Management of AML in Elderly Patients — Andrew H Wei, MBBS, PhD

DR WEI: I’m going to talk about the management of older patients with AML. And we know that, although people age 75 years and older obviously considered for palliative therapies most commonly, you can see here that the expected life remaining still remains at least 10 years among patients age 75 years and over. Therefore, new therapies with the potential to improve survival in older patients certainly have an important role to play.

In the Swedish registry, this is data provided to me by Gunnar Juliusson. You can see that in patients age 75 years and older, the proportion of patients with saving intensive chemotherapy is 50% or less and clearly declines as the patient ages. You can see that the reason for less enthusiasm for using intensive chemotherapy is that early death rates are in excess of 10% for patients age 75 years and older and 2-year survival only about 20%. Therefore, novel therapies, which are less toxic and more active than what we currently have available, certainly extremely are needed.

Obviously, the decision as to whether to use intensive or non-intensive therapy is a very complex, multidimensional question. And you can see the reason for wanting to use intensive chemotherapy is that patients have a potentially higher rate of complete remission and complete remission is generally associated with improved survival. However, for older people, you can see that intensive chemotherapy can be weighed down by poor risk AML factors, which reduce the potential for intensive chemotherapy to improve survival and also patient factors.

This assessment is a very difficult one and there are some new prognostic tools, such as the AML composite score published by Sorror and colleagues, and also the Ferrara unfitness criteria recently published in JCO. However, all of these tools are generally a guide, and, in many cases, physician expertise is often relied upon to make the final decision. And this is not an easy one at all.

So, with respect to new therapies, many of us will be aware that there are whole host of new therapies which have recently become available, particularly in the United States. For older patients with AML, there are 2 new therapies specifically approved for this indication and these include glasdegib, the hedgehog inhibitor, and also venetoclax, which targets Bcl-2. You can see that many of the new therapies approved are either targeted therapies towards either FLT3, IDH1, IDH2 or hedgehog and Bcl-2. Or, novel formulations which deliver cytotoxic moiety such as gemtuzumab ozogamicin, CPX-351, and also, the oral azacitidine analog, CC-486.

With respect to glasdegib, this was shown in a randomized, open-label study, compared to low dose AraC to improve overall survival. Low dose AraC plus glasdegib is associated with a response rate of about 28% and also, one quarter of patients manifest dysgeusia, which generally is of low grade, Grade 1 and 2. Because the response rate is about 28% compared to venetoclax plus azacitidine, which produces higher response rates, many people utilize the venetoclax-based option as their first choice for older patients with AML or considered unfit for intensive chemotherapy. This is on the basis of 2 randomized studies, one combining venetoclax with low dose Ara-C called VIALE-C, and the other randomized trial called VIALE-A, which combined venetoclax with azacitidine.

You can see in both of these studies response rate is substantially improved to 66% in combination with azacitidine and 48% in combination with low dose AraC. You can also see the 30-day early mortality rates are not increased by the addition of venetoclax to conventional therapy, and also, that overall survival is substantially improved, particularly in combination with azacitidine where you can see that overall survival has been improved from 10 to almost 15 months.

In terms of which backbone to use, you can see here that in many cases azacitidine in combination with venetoclax produces better response rate, particularly of IDH1 and 2, FLT3 and p53 mutant AML. However, for MPN1 mutant disease, you can see the reverse is true with low dose AraC plus venetoclax producing response rates of 79%.

If you look at the complete responses in the table above, you can see that venetoclax plus azacitidine has a slightly higher rate of true complete remission, being 37% compared to 28% for low dose AraC, but interestingly, the duration of remission is similar with both treatment backbones.

With respect to nucleophosmin 1 mutant AML, there is a long-term follow up available with respect to low dose AraC plus venetoclax. You can see here 3-1/2 year median follow up. And in the patients who received low dose AraC plus venetoclax, the 2-year overall survival in this older AML population was 64% and clearly defining this molecular subgroup to be quite sensitive and quite different to compared to other molecular entities.

At ASH, Dan Pollyea will present a subanalysis of the VIALE-A trial, looking at patients with IDH1 and 2 mutant AML, which makes up about 20% of older patients with this disease. You can see here that quite remarkably, the response rate with venetoclax plus azacitidine compared to placebo plus azacitidine, the response rate is increased by more than 7-fold. And furthermore, 2-year survival is 4-fold greater in the combination with venetoclax. Therefore, indicating that patients with IDH1 or 2 mutant AML, that this combination certainly produces a very substantial benefit in this patient population.

Contrast this with another subgroup of patients with AML who have p53 mutant disease. Many of us will know that a few years ago there was prominent publication in the New England, suggesting that 10-day decitabine might be relevant for patients with p53 mutant AML. However, this abstract, again presented at ASH from MD Anderson, suggest that even though there is a reasonable response rate in patients with p53 mutant disease, you can see that median survival remains below 6 months, even in combination with 10-day decitabine. Suggesting that more improved and novel therapeutic options are required for this patient with a high level of unmet need.

Another aspect of using venetoclax in combination with azacitidine is to combat the issue of myelosuppression. Many of us using this combination regularly now seem to think that perhaps we don’t need the full 28-days of venetoclax in order to get the desired effect. And so, a 21-day initiating cycle is sometimes considered. And it's important to do an early bone marrow around day 21 to verify if the patient has cleared bone marrow blasts. Because if this is the case, then it's quite important to hold the venetoclax and not recommence the next cycle of azacitidine until there’s been adequate count recovery, particularly with the neutrophil count of at least 500 per microliter. And we generally wait up until 14 days for this to occur.

If there is still insufficient neutrophil recovery by this stage, then G-CSF is very useful to try and accelerate neutrophil recovery.

Other important tips for managing these patients is to consider reducing the duration of venetoclax exposure if ongoing myelosuppression is an issue, rather than reducing the venetoclax dose. And so, reducing the venetoclax duration from 21 down to 14, and sometimes to 10 or 7 days, may be required in order to keep the patient on the current cycles of therapy.

Furthermore, if fungal prophylaxis is required, it's important to know that it's important to adjust the dose of venetoclax down, particularly if posaconazole is being used. And we often recommend a dose of 50 mg if those tablets are obtainable. If not, then alternative anti-fungals such as isavuconazole may be utilized, again with a venetoclax dose adjustment down to 200 mg per day.

This here are a number of ongoing questions with respect to targeting Bcl-2 in AML. One question is the role of monotherapy. There are some patients who start off with the combination, develop myelosuppression, and then it becomes a question whether to continue azacitidine alone or venetoclax on its own. However, what is the evidence to support venetoclax monotherapy use?

We recently published data in the JCO, showing that with a 7-day pre-phase, which was utilized in this trial, where a bone marrow biopsy was performed before and after the 7-day exposure, that in patients with MPN1, IDH2 and SRSF2 mutations, you can see that in more than half the patients treated, that a 50% or greater reduction in bone marrow blasts was evident, indicating that venetoclax monotherapy does in fact have some activity in certain molecular subgroups. However, in other molecular entities such as p53 mutant or FLT3-ITD, you can see that the effect was much less.

The second question I’d like to highlight is the potential issue of using venetoclax plus azacitidine in patients with myelofibrosis-associated AML. This does occur in a small fraction of patients with AML, and our experience has been that this combination can be very myelosuppressive. And we generally recommend to not utilize venetoclax in combination with chemotherapy type treatments in patients with myelofibrosis. And this may be because there is a reduced pool of normal progenitors that make it difficult for the bone marrow to recover after ven-based treatment. So we really do need to be cautious in this situation.

The third question, as I mentioned before, is what is the optimal dose of venetoclax in combination with anti-fungal agents? And I’ve shown here a number of different options from a slide courtesy of Naval Daver from MD Anderson. The package insert for venetoclax recommends 70 mg of venetoclax in combination with posaconazole. However, the pharmacokinetic data on a small group of patients would support the use of 50 mg of posaconazole. The issue is that 50 mg venetoclax tablets are not widely available in large numbers, and so, many people have to resort to using 100 mg. And so if that's the case, then consideration of other anti-fungal options such as voriconazole, or isavuconazole, where dose adjustments can be more conveniently adjusted should be considered.

Another question is whether to use venetoclax plus azacitidine in patients with prior MDS who have developed AML and, therefore, failed prior azacitidine. The VIALE-A trial did not, in fact, include this patient population, however, there was a presentation at ASH looking at a study which compared either venetoclax alone or venetoclax in combination with azacitidine in patients that had failed prior hypomethylating agents. And you can see here that there is a response rate of about 40% and also early responses and outcomes appear relatively promising.

Another area that is emerging is the situation where patients fail prior venetoclax in combination with azacitidine. And this is a paper published from the MD Anderson group showing that this really is a very poor risk subgroup of patients. You can see here 41 patients refractory to or relapsing after prior ven plus HMA. And you can see here that the median survival is only 2.4 months. So, people that fail to respond or relapse after this combination, really do identify a very poor risk patient subgroup where new therapeutic directions will definitely be a big question for the future.

I’d just like to close with a couple of points about future directions with Bcl-2 targeting in AML, combining with intensive chemotherapy is obviously a major question to try and bring these drugs into the domain for younger and fitter patients. We’ve recently published the study combining venetoclax with a modified intensive chemotherapy regimen commencing on day 8 after venetoclax commencement. And you can see in patients with de novo AML, response rates are almost 100%. But clearly patients with secondary AML don’t do as well.

At ASH there are 2 abstracts showing the combination of venetoclax with either FLAG- idarubicin or venetoclax in combination with 7+3. And you can see with FLA-idarubicin, quite promising response rates are achievable in newly diagnosed and also in a salvage setting. And obviously, the major question will be the tolerability of these regimens with respect to myelosuppression.

Combination therapies with venetoclax definitely very much looked at by many groups at the moment. There’s been previous reports of venetoclax showing very good response rates in combination with IDH inhibitors. And at ASH Naval Daver with present an update on venetoclax in combination with gilteritinib, which also produces a very promising response rate.

Venetoclax in combination with immune therapies and many other different types of novel combinations are currently in progress.

Maintenance therapy will also be a question for the future. This question will be made a little bit more complicated because of the advent of CC-486, the oral azacitidine analog, which has been shown to be effective in older patients after intensive chemotherapy. And obviously, they’ll be future questions trying to look at venetoclax in combination with this agent before and after stem cell transplant.

Also, combining venetoclax with oral hypomethylating agents, either CC-486 or the decitabine/cedazuridine analog, ASTX727, is also a question which has been actively investigated to try and see if there’s a more convenient way of combining venetoclax with an oral moiety to produce responses in patients with either frontline or relapsed/refractory AML.

I’d like to just highlight 2 cases which I think have useful learning points. This first case is of a 77-year-old man who was diagnosed with AML in 2016. The patient had an unremarkable karyotype and no mutations of note were identified. The patient was commenced on azacitidine in combination with venetoclax, and achieved a good response. Flow MRD was below .1%.

However, after 4 cycles of treatment, this patient developed chest pain and shortness of breath. And imaging identified that the patient had a thoracic aortic aneurysm, which the surgeons were quite concerned about, and basically, they recommended an operation which the patient refused because of his age.

And so this is a question whereby what to do in a patient that has achieved a good response with azacitidine plus venetoclax and will come across, particularly in these older patients who have an increasing number of comorbidities, situations where the decision, is do we push on with therapy and risk complications? Or do we cease therapy? And this is an example that we wanted to highlight because the patient actually decided to cease further therapy; he didn’t want the operation. And we felt it wasn’t safe to continue the ven/aza because of the risk of thrombocytopenia and bleeding if his thoracic aneurysm blew. And you can see here, this patient actually survived 2-1/2 years without any therapy. However, he did ultimately progress with relapse leukemia. We did retreat him with venetoclax plus azacitidine. And again, he responded.

And so, there’s 2 points I’d like to make with respect to this case. First of all, if the patient has had a good response, particularly with MRD being eradicated, it is possible to have quite a durable response without further therapy after venetoclax plus azacitidine.

The second issue I’d like to highlight is that you can actually have second responses by re-exposing the patient to the same regimen, particularly if there has been a treatment interruption.

However, the third issue is that because these patients were obviously very old to start with, by the time they are retreated, we often find that their even further advanced age makes it very difficult for these patients to have a very good and durable second response.

The second patient I’d like to highlight also illustrates a very common scenario that we’re finding with these novel therapies. This is a 78-year-old lady. She was originally treated in December 2014, again with AML, but this time, this patient had an IDH2 mutation.

The patient received low dose AraC plus venetoclax. Had a very nice remission. And after that 15 cycles, she said to me, Dr Wei, I’ve been on this treatment for a long time. I’ve had a really good response. I’d like to get on with my life and do something different rather than coming here every month. Howe about we consider ceasing therapy? And we had a good discussion about this, and I thought she was making quite a rationale decision. So, we decided to cease therapy.

Interestingly, the IDH2 mutation was still present at that time. So what happened to this lady?

Well, you can see that 3-1/2 years later the patient did relapse, but this patient had a 3-1/2-year period of treatment-free life, which was clearly associated with a much better quality of life than coming to hospital every month for treatment.

When she did relapse, we were able to use an IDH2 inhibitor, and again we achieved a response after the fourth cycle with treatment.

Soon afterwards, the patient progressed and we did molecular profiling again, and this time we found a FLT3 ITD mutation, which we have now started treatment with gilteritinib.

So, the learning points for this case are 3-fold. First of all, it is possible to consider ceasing therapy. There are no randomized studies at this present time. However, we are becoming aware of people who have had many cycles of therapy and where treatment can be ceased and where a good quality of life without further treatment can be achieved in some cases.

Second, when patients progress, not just the first but subsequent occasions, it's incredibly important to look for targetable mutations.

As illustrated in this case, the FLT3 mutation was present at relapse and was never present in the first 6 years of her diagnosis and ongoing therapy. And so, never give up looking for potentially targetable options is the message here.

And my last point is that with these targeted therapies, we will get very specific causes of clonal evolution. And so, we cannot generalize our previous experience with chemotherapy, whereby the clonal evolutionary profile will be quite different with these new therapies. And so, we really need to move into this mode of thinking, whereby clonal evolution is adaptable to the treatment and our treatment needs to adapt with the patient.

So, with that, I’d like to stop there and thank you, Neil, for the opportunity to talk to you all again.

DR LOVE: So, thanks so much, Andrew. That was really awesome. Very interesting. Just a couple of questions in follow up. Any thoughts or theories about why the patients with IDH mutations respond so well?

DR WEI: Yes, it's a really good question which is currently under active investigation. Clearly, there must be some elements of the pro-survival profile that are favorable for patients with an IDH mutation. The most rational explanation would be that these patients have a good level of primed Bcl-2, that is Bcl-2 with a BH3 payload already attached, that when you add venetoclax it is unleashed, and that there are no other pro-survival molecules like AMCL1 and BCL-X to impair the response. This has not been shown formerly in patient samples but is something which many groups are looking into at the moment.

DR LOVE: Another thing, and I know you presented some data on this at ASCO, but can you talk about what we know about MRD and the aza/HMA combinations and aza/low dose AraC?

DR WEI: So, MRD by flow cytometry and also molecule techniques has been looked at with respect to venetoclax in combination with low dose regimens. We’ve shown that low dose AraC plus venetoclax results in a very high molecular remission, particularly in patients with MPN1 mutant AML, where almost all patients can achieve a very good molecular response. And this can last for several years.

With respect to flow cytometry, again this has been looked at in both the VIALE-A and the VIALE-C trials. What we do know is that MRD takes longer to go down than it does after intensive chemotherapy. And so, we need to be aware of that and consider that this is more of a slow decline rather than a rapid steep cliff, as you might expect with intensive chemotherapy. And so, we’re still learning what is the optimal time point and what is the optimal threshold for long-term outcome. There is a presentation at ASH from MD Anderson, which shows that MRD suppression to below .1% with ven/aza is associated with improved long-term outcomes. However, I think MRD in AML is more complicated and I think more expanded and more detailed studies are required for us to, I guess make practice decisions based upon MRD.

DR LOVE: What’s known about, and what are your theories about the mechanisms of resistance to the ven combinations?

DR WEI: This is a really fascinating area. We looked at patients in a combined study from Alfred Hospital on Melbourne with MD Anderson, and this was published in Blood earlier this year, and two major things we noticed were first of all, patients with kinase activating mutations were a common cause of resistance, particularly FLT3 ITD and to a lesser extent, RAS mutations. Why FLT3 ITD? Well, I think FLT3 ITD is known to drive the expression of pro-survival proteins, like MCL1 and BCL-X. And so, this might be a natural mechanism of resistance, not just to chemotherapy, but also to chemotherapy in combination with venetoclax.

The other interesting area was that patients with p53 mutation, even though they might have a good initial response, were a common cause of progression. And as I showed you before, the data from MD Anderson would suggest that few long-term survivors are possible with this subgroup.

