Biology of acute myeloid leukemia (AML) and mechanism of action of venetoclax

DR LOVE: Welcome to Oncology Today — Management of Acute Myeloid Leukemia in patients not eligible for Intensive Induction Therapy. This is the second issue of a program focused on recently published and emerging clinical research, this is medical oncologist Dr Neil Love. For this program, I met with Dr Andrew Wei from the Alfred Hospital and Monash University in Melbourne, Australia. In addition to this audio podcast, there is also a video component, with Dr Wei’s presentation. To begin, I asked him to discuss the biology of AML and the rationale for the investigation of venetoclax in this disease.

DR WEI: So I must admit I’ve been working in the BCL2 area since about 2000, and so I’ve been in this area for 20 years. And I must admit before anyone used a BH3 mimetic in a patient there was always the concern how can you target a pro-survival molecule that is present in every cell and is critical for controlling the life-or-death decision of that cell. And so it was a real eye opener to us when even the less selective BH3 mimetic, ABT-263, which targets BCL2, BCL-X and BCL-W, so 3 pro-survival molecules, that this was actually tolerable in humans. And in fact the major toxicity of that drug was on target thrombocytopenia because we know the BCL-X is a survival mechanism for platelets.

So when venetoclax was developed, which specifically targets BCL2, to get around the problem of thrombocytopenia, especially in diseases like AML where baseline thrombocytopenia is a much bigger problem, this really set up the possibility of targeting tumors more specifically.

What we think is the reason why patients and their tumors can tolerate pro-survival targeting in a nonspecific manner, that is non-cell specific, is that oncogenic cells, to survive, really need to have their pro-survival molecules keep that cell alive. And so when there’s oncogenic damage, for instance, there is an increase in the proapoptotic load. And so the body wants to get rid of these cells and is trying very hard because when you have oncogenic lesions you do increase proapoptotic molecules.

However, there’s clearly a sufficient level of BCL2 or other pro-survival molecules keeping that cell alive, but just. And so by giving a BCL2 inhibitor you are now tipping a very precarious situation, where that oncogenic cell has just being kept alive by the right amount of pro-survival molecule to balance the proapoptotic desire to eliminate that cell, and you’re tipping it over the line to death. Whereas more normal cells don’t have that intrinsic proapoptotic burden, and so they can tolerate inhibition of BCL2 without intrinsic cell death.

And so I think that is the mechanism for why BCL2 inhibitor-type molecules work. The question is why don’t they work as single agents. And in the case of AML, at least, it would seem that not only is BCL2 keeping that cell alive, but other pro-survival molecules as well. We think the most important one is MCL1, but also BCL-X and possibly A1 may be relevant in certain cell context. And so there is active clinical activity trying to combine a BCL2 with an MCL1 inhibitor to see if we can now do the same thing as BCL2 inhibition with chemotherapy. Chemotherapy does the same thing except that it does it in a much more promiscuous and complex manner by inducing p53 activation.

DR LOVE: I remember one of the first people I talked to about this was Dan Pollyea, and he brought up the issue of the effect of venetoclax on stem cells, and I’m curious what your thoughts are about that.

DR WEI: Yeah, so there was an early study done by Tony Letai where he showed that human progenitors, human leukemic progenitors, were more sensitive to ABT-199, now known as venetoclax, then normal CD34 progenitors. And so functionally that’s where the story began. And then Dan Pollyea’s group showed that using CyTOF that CD34 progenitors were also eliminated with venetoclax-based therapy.

I think although some progenitors are sensitive to BCL2 targeting, obviously there are other progenitors that are not, and it differs with different molecular subgroups. So for instance, as I showed you before, patients with an NPM1 mutation have a really high response rate and very long survival, and so presumably progenitors in that setting are more ably targeted than in other settings, for instance like p53 mutation where the survival is short despite the initial response rate being quite good.

DR LOVE: So before we dive into the AML data, any thoughts about the potential role of venetoclax in solid tumors? I know it is being looked at to some extent. What are your thoughts about that moving forward?

DR WEI: Yeah, so it’ll be interesting to see how it goes. I think it’s more likely that venetoclax in combination with other drugs will be required for solid tumors. It’s always been my suspicion that in solid cancers that the main pro-survival molecule operational in solid cancer, as opposed to hematologic cancer, is BCL-X rather than BCL2, and so it could be that BCL-X-targeted therapies in combination with other conventional solid tumor approaches might be more effective. Whereas I think BCL2 just seems to be a critical pro-survival molecule for a number of hematologic malignancies, and that’s why venetoclax-based therapy has been so successful in that setting.

Efficacy of venetoclax in combination with hypomethylating agents or low-dose cytarabine for patients with previously untreated AML who are ineligible for intensive chemotherapy

DR LOVE: So we now, during this year, have seen results from the 2 Phase III studies, VIALE-A and C. Can you talk a little bit about what the thinking was when these 2 trials were designed?

DR WEI: So we know from the Phase I study, which was venetoclax on its own in relapsed/refractory AML, that the response rate was 19%, which is not bad for a single agent in the relapsed/refractory setting, but the median duration of survival and response was quite short. And as I mentioned before, it’s quite likely that in AML that other pro-survival molecules were important. There was a paper published from Stefan Glaser from Walter and Eliza Hall Institute in 2010 which showed that MCL1 was a key survival factor for AML. And so we knew that.

