iwAL 2019 | The future treatment of acute leukemias: part II

Alan Burnett: So let’s go onto this issue of MRD, which you mentioned Andrew, it’s got attractions. I think because it tells you about individual patients. Now there are two approaches. The NGS or flow, they both have challenges. What do you think the strengths and weaknesses of either technique is?

Andrew Wei: The flow cytometry is more easy to, I guess sit up in laboratories, especially without access to high tech molecular services. It’s a kind of a generic morphologic type approach to assessing the presence of blasts. However, I think it doesn’t give us the granular information that I think I find useful when using gene specific probes, particularly like NPM1, because the level of confidence and the sensitivity of the technique is extremely high, and very reproducible. It’s something which could be standardized between different laboratories, whereas we’ve continued to struggle having flow cytometry used as a general purpose MRD marker. It tends to be restricted to centers of excellence.

Alan Burnett: Yes, I think also, you’ve got, even in the MRD negative, you’ve still got a fair number of relapses going on, and it may be you just missed a sample that’s relapsing. But in addition an MRD positive, they’re not all failing either. So it’s not clear why that’s happening, but it certainly opens up the challenge of assessing treatments prospectively in the context of whether a patient was positive or negative at a certain time. What’s the flow experience in the US Mark? My impression is they’re going more towards NGS than flow, generally speaking.

Mark Levis: I think that’s clearly going to be the case. We’re busy running around talking about digital droplet PCR to identify specific mutations in cell free DNA and so forth. The flow, we’re quite comfortable with it for ALL but not for AML. Again, I’m afraid the concept of the brilliant artist localized at one or two institutions remains a problem, and so widespread use is going to have to come through usually a nucleic acid.

Alan Burnett: Yes, I think there is observer experience involved in flow, that’s much more obvious in AML context then in ALL context. What about clonal hematopoiesis backgrounds though as far as NGS is concerned? How big an issue and how can you deal with that?

Mark Levis: Oh, it’s huge. Now that we’re actually, without realizing it, doing these big panels on random people, we’re finding all manner of things and one issue that wasn’t talked about here was familial AML. It’s a lot more common than we realize. Yes, I’ve got a family history of cancer, not too much. I’ve got an aunt and a brother. Okay, you’ve got a mutation in ATM, you’ve got a mutation in DDX41 that’s actually your problem. And that is much more common than we realize. So we’re finding … And this speaks to how well we’ve advanced in our understanding of how the disease comes about. You know, this slow process over the years. Oh no, it starts at age 70 and no, no, no, it starts at age 50. It pretty much starts at birth, starts when your stem cells start dividing. The longer you’re on the planet, the more chance you have to make a mistake.

Alan Burnett: So the amount new drugs, you’ve got IDH1, IDH2 inhibitors. 10 to 12% of younger patients, half of whom with conventional therapy don’t appear to relapse. So it’s a major challenge to show that impact. So my impression is there’s a fair body of opinion that wants to jump to combos with them, or combine them with standard chemo, that’s fine. We may find an answer in five or six years as to whether that’s a good idea. If we’re lucky. It seems to me Mark, about the new FLT3 inhibitors in the next year is going to present some approvals, maybe couple of approvals.

Mark Levis: Sure. Probably in a couple of approvals, you know at least one more approval. But then we’re not going to really know how to use … but we’re already positioned, the trials are already rapidly accruing that will tell us how to use the darn things. We don’t get a readout on those for two or three years, but we’re going to get that. It’s coming fairly quickly.

Alan Burnett: Okay, but you’re going to end up, if you add quizartinib to your chemo it’s beneficial, tick. But that is a standard of care currently out there. Does your control arm have to be midostaurin, do you think?

Mark Levis: Well, but that trial is going to be accruing quickly also. I mean, we’re not going to get a readout on quantum first for, you know, probably still a couple of years. You watch, the Europeans are going to put their … get the machine going and you’ve got this gilteritinib versus midostaurin trial, roaring, roaring ahead. I have no doubt it will accrue reasonably quickly, and they’ll use event free survival as an endpoint. So, we’re going to get that readout reasonably fast. I think we’ll look at survival. Everybody looked at the ratify survival curves and said, some people said, “Meh.” But if you see a much more pronounced hazard ratio or difference with quizartinib versus placebo, there’s a biologic rationale.

Andrew Wei: The method we use to measure the presence of FLT3 today obviously not very sensitive, but you’ve done some workshops and more sensitive methods.