We also found in our Blood paper that at time of progression patients often had a biallelic p53 abnormality. And so, this suggests that dosage of p53 loss of function was increased at the time of relapse, suggesting that p53 function is a requirement for long-term outcomes and that loss of function is a barrier to the same endpoint.

Treatment Options for Patients with AML Harboring FLT3 Mutations — Alexander Perl, MD

DR PERL: I’m going to go over the role of FLT3 mutations and leukemogenesis and targeting FLT3 in the clinic for optimizing therapy for patients with FLT3 mutations.

As I think the audience knows, FLT3 mutations are common in AML, really among the most common mutations that we see, and associated with increased disease burden and increased disease aggressiveness. Patients with FLT3 mutations often have extreme leukocytosis, high marrow blast percents, proliferative disease, rapid doubling times. And we see these mutations in about a third of patients in AML, maybe a little bit lower percentages in older patients, with the vast majority of the mutations seen being FLT3 internal tandem duplications or ITD mutations. And a minority being tyrosine kinase domains or TKD mutations, most commonly residue D835. In terms of their function cellularly, both of these mutations activate the FLT3 kinase, which is a receptor tyrosine kinase involved in hematopoiesis. In leukemogenesis, this leads to an important contribution to leukemogenesis transforming cells by increasing their proliferation, blocking differentiation; increasing their resistance to apoptosis. And this is due to activation of downstream signaling, which I’ve shown in the cartoon at the left of the slide.

FLT3 mutations often are seen in patients with intermediate risk karyotype. Nearly half of the cases will have an MPN1 mutation, which is important for prognosis. And you can see these mutations in APL, although they’re not prognostic there, and targeting FLT3 is not an important molecular target in APL. FLT3 mutations in non-M-3A AML are not really static from diagnosis to relapse, so, if a patient had a FLT3 mutation at initial diagnosis to know if it's there at the time of either refractoriness to frontline therapy or relapse, you have to retest. Sometimes these mutations are gained at relapse or lost at relapse.

And then, in terms of the prognosis, as I’ve mentioned before, we’ve known for many years that FLT3 ITD mutations have been associated with poor survival, particularly when chemotherapy-only approaches are used. And the use of transplant seems to get rid of much, but not all of the negative prognostic effect of FLT3 ITD.

If you look at FLT3 TKD mutations, it's hard to show much of an effect prognostically, and I’ll tease that out a little bit in subsequent slides as to why that may be. But really, the FLT3 ITD mutation has been the kind of problem child amongst these two mutations in terms of just being a harder nut to crack in terms of finding optimal therapy.

But what we’ve learned since that time is really the cooperation of multiple mutations that best informs prognosis in AML. And the European Leukemia Net has come up with a classification scheme, most recently updated in 2017, probably updates to come relatively soon, that shows that FLT3 ITD mutations, while important for prognosis, do not alone put a patient into a risk category. You have to look at the compliment of mutations here. So, it's the combination of the presence or absence of an MPN1 mutation plus the allele burden of FLT3 ITD, shown here as an allelic ratio of ITD to wild type. And as you’ll note, there’s no effective FLT3 TKD to put you into any of these categories again, it's the other mutations that are present that will do that.

So patients can have a FLT3 ITD and being in a favorable risk category per the ELN if the allele burden is low and if an MPN1 mutation is present. This is relative uncommon, but if present is actually associated with pretty good outcomes. The challenge with this is there’s not a universally harmonized assay to say what a patient’s FLT3 allelic ratio is, and so it's very hard in current practice, outside of the context of clinical trials, to use this for decision-making. And so, in my practice I often will recommend treating these patients with transplant in remission and looking mostly at measurable residual disease if there’s a question of whether a patient with an MPN1 mutation has an optimal treatment response and trying not to steer patients to a chemotherapy-only approach who have MPN1 mutation and FLT3 ITD, unless they are MRD-negative.

There are 3 therapies that have been shown to improve cure rates of patients with FLT3 ITD AML. These include optimizing the induction dose of daunorubicin. So, on two studies of daunorubicin at doses of 90 mg/m2 versus lower doses, improved survival was seen in the FLT3 ITD-positive subset. The US ECOG 1900 study is shown at the left, and there’s a similar result from the UK study, AML MRC17. Transplant has, for many years, been an important part of management of FLT3 ITD-positive patients and in the middle one of the first studies to show improved outcomes for patients treated with transplant in first remission, not a randomized comparison, but a study that had a high allocation to transplant for patients with donors, showed really superior survival in patients who went to transplant with genotypes that included FLT3 ITD, very rich for that group. And then most recently, midostaurin added to 7+3, daunorubicin and cytarabine, has been shown on a perspective, randomized controlled trial to improve overall survival for patients with FLT3 mutations. And this benefit extends both to FLT3 ITD and TKD mutations.

Now, notably on this study a majority of patients went to transplant and there’s a lot of discussion about how relevant is the transplant for this cure versus how much is the midostaurin itself doing, really is it a package story?

And I should also point out that midostaurin is not the only FLT3 inhibitor out there. There’s been FLT3 inhibitors examined for their effectiveness in AML. I’ve given a list here just of 6 drugs that have had extensive testing in AML. Not all of these are FDA-approved; some are investigational, and some are approved but not for AML, as is the case for sorafenib, which does inhibit FLT3 but has never really been developed with that purpose in mind by its manufacturer.

Midostaurin is one of the older FLT3 inhibitors. It belongs to one of the first generation of drugs to target FLT3. It is a drug that as a single agent does not have much in the way of clinical activity in relapsed/refractory patients. And we think the reason why is, while it's quite potent in vitro, when you actually look at its potency in human plasma, measured by the plasma inhibitory assay shown at the bottom there, it's relatively weak. And the drug is not a very potent FLT3 inhibitor in this setting. It probably does inhibit FLT3 somewhat, but it's nowhere near as potent as other drugs that have subsequently been developed.

And as shown at the bottom right, it's also relatively non-selective. These are kinase binding assays that give an estimate of what’s happening in a patient in terms of how selective a drug is for its target. And as you can see comparing midostaurin to the more selective drug quizartinib, midostaurin is quite non-selective. And that's true of lestaurtinib and sorafenib, that these multikinase inhibitors really were FLT3 inhibitors by accident, rather than by design.

The more recently developed drugs, quizartinib, crenolanib and gilteritinib were specifically designed as FLT3 inhibitors as antileukemic agents rather than anticancer agents. And that's really where they have been attempted to be more potent and more selective drugs. These drugs also differ in terms of whether they will inhibit the drug in its active or it's inactive conformation. And that's what referred to as a Type 1 or Type 2 inhibitor. The important thing clinically is the Type 2 inhibitors are not active against the TKD mutations; they only are active against FLT3 ITD mutations. And that's why we can see acquisition of FLT3 TKD mutations as a cause of resistance to sorafenib or quizartinib or other Type 2 inhibitors. So, gilteritinib and crenolanib are Type 1, as is midostaurin.

And the midostaurin pivotal trial I’m going to go into detail here. It's an important study to understand. It was called RATIFY, or C10603 in the US. It was led by the now Alliance, then CALGB, and it was a worldwide study that enrolled over 3,000 patients with newly diagnosed AML in order to find 717 FLT3 mutation-positive patients who were randomized to standard 7+3, daunorubicin and cytarabine, to which 14-day course of midostaurin was added on days 8 through 21, and up to 2 cycles of induction were delivered. Patients in remission received 4 courses of high dose cytarabine with midostaurin and then a year of maintenance. And the control arm was the exact same therapy with a placebo used in the place of midostaurin.

And this study was followed for many, many years and we were awaiting the outcome, which were finally released as a Plenary abstract at ASH in 2015, I believe, and then published a couple of years later in the New England Journal, and showed a 22% reduction in the risk of death in the midostaurin arm compared to 7+3 and placebo.

So a statistically significant improvement in overall survival associated with midostaurin. And this was the first time that this had been shown from a FLT3 inhibitor, really a drug added to frontline therapy for a mutational defined subset in AML. So this became the first of the now 8 drugs that have been since approved for AML. And it's an important advance in therapy, based on this survival benefit. It is a reason that all newly diagnosed patients getting 7+3 should be screened for FLT3 mutations with a test that rapidly turns around a result that will drive this therapy. and this is why I recommend PCR testing for all newly diagnosed patients with AML getting intensive induction chemotherapy. You get an answer within a matter of days and you can choose to add midostaurin based on that, again on induction day 8.

The overall survival benefit was maintained when censored for transplant, and there was no post-transplant maintenance on this study.

Now interestingly enough, despite the lack of post-transplant maintenance on this study, when you look at the outcome of patients who were in the midostaurin arm and went to a transplant in first remission, again largely patients who gotten induction and then maybe a cycle of 2 of consolidation, but no maintenance therapy, these patients actually had better survival, a trend toward better survival than patients in the placebo arm, suggesting that there may have been a qualitative improvement in the induction outcomes relative to 7+3 alone. And we don’t have direct testing of that to say were these patients more likely to be MRD-negative. There are some data that support this that have been published since, but we’re still waiting for those data from RATIFY. But we think that this may be an important determinant of how people do with transplant and a reason to have these patients get what’s really the best therapy front line rather than wait until we know that they’re in remission and add it for consolidation. It may be important to give midostaurin during induction.

The toxicity of the midostaurin arm was very similar to the placebo arm. And there really wasn’t any important difference there, other than a higher rate of rashes, which generally were Grade 1, Grade 2. There is GI toxicity of midostaurin, I don’t want to say it's a totally non-toxic drug. This is mostly seen when the drug is given for prolonged administration and/or with higher doses. So, really in maintenance is where you would see that side effect if you were giving that. And we don’t have a randomized study that shows an improved survival for maintenance, so it has not been approved for that purpose in the US; although, it is approved outside of the US as maintenance therapy. And really, again, that's where we would see the GI toxicity. But otherwise, midostaurin was very well tolerated on this study and GI was the major side effect that was seen.

And what about relapsed/refractory patients? In this study that I was the lead investigator for, called ADMIRAL, patients with either primary or refractory AML or first untreated relapse of AML that had a FLT3 mutation at the time of study entry, were randomized on the ADMIRAL study, either to single-agent gilteritinib, again a second generation potent and selective FLT3 inhibitor with single-agent activity, against whatever was the most appropriate salvage chemotherapy choice for the patient selected prior to randomization. And the choices are shown just under the red box there.

Patients were allowed to go onto transplant, and those in the gilteritinib arm could restart gilteritinib after transplant once they were engrafted as long as they were stable and in remission and without other ongoing transplant toxicities with limited administration. There was no crossover in this study and the primary endpoint was overall survival. And the study was positive, showing a statistically significant improvement in overall survival in the gilteritinib arm. There was a 36% reduction in the hazard risk for death in the gilteritinib arm and, also, a higher complete remission with or without hematologic recovery seen in the gilteritinib arm versus salvage chemotherapy. And the survival benefit was seen both in the high intensity chemotherapy salvage randomized patients and the low intensity patients that were stratified on that.

The major toxicities of gilteritinib on this study were cytopenias, which actually were fairly similar to what was seen in the salvage chemotherapy arm in terms of their severity. The other thing that was seen more commonly in the gilteritinib arm was elevation of liver function tests, mostly asymptomatic laboratory abnormalities. There were some toxicities that were seen commonly with gilteritinib that you should be aware of which include fevers, rashes, often presenting as Sweet’s Syndrome. Less commonly, we did see a differentiation syndrome in some patients and there’s a black-box warning for that on the drug, as well at QC prolongation and lipase elevations and/or rare cases of pancreatitis. Again, rare that we saw serious toxicity with the drug and overall, particularly for the first 30 days of which direct comparison was made. There was no important difference in terms of side effects that looked any worse on the gilteritinib arm, with the exception of the LFT abnormalities.

So ongoing questions for FLT3 mutated patients and their therapy. At present, the NCCN guidelines only recommend midostaurin for intermediate risk karyotype patients who have FLT3 mutation, which begs the question of how well the drug works outside of that setting. And how much is the benefit for patients with FLT3 TKD mutations?

There was a subgroup analysis in the RATIFY study which suggested a benefit, but also very wide confidence intervals — who really benefits from the drug outside of FLT3 ITD-positive, intermediate risk karyotype?

And who really needs transplant these days? Is it everybody? Is it a subset? And if you do go to transplant, should patients get FLT3 inhibitor after transplant?

Then lastly, now that we have newer FLT3 inhibitors that may have more activity as single agents, should we be using those drugs frontline, not just for relapsed/refractory patients?

And one final consideration which is all the data I’ve shown you are for patients fit for induction. What about the large number of patients who are unfit for induction or older patients who may not benefit from intensive induction therapy? If they have FLT3 mutations, how should we treat them?

So we don't actually have data present by karyotype from the RATIFY study, but we do have it by ELN classification, which I’ve introduced previously, that, again, there’s a favorable risk group, which includes core binding factor rearrangements as well as MPN1 mutated patients, who, if they have a FLT3 TKD lumped into this group, or if they have a FLT3 ITD have low allelic ratio.

Now, the data on the left are only FLT3 ITD-positive patients. And actually, everything on this slide will be FLT3 ITD-positive patients.

The adverse risk group are the patients who don't have an MPN1 mutation, that have a high allelic ratio of FLT3 ITD, with intermediate risk being pretty much anybody else, either a high allelic ratio with NPM1 mutation, or a low allelic ratio without MPN1 mutation.

As you can see on the left, midostaurin, in red, was superior to placebo, in blue, pretty much across the board and there really doesn’t seem to be a genetic group here that sees particularly more benefit than another. So the recommendation for intermediate risk, I think now that we have this additional data, that may change. To my evaluation of these data, I think midostaurin just benefits patients with FLT3 mutations, certainly FLT3 ITD mutation across the board.

Now, what about the benefits of transplant? Now I will caution that this study was not powered to look at subgroups this small, but it is interesting to note that if you look at the favorable risk patients, again ELN favorable here, which includes NPM1 mutated patients who have low allelic ratio, in the midostaurin arm, which is shown on the right, and again here the coloring is for transplant, in blue, versus consolidation maintenance in red, midostaurin and chemotherapy are completely overlapping here. There’s not an obvious difference here. Yet, in the placebo arm, it looks like transplant does better and that's pretty much true across the board.

So the differences in terms of the transplant outcomes seem to be largely decreased in the midostaurin arm with the exception of the adverse risk group. And again, I don’t think there is enough statistical power here for us to say change our therapy only transplant adverse risk group, but it suggests that if you would answer this question this might be a way forward to say we can better allocate our patients to transplant once we know questions like are we becoming MRD-negative in these patients? Is that really who’s going to benefit relatively less by transplant? Is it really the MRD-positive patients who need it? Or, really, do we just need to think about optimizing the FLT3 inhibitor, getting more people MRD-negative, etc?

It suggests that there may be more and more patients who do not need a transplant, though I don’t think we have the final word on this yet from this study. But it's very enticing data from this study.

One thing I would point out also, is the ELN adverse patients had a relatively poor induction success rate, and that was true in both arms relative to the other subgroups. So these patients really have a high unmet need for better therapies.

Now what about the TKD-positive patients? This was recently published data from, again the same study, and if you look on the right the overall survival and event-free survival are shown for FLT3 TKD patients and on the left for the FLT3 ITD-positive patients for comparison, on the left broken down by ELN risk and on the right you can see the outcomes in the TKD patients look superb in the midostaurin arm in particular for overall survival at least.

And I think the reason why is because there’s actually a strong enrichment for ELN favorable in this group. And that largely is why, if you look historically at how FLT3 TKD patients did, and we would say it's not a negative prognostic marker, I think largely what you’re looking at is groups with favorable risk disease. Well, what happens when it doesn’t?

If we look at the green group here, which is — the coloring scheme is different — these are different publications, but on the bottom on the left there, the green curve is actually the ELN intermediate and adverse risk FLT3 TKD group patients. And they did pretty poorly on the study.

So I think we don’t have good data to say that those patients do all that well. And we should look at our most aggressive approaches for anything but favorable risk genotype. So, still consider these patients for transplant and weigh that into your decision-making. There seems to have been a benefit for midostaurin overall on the study, but, again, very wide confidence intervals. And it looks like the benefit largely is more in event-free survival than in overall survival in final analysis. But the benefit on the point estimate for the hazard ratio does favor it.

Now I mentioned measurable residual disease a bunch of times on this talk. I just want to show you those data briefly with one of the better studies to make this point clear, coming from the AML17 study from the UK, which looked at patients who were NPM1 mutated after 2 courses of induction to say, were they MRD-negative in peripheral blood by RT-PCR assay? And using that as a cutoff for MRD-positive or -negative, patients strongly grouped into 2 risk categories for relapse and for survival. Patients who were MRD-positive did much worse, had a very high relapse rate. But interestingly enough, even in patients who had FLT3 ITD-positive disease, if they were MRD-negative coming out of induction, their relapse rates were about 30% and that's shown on the left in black.

