IDH Inhibitors for AML

Transcript:Harry Erba, MD, PhD: OK, let’s move on to another topic, and that is targeting IDH1 and IDH2, and targeting CD33 with antibody drug conjugates and BiTEs [bispecific T-cell engagers]. Let’s start with IDH1 and IDH2. Raajit?

Raajit K. Rampal, MD, PhD: I think we’ve seen 2 important approvals in the last 2 years with the approval of the IDH1 and the IDH2 inhibitors for relapsed/refractory AML [acute myeloid leukemia] patients. Now, how frequent are the IDH mutations? They’re certainly not amongst the most common. They are really about 10% to 20% depending on whose series you’re reading. But, I think the important aspect is that these are oral agents that are very well tolerated and have a, I think, good clinical benefit in most patients. If we are going strictly by the CR rate, it’s 18% to 20%. It’s not in and of itself a very high complete remission rate. But the amount of patients who had some clinical benefit ranged in the 40% range. There were patients who became non-transfusion-dependent, whose platelet counts went up, sort of getting to this Bahamas idea for IDH inhibitors.

So that, I think, has been the clinical benefit. I think as we’ve been talking about for a number of these other agents, you can give a patient a pill. They can have a good quality of life. The majority of the patients will not obtain a remission. Does that mean they are not deriving some clinical benefit? Absolutely not. However, I think that we see the same sort of things that we see with the FLT3 inhibitors, where there is a rationale to combine this with other agents, to move that response right up. But I think those are things that are ongoing, and that is both in the realm of conventional cytotoxic chemotherapy, which we’ll talk about in a moment, with low-dose chemotherapies like HMAs [hypomethylating agents], and also with venetoclax. Although it is interesting that the venetoclax data do have a higher, or a very good response rate in the IDH-mutant patients.

In terms of toxicities, again, both of them are very well tolerated drugs, but there is this differentiation syndrome. It does work by differentiating cells and that’s now been well established in some preclinical studies. This is certainly not a highly prevalent toxicity, but it is a toxicity that we have to worry about.

Harry Erba, MD, PhD: Well, someone might disagree about the prevalence and incidence of the differentiation syndrome. There was an important abstract presented from the FDA where they looked at the case report forms from the enasidenib and ivosidenib phase I studies and felt that they could identify more cases of differentiation syndrome. It’s clearly something we have to make sure that clinicians are aware of when they’re treating patients with this drug either in combination or alone.

Now, some of those combination studies actually showed lower incidence of differentiation syndrome, which is interesting. But in terms of the syndrome itself, the symptoms are similar to APL [acute promyelocytic leukemia] differentiation syndrome—fever, fluid retention, weight gain, hypoxia, pleural effusions, pericardial effusions, pulmonary infiltrates, all of the hypotension, organ failure from capillary leak, and renal failure. And then, often with leukocytosis. There’s a major difference though between APL and this. In APL you support a patient through that. They will go into remission. You know it is not progressive leukemia. Here, you’re using an agent that has a 40% disease control rate, including morphologic leukemia-free state. Most of the patients aren’t going to respond, and then you have a patient coming in on this drug who has fever, pulmonary infiltrates, hypoxia, rising white count. Is that disease progression or not?

So to the credit of the investigators, it was the investigators who recognized differentiation syndrome. It was not in the original protocols. So they were being quite vigilant, and they’ve actually published on this afterwards. Amir Fathi, MD, has published in JAMA Oncology on this.

They identified about 11% of patients with this syndrome and noted that it needs to be managed with early institution of dexamethasone, very close monitoring, and hydroxyurea for the high white count. And withdrawal of the drug, maybe we’ll do it, but it has a long half-life and so maybe it is not as important.

The FDA looked at part of the case and doubled the number that were found, 19%, based on the symptoms. The trouble I have with these analyses is even if you look at the differentiation syndrome identified by the clinicians, the number of patients who went into remission was about half of the number. Not all did. There might be a biologic reason for that. Maybe we’re differentiating the IDH clone and allowing that FLT3- or RAS-driven clone to just generate a relapse. That’s a possibility. On the other hand, it might just be that we can’t really tell the difference. And in fact, when the FDA identified more cases, the number of responses went down by half. They doubled the number of cases, but they lessened the number of responses down to like 20% or 25%.

So I want to give credit to the investigators for identifying this. It is in the black box warning and we have to take it seriously, and maybe we’ll be able to decrease the incidence of it and the severity by combining with other agents like we did with APL.

Raajit K. Rampal, MD, PhD: The other major caveat, just to add to your point, that I think is important for clinicians to know, is that unlike with APL, we can see this differentiation event occurring months after the initiation of the drugs, sometimes up to 6 months. So most people have to be very vigilant about that when treating these patients.

Harry Erba, MD, PhD: So here’s the tough question. You know we’re running out of time but that’s only because we’ve never had this much to talk about before.

Alexander Perl, MD: It’s a good problem to have.

Harry Erba, MD, PhD: I think the cameras have stopped rolling actually, but we’re going to keep on talking. Enasidenib, ivosidenib—both of them have been used as single agents in the upfront setting. There was data from the Beat AML Master Trial. There was the data from the sponsor as a subset of the phase 1 studies, showing maybe double the response rate that you would see in the relapsed/refractory setting. And of course we know about data in combination with HMA. So I really want to get to the question for you. You have an older, unfit patient, however you define that, for whom you’ve decided you’re not going to give induction chemotherapy to, and they have IDH-mutant AML. HMA/venetoclax or HMA/IDH inhibitor. How do you decide? I’m going to start with Raajit.

Raajit K. Rampal, MD, PhD: I think that; well in the absence….

Harry Erba, MD, PhD: Because that’s what they want to know out there.

Raajit K. Rampal, MD, PhD: Right, and I think that’s a very difficult question. I think there are practical concerns in terms of cytopenias. Are you more likely, just from clinical experience, to get more cytopenias if you’re going to go down the venetoclax route? Yes, I think that is a potential concern. We really don’t see much in the way of cytopenias with the IDH inhibitors as single agents. And there was really not a profound safety signal seen with the combinations with low-intensity therapy.

So I think that if you are dealing with a patient who is profoundly cytopenic and perhaps isn’t going to be able to get or come in for transfusions regularly, maybe the venetoclax regimen isn’t the right way to go. I think that is one way to think about making this decision on a pragmatic level, based on the known toxicities of the drugs.

Harry Erba, MD, PhD: Naval?

Naval G. Daver, MD: I think it’s going to be tough. It’s a timing issue, right? And I think the AZA/VEN [azacitidine/venetoclax] data are more mature. There’s more of a comfort level, and 75%, or whatever, 70% is going to be hard to beat.

Actually, I think the AZA/IDH data are quite encouraging. It’s around the same, 65%, 70%, but, of course we don’t have the survival, the durability, and the other factors. So I think in that situation right now we’re still leaning towards the HMA/VEN. Now the question would be, is there a way to combine these down the line? And that’s where the research is….

Mark J. Levis, MD, PhD: Which will break our bank.

Naval G. Daver, MD: Well, that’s a whole new discussion.

Harry Erba, MD, PhD: But this is a situation where the labeled indications do help me, because I don’t know what’s the right way. So the combination of HMA/low-DAC [decitabine] with VEN is in the first-line setting. What I’d really love to see data on, and it’s probably going to come from your institutions as you collect more patients, is AZA. When you start VEN and then give these, and vice a versa, what are the response rates?

Transcript edited for clarity.