Mark Levis: That actually is key. Getting back to MRD, getting back to a nucleic acid form of MRD, that’s a component of that. If you can basically say, “Look how deep the remission is.” At least for the market that we’re targeting using this PCR NGS sandwich actually that multiple groups are developing, it’s not hard technology. You can really track this thing deep, and if your treatment makes it go deep, that’s very reassuring. It explains a lot of what we see. So in other words, the more data, the more light that’s shined on the results of that trial, will make it make people perfectly comfortable to switch their so called standard of care.

Andrew Wei: Presumably there must be more patients positive than we realize. If you had a more sensitive technique and with the lack of correlation between the level and the potential benefit of these FLT3 inhibitors, maybe we’re missing some patients that might benefit from other treatments.

Mark Levis: There is an association with the level on survival. We kind of show that just a sheer measurement. We’ll make curves separate with this assay and I think that’s an end point on all those phase three studies. That assay, done by different institutions, a slightly different methodology, but essentially in principle the same assay, is a hard end point on those. So that … we’ll get that readout.

Alan Burnett: My impression is that FLT3 by NGS has … is tricky because of the size issues. Is that being resolved?

Mark Levis: Yeah, that’s resolved. What you do is you put PCR primers flanking the area of interest, amplify it up just a wee bit, enough to make it out, compete the non FLT DNA, and then chop that up, put that on an aluminum panel, and you can get a very sensitive … a lead like burden down to 1 in 10,000, 1 in 100,000 routinely. And it turns out, not surprisingly, it’s right there at 1 in 10,000 in most patients, right after remission. A good clean remission. No, it’s right there. Of course it is. So, making that go negative will be-

Alan Burnett: Vyxeos lyposomal daunorubicin seem interesting because the phase two studies, only showed a benefit in the adversarius patients, hence its evolution as an adverse kind of targeted thing. Difficult to understand in the context, even of the experience of daunorubicin intensification. Why Vyxeos wouldn’t actually have done something in the intermediate or more favorable patient, but it didn’t. I don’t think anyone really understood why not. But clearly when it came to Europe, there was no experience of using it in Europe. When it came to Europe, the kind of doubts were expressed, well you’re only up against 7+3. So it’s clearly going to face some quite difficult challenges I think, as it moves into different situations. But it’s the technology behind it that I think anyone who’s done lab experiments, and looked at synergies or ratios, etc. Of different agents in combination, knows that different ratios produce better results, and it goes up and down, and they publish the best ratio available.

Alan Burnett: So there may well be other combinations out there if the principle of putting in a fixed ratio really is what’s making the difference. But of course there is the lyposomal itself is not a random thing. I mean it is engineered to accumulate [crosstalk 00:21:13].

Mark Levis:                       [crosstalk 00:00:21:13].

Alan Burnett: So this is not a chance, a wee lump of fat to stick drugs in. This has been carefully thought through, but it would be interesting to see whether this could be applied to other combos, and whether we’d make them more efficient. And I’m just wondering if like your venetoclax, IDH1 or 2 or something, putting a liposome could have a future. I mean is it a crazy idea?

Mark Levis: I think it sort of is because the main problem that the lyposomal gets around is it keeps the daunorubicin from your skin, and gut, and things like that, and keeps it in the blood. I don’t know that that’s … we see any toxicity from an agent like venetoclax except to the bone marrow. That’s where the lyposomal gets to.

Alan Burnett: Just tell us the story, the myeloma story, in venetoclax. If you know it because it’s obviously been put on hold in that context. So why is that different?

Andrew Wei: You’re referring to the study which combined bortezomib with dexamethason. The study was halted because of an increased number of deaths in the venetoclax containing arm. The full picture as to what was going on is not available. However, I think the suggestion is that there was an increased number of septic events, and presumably that led to increased numbers of deaths. So the question is why was the venetoclax containing arm more toxic with respect to sepsis?

Andrew Wei: One hypothesis I have is that venetoclax is a PJP inhibitor and that perhaps in combination with some drugs, I’m not sure if bortezomib is a substrate, but it could inadvertently increase the concentration of the drug in the, not just the normal cells, but the malignant cells. And normal progenitors actually have very high levels of PJP. So in fact, the normal progenitors might suffer more intensely to the malice oppression. This could also be an issue with venetoclax being combined with other cytotoxic combinations, particularly if they’re PGP pump substrates.

Alan Burnett: Now I was quite reassured to see that most people are feeling that these drugs, these new drugs, which have very, you know, they’re very interesting mechanism etc. Are moving more front line. But as they move more frontline, the practical challenges of delivering these studies increases in my view. So you were talking Andrew, about a sort of master protocol, which I think is more recognized by regulatory authorities as a reasonable thing to do. Would you like to just summarize what you’re talking about?