So, this study is important because it did not include a FLT3 inhibitor and it suggests if you adequately treat your patients with a good enough treatment response, you can get good outcomes. But it begs the question, if we had added a FLT3 inhibitor would we have done even better because we might have more patients in the MRD-negative group? And if so, what would be the best drug to achieve that?

Now one might also look at this and say, well, you can just take the patients who are MRD-positive and transplant them and you’ll make up the difference. But we don’t have a lot of data on really how well that does. And the available data that we have says their outcome’s quite poor. I will show you one study done by the BMT CTM that was looked at retrospectively with next-generation sequencing done with error correction by Chris Hourigan and colleagues, and this study was agnostic to mutational complement. So the researchers didn’t know what the patient’s had for mutations at initial diagnosis. But if they detected a FLT3 ITD at the time of transplant, all the 10 patients in which they found that died within a year. And you would expect maybe 2 out of 10 deaths from transplant. So much worse outcomes in this study with detectable FLT3 ITD mutation. Which again, argues for the idea that maybe post-transplant maintenance is a good idea. So how does that look?

Well, there’s now actually multiple studies showing a benefit for post-transplant maintenance. Two randomized studies have read out, one from Germany and more recent one from China, both showing the same result which is better protection from relapse and either trend or statistically significant improvement in overall survival. So this is a therapy that I generally do recommend in the absence of what ultimately, I think, is the most definitive way to answer this, which is to take patients not 30 to 100 days post-transplant, and transplanted maybe in first remission or maybe second or out of remission, but transplanted in first remission in what’s really the relatively lower risk group. And hopefully a group that we can cure with no additional therapy to answer the question of really whether we should be giving maintenance.

One of the problems with the SORMAIN study which, otherwise, shows a very clear protection from relapse, is that a significant number of the patients enrolled to the study had disease at the time of transplant. And as you’ve seen the prior slide, that's a very strong predictor of relapse. So I think this is a much higher risk group to be looking at, and the benefit of sorafenib, at least in that setting, I think would be more or less expected.

A phase study of gilteritinib versus placebo has completed enrollment. That study enrolled prior to transplant included MRD testing before transplant and then randomized patients post-transplant to 2 years of gilteritinib or placebo. And we should have those results in probably about a year or so. And hopefully we’ll find out whether that improves survival similarly.

Just to finally put that in context with the available data from midostaurin from RATIFY, which just shows you basically where our base line is now. On neither of those prior studies, either from China or Germany, had patients had frontline TKI. So again, a relatively lower risk group that we're studying on the CTN study, the MORPHO study of gilteritinib.

Lastly, what about bringing all of these new drugs to frontline? What would be the best FLT3 inhibitor we could give from day 1? Gilteritinib is an impressive drug in the relapsed/refractory setting, but would it work even better as a first line therapy?

So the feasibility of this has been tested and has been shown. This is a study that will be updated at ASH by Keith Pratz.

This is the last presentation that we had of these data. And basically, the take-home point from this is adding gilteritinib to 7 + 3, a number of different anthracyclines that we used, a number of different days of therapy and even dosing of the anthracycline was quite tolerable and had similar toxicity to 7 + 3 alone. Up to 3 cycles of post-remission high dose AraC were given with gilteritinib. Patients could go onto transplant, go back onto maintenance thereafter.

And the data that will be presented are the final data from this study, but at least the induction success, which has been presented before, shows responses are pretty much across the board with only 2 patients out of, I believe 33 treated that we’ve presented before, had induction failure. So, very high remission rates with this approach.

Perhaps a little bit more in the way of incomplete count recovery, but these, in general, did not limit patients to going on to consolidation. So very promising regimen and again, we will have comparison from future studies to 7+3 plus midostaurin and high dose AraC plus midostaurin, both a US study and a European study will answer that question.

There’s also a quizartinib frontline study compared to placebo called Quantum-FIRST. That has completed enrollment and should read out in about a year or so.

And lastly, a study testing crenolanib against midostaurin, a similar design to the gilteritinib study. Again, active control, Phase III trial. And that is still enrolling at present.

In about the last minute that I have here, I’m going to just talk briefly about patients who are unfit for induction chemotherapy, to show the VIALE-A data with venetoclax and azacitidine, a very impressive result showing that a drug targeting Bcl-2 added to hypomethylating agent not only more than doubled response rate, but substantially improved overall survival in a population that has been very difficult to improve survival with any combination regimen. This study showed an unequivocal improvement in overall survival with venetoclax plus azacitidine, but does that extend to patients with FLT3 mutations?

And the answer from this is it looks like it does. So I’ve highlighted in the FLT3 mutated group the response rate which is pretty much the same as the entire study, 72% CR/CRi rate in the venetoclax arm. And the point estimate for the hazard ratio is exactly the same as the entire study population. And, while the confidence interval intersects 1, the study is not powered to detect these differences in these subgroups. And I think across the board, the consistent pattern is that there’s really not an identified molecular subgroup that does worse with venetoclax. And certainly, the FLT3 mutated patients relative to azacitidine, seem to do better. But what it doesn’t answer is should we be using a FLT3 inhibitor front line? And I think the problem is right now, we don’t have enough data to say that that's better.

So my practice has been to use venetoclax/azacitidine for these patients as long as you can safely administer the venetoclax. And that means the ability to cytoreduce them to appropriate levels that were studied on this trial, which required a white count less than 25,000. So, in patients who do not have rapidly proliferative disease that can’t be controlled with hydroxyurea, I think it is more than appropriate – it is proper to treat these patients with venetoclax/azacitidine. In patients who cannot be cytoreduced easily, I am using FLT3 inhibitors in that setting. It's off label, but I think it's reasonable. And again, there’s data with midostaurin, data with sorafenib, data with gilteritinib. We don’t have a preferred approach at this point. But at present, that is not my preferred strategy. I would use the data from VIALE-A to guide your therapy.

So, just in conclusion, therapeutic advances have improved survival for FLT3 mutated patients. And FLT3 inhibitors are an important part of that improvement.

Intensively treated patients, regardless of their genotype, I think should receive midostaurin. I don’t think this matters based on karyotype. I don’t think this matters based on the type of FLT3 mutation or other co-mutations. The possible exception is patients with core binding factor rearrangements who would receive gemtuzumab ozogamicin. Because we really don’t know in that setting what’s the best third drug to add. And in my practice, I’m adding gemtuzumab, honestly. It's very rare to see FLT3 ITD in a core binding factor rearranged leukemia. And I think the more important question is just are you achieving an MRD-negative state at the end of induction, testing after — ideally after 2 courses if you’re not in an MRD-negative state after 1.

And randomized, controlled trials will clarify second generation tyrosine kinase inhibitors are superior to midostaurin in frontline 7+3-based therapy. We don’t have those data yet, so I would not do that outside of trials.

Transplant is still an important part of the management of patients with FLT3 mutations, both ITD and some patients with TKD, particularly MRD-positive patients who I think really should look at maintenance strategies post-transplant. And for patients who are FLT3 ITD-positive, with the possible exception of patients with the lowest allele burdens who are MPN1-mtuated at initial diagnosis and become MRD-negative at end of induction.

Post-transplant maintenance I think is an established modality now, but yet, it's hard to give sorafenib in this setting. The toxicity seen on the SORMAIN study has not generally been my experience with that drug and I think until we have the data with gilteritinib, we’re still waiting to see what the best strategy for these patients is. And that may ultimately be the way to go.

And venetoclax/azacitidine I think should be preferred approach for unfit patients with FLT3 mutations.

So my first case is a woman that I recently treated who is 66 years old and presented with a couple of months of recurrent tonsillitis, arthralgias and bone pain, high fevers, like 103, 104 fever. Wound up getting admitted to the hospital with leukocytosis, concern for a pharyngeal abscess. She actually didn’t have any drainable fluid collection. Was very anemic. Had thrombocytopenia. But otherwise, good organ function. Normal heart, ejection fraction. Normal LFTs and normal renal function.

She underwent a bone marrow biopsy, which was hypercellular with prominent dysplastic changes in the erythroid and megakaryocytic cells. Monocytosis and left shifted myeloid cells. The aspirate was hemodilution, but 11% blasts were counted by flow. There was no manual count done, and her karyotype was notable for a trisomy 8.

She had her bone marrow biopsy read at as CMML-2. And later on, her sequencing came back showing mutations in NPM1 with a variant allele fraction of 42%, a PTPN11 mutation and FLT3-TKD mutation, D835Y at an allele frequency of 10%.

She has baseline diabetes and some obesity. She is retired, but otherwise fit and had a performance status of ECOG PS1.

So, what would be the best treatment for a woman like this with the possibilities including single-agent azacitidine? Venetoclax plus azacitidine? Azacitidine and midostaurin? CPX-351 and then a transplant? Or 7+3 and midostaurin? And she was sent to me with this question of really, what should we treat her with?

So I actually chose to re-aspirate her because we really didn’t have a diagnosis in my mind. You can’t make a diagnosis of AML from flow alone when you have a blast percentage of 11%, in particular you don't want to use the flow percentage to tell you whether the patient meets a cutoff at 20% blasts. And with the mutational complement, including MPN1, I was very suspicious that this actually a de novo AML. And indeed, that's what we saw on the aspirate, but quite honestly, even she had shown 11% blasts on a 200 cell or 500 cell aspirate count, I probably would treat her as AML with this presentation, a de novo presentation, well until suddenly she’s unwell.

Ultimately, what we found were constitutional symptoms, not so much just infection, but recurrent infections that would limit her from getting back to full activities and likelihood of relatively low response rate to azacitidine alone, as well as potentially better outcomes with induction chemotherapy. And there’s no head-to-head study that's says what the right thing to do here is. But from case series, including the one I’ve referenced below, which is a 30-something patient series from MD Anderson, which showed better outcomes in patients treated intensively with NPM1 mutated MDS that show, to my satisfaction, that these patients really should be treated as AML.

Now, she has a FLT3 mutation, which also gives us the option of adding a FLT3 inhibitor. And with the 22% blasts, I use that as evidence of sufficient data to extrapolate the data from RATIFY to an older patient population.

There is a study from Germany, from the AMLLG using a very similar regimen to what was tested on RATIFY 7 + 3/midostaurin. Slightly different dosing schedule of the midostaurin, but otherwise had basically the same results in terms of a high remission rate, strong protection from relapse and that study was a Phase II, so compared against historic controls, but showed much better outcomes, including better outcomes in patients age 60 to 70, in what had ever been seen with that Cooperative group with prior therapy lacking a FLT3 inhibitor.

I gave her basically the same therapy that was done on RATIFY with a dose modification for her cytarabine for her age. Her NPM1 mutation was present, in remission, after 1 cycle of induction. I retested her after first consolidation and she’s now MRD-negative. And we’ve continued with consolidation high dose AraC.

We probably will not transplant her if she has had an MRD-negative result. She’s risk-averse to transplant and at her age I don’t think that's outrageous. And we’ll consider oral azacitidine maintenance. I don't know what the right answer is as to whether we should use oral aza maintenance, midostaurin maintenance, no maintenance in this setting, if she’s somebody that — she might reassess the benefits of transplant at a later date should we relapse. And on the QUAZAR study, oral azacitidine did improve survival in terms of medians, but it didn’t clearly improve the plateau ultimately. And so, whether this is more likely to cure her, I can’t say.

So, with that unknown, we’ll consider this. We’ll cross that bridge when we get to it. But I likely will give her oral azacitidine.

Now the second case I’m going to present is a somewhat complicated case from a few years ago, someone I’m currently treating now, but first presented in 2017 with dyspnea on exertion for only a couple of weeks’ duration. And she came in and saw her primary care physician who drew some labs and was shocked to see her white count was 300,000. Sent her immediately to the Emergency room. She was emergently leukapheresed for a new diagnosis of acute myeloid leukemia with myelodysplasia related changes on her marrow biopsy, and a high risk karyotype with an isochromosome 17q. This leads to deletion of the p53 locus.

And not surprising, given her very high white blood count and aggressive presentation, she had a FLT3 ITD with an allelic ratio of .05. So again, high allelic ratio. And no other mutations on an NGS panel. So, just the karyotypic abnormality in FLT3 ITD.

We gave her 7 + 3. This was prior to the approval of midostaurin. She was refractory after 2 induction cycles. And her genetics at the time of primary refractory disease were unchanged from initial diagnosis.

She was enrolled on a Phase III study of a FLT3 inhibitor versus standard chemotherapy. She was randomized to the control arm. She didn’t respond. This was not atypical on those studies. Instead, I gave her sorafenib and azacitidine, which was effective at controlling her disease, clearing peripheral blasts, knocking down the marrow blasts to less than 10% and basically taking her from being transplant-ineligible to being transplant-eligible. But I would point out that she was still FLT3 ITD-positive by PCR at the time of transplant.

She underwent a myeloablative transplant from her HLA identical sibling. She actually did very well during transplant and grafted with full donor chimerism. Had no detectable FLT3 ITD in her marrow done about a month or 2 after transplant. Was started on post-transplant sorafenib maintenance on day 50 and did well for 2 years.

However, 2 years post-transplant, she had developed chronic graft-versus-host-disease and she needed treatment. And because of symptoms she was having, she had a pulse of steroids. And started on photopheresis during which time we were able to ween down the steroids substantially, and during that time her leukemia relapsed while still on sorafenib. We repeated her genetic testing and she remained FLT3 ITD-positive. She did not have anew FLT3 TKD mutation, which can occur on sorafenib.

And so the question is what do you do with patients who have a relapse like this? She is far out from prior transplant, but she does have GVHD, so she might not tolerate intensive chemotherapy. Is there any molecular test or other test right now that would alter our recommendation what we would do? And are there resistance mutations we should look for outside of FLT3 TKD mutations here?

And what should be the right therapy for her? I’ve shown data that gilteritinib is superior to salvage chemotherapy, but relatively few patients enrolled on that study, either with prior sorafenib or 2 years out from prior remissions.

So there aren’t actually a lot of data to guide us here.

I’ll talk briefly on resistance to FLT3 inhibitors, which is something I studied. And I would point out that after transplant immunologic loss of graft versus leukemic effect may be a relatively common cause. And patients can respond to donor leucocyte infusions and/or tapering of immunosuppression, though it's very hard to get patients to go into remission alone if they have a FLT3 ITD mutation because they have such proliferative disease. They need something to control their clone.

With this in mind, we wanted to give her something that would be active against her relapsed disease, not knowing whether she would respond to a different FLT3 inhibitor. We know that patients can develop new mutations in FLT3, commonly during sorafenib they develop TKD mutations, such as D835 mutation. Rarely on gilteritinib we can see a gatekeeper residue, F6901L. This is structurally analogous to the T315I mutation that you see with CML and limits activity of Bcr-Abl inhibitors. So, it's structurally the same and functionally the same; it limits the activity of gilteritinib, although higher doses of gilteritinib may work against it.

More commonly what we see are clonal evolution with new off-target mutations, which, if we look across the board mutations in RAS or in the RAS pathway, which is downstream of FLT3 are the most common mutations we see. And we actually can see patients with Bcr-Abl mutations. And so, I always look for Bcr-Abl in terms of seeing a karyotype or FISH, to see if that's popped up.

Lastly, patients can lose their FLT3 mutation altogether, and that’s actually one of the more common ways to become resistant to midostaurin in terms of FLT3 targeted therapy as it's just no longer FLT3 mutated. And what to do in this setting really is uncertain. We don't have any trial data to guide us to say what the right answer is. I can tell you what we did for this patient.

Knowing that only about 10%, 12% of patients on the ADMIRAL study had a prior TKI and only half of these had sorafenib, we did not have enough luxury of time to see if she had a RAS mutation and even when we see RAS mutations in this setting, they don’t necessarily predict resistance to gilteritinib.

This patient enrolled on a study combining venetoclax and gilteritinib. And she rapidly entered a second remission. And she has remained in that remission now more than year. It's not to say she hasn’t had other issues along the way; we have had to treat the graft-versus-host-disease. She’s had some infectious complications of that that have been actually more of a problem for her than the leukemia. But in terms of leukemic control, fingers are crossed, but we’re 14 months out on this therapy and I literally saw her yesterday in the clinic, and she remains with normal blood counts, transfusion-independent, with full donor chimerism, doing very well.

So, we can get prolonged responses and survival to salvage therapy with some regulatory with new therapies. And I think that gilteritinib is an excellent drug in this setting, but combinations are going to be important in terms of salvage therapy and as we think of how best to use gilteritinib. It’s not a magic bullet for FLT3 mutated AML. And alone, it is not a cure. That's why I think coming up with rationale combinations is very important and then moving those combinations early in the course of therapy to design the best frontline combinations for patients. And, hopefully, ultimately, decrease the intensiveness of treatment, so that while at present we’re still relaying on transplant heavily for patients who really need it, I hope fewer and fewer patients with FLT3 mutations require transplant in their therapy.