And when the Phase I study showed modest results, when it was made to go to the first-line setting in AML, as with many other agents, generally the combination is with low-dose ara-C or with hypomethylating agents. However, we also knew that low-dose ara-C and hypomethylating agents would induce some proapoptotic activity which might augment the combination with venetoclax. And so there was some rationale, but also some degree of pragmatism about the combination and the way it was operationalized in the first-line setting for older patients with AML.

What wasn’t expected, I think, was the remarkable response rates that were observed in the first few patients that were treated with this combination, and I think the final results from the VIALE-A and C studies showing a 66% and a 48% response rate in older AML patients, delivering responses similar to intensive chemotherapy, really was quite a groundbreaking change for older patients who previously were really treated with palliative intent.

DR LOVE: Can you kind of put a more clinical background to these 2 trials in terms of how is plays out in practice and compare the typical course of a patient who is treated just with a hypomethylating agent compared to now with venetoclax clinically? What you see in terms of speed of response and depth of response and duration.

DR WEI: Yeah, it’s been quite a remarkable change because for decades, first of all, clinical trials have been performed in the older AML setting. And sometimes clinical trials could produce double the response rate compared to low-dose ara-C or hypomethylating agents, and yet overall survival was not improved. And we think that’s most likely because of the more complex genomic makeup of older patients with AML, that even if you achieve a response all you do is select for more aggressive clones, and they progress, and the patient’s now resistant, and survival is not improved.

When we treat an older patients with AML we’re presented with several problems. First of all, 1) the patient generally has more intrinsically-resistant disease. 2) The patient can’t tolerate high doses of chemotherapy because of usually associated comorbidities. And I can tell you that many older patients just don’t want to spend a lot of time in hospital, and many of them are not prepared to take the same risks as younger people in terms of going through intensive chemotherapy, experiencing potential side effects, even if it means a few months of life extension.

What we noticed in the past was that if we used say for instance azacitidine that we would have a response rate of under 30%. And sometimes we would have to wait 4 to 6 cycles to see that response. What that means for the patient is that for 4 to 6 months they are rendered severely cytopenic, at risk of infections, and also requiring multiple blood transfusions from week to week. And so this cumulative risk of having infections really is the problem. That risk of infection means hospital time and increases the risk of mortality.

However, with venetoclax-based regimens we find that the speed of remission is remarkable, and so the majority of patients achieve hematologic response within the first cycle of therapy. And what that means is that the patient now has a remission. They’re now, in many cases, transfusion independent, and the risk of dying from an infection is much lower because the neutropenia has been resolved with the response. And that, I think, is the most important change with these therapies. Not only does it increase the response rate, but it does so incredibly rapidly and at no increase in early mortality.

DR LOVE: I’m curious how you approached these patients prior to venetoclax. I’ve seen data. It’s hard to believe from the US and the SEER data suggesting a lot of people weren’t even getting treated, and when you think about the numbers and the typical course you just described, you can kind of understand that some patients might want to go directly to palliative care. I’m curious what it was like when you started to see, after all this work in the laboratory, as you started to treat these patients. If you have any memories of the first couple patients that you treated.

DR WEI: Well in fact the first patient that I treated with low-dose ara-C with venetoclax was in 2014. And she was an older lady, didn’t want to have intensive chemotherapy, was 78 years of age, and she went into remission with the first cycle of therapy, didn’t lose her hair. And we thought gee, this was good. Obviously, this was followed up by many other patients that did the same thing. But this first patient that we treated, is actually still alive in 2020.

And so obviously that’s an extreme case, but to have a patient live for more than 6 years now without receiving intensive chemotherapy was just not a situation that we could previously imagine.

And so I think prior to that we were all very similar, that older people in their 70s and 80s, we really didn’t have expectations beyond a few months. And for that reason many people elected to not have any therapy at all. So I can certainly understand where the SEER data is coming from. It’ll be really interesting now, over the next decade, to see what proportion of patients now actually receive active therapy and what are the gains, I suppose, not just in specialized hospital, but in community-based settings. Because I think will be the true test of how well this treatment is 1) adopted and 2) delivered, and 3) managed in terms of the potential toxicities.

DR LOVE: I can remember when we first started presenting some of the early data a couple years ago the audiences of our live meetings are like really? We’re going to be treating with this? And now they’re using it all the time. It’s really amazing, we had a case presented the other day of a patient who was on a trial. It was actually venetoclax and I think it was bortezomib in myeloma, and the doc was presenting it, and he said yeah, the patient went on this trial in 2015 of ABT-199 I think it was called. It didn’t have a name in 2015. Which is hard to believe.

DR WEI: That’s right. It’s been one of the most rapid developments that I’ve seen in AML. And it’s just been a whirlwind tour. That said, I think there’s still work to be done. As I mentioned before, patients with FLT3 and p53 mutations, they still have a response, but we still need to improve the longer-term outcomes in these patients, as well as others. So I really do see this as obviously a very groundbreaking change. But it’s still the first step. But I think it’s a very positive step forward.

We can use this framework to try and improve the responses further, either by combining ven/aza or low-dose ara-C than with other agents or combining this combination with other combinations in sequential or tandem regimen, or considering the use of allogeneic stem cell transplant even in older people because generally once they’re in remission their condition is markedly improved. And so for some patients in their 70s, where we would never consider transplantation, it is now possible and has actually been demonstrated now already in several small series that this is possible, and that longer-term outcomes are possibly further improved by this practice.

