the shortcomings of the response criteria, that you record the longest duration of transfusion dependence only 1 time. But there are patients who will go 12 weeks and then just get 1 blood transfusion, and then go another 12 weeks not needing transfusion. But the criteria do not recognize that. So, they try to look at this, and when you look at the lump of those, there are patients who went more than almost 90 weeks in periods where they had transfusion independency.

I think the data are exciting. Again, the indication or the approval will be in ring sideroblast lower-risk MDS. Down the road there could be implications for testing this outside that setting. In the original studies, any patients with splicing mutations were responding. There could be responses in non-splicing mutations. There are efforts obviously doing it in combination. I know about stuff down the road, but hopefully it will be a new option for our patients.

Mikkael Sekeres, MD, MS: I hope so. I don’t know if you’ve had experience using the drug, Ellen or Jamile.

Ellen K. Ritchie, MD: No, I haven’t.

Jamile M. Shammo, MD: Not yet.

Mikkael Sekeres, MD, MS: It’s a shot. It’s every 3 weeks. It was in patients who either stopped responding or didn’t respond to an ESA [erythropoiesis-stimulating agent] or didn’t have a good chance of responding to an ESA, who were transfusion-dependent. Personally, I would envision this as starting an erythropoiesis-simulating agent like erythropoietin or darbepoetin alfa first, and then when that stops working—it always stops working—or if it doesn’t work, then trying luspatercept. The transfusion-independence response rate of 37% is generally better than that we see for other treatments for MDS. In my personal experience, the adverse effects are manageable, and then moving to other therapies.

Jamile M. Shammo, MD: That’s going to be a very good option for patients who may be ineligible, intolerant, or even refractory to erythropoiesis-stimulating agents because this was the patient population in which the drug was studied. On a different note, such as with some of the data that were presented relative to mutational profiling and what is the difference between those who had responded and didn’t respond, I found it very interesting with the Uwe Platzbecker, MD, abstract that presented at ASH [the American Society of Hematology 2019 annual meeting]. This is because, obviously, most of those patients had ring sideroblasts. Over 90% had a SF3B1 mutation, and the variance allele frequency was rather high, about 40%. So, what was the difference between them? It seems as though the mutational profile didn’t really matter between those who respond and those who didn’t. Again, it’s very interesting to see that responses will occur irrespective of what you might find on NGS [next-generation sequencing]. Again, perhaps long-term data would be a little bit more helpful in that sense, but it’s comforting to know.

Ellen K. Ritchie, MD: Or maybe just working on a completely different pathway than what we’re looking for in NGS, may be what’s going on there. When I first started doing next-generation sequencing it was 7 genes, then it went to 14 genes, then it went to 20-some genes. Then it went to 44 genes, and now our institutional panel is 55 genes. So, we’re not covering everything that we may need to know, and I’m expecting that 100-gene panel to be the next myeloid panel as we incorporate more mutations that might be involved in these pathways.

Mikkael Sekeres, MD, MS: Which gives opportunities for therapy.

Transcript Edited for Clarity