DR LOVE: So, it seems that she kind of had primary refractory disease, how often do you see that? And were there any suggestions, clues, at the beginning that you might see that?

DR PERL: The major thing that predicted that was her karyotype, which was a high risk karyotype. So she had a isochromosome 17q, which is a deletion of the p53 locus. We don’t always see that with a complex karyotype, but it fits into the ELN adverse risk group. And so, the data from RATIFY said that the CR rate in that group was around 55% to memory. So we see that relatively commonly, unfortunately. And whether her presenting white count at 300,000 predicts that or not, I think is anybody’s guess. She certainly had 3 factors kind of arguing for more aggressive disease, which is presenting in leukostasis, high risk karyotype and high allele burden of FLT3 ITD. So, kind of the worst to the worst in terms of leukemias that we see.

DR LOVE: So, another question I was going to ask you is you talked about the patient unfit for intensive chemo but with a high white count, so, therefore, not really a good candidate to start with aza/venetoclax or HMA/venetoclax, where you start out to cytoreduce them with a FLT3 inhibitor. So, is that just you cytoreduce them and then you go to azacitidine/venetoclax? Or you just keep them on the FLT3 inhibitor?

DR PERL: This is more the art than the science here because we don’t have a lot of prospective data. I can tell you what I’ve done, which is if I can cytoreduce patients with
hydroxyurea and they respond, they usually respond rapidly. But every so often when starting patients who we give hydroxyurea to, and their mouth just gets horrific mucositis and they can’t swallow pills. And that limits us using an oral agent going forward. So we don't want to really overdo it that setting. We want to be cautious.

FLT3 inhibitors cytoreduce very quickly and just in a few days they can get your white count down from something very high to single digits. Gilteritinib, typically you clear your peripheral blasts within about 3 to 4 days of starting that drug and it's very rare that we see any circulating blasts a week into therapy. In fact, if we still have a white count that's over 5 a week into therapy, I’m worried that the patient isn’t going to respond at all. So that can be a little bit of a predictor of how patients are going to do.

But I give hydroxyurea across the board to these patients and see how they do. If the white count comes down then I think about giving them ven/aza, as was done on VIALE-A. If they don’t come down, I don’t waste a lot of time. You can use leukapheresis in this setting. But it's not an entirely pleasant experience to get this huge catheter in your neck, 80-year-old patient. You have to be very careful with these kinds of things. You want to be very gentle in patients who are frail, who are very fragile and who can be really tipped over. You don’t want to undertreat them though because they can be quite sick with leukemia. But the remarkable thing is I really have a number of patients who presented quite sick with leukemia and we put on FLT3 inhibitor plus azacitidine, they rapidly went into a complete remission, because transfusion-independent, within about a month of starting therapy and maintained that response for a year, 2 years, 3 years, and needed no additional treatment. And really had minimal supportive care needs once they were responding.

So it can be a very gratifying therapy when it really works. But we don't have big studies of it. And so, it really becomes anecdotal medicine more than anything else. But that is a very reasonable alternative.

There’s a big study comparing aza/gilteritinib to azacitidine alone that's mostly enrolling ex-US because I think very few people will treat those patients with azacitidine alone here. But that will be answered by a trial that's currently enrolling.

DR LOVE: So a few random questions, CPX + FLT3 inhibitors? What do we know about it? Would you utilize it?

DR PERL: Not a lot yet. I think it's reasonable to try in the context of a trial. Lots of reasons to think it could work. Again, I think one of the benefits of CPX is just the pharmacokinetics of the anthracycline exposure are different. And we know that small differences in daunorubicin concentrations or duration of exposure can impact induction success and survival. And so, I was not surprised to see that a head-to-head comparison of CPX against standard 7+3 led to better survival there. But will there be additional benefit from the FLT3 inhibitor? I don’t think we know. And it may be that you’ve already eaten up all of that benefit by just using a more intensive front line therapy, but it could be better. The one thing I’d worry about is we see a little bit slower neutrophil count recovery when we add a FLT3 inhibitor. And we also see longer nadirs with CPX. I don’t know what would impact the safety. And that's why I’d be cautious about doing that outside of the trial.

But, in general, midostaurin doesn’t add a lot of toxicity otherwise. So, I think that should be a pretty easy trial to do.

DR LOVE: So I can’t believe I never thought of this question until now, but it's so obvious, but is midostaurin a TKI?

DR PERL: Yeah, that's it's mechanism of action.

DR LOVE: So why does it end in “staurin” instead of “in”?

DR PERL: So it's a derivative of staurosporine, which is one of the original small molecule kinase inhibitors out there. 7 benzoyl staurosporine is its chemical name. So that's where the name comes from.

It was developed originally to inhibit protein kinase C. And that's why it was known chemically as PKC412.

DR LOVE: Cool. Another question. Relevance of allele ratio?

DR PERL: One of the first ways to assess for the aggressiveness of FLT3 ITD-positive AML was to look at the allelic ratio, and this dates back probably 20 years or so, to the first descriptions from large Cooperative group studies of FLT3 mutations, how common they were, how aggressive they were. And the big question was, was 1 copy of the mutation versus no wild type predictive of a more aggressive disease? And there wasn’t a quantitative assay back then, you were looking at just ethidium bromide staining gel on a PCR reaction. So they would get kind of a guesstimate. And a later advance was to put a fluorophore on capillary electrophoresis as a read out and then you could actually quantify it. And when you did that you could get a ratio of the peaks of the PCR products that were amplified, and you could say numerically how much mutant is there and how much wild type was there. And that gave you really 3 things.

One was it told you with much greater precision, do you have 1 ITD or more than 1 ITD, because you could see little peaks more easily. And you could better say that there is this much mutant signal and this much wild-type signal and compare the 2. And what became clear the scene on multiple studies was the more mutation you saw that was ITD, the worse the patients did. And so, that was a consistent finding across studies. But what wasn’t known at this time, which was like 2001, was any of the other molecular genetics of AML. NPM1 mutation was first described in 2005. The first sequence of an AML genome was 2008. We really knew almost nothing about other mutations in AML. So there really wasn’t any way to say what’s the best way to prognosticate? All we knew was FLT3 ITD present, absent. And then when NPM1 came along, present, absent. And now we're much better at making these assessments.

The problem is we almost have too much data now with these NGS panels and the NGS panels aren’t linear in terms of the amount of ITD signal. So, a PCR is actually better here because the PCR products amplify more linearly than you see on the next-generation sequence. So you can’t really use an NGS panel to say what the allelic ratio is. At least, I would not recommend using it because it's not accurate. If all you see is mutation, that’s easy. And if it's anything but that scenario, you really can’t say, in particular, because large ITDs, meaning the amount of sequence that's inserted into the gene causing the mutation, tend to be undercounted by next-generation sequencing. And that actually can happen in PCR, too. But the bottom line is, it’ll make it look like it's a lower allele burden, even when the patient could have a high allele burden. And so, you would underestimate the risk of that leukemia by the next-generation sequencing, and undertreat the patient if you said don't get a transplant.

DR LOVE: So, for the doc in practice, is it either it's there or not, allele ratio doesn’t change it —

DR PERL: It's there or not in my estimation. I would not use this as a decision, with the possible exception of, if you have somebody that you think probably should go to transplant and they go to the university center and they re-estimate this and they look at the primary data and they're comfortable with this and it's on a validated test and they’ve got a transplant opinion, or it's not a great transplant donor search, there are reasons we might not transplant a patient who’s FLT3 ITD-positive. But I always recommend telling these patients for an evaluation. Because more often than not we’re going to say, yeah, we think they should be transplanted, still in this day and era.

If you look at the data that I just showed from RATIFY, the ELN favorable patients do the same in terms of chemo versus transplant, but they don’t do worse with transplant. And if we can’t tell who these patients are, we’re cautious to say we want to undertreat this group.

DR LOVE: So, I was working with another faculty member, Dr Stein, and he was talking about IDH inhibitors and he used the term “differentiation therapy,” which I hadn’t actually heard anybody use the term, although it kind of made sense to me. So, maybe it's a little bit far from field, but I am curious what your thoughts are about that term. And he also brought up something that you mentioned, too, in your talk that you can see differentiation syndrome with FLT3 inhibitors. So, first of all, do you like that terminology for IDH inhibitors? Do you think it accurately describes how they work? And to what extent do you think that FLT3 inhibitors function through differentiation?

DR PERL: First off, thanks for the question. This has been an active area of research in my group and something I’ve published on previously.

So, first of all, with respect to IDH inhibitors, I absolutely do think they induce differentiation of the leukemic clone as a major mechanism of response. And we know that, based on studies that were done during clinical trials of IDH inhibitors that show that the allele frequency for the IDH mutation was maintained when patients responded to the drug. So when their blast percent of the marrow dropped from, say, 50% to less than 5%, a patient goes into remission, the allele frequency for the IDH mutation remained exactly the same as pre-therapy. and it's not to say that's the only way that patients can respond. But it's the predominant way that patients respond to that group of drugs that target IDH mutation.

Now, it turns out that we had actually seen the exact same thing in studying patients with FLT3 mutations. Mark Levis at Johns Hopkins and I had collaborated on a study extending work that he had done in his laboratory looking at the effects of FLT3 inhibitors on the differentiation block that occurs in leukemogenesis. And prior preclinical studies done in the laboratory had suggested that FLT3 inhibitors would release that block and cause the blasts to turn into neutrophils in the laboratory in a test tube.

And we actually saw the same thing in the patients. We collected blasts from the patient. We collected neutrophils from the patient prior to therapy. and then we did the same thing after they were on the FLT3 inhibitor. At this point, they didn’t have any blasts, they just had neutrophils. And the neutrophils prior to therapy with the FLT3 inhibitor had no FLT3 ITD mutation. And the neutrophils after therapy were chock full of it and the blasts were absent.

So we at least know that the neutrophils in those patients that we saw after they responded to the drug were derived from the leukemic clone.

And so, the question then is, if the drug can induce differentiation of a leukemia cell into a terminally differentiated neutrophil, is that going to make the patient sick? We know that that happens in acute promyelocytic leukemia in a subset of patients treated with all-trans retinoic acid, with or without arsenic, and that can cause significant morbidity and in some cases mortality, and it has to be treated with corticosteroids, in some case ventilatory management if they get really sick, and very careful management of the patients because they can get so ill.

We do actually see a differentiation syndrome with regularity to IDH inhibitors. And we see it less commonly to FLT3 inhibitors, but we do see it.

We do see Sweet’s syndrome with some regulatory, which, I’m not sure if that's a differentiation syndrome or just that's related to the patient population or something to do with the treatment. But it has some of the features that we see, again, neutrophils in the skin or other organs, inflammatory symptoms, high fevers. And if it's just a skin rash I may treat that with topical steroids, but if patients develop pleural effusions, paracardial effusions, organ dysfunction, which I have seen on rare occasions, single digit percentage, we use corticosteroids in those patients. And there is a black-box warning on gilteritinib for that syndrome. So it's important to recognize it. It happens more commonly with IDH inhibitors. So, yes, it's important clinically.

DR LOVE: So, one final question. In what situations, if any, do you utilize azacitidine/HMA in MDS?

DR PERL: Yeah, we do use it mostly in patients where we want to get the patient to transplant and the rate-limiting step is they still have a persistence of blasts or an increase in blasts while on azacitidine therapy. We saw them — initially they had MDS EB1, 6% blast, but we don't have donor. And now we have a donor, we’re ready to go to transplant. We do a marrow prior to transplant and they’ve got 15% blasts and we’re worried both because they have more blasts now than they had then, but, also, it looks like disease might be getting more aggressive. So that's a case where we’re worried that they’re evolving to leukemia and we’ve kind of used a leukemia focused approach. And the right answer there, we don’t really know.

Should we use just standard AML induction, or should we just say use aza/ven because that is actually, in the right patient, AML induction? And that has been my approach in this setting usually as a bridge to transplant. The marrow blast clearance rates are actually pretty good. The problem with the approach is that azacitidine/ven has not generally led to full count recovery. So if you’re trying to use it for a long-term MDS focused management of these patients, while it does have activity in patients without response or with progression on HMA alone, and that's one of the important things to know about the study that's been presented, it rarely leads to these patients becoming transfusion-independent and that can be a rate-limiting step. It's also important to limit the number of venetoclax doses, as was done on the study. It's not like AML, where you would give 28 days and then potentially 2 cycles of it until you achieve remission if the patient at day 28 is not in best response. Because you’ll make the patient incredibly cytopenic. And that could really increase their infection risk and supportive care needs.

So, yeah, there are times when we use it largely in the bridge to transplant setting.

Management of Newly Diagnosed and Previously Treated AML with IDH Mutations — Eytan M Stein, MD

DR STEIN: So thank you for inviting me to speak today. We’re going to be talking about the management of newly diagnosed and previously treated acute myeloid leukemia in patients with mutations in isocitrate dehydrogenase.

So as background, I think it’s important to recognize that when we think about treating acute myeloid leukemia, up until very recently the only real way we had to treat patients who did not have acute promyelocytic leukemia was with therapies that caused cell death. Now we all know that in patients with acute promyelocytic leukemia, that special subset of acute myeloid leukemia, when you give those patients all-trans retinoic acid, and these days we now give them arsenic trioxide as well, you can get myeloid leukemia cells that are blocked in their differentiation at the promyelocyte stage of development to differentiate into normal, healthy cells. And the power of this is that, as the audience all knows, when we give treatment with differentiation therapy for APL, in this case with arsenic and ATRA therapy, what ends up happening is that you can actually get about 100% cure rate.

So the idea that differentiation therapy is extremely powerful and has a huge place in the treatment of patients with acute myeloid leukemia is really proven by the photomicrographs you see here, which is showing some of the first data that showed how promyelocytes can turn into healthy, mature neutrophils in APL. And I think the question has always come up whether there are other subsets of acute myeloid leukemia where patients might also benefit from differentiation therapy.

This is what’s called a circos plot looking at a large clinical trial done by the Eastern Cooperative Oncology Group to understand the genetic heterogeneity of the patients that were treated on that trial. And let me just walk you through how you read these things if you don’t know. If you look at the circumference, the circumference are all the different mutations that the patients in the cohort had. So you can see that there were many patients in red who had DNMT3A mutations. There were also patients who had IDH1 and IDH2 mutations, et cetera, et cetera. And the ribbons that connect the different mutations show the relative frequency of cooccurring mutations.

So for example, DNMT3A-mutant patients very commonly will have mutations in NPM1. However, they may not have mutations in something like KIT. Similarly, you can see the patients with IDH1 and IDH2 mutations also commonly have mutations in NPM1, but they never had mutations, at least in this cohort, in TET2.

It’s also important to realize that although the most common mutations that we see in patients with acute myeloid leukemia are FLT3 and NPM1, if you combine the mutational frequency of patients with IDH1 and IDH2 it ends up being about 15% to 20% of patients with AML, so about a fifth of those patients. So it is a very, very important subset of patients with AML that needs to be addressed with some sort of therapy.

Now mutations in IDH1 and IDH2 lead to the development of AML in a very, very interesting way. So the normal function of IDH1 and IDH2 is to convert isocitrate to alpha-ketoglutarate. And the difference between IDH1 and IDH2 is where the enzyme resides. So IDH1 resides in the cytoplasm. IDH2 resides in the mitochondria. And as I said earlier, in both mutations the activity of the enzyme is to convert isocitrate to alpha-ketoglutarate. When you have a mutation in either IDH1 or IDH2, what happens is that the end product of that normal reaction, that is alpha-ketoglutarate, gets enzymatically converted to what we call an oncometabolite, which is beta-hydroxyglutarate.

And when you get increased levels of beta-hydroxyglutarate in the cells, which is what happens when you have one of these mutations, that leads to a block in myeloid differentiation. So you get the accumulation of myeloblasts, and then the hypothesis was, which has now been proven based on the clinical studies I’m going to show you, is that if you can block that mutant enzyme and lower the levels of intracellular beta-hydroxyglutarate, also known as 2HG, what ends up happening is that you then get a removal of the block in differentiation. The brakes get taken off, and those cells will start differentiating into normal, healthy neutrophils. And this was shown in laboratory studies, and I’ll show you the clinical studies in a second.

Now how do IDH1 and IDH2 mutations differ? So what’s interesting is that IDH2 mutations are primarily seen in patients with hematologic malignancies. You can see in this data set that they occur in about 15% of patients with AML and about 5% of patients with myelodysplastic syndromes and myeloproliferative neoplasms. IDH1 mutations, on the other hand, occur primarily in patients with solid tumors, so they occur in 70% of patients with low-grade gliomas and secondary glioblastomas and about 50% of patients with chondrosarcomas. They occur in fewer patients with acute myeloid leukemia, only in about 7.5% of patients.

And this may be why there’s really only 1 IDH2 inhibitor that has been clinically developed, that’s a drug called enasidenib, while IDH1-mutant inhibitors are being developed not only for acute myeloid leukemia and hematologic malignancies, but also for solid tumors as well. And though there’s 1 approved drug for hematologic malignancies there are efforts to develop other compounds as well.