DR LOVE: Actually when you think about it too, in addition you’ve seen the explosion of venetoclax in CLL, so in 5 years we’ve gone from a drug that didn’t have a name to a drug that people in general practice are using like water nowadays.

Design and results of the VIALE-C study investigating venetoclax with low-dose cytarabine for older patients with previously untreated AML

DR LOVE: Let’s get back to the 2 big trials that looked at this strategy, and maybe you can talk a little bit about how the trials were designed, the eligibility, and how the results compared, indirectly.

DR WEI: Sure, so VIALE-C was low-dose ara-C, for C, in combination with venetoclax, and VIALE-A was with azacitidine. The VIALE-C study was based on data that we had in 2016, because the VIALE-C study began in 2017. And at that stage it was projected that the median survival for low-dose ara-C/ven was going to be about 12 months. And we knew at that time that low dose ara-C alone had a median survival of about 4 months.

And so the study was designed to look at an improvement in survival from 6 to 12 months to be a little bit conservative. However, what we didn’t appreciate at the time was that the follow up from the Phase IB studies was really short, and that with further follow up the median survival was not so much 12 months, but more towards the 10-month level. Furthermore, patients that received hypomethylating agents, with further follow up, even though there was a response rate early on of about 43%, the survival from those patients was not similar to patients that had not received prior hypomethylating agents.

So if the study had been powered to a hazard ratio say of 0.7 or thereabouts, then I think the study would have been far more positive. Furthermore, we also know now that because the study was completed, and the event rate was reached 6 months after the study closed, the follow up time at that stage was only about 12 months. And as we know now, one of the really interesting features of venetoclax-based therapy is that a proportion of patients, maybe 20%, have long-term survival beyond 2 years.

And if the follow up is only really short, you just don’t see the benefits there, and it can’t be accounted for by the survival curves. So I think in hindsight a slightly larger study, perhaps powered independently of HMA exposure, and third with longer follow up I think would have given us a much clearer result.

Furthermore, in the control group for the VIALE-C study, 22% of patients in the control arm were salvaged with intensive chemotherapy versus only 8% in the venetoclax arm. And so again, confounding therapies in the control arm could have brought the 2 survival curves closer together. And we’ve now seen this in the enasidenib plus azacitidine study, where 21% of patients in the aza-alone arm received enasidenib salvage. And again, this confounded the overall survival endpoint potentially.

Overall survival in the VIALE-A trial evaluating azacitidine with venetoclax versus azacitidine alone for patients with previously untreated AML who are ineligible for intensive chemotherapy

DR WEI: In the VIALE-A trial, however, it was a much larger study, 431 patients. The hazard ratio targeted was 0.7, so more conservative, and there was longer follow up. And the results from the trial were much more clearly positive with a clear separation. And I think one of the most interesting aspects of the VIALE-A trial was that at 2 years, or 24 months, about 35% of patients are still alive, indicating again that there is a subgroup of patients that have prolonged survival. And I think we really do need to try and understand who these patients are, what does the MRD tell us, and can we try and produce these types of longer-term outcomes in other populations receiving venetoclax plus azacitidine.

DR LOVE: So the hazard ratio with the VIALE-A study for survival was 0.66. What is the hazard ratio with VIALE-C?

DR WEI: In the updated analysis 0.75.

DR LOVE: Very similar.

DR WEI: 0.7 in the updated follow up and 0.75 in the initial analysis. So it was similar, but unfortunately not powered sufficiently because of the smaller sample size, and also the inclusion of patients with HMA exposure.

DR LOVE: So speaking of that, I am curious right now how you approach patients who’ve, for example, had MDS, they’re on HMA and then develop AML?

DR WEI: Yes. That remains a major area of clinical unmet need, I believe. So with the low-dose ara-C plus venetoclax option there was a reasonable response rate. However, longer-term survival remains problematic. In the VIALE-A trial patients with HMA exposure were excluded from that trial.

However, there has been a Phase I experience performed separately to the VIALE-A trial, presented at ASH in 2019 by Amir Zeidan. And he looked at patients who had prior HMA exposure and received HMA plus venetoclax. And obviously you would wonder whether patients who had already failed HMA wouldn’t respond to the combination with venetoclax. However, his experience was that there was a response rate of 40% and that at 6 months a substantial proportion of patients remained alive.

And so I think at this stage it remains a reasonable option to use HMA/venetoclax even in patients that have failed prior hypomethylating agents. But, if you do achieve a response, I think my advice would be to take those patients to transplant if that was possible because obviously that gives us potentially a better option of sustaining that response rather than waiting for I guess what would be an inevitable relapse.

Mechanisms of resistance to venetoclax-based regimens in patients with AML

DR LOVE: What do we know about the mechanisms of resistance to venetoclax combinations? And are there ways clinically to identify these mechanisms, for clinicians at this point?

DR WEI: Yeah, it’s very interesting. So we know that in CLL with longer-term follow up that BCL2 mutations have been identified. And that’s because on the surface of BCL2 there’s a hydrophobic canyon, and the venetoclax drug binds into this canyon and blocks the binding of other BH3-only molecules, thus neutralizing the effect of BCL2.

