Combining the novel anti-CD47 antibody magrolimab with azacitidine in patients with TP53-mutant AML

Bookmark and Share
Published: 26 Jun 2020
Views: 1218
Dr Naval Daver - MD Anderson Cancer Center, Houston, USA

Dr Naval Daver speaks to ecancer in an online interview for the virtual EHA 2020 meeting.

He describes the background, study design and results of this phase 1B trial, which investigated the combination magrolimab and azacitidine in patients with untreated acute myeloid leukaemia (AML) and a TP53 mutation.

Dr Daver concludes that this novel treatment combination is well tolerated and effective in patients who are unfit for chemotherapy, and follow up studies are planned to investigate this further. 

ecancer's filming has been kindly supported by Amgen through the ecancer Global Foundation. ecancer is editorially independent and there is no influence over content.


The abstract I presented at the EHA 2020 meeting was a study that we looked in older patients with acute myeloid leukaemia who were ineligible for intensive chemotherapy, implementing a novel combination of azacitidine with a CD47 antibody called magrolimab. This study was based on extensive preclinical data that has been done and published by the group at Stamford looking at the use of CD47 blockade in patients with acute myeloid leukaemia.

CD47 is a very important immune checkpoint present on the surface of macrophages that when engaged inhibits the macrophage from performing their routine function of phagocytosis of foreign cells which includes tumour cells. The preclinical data showed that by blocking CD47 using magrolimab which is an IgG4 CD47 antibody we could get a high degree of activity, especially a high degree of synergy in combination with azacitidine, in preclinical acute myeloid leukaemia mouse models.

This led to the combination of the azacitidine with the CD47 antibody in the ongoing phase Ib study. The study has two arms, one of them is looking at patients with acute myeloid leukaemia and these are patients who are older and considered to be not suitable for induction chemotherapy, either because of age alone or because of other comorbidities that would put them at high risk from toxicity mortality from the induction chemotherapy. The other arm is looking at patients who have frontline myelodysplastic syndrome, this included patients with intermediate, high, very high IPSS risk with MDS. I’ll be focussing on the AML data, my colleague Dr Sallman presented the frontline MDS data. Both these datasets are looking very encouraging at this time-point.

So in acute myeloid leukaemia one of the reasons we think that there is a distinct synergy between magrolimab CD47 antibody and azacitidine is because azacitidine actually has been shown to upregulate pro-phagocytic signals on the surface of leukaemia cells. This includes signals such as calreticulin and other pro-phagocytic signals and we’re looking at this in detail in the ongoing study. Usually in the presence of the upregulated pro-phagocytic signal azacitidine there was a much higher synergy and activity with the CD47 antibody.

So at this time we had a total of 29 patients that were evaluable at the cut-off for this abstract. This was enriched by an older patient population, the median age was 75 and the upper age extreme was 89. So these were quite old patients. Also important among the patient characteristics was that we had an especially high predominance of patients with poor risk cytogenetics – 72%, or three-quarters, of the patients had poor risk cytogenetic as per the ELN and about half of the patients had TP53 mutation. This was partly by design, about five or six months ago when we started seeing encouraging responses and duration of response in TP53 we amended the study to actually only focus and enrol frontline TP53 mutated AML which is why we see much more TP53 AML in this study.

The safety is probably the thing to me that is most exciting at this time about this agent, given that we have early follow-up. The combination in general was very safe, in fact I would say very similar to our experience over the last decade with azacitidine alone. We did not see any immune mediated toxicity such as organ inflammation, immune itises. We did not see any severe or prolonged or unexpected cytopenias. The one thing that we do see which was predicted and expected based on the preclinical data was an on-target anaemia. This is because CD47 actually is heavily expressed on the surface of older red cells so when you give the CD47 antibody there is an initial clearance of the older red cells and then the younger red cells, the reticulocytes, have come in, often expressed low or no CD47. So to try to mitigate this anaemia we’ve used an intra-patient dose ramp up escalation schedule. With that we did see anaemia but it was transient, reversible and no patients had to come off study. There were no severe effects from the anaemia and it was quite manageable. Usually by day 12-14 the anaemia regresses consistently in most patients.

Also in the safety, especially in this high risk group that I think is really most critical, the sixty day mortality was zero. This was true in both the AML and MDS studies. So aside from response and survival the third parameter we usually consider important in older AML is sixty day mortality and that was very favourable in this study.

From an efficacy point of view we had 25 patients who had completed the time period of at least two cycles. A couple of patients are still too early at the time of the data cut-off. The overall response rate was 64% which included a true CR/CRi rate of 56% and then a few additional MLFS PRs. This is encouraging, definitely better than the expected CR rates of 15-18% that have been shown with azacitidine decitabine alone in phase II/phase III studies in similar AML populations.
More importantly is the durability of the response. At this time we see that the median duration of response has not been met with a median follow-up of about 8½ months. Also, importantly, we looked at some deeper response parameters including MRD by flow cytometry and cytogenetic responses. We see that about 50% of patients have a cytogenetic response and 50-60% have MRD negativity, indicating that these are, in the majority, quite a deep response.
We looked at survival. The six month projected survival is 91% and at this time the median survival has not been reached and the survival curve looks quite encouraging.

The last thing which I think is probably going to be very important for the development of this drug in AML is that we actually did have a good number of TP53 AML patients. In this study there were twelve in total out of the 25 evaluable and among the TP53 the CR/CRi rate was 75%, the median duration of response was not reached at this time with a median follow-up of about 8-9 months. This compares favourably to azacitidine venetoclax which is a new standard for frontline older similar AML where the response rate CR/CRi is about 47-50% and the median duration of response is around 4-5 months.

So, again, early days; we need to build on this. The study is only enrolling TP53 AML, we should have more patients enrolled by the end of the year and hopefully we’ll get an update. But the future development of the azacitidine magrolimab in the AML is going to be focussing heavily on the TP53 mutated AML population. Hopefully we will have some interesting activity in this subset which continues to remain probably the most difficult subset in AML and MDS. Thank you.