Okay, so let’s talk a little bit about enasidenib. Enasidenib is an inhibitor of mutant IDH2. And we’re first going to talk about the use of the drug in patients with relapsed and refractory acute myeloid leukemia. So there was a large clinical study done, a very, very large Phase I/II clinical study, that simply gave enasidenib, which is an oral drug taken once a day, to patients with relapsed and refractory acute myeloid leukemia.

These are the results of the initial Phase I/II study. Now what you’re seeing in the 2 columns here is that we looked at the patients who received 100 mg a day of enasidenib, which is the current FDA-approved dose, those 214 patients. Again, this is a Phase I/II study, but a very large Phase I/II study, and then we looked at the results in all patients treated with relapsed and refractory acute myeloid leukemia regardless of the dose that they got on the Phase I/II study.

And you can see that in this trial the overall response rate, whether you look at 100 mg or whether you look at all patients at all doses, was about 40%, with about 20% of patients achieving a true complete remission. A little bit less, about 9%, achieving a complete remission with incomplete count recovery, and about 4% to 6% of patients achieving a true partial remission with a decrease in blasts, but also an improvement in hematologic parameters, looking at the absolute neutrophil count, and also looking at the platelet count, that it went up over 100,000.

What’s interesting about any differentiation therapy is the kinetics of response of differentiation therapy is not the kinetics of response with a cytotoxic. So for patients to respond to differentiation therapy they have to remain on the drug really for a prolonged period of time. This bar graph here really demonstrates that.

When you look at cycle 1, so that’s all the way on the left-hand side, and you look at the number of patients who achieved a complete remission by the end of cycle 1, in red, you can see that it’s very few patients. In the majority of patients it takes 3, 4, 5, even 6 to 7 cycles before they will achieve a complete remission. And this gets to the really important clinical point that when you’re treating a patient with an IDH inhibitor, in this case enasidenib, you can’t expect that drug to work after 1 or 2 cycles. It’s very similar to what we used to see, or what we see in patients getting ATRA for acute promyelocytic leukemia, where if you do a bone marrow biopsy at the end of the first cycle of therapy, they’re unlikely to be in a true complete remission at that point.

So how those response rates translate into survival is depicted here. Again, this was not a randomized study, so the median overall survival of all of the patients treated on this Phase I/II study was 8.8 months. And if you break that down into the overall survival stratified by best response, you can see for those patients who achieved a complete remission the median overall survival is 20 months. For those patients who don’t have a response, the median overall survival is 7 months, and for those patients who have a CRI or a partial remission it’s in between. The median overall survival was 13.8 months.

So the interesting part of this is that the approval for patients with relapsed and refractory AML was not an accelerated approval for enasidenib. It was actually a full approval, and that full approval was based not only on the remission rate, but also on the duration of response and the rate of transfusion independence, which was thought to provide clinical benefit.

As a follow up to that study, there’s a randomized Phase III study that was done called the IDHentify study. So the IDHENTIFY study simply asked the question in an open-label fashion that if you take patients with relapsed and refractory acute myeloid leukemia and you randomized them to get either enasidenib or investigator’s choice of standard of care, with those choices being a hypomethylating agent or intermediate-dose cytarabine or just supportive care alone, do you improve the median overall survival.

And we don’t have a lot of data about the results of this trial outside of this press release back in August, and I’ll just read it. It says, “The Phase 3 IDHENTIFY study evaluating enasidenib versus best supportive care with conventional care regimens, et cetera, et cetera, did not meet the primary endpoint of overall survival in patients with relapsed or refractory acute myeloid leukemia with an IDH2 mutation.”

I think what’s going to be important is that we don’t know any details about this study. That is we’ve got this 1 blurb. We don’t have any clinical data that’s been released yet. And I think it’s going to be really important to help us understand who in this clinical trial did not benefit from enasidenib and who in the clinical trial may have benefitted from enasidenib. Is it only those patients who were randomized to intermediate-dose ara-C compared to enasidenib that didn’t benefit? It’s hard to believe the patients randomized to supportive care alone wouldn’t benefit from getting enasidenib and not getting active therapy. So more to come on this.

Okay, moving on to the IDH1 inhibitor ivosidenib. So this was also explored in patients with relapsed and refractory acute myeloid leukemia. Similar to the data with enasidenib, it was explored in a Phase I/II study looking at what they call the primary efficacy population, and that was the patients who received the FDA-approved dose of the drug of 500 mg once a day. And then looking at all patients, that’s 179 patients, who received any dose of the drug. And similar to what we saw with patients who received enasidenib, the overall response rate is just about 40%. That’s the last row down there looking at overall response.

And looking at the rate of complete remission, the complete remission rate is 20%. And then in the ivosidenib study they defined a new remission category called CRh. CRh basically takes the criteria for CR but lowers the need for a platelet count above 100,000 or an absolute neutrophil count of above 1,000 to a platelet count above 50,000 and an absolute neutrophil count above 500. And if you use that relaxed criteria and combine that with the CR criteria the rate of CR or CRh, combining them together, is about 30%.

Looking at the overall survival by best response, looking at the CR plus CRh patients, the median overall survival is 19 months. In the non-CR/CRh responders the median overall survival is 9.2 months. And in the nonresponders the median overall survival is 4.7 months. You look at all the patients grouped together, the median overall survival is 9 months. So this is very, very similar to what we saw in the enasidenib study that I just showed you previously. Median overall survival in this single-arm study is just about 9 months.

Now before we get to this slide here, there is a randomized Phase III trial that is exploring the use of ivosidenib in patients with relapsed and refractory acute myeloid leukemia. That randomized Phase III trial is randomizing patients, it’s called the AGILE study. It’s randomizing patients to receive the combination of ivosidenib and azacitidine or azacitidine alone. I’m going to show you data from an earlier version of that trial in a little bit.

Now ivosidenib and enasidenib, the IDH2 inhibitors, are only approved for relapsed and refractory acute myeloid leukemia. Ivosidenib is also approved, in addition to relapsed and refractory IDH1-mutant acute myeloid leukemia, is also approved for patients with newly diagnosed acute myeloid leukemia. And what’s interesting is that that approval is really based on an arm of the original clinical study that we did in 34 patients with newly diagnosed IDH1-mutant acute myeloid leukemia. And that approval is based on the remission rate that you see on the lower left here showing that the rate of complete remission was 30.3%, the rate of CRh was 12.1%, and then the overall response rate ends up being 54.5%. That includes patients with partial remissions.

So if you just include the patients with CR or CRh that ends up being about 42% of patients who have a response. Median overall survival, again in this very small single-arm trial, is 12.6 months, but that was enough for the FDA to grant this an approval for newly diagnosed patients with IDH1-mutant acute myeloid leukemia.

Now with both of these agents, whenever you give a drug that causes myeloid differentiation, like ATRA with acute promyelocytic leukemia, there is a black-box warning for what we call differentiation syndrome. This is the black-box warning for enasidenib. It’s identical for patients given ivosidenib. And again, to read it to you “patients treated with enasidenib have experienced symptoms of differentiation syndrome.” What are those? Those can include fever, dyspnea, respiratory distress, pulmonary infiltrates.

So essentially what it is is, it’s a capillary leak syndrome being caused likely by cytokine production from the maturing myeloid cells. And how this looks clinically was explored in this study that we did, published in JAMA Oncology, looking at the top signs and symptoms consistent in patients with IDH inhibitor-induced differentiation syndrome. You can see that the majority of patients had dyspnea, unexplained fever, pulmonary infiltrates, hypoxemia likely causing the dyspnea, and acute kidney injury. And some patients had pleural effusion. Fewer patients had bone pain or arthralgias, lymphadenopathy, rash or DIC, and edema.

If you graph this out for 1 patient, and this is a patient actually of mine that I treated at Memorial Sloan Kettering. Just to walk you through this because it’s a little bit complicated. So at the top here what we’re looking at is the white blood count of these patients, specifically the differential. And in orange is the absolute neutrophil count. On the bottom panel we’re looking at the proportion of various myeloid lineage cells in the bone marrow at the same time. And here what you’re seeing is in the green is the mature granulocytes.

So what you can see on the top is the myeloid cells, the neutrophils start to increase both in the peripheral blood and in the bone marrow, and in the red is the myeloblast decrease. This patient develops, in the CT image at the very bottom, differentiation syndrome. You can see on the left-hand side of this CT scan the patient has some pulmonary infiltrates. If you look at the next CT scan, on the right-hand side, you can see that after the patient has received the treatment for differentiation syndrome, which is dexamethasone 10 mg twice a day, or an equivalent steroid, those pulmonary infiltrates have gone away.

This patient had a variety of episodes of differentiation syndrome. Sometimes it doesn’t just happen once, it can happen multiple times, as the cells are differentiating. And the patient needed repeated episodes of steroids to get them through their differentiation syndrome.

What’s interesting also is that if you look at the response among patients with and without IDH inhibitor-induced differentiation syndrome, it doesn’t really change whether you have differentiation syndrome or if you don’t. And what I mean by that is if you have differentiation syndrome, you’re not necessarily going to be a responder. And if you don’t have differentiation syndrome, it doesn’t necessarily mean you’re going to be a nonresponder. So the presence or absence of differentiation syndrome is not tied to response. So if you had differentiation syndrome, your overall response rate was 45.5%. If you did not have differentiation syndrome, in those patients the overall response rate was 37.5%, which in this study was statistically equivalent.

Okay, so after we did these studies where we looked at IDH inhibitor as a single agent, both in relapsed and refractory and newly diagnosed acute myeloid leukemia, we then moved to try to combine these agents with standard of care therapy. So in the first study that I’m showing you here, that was just published in the Journal of Clinical Oncology, this is a trial that combined ivosidenib with the hypomethylating agent azacitidine. Now we’re using, obviously, a lot more azacitidine and venetoclax. When this study was designed venetoclax wasn’t approved.

But what you can see on the left-hand side here is that the overall response rate for IDH1-mutant patients treated with the combination of ivosidenib and azacitidine was very, very high at 78.3%. The rate of CR and CRh was also very high at 70%. Of course, the caveat is that this is only in 23 patients that could be highly selected. We’re not sure whether this is going to persist in the larger AGILE study I told you about before.

On the right-hand side, in this so-called swimmers’ plot, you can see that many of these patients who achieved a complete remission, which are the patients with the blue arrows. These patients have remained on study for a prolonged period of time, many of them over 2 years.

Now what happens if you do the same study, and you take the IDH2 inhibitor enasidenib and you combine that with azacitidine? But instead of doing that as a single-arm study, this was done as a randomized Phase II study, on the right-hand side. So ignore the stuff on the left-hand side. If you just look on the right-hand side here you can see the patients in this study with IDH2-mutant acute myeloid leukemia, newly diagnosed, not a candidate for intensive chemotherapy, were randomized to receive azacitidine alone, that happened in 33 patients, or the combination of enasidenib, the IDH2 inhibitor, with azacitidine.

And this was a 2:1 randomization, so 68 patients received the combination, 33 patients received azacitidine alone. The primary endpoint here is the overall response rate. Secondary endpoints were overall survival and event-free survival. This is just the disposition of all the patients treated on this study. Again, 68 patients went on enasidenib with azacitidine, 33 patients went on azacitidine alone.

And if you look primarily at the event-free survival and overall survival, what you can see is that the overall survival between the patients who received enasidenib and azacitidine compared to azacitidine alone was actually exactly the same. The event-free survival favored the combination of enasidenib and azacitidine, but it was not statistically significant, with a p-value of 0.12, compared to the patients who received azacitidine alone.

Now I think this gets to the question of what is the best way to give azacitidine with enasidenib. So should you start giving azacitidine with enasidenib together? Or maybe what you should do is you should start with enasidenib, see how your patient does. If the patient does well with enasidenib alone, that’s great, and if they don’t perhaps you can add on azacitidine as a salvage therapy to that enasidenib. And that is actually the subject of an abstract that’s going to be presented at ASH as part of the Beat AML trial run by Leukemia & Lymphoma Society.

Okay, what about if we combine enasidenib, again the IDH2 inhibitor, just repeating it because the names are similar, or ivosidenib, the IDH1 inhibitor, with induction and consolidation chemotherapy? So this is a Phase I/II trial that was actually, I’ll show you the results in a second, just published in Blood. And it was a trial that combined either ivosidenib if you had an IDH1 mutation or enasidenib if you had an IDH2 mutation with standard induction chemotherapy. Patients who achieved a complete remission could go on and get consolidation with ivosidenib with consolidation chemotherapy or enasidenib with consolidation chemotherapy. And finally, for those patients who didn’t have an allogeneic bone marrow transplant, they could go on and receive maintenance therapy with ivosidenib or enasidenib.

I think an important point here is that patients who did get an allogeneic bone marrow transplant came off study, and they did not receive maintenance therapy with the IDH inhibitor.

So these are the response rates in patients who were treated on this trial. So in the top row, let’s just focus on the patients who received ivosidenib who achieved a CR, CRi, or a CRp, you can see of all the patients treated 77% of them achieved a complete remission of some sort. In the enasidenib-treated patients, 74% of those patients achieved a complete remission of some sort. If you look at a strict CR, 68% of patients who received ivosidenib with chemotherapy achieved a true complete remission, 55% of patients who received enasidenib with chemotherapy achieved a true complete remission.

What I was very excited about when I looked at this data are these survival curves. So again, all the caveats of a single-arm Phase I study with an expansion group, small numbers. Who knows if this is representative in a large group of patients? But just looking at the survival curves here you can see that out past a year and a half the patients both treated with ivosidenib and the patients treated with enasidenib did very, very well, with survival rates upwards of 60% or 70%. This is not meant to compare ivosidenib and enasidenib. These were 2 separate arms of the same trial. It is, I think, very exciting.

And this data has led to a large placebo-controlled randomized Phase III trial being conducted in Europe, which is comparing in a placebo-controlled fashion the addition of either ivosidenib or enasidenib to induction chemotherapy. And obviously the control arm is using placebo.

So I wanted to go through a few cases. I’ve got 2 cases here to discuss with you. These are real cases from patients I take care of in my clinic. So the first case is an 86-year-old woman with newly diagnosed acute myeloid leukemia associated with mutations in IDH2, RUNX1, and DNMT3A. She came to see me. Her physical exam was normal. Her performance status was 1.

But the white count was 3, with an ANC of 0.5. The hemoglobin was 8, and the platelet count was 13.

So this was in the age where aza/venetoclax was actually approved already, so we started her on treatment with azacitidine and venetoclax. In this case I didn’t start her on an IDH2 inhibitor, 1) because it’s not approved for this indication. I could have done it off label, but I thought that the combination of azacitidine and venetoclax might work a little bit better in this patient. And she did achieve a complete remission with the presence of MRD after 1 cycle of therapy.

Unfortunately, this was right around the time of the COVID outbreak, and she actually developed COVID pneumonia, ended up in the hospital, was really touch-and-go for a while. She was never ventilated but was on high-flow nasal cannula, and this was before we knew anything about steroids or really knew what to do with these patients.

But remarkably she actually survived. And after she recovered, she really had a prolonged recovery, she received another cycle of azacitidine and venetoclax, but then she actually relapsed with 30% blasts. And because of that, and because of the persistence of the IDH2 mutation, we started her on enasidenib 100 mg once a day. She had the brief onset of differentiation syndrome about a month into therapy with lower extremity edema and shortness of breath, and we gave her steroids.

She did achieve an MRD-negative complete remission at cycle 3, day 1, and actually remains on enasidenib and is feeling very, very well and things are back to her normal state of affairs and her normal life. So really a remarkable woman who not only achieved a complete remission once, achieved a complete remission twice, once on enasidenib therapy, and actually survived COVID pneumonia at the beginning of the pandemic.

So my second and last case is a 76-year-old woman who presented with fatigue, shortness of breath, and a petechial rash. On exam we confirmed that she had scattered bruising and a petechial rash. Her labs were really truly awful, with a white count of 6, mostly blasts, hemoglobin of 6.4, and a platelet count of 5. She had a bone marrow biopsy that confirmed a diagnosis of acute myeloid leukemia with 55% blasts and mutations in IDH1 and DNMT3A.

So again, for this patient I suggested perhaps she could go on azacitidine and venetoclax. She was really adamant that she did not want to have to come into the infusion center month after month to receive IV therapy with azacitidine, or subcutaneous therapy with azacitidine, for 7 days in a row. She lives an active life, and she wanted to go on something that was oral which was easier to take. So because of that I started her on ivosidenib 500 mg once a day.