As you can imagine, cancers are very clever, and tumors have developed mutations in this canyon which prevents venetoclax from binding. In AML however, we have not identified such BCL2 mutations yet, and so other mechanisms must be operational. And so in AML the most common mechanisms, we think, are severalfold, which include either an increase in other pro-survival molecules, particularly MCL1, which venetoclax can’t bind to and can keep the cell alive. And other pro-survival molecules like BCL-X, and also another one called BFL1. This is just starting to be studied now, but it’s possible that an increase in BFL1 might also be a mechanism of resistance to venetoclax-based therapies.

Other mechanisms include p53 mutation, which perhaps might be due to the inability of the chemotherapy component to adequately increase the proapoptotic burden because p53 is often an important component of those drugs working adequately.

And furthermore, at the recent American Society of Hematology Meeting, they published some data showing that downstream BAX mutations can also cause venetoclax resistance. We found that in venetoclax-treated patients 14% of patients had BAX mutations emerge compared to no patients who relapsed after conventional chemotherapy. And furthermore, we did laboratory experiments that showed that if you deleted BAX from cells that not only would venetoclax not work, but also other BH3 mimetics such as MCL1 inhibitors on their own.

And so the question now becomes how do we design strategies that would reduce the, I guess, resistance mechanisms associated with venetoclax. Several possibilities are being explored, by combining venetoclax with ways to target MCL1 in particular, either through direct inhibitors of MCL1 or indirect inhibitors of MCL1, such as CDK9 inhibitors. So both of those types of studies are underway.

With patients with p53 mutation, which represent another common and key mechanism of resistance, new trials are being looked at which, for instance, combines azacitidine and venetoclax with a drug that might overcome p53 mutation, such as the anti-CD47 antibody called magrolimab. And because magrolimab doesn’t increase the myelosuppression which is known to be associated with ven/aza, this would seem to be a really interesting triplet combination.

And so I think the future, and this is a great example of ven/aza plus immunotherapy, targeting a completely different mechanism that might be completely blind to the intrinsic molecular complexities which might be associated with ven resistance, I think, represents a really attractive option for the future.

Predictors of response to venetoclax-based therapy for AML

DR LOVE: What do we know about predictors of treatment benefit to venetoclax combinations? I’ve been kind of hearing intermittently about BCL2 assays. I know it seems weird, but in myeloma t(11;14) patients seem to respond. There’s some other predictors they’ve looked at with myeloma. What about in AML, what do we know about predictors of treatment benefit?

DR WEI: So although there have been some studies published looking at these laboratory techniques such as BH3 profiling which might predict response. In practice, this is not widely available. And so what we have available are obviously mutation profiling. And so so far we know that patients with NPM1 and possibly IDH and SRSF2 mutations seem to be not only sensitive to the single agent, but also sensitive to the combination with either low-dose ara-C or azacitidine.

With the Phase IB studies we know that low-dose ara-C plus venetoclax produced a response rate that was only percent in NPM1-mutant AML. And at 2 years the survival in this NPM1-mutant subpopulation was 64%. We also know from an institutional study from MD Anderson that NPM1 mutations also do very well with azacitidine plus venetoclax.

One question, however, arose from the VIALE-A study in that the hazard ratio for survival amongst patients that had NPM1 mutation crossed unity. And so this brought to question whether NPM1 mutation was in fact a harbinger of good outcomes with ven-based therapy. And I’ll just point out that even in the VIALE-A study I think the follow up is still inadequate to see the longer-term benefits of venetoclax-based therapy. So I think it remains an open question in that area.

We also know from the VIALE-A study that IDH mutations had really positive benefits with ven/aza, with a hazard ratio, I think it was 0.34, in IDH-mutant AML, showing quite substantial improvements or reductions in the risk of death by literally 66%, which is quite remarkable.

In terms of other subgroups that might benefit from venetoclax-based therapy. In the preclinical literature there’ve been a number of interesting observations being made, but I think it remains to be determined whether these pan out in practice.

MRD response, early MRD response may also be a more generic predictor of longer-term outcome. And new therapies trying to improve the MRD negativity rate are being explored. So for instance at the American Society of Hematology Meeting Tapan Kadia from MD Anderson looked at this new approach of combining low-dose ara-C/ven with cladribine for 2 cycles, followed by ven/aza. And he found that he could achieve a response rate in older AML of 93% and an MRD negativity rate by flow of 89%. And so if that’s true, then it will be really interesting to see how this compares with low-dose ara-C/ven or aza/ven alone.

Therapeutic approach for older patients with AML harboring IDH or FLT3 mutations; role of venetoclax-based therapy

DR LOVE: Can you talk a little bit about what we know in terms of aza. You mentioned IDH mutations, IDH inhibitors, and the different strategies that have been looked at, including hypomethylating agent plus IDH inhibitor without venetoclax. Can you kind of talk a little bit about what we know about these patients and right now what you think the optimal therapy is in an older patient with an IDH mutation?

DR WEI: Yeah, so when we have an older patient the first question is should we just commence ven/aza, for instance, or should we do molecular profiling and hold off treatment to identify an IDH mutation? And that would be appropriate if we had an alternative regimen that would produce an equal or better response.