Andrew Wei: Yeah, so at the moment we have more partitions, and segmentation of AML, which means doing subgroups studies now becomes more challenging. Bob Lowenberg discussed the HOVON approach where they’re getting together many countries to do an IDH, a randomized study, and this will require 7,000 patients to be screened to find the appropriate number of patients to get an adequately powered, event free survival end point. For us folk in Australia where we don’t have-

Mark Levis: 7,000 people period.

Andrew Wei:  … 7,000 with AML we have to come up with more, I guess innovative approaches to do trials where we can get useful information. And the thing that’s always puzzled me is that to do all these different combinations you have to literally open six, seven, eight, nine separate trials, which drives our governance office and our coordinators spare. So every patient relapses ultimately, if they’re not cured obviously, and they will have a different clone that might be targetable.

And the point of the master protocol is to find some way where we can unify different drug combinations into a package whereby we can recycle the patient onto new combinations when they fail. Combination A, move them onto combination B, rather than a new study. And to provide pharma with a way of integrating their new drugs more seamlessly into a master protocol. Similar to how you ran the pick a winner type of study.

Alan Burnett: So, I mean they could be in C, they could be D, they could be in E. So the patients that hit randomization E, could vary quite a lot in terms of what they’d been through, and you know, whether or not their disease evolved. How do you sort of control for that?

Andrew Wei: It’s no different to how we do relapse refractory studies anyway. There’s people that have 1 to 15 prior therapies, but I think the most informative information is, what was the target? What happened to the target? And did doing something to that target have a patient benefit in terms of prolongation of remission? And if the patient didn’t have a benefit, what other things were going on at a mutation, or non mutational level that might have been causing resistance? I think we can gain lots of information from looking at individuals, not necessarily control, because of prior clinical other factors.

Because what we’re looking here is purely the drug versus the target, and each patient has often five or six different mutations, and so each patient actually gives us five pieces of information. If we can recycle that patient onto multiple drugs, then we can get much more data than we can from just putting one patient onto one trial, looking for one survival endpoint, and then again trying to find another 7,000 patients do it all over again.

Alan Burnett: Yes, this is similar, not the same as, but similar to the Beat approach where they look at the molecular signature-

Mark Levis: Try to classify every patient and child, literally not just molecular signature but even maybe even in vitro assay.

Alan Burnett:  Yes. Now I’ve not heard too much about that program in terms of how it’s expanding. [crosstalk 00:00:27:31].

Mark Levis: I think all they can see is-

Alan Burnett: These interventions are not randomized, are yours randomized?

Andrew Wei: No, not randomized because-

Alan Burnett: So they’re exploratory?

Andrew Wei: Exploratory, yes.

Mark Levis: I think that’s what this is too. And I think that to some degree there was a sense … part of it is limited by what drugs are available. So, well it could be this, it could be that, it could be this, what drug do we have? That one. Well that’s what they get. So you know, it may be limited by that in a lot of ways, but if they get enough drugs, even then the lack of randomization, the uncertainty of, it’s really a big pilot trial more than anything. Then my one concern with Beat AML, I’ll be honest, is they don’t quite pay attention to the pharmacology. The drug in the dish is not the same as the drug in the patient, and the drug interactions that you’re talking about. The lack of control over that concerns me that there’ll be more noise than signal.

Alan Burnett: So just reflecting on the meeting as a whole, how did you kind of … were there advantages, disadvantages, in the meeting in terms of the general format? Speakers were invited to talk for short periods of time with long discussion periods. There was quite a lot of intermingling out off the formal partner meeting. Did you think this worked as a meeting? Did you find it beneficial?

Andrew Wei: I can honestly say it’s probably one of the most enjoyable meetings I’ve had in the last 12 months. Mainly because it wasn’t as packed and rushed as many other meetings. I think the short time format forced us to be more diligent about choosing what was relevant, and what was provocative. Then having very relaxed discussion amongst a smaller group of people is really a format that is almost a bit of a dying breed of meeting at the moment. So I really enjoyed it and got a lot of good interactions with pharma. I think the pharma people had very good access to, you know, KOL’s here and we had good access to them, and good access to catching up with good colleagues.

Mark Levis: I think the whole concept of get up and say a few quick words and then start the argument is more fun. In fact, I would expand more on that. I would, you know, have a gong. Okay, you you have 10 minutes. Talk fast because we really want to start arguing. That’s the fun-

Alan Burnett: I think the other thing is that, you know, you’ve got quite a number of disciplines, preclinical and clinical involved, which is interesting. Also I think the age of the KOL’s and the speakers has dropped from previous meetings that I go to.

Mark Levis:  Some of them are fresh out of high school, looks like.

Alan Burnett: So you look forward to the next year meeting. Okay.

Mark Levis: Yes.