At cycle 4, day 1, which is when her blood counts improved, I did a bone marrow biopsy. She achieved an MRD-negative complete remission. And what’s amazing about this patient is that she has now been in a continuous MRD-negative complete remission for 5 years. I actually just spoke with her. What’s amazing, so this is a woman with acute myeloid leukemia who has been controlled on an IDH1 inhibitor, and in the course of being in complete remission has had 2 hip replacements, has had a broken arm, has had all the kinds of problems that older adults can get, yet her acute myeloid leukemia is not a problem. And I consider that really a sea change in how we think about the survival of patients with acute myeloid leukemia when they’re older.

And with that, I want to thank you for your attention, and I’m happy to chat and take any questions.

DR LOVE: So that was really awesome. I learned a lot. I want to just ask you a few follow up questions kind of going in reverse. So starting out with your second case, I’m curious, any speculations or anything in this patient in terms of what kind of patient you would see this incredible response to as opposed to people who don’t respond?

Any combination there or any predictors of why you might see a great response like this?

DR STEIN: Yeah, so I think it’s the patients who are mutationally simple that do the best, who have an IDH1 or an IDH2 mutation. So in this particular patient she only had a mutation in DNMT3A, and IDH1. We now understand all sorts of things about how leukemia gets initiated, so the DNMT3A was probably the clonal hematopoiesis, the underlying soil, upon which the IDH1 mutation developed, which led to her acute myeloid leukemia. So if you knock out the IDH1 mutation you go back to that process of clonal hematopoiesis.

But what’s interesting here, actually, which I didn’t say previously, is that this patient, actually her DNMT3A mutation also disappeared. So it’s something like when you got rid of that clone with the IDH1 mutation you also got rid of the DNMT3A mutation.

We have noticed not only with IDH inhibitors, but really with all targeted therapies for acute myeloid leukemia, that patients who have RAS pathway mutations don’t seem to respond to these drugs very well. I mean they respond, but the response rate is much lower. So patients with IDH1 or IDH2 mutations, if they have a RAS pathway mutation, that’s a common mechanism of escape from IDH inhibitors.

Patients with FLT3 mutations, if they have a cooccurring RAS pathway mutation, they don’t respond very well. Cathy Smith has shown that. And there is increasing evidence that in the aza/venetoclax group of patients, those patients, if they have RAS pathway mutations, they also don’t respond very well. So RAS is a common way of escaping targeted therapy or really all therapy for acute myeloid leukemia.

Just to make 1 other comment about that, the flipside of that is that induction chemotherapy, so real, true induction chemotherapy, 7 + 3, actually is very good at clearing RAS pathway mutations. So if you look at RAS-mutant patients with newly diagnosed AML, and you give them 7 + 3, and then you look and see what happens after they get 7 + 3, the RAS-mutant clone is gone, in almost all cases.

DR LOVE: Wow.

DR STEIN: So it raises the question of whether in patients who develop resistance in these targeted therapies, it’s some modification of induction chemotherapy that can be given to get rid of that abnormal clone.

DR LOVE: So a couple other questions. So I assume the reason that these drugs have been studied with aza or HMA alone is because this is kind of before the aza/HMA thing? I mean it seems like it makes a lot more sense to think about aza/venetoclax plus these drugs.

DR STEIN: Yeah.

DR LOVE: Is that the direction?

DR STEIN: Yeah. That’s exactly what happened. Now there is a small study that was done out of the MD Anderson Cancer Center that was presented at ASCO looking at the combination of aza/venetoclax and IDH1 inhibitors. I think it just started as a venetoclax/IDH inhibitor, and then it turned into an aza/venetoclax/IDH inhibitor study.

Yeah, I think the big question about those studies is that the response rate with aza/venetoclax is so high already, especially in IDH-mutated patients, that the question really is how much bang for your buck are you getting for that IDH inhibitor, and should you reserve it for the time of relapse? I mean, it’s hard to know in the absence of a randomized study. But those studies are being done.

Tailoring Induction and Maintenance Therapy for Younger Patients with AML without Targetable Tumor Mutations — Mark Levis, MD, PhD

DR LEVIS: I’m going to talk about tailoring induction and maintenance therapy for AML patients without targetable tumor mutations.

So I want to start with a case. This is a patient of mine. And he was a 70-year-old guy, just retired; it always seems to work that way in our field. And he’s traveling around the country with his spouse. Very active. I mean, yes, he’s 70, but boy he looks and acts much younger. And he actually noticed while hiking, fatigue. And when he was out — I think it was in Colorado somewhere, he actually had an infected tooth. Didn’t respond to antibiotics. He went to a local healthcare provider. And so, things were going wrong. It was kind of the usual story. He gets back to this area. Sees his internist who orders some bloodwork and notes these abnormalities, the low white blood cell count, low ANC, explaining why he’s having a problem with infections. Low hemoglobin, which is why he’s tired while hiking in Colorado.

So he ends up with a hematologist who does a bone marrow biopsy and notes there’s AML with dysplastic granulocytes, 60% blasts. And this molecular analysis is completed. And I always describe the monosomal or complex karyotype as something that takes up a whole paragraph on the report form. And this isn’t a pretty one at all. There’s a loss of 3. Remember a monosomal karyotype means you’ve lost one whole chromosome and you have one other abnormality. This is both monosomal and complex. This is as bad a karyotype as you could want to have. And now, throw in the usual complement to this karyotype, which is a TP53 mutation, with a pretty decent VAF. And if you’ll note on the karyotype there are abnormalities with chromosome 17, a short arm. This is exactly where TP53 is found.

So, this is a terrible prognosis. We have a 70-year-old patient with a leukemia that very few people survive even when younger and what do we do? And there’s nothing targetable there. There’s no FLT3. There’s no IDH or something that will make our lives easier.

So maybe we should give him intensive chemotherapy, and for that we have a bold answer, no. Don’t do that. The data are pretty clear on that. These are a couple of different studies looking at patients treated with intensive chemotherapy regimens. And it's broken down as to whether or not you have a TP53 mutation or not, a complex karyotype or not, or both. And our patient that I’ve just shown you is both. And on this graph on the left-hand side in the New England Journal article, that combination is the blue line here, which pretty much goes to ground rather quickly. You’re looking at a median, very short median survival. You can estimate that, it's in the 6-month range, something like that. And over here in the European study, we see the same result, a complex – and in fact they’ve even broken down into this monosomal karyotype with a TP53 altered again — they’ve chosen the blue curve here — rapidly plummeting, with a median survival of really less than 6 months. That's if you’re trying to treat this thing with intensive chemotherapy. So we would wave off, even though this patient is fit, could tolerate intensive chemotherapy, probably a bad idea. No evidence that that's going to help him.

So what do we do? He’s referred to me. Supportive care, we may well do it. We could give you transfusions. We could give you antibiotics and just hang on. We could refer you to a clinical trial, which is always the right answer when you have that available. We could give decitabine. There’s a report out in the New England Journal that patients with TP53 mutations do better with 10-day decitabine. Now I have this direct from word of mouth from one of those authors, they’re kind of backtracking on that claim and that's what most of us in the field believe also — 10-day decitabine does have activity in some of these TP53 mutation patients. But maybe not as much as that New England Journal article led us to believe initially.

Certainly, azacitidine, a conventional course, or the new toy that everybody has heard about, aza or decitabine plus venetoclax; certainly, not unreasonable.

So as it turns out, we had a clinical trial of the HMA plus venetoclax. And fast forward, that patient was actually eligible for the Phase II study that preceded this one, VIALE-A. So let me tell you about VIALE-A.

So, patients with newly diagnosed AML, age 75 or older or unfit, so actually this patient is not eligible for that, not in the least technically. He is fit and he’s 70. Randomization 2:1. Patients were randomized to either get azacitidine plus venetoclax or azacitidine plus placebo with an endpoint of overall survival. And a good number of patients were enrolled. We participated in this trial, a couple of years of enrollment. And there it is, a lot of these patients had these poor risk cytogenetics, older patient population, of course. And here’s the bottom line from that trial very recently published, is that azacitidine plus venetoclax was a clear winner in overall survival.

Now, is it as dramatic as we might say — remember, this is 431 patients all ground up. There are patients that do well and patients who do less well. So there’s a lot of stories within this trial. And in fact, I’ve got a 85-year-old who I think is that line right there, still working at age 85 on this trial. So, when you really see this regimen work well, it's quite dramatic.

Alright, so as it turns out, this patient was started on aza/ven. And the azacitidine was given as shown here, IV 7 days. And it can be given subcutaneously purely as an outpatient. We admitted the patient, which I tend to do in a new leukemia patient, but not everybody does, and then we scale up the venetoclax because every now and then you get this tumor lysis that others have talked about of course.

So, after the first cycle, blasts in his marrow fell to 5%. And after the second cycle to 1%. Now we always could detect them. They were right there by flow cytometry. The pathologists would say, yeah, there they are, 1%. That's the leukemia. But technically a remission. What’s interesting though is that his counts rather dramatically improved and he went right back to his active lifestyle; he got his boat out, I remember him telling me, and sailed around all that summer. Very active.

Shown here are his blood counts as an illustration of this. So his hemoglobin in red, when he started the trial is right down here keeping from hiking in Colorado, and really, within a couple of months, was right back up into a range where he’s able to do whatever he wants to do. Look at the ANC rocketing up after his second cycle into the normal range.

So, what we’ve done is — I have not given him a curative therapy. I haven’t technically even put him in a proper remission because I can still detect his leukemia. But look at this, after the 1 week of admission, which is really protocol mandated — a lot of people wouldn’t even have done that — he’s out sailing his boat with normal counts. So this is sort of a quality-of-life thing. Giving the man 7+3, which technically he was eligible for, seems like a bad idea in view of this kind of data. So, taking an unconventional approach, no, he technically doesn’t need label criteria because he’s fit and under age 75. And yet giving this man 7+3 seems like a crime when we could do this.

So, he completed his 12 cycles total actually and died 18 months after his diagnosis. But I want to emphasize, 12 of those 18 months were feeling totally normal. Which just isn’t something that you're going to do in somebody that you try and treat in kind of an old-fashioned way with intensive chemotherapy, which a lot of people, we’re still arguing that I should do. He’s fit. You should give him what’s well described.

So, no, I think this is an unconventional approach. It's technically off label. But, in fact, I don’t think anybody would hesitate about doing this right now knowing what we know about these combination regimens.

And, again, even with this regimen of aza/ven, HMA/ven, TP53 mutations are trouble. This is looking back on data from the earlier trials showing that these TP53 mutant patients are not going to have the same response rate, and in the same publication the duration of response is shorter as well.

So, I was happy to have a response of 12 months, which, technically, isn’t all that impressive for the regimen. But for this particular subtype of patient, this is key. And we have to kind of change our thinking about this particular subtype of leukemia. Mind you, I think we are making progress with TP53, there’s a couple of new things coming out for it. And I think this will change going forward. But right now, this is the best we have for this, which is a lot better than what we had just a few years ago.

Next case. Again, a guy in his seventies, still working. He was running a couple of companies. Hardly time to stop to be told that he had something wrong with his blood. He was noted to have a mild anemia 3 years prior from when I saw him, and didn’t require any treatment. And in fact was diagnosed with MDS with that.

So, he finally develops dyspnea on exertion and his CBC at that point shows what really is frank leukemia, white blood cell count rising to 26,000, platelets dropping and hemoglobin of 8.3. Mutation analysis, okay, this is kind of this intermediate risk. The karyotype is normal. Those mutations are not targetable. Like an NRAS, we love to target NRAS, but we haven’t figured out how to do it yet. STAG2, that just says he had MDS. These were his old MDS mutations, TET2 and STAG2.

Okay, what do I do know with him that he’s got AML? He’s fit, certainly. Well, again, I’m going to give him azacitidine. Now this is not the whole story, but that was azacitidine and venetoclax. This was the hot new regimen. So, of course he got that. This was actually a few years ago, and he flew into remission. Everything went well. And this guy even had a clean bone marrow biopsy. We couldn’t detect anything. His mutations cleared. And basically, he went back to work and stopped paying attention to all this.

However, we did a follow-up bone marrow biopsy about a year later and we could see those mutations emerging and dysplasia was pretty clearly evident. So we persuaded him to undergo and allogeneic transplant; he had actually been inquiring as well. And we said, okay, we’ll transplant you. We used stem cells from his grandson. Engrafted beautifully. Count’s normal. Returned to work. And then in early this year, 2-1/2 years after the transplant, he’s now 78, he’s getting kind of tired of this, although he didn’t come in too often and he doesn’t need to, his counts have been fine. However, finally, just before old COVID hit, his CBC dropped, platelet counts 70,000. He’s got 5% circulating blasts. Here comes the monster back. Okay, we did not cure him.

We started this journey 5 years prior and he’s spent of that time outpatient, in remission, but, nonetheless, here they are. Here’s the STAG2 and TET2 back. He’s losing his donor chimerism. And so, this has been a well-described technique, which I’ll go over, initiate azacitidine in this circumstance. And we get his counts to normalize after a couple of cycles. The patient wants to know, look, I’ve been doing this for 5 years. I’m a busy guy. What can I do? Do I have to keep coming into clinic? I don’t want to come in once a month or once every 6 weeks and get azacitidine. I get it. I’m not going to get cured, but I really want to be outpatient.

So, is there something I can give him to maintain this state? Is there anything we know about maintenance therapy in AML? Well, people have been looking at this for years, of course, I’m just going to show you a few more recent studies. This is from the Phase III HOVON97 trial, started way back before the turn of the century, for patients with newly diagnosed AML, age greater than 60. Everybody got 2 courses of intensive chemotherapy and they were in CR1. Now, that's not complete therapy. Nobody would regard that as complete therapy, but these patients weren’t actually — it wasn’t clear that giving them more therapy in this older population was going to benefit. So they were randomized to get either just observation or 12 weeks of azacitidine, this kind of lower dose, 50 mg/m2, 5 days. Kind of a low dose thing.

And you can see the results here. Yeah, it seemed to keep many of them in remission, as shown in disease-free survival. Was not in any way an obvious benefit in overall survival, as shown here. So, kind of meh, result there.

And then there’s this interesting RELAZA trial. So that kind of applies to our patient. He relapsed after an allogeneic transplant, losing chimerism, but in early relapse. Can we do something to push him back into remission? And it turns out azacitidine can do that. Again, as shown here, this patient being pushed back and forth into early and more advanced relapse with azacitidine. Again, this is exactly my patient. But can I give him something to keep him, that he doesn’t have to keep coming in and getting treated in the clinic?

And to summarize, there’s lots of data on maintenance therapy in AML, and I think this recent review summarizes it best. The benefit of the benefit of maintenance seems more apparent after suboptimal induction and consolidation. The older patient who’s just not getting any more treatment, but you know they haven’t really gotten a curative regimen. And, again, may be relevant to patients who can’t tolerate consolidation, older patients who get into remission and are pretty much done at that point.

So, back to my patient here. Our options are we could just watch until his counts change again and hit him with more azacitidine. We could continue azacitidine indefinitely. Are there any other options?

Well, let’s look at azacitidine. So what about oral azacitidine? This guy has actually heard of oral azacitidine. He’s been asking me about it for about the last year: What about the pill version of this thing?

So this is from the Phase I study, which was published back almost 10 years ago. And the formulation was, again, it is the same drug, azacitidine. There’s nothing funny about that, but really a high degree of inactivation with intestinal epithelium because the enzyme that deactivates this, this cytidine deaminase is quite well expressed in there. Forty-five patients who are eligible for azacitidine treatment. We treated the first week with subQ aza, and in the second week with oral azacitidine, known as CC-486, and they reported an overall response rate of 73%, for what that's worth. This is the curve you’ve got to focus on. This is the PK curve. So, in the same patients, they had the blue curve, which is the pharmacokinetics of subcutaneous azacitidine and the yellow curve in the same patients when they got oral azacitidine. And you can see that in fact you just don’t get — you get a quarter to a third of the peak levels. Both are gone very quickly. This is measured in the hours, by the way. This is not days. This is hours. The drug is very rapidly cleared.

So, trying to move this to substitute for subQ or IV azacitidine, the thought was well, okay, we have to treat it for longer. So, rather than 7 days, we’re going to do 14 days — 14 days oral we hope is the same as 7 days subcutaneous.

So in comes the QUAZAR study. Now, we’re going to take that population I just defined, older patients, who are in remission but pretty much are done with therapy. They haven’t gotten optimal therapy. So these are patients over age 55. They’ve gotten remission. And most have received no or 1 cycle of consolidation, again incompletely treated patients. Mostly older. Endpoint just overall survival. Can we keep them alive with an oral regimen, 2 weeks of azacitidine out of out of every 4 and just repeat? Yes, it turns out that actually will buy you a better overall survival. Outpatient therapy. Is this a homerun? Well, you can shrug about it, again like everything in oncology, advances are incremental. But, yes, this is a believable curve. I rather it be on the blue line. There’s a pretty good median — the overall survival at 24 months — a median survival of 24.7 versus 14.8. That's pretty believable.

So getting back to my patient, he’s a 78-year-old guy in early relapse after his transplant. Counts are normalized. He begins therapy with oral azacitidine. He’s been on that now really — almost since the drug was approved. I swear he was watching for the approval and wanted to go on this thing.