So the literature at the moment stands where a comparative study has been done looking at enasidenib plus azacitidine versus azacitidine alone. And overall survival between those 2 arms, in contrast to the VIALE-A study, was no different. However, there are a couple of caveats with that study. First of all, in the control arm 21% of patients receiving azacitidine alone were salvaged with enasidenib, the IDH2 inhibitor. Furthermore, the 2-year survival in the control arm was 45%, which is quite different from the azacitidine control arm in VIALE-A, where the 2-year survival was much lower than that. And so we don’t know yet whether the post-study salvage therapy with IDH2 inhibition really confused the interpretation of the survival endpoint.

If you look at the 2 investigational arms, enasidenib plus azacitidine and azacitidine plus venetoclax, in fact the 2-year survival is almost identical. Furthermore, we don’t know yet, but it’s possible that an IDH inhibitor plus azacitidine may cause less myelosuppression and perhaps be better tolerated than ven/aza. So I think what we really do need in the future is perhaps a randomized study between enasidenib plus azacitidine versus venetoclax plus azacitidine to determine which is the better drug in terms of tolerability, as well as efficacy.

DR LOVE: What about the triplet?

DR WEI: In older people, however, it may be appropriate to consider, say, an IDH inhibitor on its own in people who are really frail, and we don’t think might be suitable to receive ven/aza. I think in that situation it is worth doing molecular testing and considering perhaps a single-agent oral option in terms of an IDH inhibitor.

With respect to IDH1 inhibition the results of the randomized front-line study is still not available, and so it still remains possible that we may see positive benefits from an IDH inhibitor plus aza compared to aza alone. And as you mentioned, the next step forward might be the possibility of combining all 3 agents together in terms of an up-front triplet.

Obviously, we have to be aware of 2 considerations. The first is whether this would be equally tolerable. What we don’t want is an increase in myelosuppression or increase in toxicity without improvement in long-term survival. And what we don’t know is whether it’s the right strategy to combine all these agents up front, particularly for the IDH-mutant subgroup, where we know that long-term survival is pretty good with azacitidine plus venetoclax alone.

Maybe the better strategy might be to select patients which may not do so well with that combination, whether that’s people that don’t have MRD negativity or whether that’s people that might have persistent IDH mutations on ven/aza. It remains to be determined. But I think it’s either going to be an everyone gets it up front or we do a more selective strategy of identifying patients that may not have an optimal outcome and then add in an IDH inhibitor. Obviously, the second option would be economically more efficient, but obviously those 2 strategies remain the 2 options which people will hopefully investigate in the future.              

DR LOVE: What about venetoclax combinations in patients with FLT3 alterations?

DR WEI: This is a really interesting area at the moment because what we know is that ven/aza, for instance, seems to produce improved response rates, both for patients with FLT3-TKD and also patients with FLT-ITD mutations. However, we know from a subgroup analyses presented by Marina Konopleva at the recent ASH Meeting that patients with TKD mutation have an improved median survival with ven/aza compared to aza alone, of 18 versus 10 months. However, for FLT3-ITD overall survival does not appear to be substantially improved with the combination.

So the question then becomes should we use a FLT3 inhibitor up front for these older patients? What we know is that preliminary data from the LACEWING study, which is azacitidine plus gilteritinib versus azacitidine alone, was actually closed early by the Safety and Data Monitoring Committee because of the lower likelihood of the study being positive. And so we don’t know at this stage whether it is appropriate to consider a FLT3 inhibitor in the front-line setting.

There were definitely studies which are investigating the possibility of incorporating a FLT3 inhibitor up front with ven/aza in a triplet combination. But again, the possibility of myelosuppression, I think, needs to be considered and the regimen optimized before we can use this in practice.

The other possibility is to consider what’s called a tandem approach, that is using ven/aza in the first cycle, and then in the second cycle bringing in a venetoclax-based combination. Whether it’s with azacitidine or venetoclax plus gilteritinib remains the open question because we know that venetoclax plus gilteritinib, as presented by Naval Daver at the recent ASH Meeting, produces really high response rates in patients with relapsed/refractory FLT3-mutant AML. So whether a tandem approach might be better.

But the actual issue of using a FLT3 inhibitor up front brings other questions into mind. For instance, is the tumor really dependent upon the FLT3 mutation in the first-line setting? Or should you consider this in a salvage setting where perhaps the FLT3 mutation has been selected, and the tumor’s now much more dependent on the FLT3 mutation and therefore more responsive to the FLT3 inhibitor. And where salvage studies such as the ADMIRAL study, have demonstrated improvements in survival with this approach. It does bring into question whether the situation is different in the front-line setting. So that remains an open issue.

Perspective on the role of antifungal prophylaxis for patients receiving venetoclax-based therapy; dosing venetoclax in combination with antifungal agents

DR LOVE: So let’s talk a little bit about some of the practical clinical questions that come up, and one I hear a lot is the issue of antifungal agents in patients who are receiving venetoclax combinations. Can you talk about what we know about that and what your recommendation is in terms of dosing?

DR WEI: Yes. This still remains a very complex, and for me, a vexed question. And so when we commenced these studies what we did know was that ven/aza or low-dose ara-C/ven produced quite prolonged periods of myelosuppression and neutropenia, similar to what we observed with intensive chemotherapy. And so our early practice was to consider antifungal prophylaxis for the majority of patients.

However, in the VIALE-A study a subanalysis was again presented at the American Society of Hematology, and I was actually surprised to look at the data and find that in fact not as many patients received antifungal prophylaxis as what I thought they should have received. And furthermore, comparison was done of people that did receive and did not receive antifungal prophylaxis. And interestingly there did not appear any difference in the rate of fungal infection or survival between the 2 groups. And so I think it remains an open question as to whether we need to use antifungal agents or not. I’m certainly still worried by the prolonged neutropenia which is induced by these agents.

And so I think what we may need in the future is perhaps a more nuanced approach that is looking at the patient in front of us, looking at their comorbidities, the potential ability to tolerate neutropenia, whether they presented with severe neutropenia in the first place, how long it takes for them to recover after each cycle. And I think we may need to use a more individualized approach rather than a one-size-fits all.

The next question is what dose of venetoclax to use if you are considering use of an antifungal agent. And again, this is still a very complex question. The product information for venetoclax recommends 70 mg. However, in the Phase IB and the Phase III trials venetoclax was used at a dose of 50 mg in combination with posaconazole. We also know from PK studies that 50 mg perhaps is the best dose of venetoclax to use in combination with posaconazole. And so the question is how do we do this, especially when 50 mg tablets aren’t widely available because the drug usually comes in 100 mg bottles.

And so if you’re using 100 mg of posaconazole in combination with venetoclax, then this was not the dose which we did the trials with. And we don’t have long-term survival outcomes or even tolerability data of 28 days of ven at 100 with posaconazole. And so other options include using alternative antifungal agents such as voriconazole, where 100 mg could be considered, or using echinocandins. And so these are all different options. I don’t have definite evidence-based guidance as to which one should be used, but hopefully I’ve illustrated some of the complexities and issues we’re currently thinking about.

Dose modification to manage venetoclax-associated cytopenias in patients with AML

DR LOVE: So it was really amazing for us to watch as these venetoclax combinations moved into general practice. We would have working groups with general medical oncologists to try to figure out what the issues were. And one thing that came out like almost immediately was the issue you really went through in great detail, and it was really, really helpful, I recommend people check out in your talk, which is the issue of cytopenias, particularly at the end of the first cycle. And it seems like we’ve learned a lot about this in the last year or 2. Where are we today? And what kind of advice would you give to an oncologist in practice?

DR WEI: Yes. Another really complex and open question. So we know that the FDA recommends 28 days of venetoclax during the first and subsequent cycles if the patient hasn’t experienced problematic cytopenia. However, we also know that pretty much the vast majority of patients eventually go to a 21-day schedule or less. And an abstract was presented at the recent ASH Meeting showing that there was no survival difference amongst patients that received a 21- or a 28-day schedule.

Furthermore, it’s always been I think a biological conundrum as to whether it’s actually biologically rationale to use venetoclax beyond its direct combination with a cytotoxic component because we know that the single-agent venetoclax perhaps doesn’t have substantial activity on its own. And so it always has been a question whether this monotherapy tail after the combination has 1) biological and clinical rationale.

Furthermore, we also know that there have been other studies performed. For instance, there was a study recently looking at cladribine plus low-dose ara-C/ven which actually started off with a 21-day schedule and then went to a 14-day schedule with ven/aza. And outcomes in that setting appear to be quite good. And we also looked at a study of venetoclax in combination with intensive chemotherapy, where we only had a 7-day overlap with intensive chemotherapy preceded by a 7-day monotherapy tail to get the dose ramp up done. And again, response rates were very high in that setting. So I think in the future the question will be do we need 28 days or could 21 or even 14 days of venetoclax be sufficient to produce the same outcomes with less toxicity.

What we do, however, in practice at the moment is that we do a bone marrow on day 21 rather than day 28. And if the bone marrow shows that the blasts have been cleared, because of that data I mentioned before of not being a difference between 21- and 28-day schedules in terms of survival, we actually stop the venetoclax on day 21 and then continue with a 21-day cycle subsequently. If there’s still persistent disease at day 21, then we wait a week, give the patient a break, and then go on immediately with cycle 2.

DR LOVE: And again, 21 or 28 days in that situation?

DR WEI: Yes, good question. My feeling would be to look at another 21-day marrow, and if there’s been no reduction at all in the blast count I would be considering alternative therapies. If there’s been a nominal reduction in the blast count, say from 50% down to 30%, then I will continue because we do know that delayed responses can be achieved in about 15% to 20% of patients if you continue on to between 3 and 4 cycles rather than ceasing after 1 or 2.

Monitoring patients receiving venetoclax; perspective on the optimum duration of therapy

DR LOVE: And I know there’s no typical clinical course of any patient in oncology, but what do you usually see in these older patients who get, for example, aza/venetoclax? No IDH, no FLT3, just a typical patient being treated in the community. When do you see the blood getting cleared? What do you see in the first few months? And what do you see in the first few years?

DR WEI: Sure. So it’s really interesting. So the first consideration we have is because of the high response rate, and we know that some people can have really long-term outcomes, we treat these patients as being really precious, and we look after them. And so we first make a decision should we treat this patient as an inpatient before we discharge them from hospital with normal blood counts and really reduce the risk of febrile and septic complications. And so that’s the first thing. If the patient’s really robust, lives close by, has great social support, then yes an outpatient monitoring approach after the dose ramp up is a consideration.

However, we don’t want to lose the patients from TLS because in some patients who are really hyperresponsive to venetoclax, and these include patients with NPM1 or IDH mutations, sometimes we’ve seen potassium levels above 6 four hours after the first dose of venetoclax. And so therefore we always admit our patients for close TLS monitoring because you just don’t know who’s going to be that patient who’s hypersensitive to the regimen.