So, AML patients without obvious targetable mutations clearly can still benefit from these newer induction regiments with targeted agents, and venetoclax with azacitidine or decitabine is targeted. It's got a broad applicability. Maintenance regimen with an agent — with a hypomethylating agent seems to be most effective in AML patients who’ve been able to complete a standardize course of intensive chemotherapy. Oral azacitidine appears to offer a more patient-friendly version of this type of maintenance. And I think the real question going forward is can this oral azacitidine or by the way, there is an oral decitabine, be substituted for the subQ or IV versions in those venetoclax-based inductions that you’ve been hearing so much about.

DR LOVE: So just to follow up on this. I guess I’m wondering in this study, the QUAZAR study, as well as other studies of maintenance, I had assumed that most of these people got chemotherapy, not HMA/venetoclax. Is that the case?

DR LEVIS: They did. Oh, they did. These patients got incomplete chemotherapy. So they got 1 or 2 cycles of induction and then quit. We don’t think that’s curative for intermediate or favorable risk AML patients. You’re supposed to get 7 + 3 and 4 cycles or 3 cycles of consolidation. These patients previously got 7 + 3 and a lot of them stopped right there.

DR LEVIS: I think any maintenance regimen where we incorporate this oral azacitidine would still probably include venetoclax.

Therapy in AML, as far as I can tell, is still only of benefit just for someone who has been, call it incompletely treated. Aza/ven or decitabine/ven, is in itself, a maintenance therapy. I have patients who are on it like a maintenance therapy and they’re on it for months and months and months. Those are patients who are responding.

What do you do though, when somebody really is either tired of doing this? They don’t want to get the subQ. Again, my patient who relapsed after his transplant; he’s just kind of tired of coming in. He’s been coming into oncologists for 6 years and he’s wondering if there’s an option. So he kind of leaped all over the oral azacitidine.

DR LOVE: So typically when you give HMA/venetoclax, for how long do you continue it? Indefinitely?

DR LEVIS: I continue it indefinitely. Yeah. Now I space out — so the patient’s, like this guy, he was getting aza/ven for every 8 weeks until he started progressing. I do space it out to make it nicer, but I keep that as maintenance regimen.

DR LOVE: Anything there you want to comment on.

DR LEVIS: Again, there’s no question that there’s a benefit for this — I don’t think there’s any question that oral azacitidine benefitted these patients. It's the patients — it was the selection of the patients that I think people need to be aware of that, that you’re selecting incompletely treated patients.

DR LEVIS: Well, making the broad characterization that we’ve demonstrated a benefit to maintenance therapy for everybody is an incorrect statement. I think a lot of people looked at QUAZAR and said, oh, everybody who’s in remission who might relapse should go on oral azacitidine. I do not believe there’s any evidence of that whatsoever.

I think there are multiple studies using HMAs in the maintenance setting, but it's almost always somebody who has been incompletely treated, where you can show a benefit. Which doesn’t seem that surprising. That just tells you though that — again, I remain pretty excited about oral azacitidine. I think we’re going to be able to substitute this drug for the subQ or IV version. I think it works just as well. I think the maintenance regimens that worked with azacitidine when they showed that aza subQ or IV worked as maintenance. Now you show that the same oral drug works as maintenance, that's the equivalence I was after. So now I’m going to argue that you’re going to be able to substitute the oral into the other one in the serious regimens, the up-front regimens.

DR LOVE: So, can you talk a little bit, you referred to the possibility that maybe the oral azacitidine could be utilized in the HMA/venetoclax treatment paradigm. How is that being looked at in clinical trials right now, or is it?

DR LEVIS: So far it is not. That much is very clear. Well I asked the drug company sponsoring this, what’s being done about that? And they said, currently nothing. They were focused on getting their label done before they started talking about off-label activities.

In terms of a clinical trial, I’m unaware of any trial that has fired up using oral azacitidine combined with venetoclax. I have no doubt it's on everybody’s minds right now and it will be done off label I’m certain. Because I’m already — the label remains more or less vague. No one says you can give oral azacitidine as a maintenance therapy, as I’m doing with my 78-year-old guy. I don’t think anyone would bat an eye if I decided to give that patient venetoclax along with that oral azacitidine. I may well try it if he progresses with blasts.

DR LOVE: What would you like to see in terms of clinical trials to address the question? What kind of clinical trial do you think would be — I mean do you think you need a Phase III trial?

DR LEVIS: Well, I think a small Phase II, but randomized would do. Yes, I do. I think you're going to have to do aza/ven subQ/IV versus — and they’ll probably just do it subQ — and compare that with oral. Absolutely. I think that's got to be the next order of business.

DR LOVE: Are you using subQ —

DR LEVIS: But again 14 days…

DR LOVE: Are you using subQ right now with HMA/venetoclax strategy?

DR LEVIS: I tend to use subQ more frequently. If I bring them into the hospital for their first week, I use IV. But I almost always move to subQ. SubQ is, although if the patient is really skinny, they don't like the subQ. The shots can be troublesome. They have to have a little bit of cushion to inject.

DR LOVE: How do you see the HMA/venetoclax strategy being moved into the upfront setting of the younger patient?

DR LEVIS: That needs to be another clinical trial and it really has to focus on MRD. Again, remission is remission, but if you’re not really looking at good MRD parameters — and here it's going to get tricky because our best MRD marker is MPN1. Do you want to know the subset of AML that responds best to aza/ven? It's MPN1. So, if you pick out subsets of AML patients that you give aza/ven to, it will do just as well as 7+3, I’ll bet. Now, that said, I’d rather give aza/ven to a TP53 complex karyotype young patient. I might well like giving aza/ven to an MPN1 mutated young patient. Am I going to want to give aza/ven to a FLT3 MPN1 patient? Not so sure. The FLT3 is a problem.

So, like anything, AML is going to break down into its subgroup and we're going to have to do the trials, literally randomized, to tell us what to do. I would have no ethical problems treating a younger patient with aza/ven though. None whatsoever. Because it could well be as good or better. And I’ve had some young patients refuse the intensive stuff, they said I’m not going to lose my hair. Give me the aza/ven.

DR LOVE: Practical question. How do you approach the patient with MDS who’s been receiving an HMA, who then develops AML?

DR LEVIS: With a heavy heart and sagging shoulders mostly. Sure, we look for the targeted agents. They may respond if they have an IDH mutation. In fact, I’m doing that right now by the way. I’ve got a patient who’s beginning to lose response to aza, and he had an IDH2 the whole time. And he’s actually on just ANA alone. He never made AML. His blasts went to 15%. All of a sudden, he’s responding beautifully to ANA. So we should move them earlier before they do that. In fact, I think the bigger question in the field is why did that patient get treated with aza alone? Why didn’t you think about moving aza/ven sooner, before they met criteria for AML?

DR LOVE: So, does that mean you use aza/ven to treat MDS?

DR LEVIS: I am doing that now.

DR LOVE: Really?

DR LEVIS: Well, what’s the definition of AML? Oh, it's 20% blasts. I see. So, 19.9 isn’t, and 20.1 is. Really? So what about 11%? What about 9%? What percent?

So I actually am treating right now, I’m going to go see her this afternoon, a patient who’s got somewhere between 5 and 10% blasts, it's a little sketchy when you’re diagnosing that. Yes, she’s young. I’m putting her on aza/ven. I want to clear remission. Aza alone is not going to give me the response I want before going to transplant. So I chose aza/ven and I know others are doing that, too.

DR LOVE: What happens when you add ven into a patient who’s going from MDS to AML on azacitidine?

DR LEVIS: Sometimes it works. More often it doesn’t.

DR LOVE: Is it worth doing?

DR LEVIS: Oh, yeah, we do it. Because every now and then it works like a charm. I’m seeing another patient that I did that with. He was on aza for 6 months. Kind of didn’t get much of a response. I added ven and he flew into remission. And by the way, he was never technically AML, he was 15% blasts the whole time. But I kind of declared him, oh, well, this is 15- go 20%. I hereby declare this AML and treated him as if he were — I just added ven after 6 months of aza. He got an immediate response. That's not the norm, but it can happen.

Other Novel Agents and Investigational Strategies for Patients with Acute Myeloid Leukemia — Daniel A Pollyea, MD, MS

DR POLLYEA: I’m real excited to be presenting this really exciting topic, the title of my talk is Other Novel Agents and Investigational Strategies for Patients with Acute Myeloid Leukemia. And there have been so many advances recently, it's hard to sometimes categorize them all. So in this talk we’re going to talk about some that are often put into various number of different categories, but really promising hopeful strategies, all of them, and already having an impact or we believe soon to have that impact.

So, the first subcategory of AML that has become crucially important to talk about is the category of patients that we define as having secondary AML. And you’ll see in a few slides why this is so important to know and understand what this means. But before we get to that, how exactly are we defining secondary AML?

So, secondary AML is defined as AML that evolves from antecedent hematologic disorder. Usually, when we’re talking about AML from an antecedent hematologic disorder, we’re talking about MDS that evolves to AML. And in fact, upwards of a third of MDS patients will ultimately develop AML. But you could also get to AML from other antecedent hematologic disorders like myeloproliferative neoplasms or even sometimes less commonly, nonmalignant antecedent hematologic disorders like aplastic anemia can evolve to AML. So, that's one subcategory when we're talking about secondary AML.

The other subcategory is what we regard as treatment-related AML. And the World Health Organization has been fairly specific about this, and treatment-related AML is AML defined as arising out of a history of exposure to alkylating agents and/or toposomeriase-2 inhibitors and/or ionize radiation that's occurred to a large field, including the bone marrow.

There are other implicated therapies that are relevant with respect to treatment-related AML. Less is known about these associations, but antimetabolites, anti-tubulin agents, we know can have this affect or this future consequence.

So these two elements together are what we package together as one entity to call secondary AML.

And when we look at secondary AML outcomes, they are not as good as patients who have what we would call de novo AML. And there’s a variety of reasons for this. There are a multitude of other associations that come with secondary AML that make them higher risks. These patients are often older. They have more other adverse biological risk factors like poor-risk cytogenetics, TP53 mutations. They typically respond less well to conventional treatments like standard induction chemotherapy or like hypomethylating agents. As a result, because of the biology of the disease and this resistance to traditional treatments, these patients do more poorly with respect to overall survival. And we really consider a secondary AML patient to have an incurable disease without an allogeneic stem cell transplant.

And so, this survival curve shows this. So the solid line are de novo AML patients. You can see in a large group by all ages, nothing great. Nothing to brag about there for sure. Lots to improve upon. But when you look at this subgroup of de novo AML compared to either treatment-related AML or AML from an antecedent hematologic disorder, you can see a significant decrease in survival, even below the very poor outcomes we already see with de novo AML.

So, why is it so important to consider this entity? This has become crucial to recognize because of the recent approval of the drug that we call CPX-351. CPX-351 is a liposomal formulation of the standard 7+3 regimen. It consists of a single drug that is made up of a fixed molar ratio between daunorubicin and cytarabine in a 1:5 relationship. So instead of giving 7 + 3 with an infusion of cytarabine and intermittent doses of an anthracycline, it's already all put together, packaged together. And originally, when this was given in comparison to 7 + 3 and looked at all comers, there were no differences in overall survival in the initial clinical trial that was done.

However, it was interesting that there was a subgroup that wasn’t a preplanned analysis of this study, but a subgroup of patients that they looked at that they defined as having secondary AML, based on the definition that we invoked before, and when you looked at that subgroup there was a pretty significant survival benefit. And so, the decision was made to confirm this finding in a follow-up study in which that was the population of interest.

And, here is some data from that study. When you look at secondary AML patients, newly diagnosed, and fit for standard induction chemo, who are randomized to either receive the CPX-351 or the 7 + 3, this is the breakdown with respect to the toxicity profile.

So I think it's important to recognize that the CPX-351 has a toxicity profile that's very similar to 7 + 3 with respect to hematologic toxicity, infection, other well-known adverse events associated with the standard chemotherapy because it is those elements of chemotherapy just packaged up differently.

So, with respect to safety, pretty equivalent there. But when you look at response rates you could see differences. So, with the experimental arm, the patients who received CPX-351, there were better responses with respect with overall response rate and pure CRs.

And then, the endpoint of the study, I’ll focus your attention here on the survival curve, was met. So they confirmed what was seen in this previous observation in which an improvement in overall survival was seen in the patients who received CPX-351 versus 7 + 3. Median survival, about 10 months, compared to about 6 months for this population.

A really interesting observation was also made, and I’ll call your attention to this other survival curve, that patients who achieved a remission in either arm could of course go to an allogeneic stem cell transplant if that was felt to be appropriate by their treating physician. And an interesting observation here, because these are all patients — all the patients who went to transplant theoretically had achieved a remission regardless of which arm they were randomized to. And we often think of a transplant as this great equalizer. If you can get to a transplant, it kind of doesn’t matter what happened that came before, outcomes are — the starting point is level — the playing field is leveled. And that wasn’t the case here.

So if you achieved your remission in the setting of CPX-351 and went on to a transplant, you can see you had a much better chance of a good outcome, and ideally, some of these patients or many of these patients were cured, compared to if you went to the transplant with the 7+3 regimen. A variety of hypotheses about why this may be the case. The suspicion is that the CPX-351 allowed for deeper remissions which then resulted in these better outcomes after the transplant. Unfortunately, those correlative studies were not able to be done or were not done. So it's sort of still just a theory, but a pretty compelling one.

So, since FDA-approval, most of us would consider CPX-351 as a treatment for this very specific population of patients, newly diagnosed patients who are induction-eligible and have secondary AML. And in particular thing about this, if you’re seeing the patient and you think, yes, this is a patient that would be appropriate for a transplant in first remission if I can get him or her to that point. I think that's really the ideal scenario for this drug.

A couple of other caveats. So how do you know if they have secondary AML? Well, if you treated for the patient for her breast cancer and now she has AML, then you know she had secondary AML with respect to the treatment-related AML. But how can you tell if they had an antecedent hematologic disorder if that was never clinically obvious before evolving to AML? Which many times that's the case. People harbor asymptomatic MDS with some mild cytopenias. No one knows or is aware. Then they develop AML.

There is a way to pretty reliably tell if a patient’s AML came from an antecedent hematologic disorder based on the genotype. The gene mutation profile of the patient. So a patient with AML who has certain gene mutations, you can be pretty sure evolved their AML from an MDS.

And so, it amounts to doing that testing. That testing takes some time. And as AML doctors, we're often very impatient to get started with treatment. This is a medical emergency. But what’s been coming out recently and was published by the German Group in Blood, is that delay in treatment of up to 15 days or maybe even more, to collect this information so that you can make an optimal treatment plan for the patient does not negatively impact outcomes. And so, just keep that in mind if you’re weighing how to get the information you would need to know if you should apply a treatment like CPX-351 to a patient.

And then final point, and this is just some emerging data, more coming recently from ASH, a couple of abstracts including this one. Patients with a p53 mutation do quite poorly with CPX-351. They do quite poorly with intensive chemotherapy as well, and CPX-351 as we said is chemotherapy. So if you happen to know that your patient has that mutation you may want to think about other options. And I think that's something that hopefully, very soon, the field will be able to address a lot better than we can now.

So, there is a multitude of investigations going on. We have just such a great community of clinical investigators in our field that are looking to pair CPX-351 with any number of targeted therapies and non-targeted therapies in a variety of different settings, first line, relapsed. Even in MDS and MPN. So, we anxiously await more information about how CPX-351 can continue to move the field forward.

Other options. So, as we discussed, a patient who’s appropriate for CPX-351 is a patient who is appropriate for intensive induction chemotherapy. I showed you that the toxicity profile is very similar, and we know the toxicity profile for standard induction chemotherapy is quite rough. If you have a patient with secondary AML who can’t tolerate intensive chemotherapy and therefore, is not a good candidate for CPX-351, what to do for that patient?

There’s nothing so specific like with a labeled indication as there is for CPX-351, but I call your attention to this recent data that Courtney DiNardo presented on behalf of the Phase III VIALI-A study, which has been subsequently published in the New England Journal, comparing azacitidine plus venetoclax to azacitidine alone for newly diagnosed, unfit for standard chemotherapy patients. And what you can see here is that certainly, when you look at de novo and secondary AML patients, the preferred therapy is azacitidine plus venetoclax. And this is showing the response rates here. You can see that azacitidine alone patients, they do quite poorly in general and they probably do worse when they have secondary AML.

I alluded to that in the beginning, that most conventional treatments, like induction chemotherapy and hypomethylating agents, they don’t do as well with secondary AML compared to de novo AML patients. But I just call your attention to the relative similarity and response rate for de novo versus secondary AML when it comes to venetoclax regimen.