Second thing is to obviously consider who should stay in hospital. Third is who should get prophylactic antibiotics and antifungal agents. Again, we don’t have clear answers to this, so I think I have to refer patients to institutional practice for that.

And that what we normally notice is that once patients get into response, then their ability to tolerate the treatment subsequently is much better. And so we find that the degree of cytopenia is far less severe, far less prolonged, especially if the venetoclax dosing is truncated to 21 days or less. And the patient can tolerate this really well as an outpatient. If your hospital is set up to do it, giving the low-dose ara-C by hospital in the home is really attractive, and also azacitidine can also be given by hospital in the home if your hospital is set up to do it. And that means the patient can receive the bulk of their therapy in the home setting, which I think is fantastic.

And so the next question becomes after several cycles of therapy how long should we go, and should we use both agents in combination or should we drop 1 of the agents. I strongly believe that the best gains from this therapy are through the direct combination because we know that ven alone and HMA or low-dose ara-C alone really are not that effective. So it makes no biological sense to me to use single-agent therapy.

However, if you can deliver a certain amount of therapy, we don’t know how much, in my practice it’s an aim of 12 cycles, if I can get the patient through 12 cycles, and if the patient is MRD cleared, no cytogenetic abnormalities and no clear evidence of disease in the bone marrow, then I do have a discussion with the patient as to whether we might consider ceasing therapy and watching closely or not. There’s 2 reasons for this. 1) I do have patients who have had long treatment-free intervals, where they clearly have benefit from that approach. And second I do have patients that have re-responded to the therapy after a treatment-free interval. And third, we also know that in patients that have received ven/aza continuously and the progress, that the median survival is very poor in that setting.

So I think a really important future question is whether a treatment-free interval perhaps preserves sensitivity to venetoclax and maybe other therapies versus continuing therapy where you might select for a really resistant subclone, which is really untreatable at that stage. So I think there will be data presented maybe sometime this year looking at the comparison between a treatment free versus a continuous treatment approach. I think this will be really important data for the community.

Spectrum of side effects associated with venetoclax-based therapy

DR LOVE: Any quality-of-life side effects that you see with venetoclax in AML? I haven’t heard about too much, maybe GI problems. Anything that’s been seen in the trials or you’ve seen clinically?

DR WEI: Yes. Certainly venetoclax/aza or low-dose ara-C/aza is certainly not without its price. Those prices include an increased risk of febrile neutropenia, which was substantially higher in the ven/aza compared to the aza arm of VIALE-A. Patients may also experience gastrointestinal side effects, nausea, sometimes vomiting, constipation. That also needs to be considered. And some patients just report lethargy or tiredness, which is a further consideration. So I think we need to question these patients carefully.

The clinical trials show that there was no decrement in quality of life, and in fact there was an improvement in quality of life in people receiving venetoclax, presumably because of the remission, and I guess the quality-of-life gains from not having transfusions and not having infections and having a normal neutrophil count. But there can be side effects. There was a high level of electrolyte disturbances reported in this study, and I wonder whether this was something to do with the tumor lysis prophylaxis in the initial phases because of all the hydration and induction of electrolyte disturbances.

So patients need to be watched carefully. But if they are watched carefully then obviously the gains are hopefully worth it for a good number of patients.

DR LOVE: So I’m still kind of mulling over what you were saying about TLS because I honestly don’t recall anybody else saying what you said. And maybe I didn’t ask the question correctly, but I kind of feel like the AML investigators have kind of dismissed TLS when I ask them about it. It’s not like CLL, et cetera. But you kind of are telling a different story. Do most investigators talk about it? Because I haven’t heard anybody say what you said before. A) That it occurs, and B) that you can’t predict it.

DR WEI: Yeah, so I think the rate of TLS, in truth, is low. So it would be 5% or less. However, that was in the context of a very well-controlled clinical trial. And if it’s 5%, then if you treat enough patients you will see it. And so I think experience and volume of practice is important in making that conclusion. And furthermore, it’s really important that the practices in the trial were adhered to to keep that rate low. So I think if we start treating patients as an outpatient, if we start treating patients that have higher white cell counts than recommended at baseline, and we start treating patients that have a high intrinsic risk of TLS, which include things like a high LDH, a high uric acid, and renal impairment, we will inevitably see more complications.

Furthermore, we do know that there are some patients, such as those with IDH and NPM1 mutations, which in my experience seem to be more predisposed to having these TLS events. And so I just wanted to point out that we should remain vigilant. We shouldn’t be too carefree with this regimen because the reason we saw a low rate of TLS was because of all the precautions put into place in the trial, whereas when they started the CLL studies they didn’t know this. And so that’s why I think CLL got a bad name with TLS, whereas in AML we were much more prepared and prevented most of these problems.

DR LOVE: But it really is interesting too because in CLL I hear stories about accelerating the pace to ramp up so patients don’t have to spend as much time in the hospital. And these people literally, they get biochemical TLS. I mean it’s for real. It’s not just in the trial. And I think that what you’re saying makes a lot of sense.

DR LOVE: So what about the post-venetoclax combination space, so to speak, in AML? Is there anything effective? Is it palliative care?

DR WEI: So I think it’s really important to look at the possibility for new clones emerging in patients that fail or relapse after venetoclax-based therapies. We’ve seen multiple times that patients either have a persistent IDH mutation, which can respond to an IDH inhibitor, or an emergent or an expanded FLT3 mutation, which could make them amenable to a FLT3 inhibitor either alone or in combination with other therapies.

And so I think recharacterization of the molecular landscape because we’re essentially treating a moving target. What the patient had at diagnosis is not going to be what they have at progression because the treatment will select for what is resistant. And if that resistant clone happens to have a targetable option, then I think it would be a tragedy for the patient not to receive that, and worse for the physician to not know that that was there. So retesting is important.

In terms of whether generic strategies in the future might be more broadly applicable to patients with ven/aza failure, I think this remains one of the key future questions in the field, and I think it will be a major source of investigation. A bit like HMA failure, which I think still remains a major problem. In both settings I still feel that identifying patients who are not likely to do well long term and considering an early transplant in first remission at this stage remains something that we can do. And whether we can therapeutically augment this in the future with novel therapies, I think, will be a future question.

Role of venetoclax in therapy for younger, fit patients with AML

DR LOVE: What are the clinical scenarios right now where you think about using venetoclax combination in younger patients?

DR WEI: Yeah, so we have always wondered, for several years now, whether the benefits to older patients can be translated to younger patients. With younger patients, obviously we’re competing against a much better control option, which is in the form of intensive chemotherapy and often intensive chemotherapy followed by allogeneic transplant. And so the question is how do we fit this new drug venetoclax into that conventional paradigm.

I think there are a number of ways of doing it. So for instance, let’s take the patients that we know molecularly do best with a venetoclax-based option, and those will be patients with an NPM1 or an IDH mutation. So one option is could we identify those patients who receiving intensive chemotherapy but have persistent molecular of flow MRD and therefore are not responding ongoing to intensive chemotherapy and give consolidation or maintenance therapy with these new agents in the forms that they’ve already been validated with, that is ven in combination with HMA or low-dose ara-C. And so there is a maintenance study looking at ven/aza in the maintenance setting.

The other options include can we incorporate venetoclax into the conditioning therapy for transplantation, and there’s been some interesting work done by Jacqueline Garcia showing that this is 1) feasible, and 2) showing interesting median-term outcomes with transplant. The third, and I think the most difficult, is whether we can combine venetoclax with intensive chemotherapy. Obviously, the key issue here is whether myelosuppression will become an impenetrable barrier with respect to tolerability.

So to look at this, we performed a study which was published in JCO where we took patients aged 65 years and older, and we took a very cautious approach because no one had done this before. We don’t know whether it was going to be feasible or not. And so we did a dose-escalation study looking at venetoclax at dose ranges from 50 mg through to 600 mg in combination with a truncated form of intensive chemotherapy, 5 and 2. And we found that 1) it was tolerable way up to 600 mg per day. Second, we found response rates to be very good, 72% in elderly AML. And in de novo AML 97% of patients went into remission.

What we did find also is that time to count recovery was no longer with venetoclax in combination with that schedule of chemotherapy. And so that’s prompted other groups to look at venetoclax in combination with 7 + 3, and there was a paper presented at ASH by Richard Stone to that effect. What they did find, however, was that the septic complications were problematic in older people with full 7 + 3 chemotherapy, and so they’re focusing their studies on patients under the age of 60. And furthermore, I think they found that doses of venetoclax up to 400 mg were the best-tolerated, and so higher doses than that are not currently being explored.

And so we’re still trying to find the optimal dose schedule. And I still caution people with respect to using venetoclax in combination with particularly anthracyclines because I worry about the potential for stem cell toxicity and also the possibility of gut toxicity in that setting as well. So I think although it’s really enticing to look at these agents in combo with intensive chemotherapy, we have to do it in the context of trials, and we have to do it in a very careful and controlled manner to make sure that we don’t ablate the benefits through inadvertent toxicity.

DR LOVE: What about the use of venetoclax combinations in younger patients who don’t do well with chemo, for example those with complex karyotype?

DR WEI: Yes. This is another really interesting setting because we know that with complex karyotype, many of which have p53 mutations, that the response rate with ven/aza is about 55%. We look at the alternative options, which would be either CPX-351 or intensive chemotherapy, where I don’t think the response rates would perhaps be as good as that. And so it is attractive to consider ven/aza as an option for people that we know won’t do particularly well with intensive chemotherapy. And there’s been talk for many years as to whether to perhaps do a trial comparing 7 + 3 with ven/aza in people with poor-risk AML fit for intensive chemotherapy.

I still think that despite the response rate, which looks good, we know that the longer-term survival outcomes are still not great. And so I think I still consider ven/aza in poor-risk patients to be a bridge to a transplant rather than a treatment in itself. And so it begs the question can we use other novel agents to improve the response further. And perhaps the most attractive agent at the moment would be magrolimab because of its reported high response rates in patients with p53 mutation in combination with aza, which I think approaches 75% and is now going to be the subject of a pivotal study to see whether magrolimab can be registered in AML for that purpose.

And so because magrolimab’s main toxicity is on-target anemia rather than neutropenia and thrombocytopenia, I think the attractions of combining ven/aza with magrolimab are very high, particularly for this poor-risk group. And a study looking at this triplet combination has already commenced at MD Anderson, led by Naval Daver.

DR LOVE: This concludes our program. Special thanks to Dr Wei, and thank you for listening. This is Dr Neil Love for Oncology Today.