So this is a principle that we’ve been learning that most of the subtypes of AML that we have traditionally recognized are agnostic with respect to response to a venetoclax-based regimen. And that is quite different than our expectations for the more conventional treatments, as you can see here, with azacitidine alone. So keep this in mind. Venetoclax-based regimens are very effective across the board and even in subsets like secondary AML that traditionally don’t do well.

Other treatment approaches. So one to think about that would fall into this category is patients with CD33 protein expression, and that's 90% plus of AML patients. For these patients, can you consider incorporating the treatment gemcitabine ozogamicin into a therapy. And this is an antibody drug conjugate that targets CD33 and has a payload of a very toxic substance called calicheamicin. And when the CD33 cell is engaged by the antibody, then it allows for the calicheamicin to enter the cell of interest, there’s a binding domain that sort of dissolves, and you release this very toxic substance into the CD33-positive cells specifically.

And gemtuzumab has an interesting history in our field. It was approved all the way back in 2000, based on a 30% overall response rate. It was withdrawn subsequently from the market because of concerns about efficacy and toxicity. And then more recently, it was reapproved. So kind of an interesting history there.

So, because it's been around so long, we have several Phase III studies with quite a few patients for AML studies. These are huge numbers. And you can see that there are patterns that emerge with most. And most of these gave gemtuzumab plus intensive induction chemotherapy, usually in the upfront treatment setting. And in many cases, what you would see were no differences in response rates, although sometimes there were some response rates seen, but across the board, no differences in response rates, but if you waited for the survival or event-free survival data, you saw the impact there. And so, with the exception of the SWOG study here that didn’t show that, the others’ really did show that there was a survival benefit even without a difference in response rates. So an important clinical trial evaluation question there.

Now, because so many patients have been treated with gemtuzumab in the AML field, the field had the ability to actually do a meta-analysis of this drug which is quite impressive when it comes to such a rare disease. But these investigators did a great job collating data from all those clinical trials and more, and really some patterns began to emerge when you looked at this.

Now, these were all newly diagnosed patients, usually fit for intensive induction chemotherapy, that received some element of induction chemotherapy plus gemtuzumab. And the pattern that emerged is that patients with favorable risk disease, and usually what we’re talking about here, this is defined as patients favorable by cytogenetic subtypes, usually the core-binding factor AML, the 821s and the inversion 16s, you could see there’s a quite significant survival benefit for these patients with the favorable risk disease.

On the other end of the spectrum, there was no benefit seen in patients with adverse risk disease.

And then in the intermediate group there was a significant improvement in survival. You can see that relative to the favorable risk group, not quite as big, but still there.

And I think this has led to the general practice approach, my practice, and most of my colleagues that I know, that will offer gemtuzumab plus chemotherapy to a newly diagnosed AML patient with favorable risk biology. We will not do that for adverse risk patients because these patients would only get the negative effects, the toxicity with benefit. And most people I know do not so this in the intermediate risk setting for a variety of logistical reasons, mostly being that a lot of these patients will go on to get a transplant. And one of the main toxicities with this therapy is veno-occlusive disease, which is worsened in the post-transplant setting. And so, a lot of people are a little cautious to give gemtuzumab to patients who might go on to get a transplant. Favorable risk patients usually don’t need a transplant or aren’t given a transplant in first remission.

So, that's how this has shaken out in general terms and practice patterns in the United States at least.

And then what about single-agent gemtuzumab? This is some data from the pre-venetoclax era. When you look at a newly diagnosed unfit for intensive chemotherapy older patient who gets gemtuzumab versus basically supportive care, there’s maybe a modest improvement there. I think we have other things that we're more excited about now. And then what about in the relapsed setting? Gemtuzumab has some activity, but I don’t think any of us think that this is anything except kind of a treatment of last resort in many cases.

So, moving on to a really exciting and emerging field, other people that have been working in solid tumors had been experiencing or getting experience with immunotherapy for quite a while. In AML, we’re maybe at a different point and there’s definitely a lot of differences between AML and solid tumors, but there are a variety of opportunities here I’d like to discuss. I already touched on antibody drug conjugates. I’d like to talk a little bit more about those. And then some different T-cell-based therapies that are excited in AML.

So, antibody drug conjugates, like we talked about gemtuzumab, are pretty pervasive in our field, at least in the clinical trial setting. So, we have gemtuzumab of course, which is approved. Other strategies to target CD33 from an antibody drug conjugate perspective. CD123 is a compelling target. A couple of different approaches there. And CD123 is a potentially stem cell directed therapy. Others, CD25, CD30, CLL1, and this isn’t an exhaustive list, but you can see a lot of different attempts and interests at this mechanism.

What about multi-valent antibodies? So, the idea here, and we borrowed this idea from the great success of blinatumomab in ALL, is can you conjugate an AML antibody of interest? So we talked about CD33, CD123, any target of interest, can you conjugate that not with a poison that would act as a targeted chemotherapy type approach, but instead with a T-cell engager or an NK-cell engager? So you deliver your AML cell of interest in close proximity to the patient’s own immune system and allow the patient’s own immune system to do its job and gobble up that target cell.

So, really exciting prospect. And again, blinatumomab is the proof of concept at least that this can work in a disease that targets CD19 and conjugates to CD3 and is extremely successful.

And there are some really promising attempts at this. Flotetuzumab is one version of this that exploits CD123 with CD3. And you can see a recent publication from Blood. It just shows a nice blast reduction in a majority of patients who got this therapy. And some of these patients had real responses. Perhaps there’s a signal that this therapy is a little bit more active in the refractory setting than in the relapsed setting, a little bit of a subtle distinction there in AML. But I think a lot of promise for that therapy and other related strategies here.

What about checkpoint inhibitors? This is really an area that our solid tumor colleagues are well versed in and very experienced with. Some really brilliant investigators in our field, like Naval Daver, whose paper I’m citing here, have been taking the lead in this strategy. And so, both PD-1 and PD-L1 inhibitors had been given in relapsed/refractory AML. Most strategies are now pairing these therapies with azacitidine for a variety of reasons. And I think there is some promise here. There are some early signs of responses. And I think the field is very interested in working to identify those patients most likely to respond to this type of an approach and also, thinking about ways to combine this therapy that maybe make it even more effective.

So, there’s ongoing work with pembrolizumab, durvalumab, all different kinds of combinations and settings. You can bet in our field that folks are all over every single potential scenario. Lots and lots of smart people that work in our field that have such great ideas to keep pushing this forward.

So, a little bit of a different approach, but still immune-mediated is this rather brilliant concept that has led to this drug called magrolimab that we’re all very excited about. So, the idea here is that in general the human body has developed a variety of ways to recognize a cell that is self, versus a cell that is non-self. And it also has to have ways to recognize a cell that is self, that needs to be recycled or destroyed. And CD47 is one of the arbiters of those decisions. And so, when CD47 is expressed it's a signal to macrophages to leave that cell alone. We like to say it's a “don’t eat me” signal. And that's just a normal process that your immune system has evolved over many, many millennium, and it's quite efficient process. The problem is, is that cancer cells, in this case blood cancer cells, have figured this caveat out and they’ve highjacked the system by inappropriately expressing CD47, which allows them to be under the radar.

And so, some very brilliant folks at Stanford, who then formed a company to investigate this further, said if we can block CD47, we can re-expose these bad tumor cells to the normal immune surveillance system, in particular macrophages.

And that's how magrolimab works, that's the mechanism. And I think we’re all very excited about this. What’s the clinical output of this? Well, when you look at magrolimab plus azacitidine in newly diagnosed AML patients unfit for intensive chemotherapy with relatively small numbers, we’re seeing good response rates. And in particular, as is highlighted here, even potentially better than expected results in the p53 cohort of patients that have unusually poor responses to conventional treatments as we alluded to earlier.

And then, with respect to durability, still early days, but there is some real hope that this is potentially a durable therapy.

CAR T cell approaches in AML. We would love to be where our ALL colleagues, and even our lymphoma and myeloma colleagues are in AML. I think there’s a lot of potential here for these strategies to be relevant. No real published data yet to speak of, at least recently, but a lot of hope that this can be something that we can exploit and figure out for AML patients.

As you would expect, new therapies have new toxicities and we have to be ready for that. So, the antibody drug conjugates, these have toxicities specific to the target. So, CD33 targeting can cause prolonged cytopenias. I alluded to the veno-occlusive disease before.

These multi-valent antibodies conjugated to CD3, again there’s a lot of specificity to the target, but you can expect cytokine release syndrome-like symptoms to be associated with these drugs.

Checkpoint inhibitors, our solid tumor colleagues have made us well aware of the unique toxicities here, multi-organ autoimmune toxicities, really any organ tissue is possible, as we’ve seen. And these can be very, very subtle and insidious.

And then with CD47, there is the potential for hemolytic anemia, and there’s ways that have been worked out, at least with magrolimab and with some of the other CD47 antibodies, to really mitigate this. But that's a potential concern.

And of course, the CAR T cell therapies for those of you who’ve experienced them have potential, very significant toxicities that require intensive care unit level care and cytokine release syndrome.

So, just going to take you through some cases to discuss some of the real-life situations that we’re facing and some of the tools that we have to deal with them.

So the first case is a 79-year-old healthy male, newly diagnosed AML. At diagnosis, he had intermediate cytogenetics with a trisomy 8, that's not good or bad, and mutations in ASXL1, BCORL1, TET2 and RUNX1. So, you have 2 mutations there, ASXL1 and RUNX1, that are on the ELN risk group list of bad mutations. So this is a patient who, based on this mutational profile, would be classified as adverse risk.

He was started on venetoclax plus azacitidine and achieved a complete remission with incomplete count recovery after the first cycle. He continued therapy and did quite well for over a year. But 10 cycles in, unfortunately, on a routine bone marrow biopsy with relatively normal blood counts, somewhat surprisingly was noted to have relapsed disease with 10% blasts.

Repeat sequencing was done to see if there was something that had been acquired that would shed some new light or information on the situation or even maybe give us a targetable opportunity, like a FLT3 mutation. Unfortunately, those opportunities did not present.

He stopped the azacitidine and the venetoclax. Had 2 cycles of docetaxel and then disease progression. Again, we always check for any mutations on every one of these intervals because the acquisition of something targetable can happen and can give us an opportunity we didn’t have.

It was noted that this patient was CD33, as most AML patients are. So, we did try a course of gemtuzumab, and it was 3 doses given over a course of 28 days. Repeat bone marrow biopsy now at 50% blasts. Again no mutations to speak of. And at this point, we were, outside of clinical trials, out of options. And this patient opted for hospice and passed away shortly thereafter.

I think this illustrates a major challenge in the field, is that venetoclax has given us this incredible opportunity to have this really stratospheric improvement in outcomes for patients that previously, very recently, in the recent past, had very few options. But once a patient progresses on venetoclax options are really limited. So targeted therapy if you can get one, a FLT3 inhibitor or an IDH inhibitor makes sense. Most patients just statistically don’t have one of those mutations and that's where it gets really challenging. There isn’t a whole lot out there from a conventional treatment standpoint to offer these patients. These are patients that really should be enrolled in clinical trials. But we also know from data published by the MD Anderson group that relapse after venetoclax-based regimen has a very brief survival, something on the order of 2 to 3 months. So, really, really grim prognosis when that happens. A lot of need here to develop new drugs for this specific situation.

And then the last case, 67-year-old woman with hyperlipidemia, hypothyroidism. A 4-year history of low-grade MDS that was well known, followed expectedly. Never needed treatment; just some mild cytopenias.

She comes in with 2 weeks of fatigue, weakness, easy bruising and dyspnea. And like we talked about from the beginning, upwards of a third of MDS patients will progress or evolve to either higher grade MDS or AML. And that was definitely the suspicion when we heard of this patient’s symptoms. The bone marrow biopsy in fact confirmed this. She did have AML with a complex monosomal karyotype mutations in TP53 and PTPN11.

She was felt to be a suitable candidate for induction chemotherapy. And so our choice was 7+3 versus CPX-351. This is a patient what we would have wanted to transplant if we could get her into a remission. And so, for all those reasons, the fact that she had secondary AML and there’s a survival benefit compared to 7+3 and the really promising long-term outcomes from CPX-351, patients who go on to get a transplant, CPX-351 was prescribed. She got this therapy according to the package label.

She did experience a complete remission. She had detectable residual disease after a prolonged period of count recovery. These patients are extraordinarily difficult to get into any type of remission with any therapy, any conventional therapy, and that's suggested by the monosomal karyotype and the TP53 mutation. And then, while this patient was contemplating whether she wanted to pursue a transplant in first remission, she’s unfortunately admitted with a neutropenic fever and passed away fairly rapidly from a septic event. And this was about 10 weeks after initial diagnosis.

So this is such a difficult disease for so many reasons. We have all these new and exciting tools and we are getting better as result, but, still, lots and lots of challenges, lots of opportunities and lots of room to improve.

DR LOVE: I just wanted to actually ask you a few questions. There was a trial that got stopped that we just wanted you to make a comment on.

DR POLLYEA: The pracinostat study got started a while back. Again, in the sort of pre-venetoclax era. These were, I think, mostly older patients that were not fit candidates for intensive chemotherapy and a reasonable thing to do. Hard to know, in light of the way the field has changed, how this would fit into the current landscape, given that now the control arm would be venetoclax plus azacitidine. But we did hear a couple of months ago that there was — an unlikely scenario where these patients would have an efficacy level that would make it suitable to continue. So, for that reason, it was stopped. And we don’t know anything more.

DR LOVE: Another question I want to ask you. The second case, the lady who tragically died, was she on active treatment at that point?

DR POLLYEA: So, this is a patient who had completed her treatment with the CPX-351 and we were waiting for her to more fully recover counts. But I mean she hadn’t gotten any actual treatment in 3 to 4 weeks. So she was still recovering from —

DR LOVE: But she was still — So she still had recovered?

DR POLLYEA: Correct. Yeah, and still pretty immune compromised.

DR LOVE: Wow.

DR POLLYEA: I mean, 67, intensive chemotherapy in the setting of a preexisting MDS, the bone marrow just don’t pop back up like a 25-year-old de novo AML patient. So this is something we really struggle with.

DR LOVE: A couple of other things. I noticed on your list of new trials there are a couple, that I thought was pretty interesting, palbociclib? The CDK4/6 inhibitor?

DR POLLYEA: Yeah. I noticed that, too. And I know the breast cancer folks have a lot more experience in that than I do. I’ve seen some of these papers and it's pretty compelling that there may be a signal for activity there in AML/MDS. And I’m not too sure the details on the combination with CPX there. But I know some of my colleagues are pretty excited about that mechanism.

DR LOVE: That's interesting. And another one that was kind of interesting was combining it with ruxolitinib and MPNs?

DR POLLYEA: Yeah.

DR LOVE: I mean that seemed kind of interesting.

DR POLLYEA: MPNs have such high need and I think we’re all willing to try whatever we can. Ruxolitinib is sort of the standard backbone for anything MPN related. So, it's as good of an idea as any in MPN. We really need a lot of good stuff there.

DR LOVE: I mean has like 7+3 or chemo been looked at in MPNs?

DR POLLYEA: Traditionally, that doesn’t work very well to the standard 7+3 or intensive chemotherapy. But nothing has worked very well. That microenvironment of the MPN patient is such a hostile place. It's just full of fibrosis. And I think there’s some hope that because CPX-351, that it may hang around longer in the bone marrow it may have different pharmacokinetic properties there, that may be able to overcome some of those bad risk features?

DR LOVE: Another question about CPX. I’ve heard it can be administered outpatient? Is that something you ever consider. I assume you never do 7+3 as an outpatient?

DR POLLYEA: Right. I mean we don’t. But I will say, there are some places, a few, that have worked out a protocol to give 7 + 3 as an outpatient. It's a very labor-intensive thing to do. And it's not unreasonable for those centers. There’s a lot of interest in doing this for CPX for a variety of reasons, and I know a lot of my colleagues do that or sometimes they give the day 1, 3, and 5 outpatient and then admit the patient for the period of cytopenias because it's such an intensive period for transfusion support, etc.

And so, I think places that have been working with this longer have more experience, have a little bit more confidence in that doing that. I believe that there’s a clinical trial specifically looking at the feasibility of outpatient administration, just to give us all a better idea of how — what the complications are there. But a lot of people do do that. Yeah.

DR LOVE: When you use CPX, is it completely with the idea of better efficacy? Or any thought about in any way of less toxicity?

DR POLLYEA: I think that there is a theme or an assumption maybe that it is less toxic. And when you look at the data, I don’t really see that myself. Patients usually do not have alopecia, which is very interesting, with it. And so, that is a toxicity that's very important to patients that can most times be avoided with CPX-351. Otherwise, in my view, my read of it, in my experience, the toxicity profile seems pretty similar to 7 + 3.

So, I encourage people who are considering whether to give it or not to not really let that factor in, unless the patient’s number one goal is to not lose their hair, which, we have those patients, and those things are important. So, outside of that specific example, I would say consider it very similar from a toxicity profile.