I’m delighted to present this on behalf of the co-authors. Work of this scale is a team effort and there’s many people involved with this, there’s a large author list you can see. I should also say at the start I didn’t have room for my NIH disclosure form. I work for the US government, I do not speak for the US government, this is personal opinion and it’s not the official policy of the Health and Human Services or of the US government.

Really this study is predicated on a misperception of how we grade responses in acute leukaemia. Our first intent or the first endpoint we’re going for when we’re treating patients intensively with acute myeloid leukaemia is for a complete response. One of the known issues or one of the secrets of the past sixty years is there’s nothing complete about that complete response; patients remain at risk of relapse even after achieving the best possible response to initial therapy. So the typical treatment course for someone who is young enough to receive intensive therapy, young enough and healthy enough to receive a further therapy, is to take people that are highest risk, so in a complete remission, no evidence of leukaemia by cytomorphology and count recovery, to additional therapy to prevent a subsequent relapse. That additional therapy is a bone marrow or stem cell transplant and is effective in some but not all patients. So despite the fact those patients go into a transplant in remission there’s still a significant relapse rate post-transplant. So the purpose of the therapy is not met.

One of the questions with that therapy is how much therapy we do need to do with a bone marrow transplant. Really you can lump it into two main groups – do we do myeloablative, more chemotherapy, or do we do a reduced intensity regimen which makes the bone marrow transplant more tolerable but may not have the same anti-leukemic efficacy? So that was the thinking behind the study.

The purpose of the study is a higher intensity preparative regimen or conditioning in patients in AML in a conventional complete remission, so received their initial therapy, have met that highest endpoint, but have genomic evidence of residual disease before the transplant. Is that going to reduce relapse rates and improve survival compared to a reduced intensity conditioning regimen? 

We based this on a large randomised phase III study that was conducted between 2011 and 2014 by the Blood and Marrow Transplant Clinical Trials Network called the 0901 study. This was a large study that took adult patients in remission from AML and randomised them to either that myeloablative, or MAC, conditioning, the high intensity conditioning, or the reduced intensity conditioning. That study was actually stopped early, as many of you will know, it was reported in the JCO in 2007, but the study was stopped early because there seemed to be a survival advantage for those who were getting the myeloablative, the high intensity therapy. So our question was was the increased relapse rate seen in the reduced intensity arm equally true for all patients who got reduced intensity or was it just for those patients who had evidence of residual disease by high sensitivity tools?

I’ll just introduce here the idea of measurable residual disease. My one ask is that when you report MRD you call it measurable rather than minimal. This can be high burden disease and people are dying of this MRD and so it’s really just disease we can measure which is just more with higher sensitivity tools than the 1960 response criteria that we’ve previously been using. So in this case we used genomics, we used ultra-deep sequencing. So looking in patients who are in a conventional complete remission by cytomorphology to see just under the surface there is there still evidence of residual leukaemia burden and are those the people who do not benefit from reduced intensity?

So the headline results are here. This is a survival analysis, overall survival, and we now have 49 months of follow-up in survivors. So for those who had genomic evidence of residual AML before transplant receiving reduced intensity conditioning compared with the higher intensity form in multivariate analysis was associated with an increased relapse with a hazard ratio of 6, decreased disease free survival and decreased overall survival. In those who had no evidence of residual disease before it was actually a very even balance between reduced intensity and myeloablation by overall survival.  There are a couple of factors to consider here – there’s the mortality from the transplant itself, from the conditioning, and those mortality rates are high in the myeloablative transplant, so up to 30% mortality from the procedure itself, versus the reduced intensity where there’s a higher relapse rate. Those things counterbalanced in people who came in without evidence of leukaemia in the transplant. So in the NGS negative population there’s absolutely no difference in overall survival between those two approaches. It’s only in the people who have evidence of residual disease who seem to benefit from getting myeloablative transplant.

So the conclusions, the detection of an AML associated variant using ultra-deep next generation sequencing in the blood of AML patients in a conventional remission prior to transplant is associated with an increased relapse rate and inferior overall survival in those randomised to reduced intensity compared to those receiving MAC.

Prof Anton Hagenbeek
That’s the next speaker coming up. Okay, this concludes your presentation now. Thank you very much. Questions?

Neil Osterweil
Hi, Neil Osterweil with Hematology News. Last November in The New England Journal Washington University and St Louis researchers reported on genomic and epigenetic changes that seem to occur in patients with AML that seem to occur after transplant and think that these changes might be addressed by possible therapies, changes in both adaptive and innate immunity. Have you seen that or are you looking at that?

Dr Christopher Hourigan
Yes, so we were limited in this particular example of this phase III trial by sample availability and so the only samples that were collected on this trial that were accessible to us were immediately prior to transplant blood sample. The reason we did this particular analysis was these samples, it was a whole blood tube shipped at room temperature and put in a -80o freezer so not good for viable cells, not good for looking at the kind of correlates you’re looking at. We didn’t have any post-transplant, what we would have loved to do is look at exactly those kinds of things. So this was Tim Ley and his group’s work looking at the HLA loss and is there an immune escape responsible for relapse. We don’t have any of those post-transplant samples to look at but we would love in other situations to try and combine this idea of residual leukemic burden with the idea of immune surveillance and see if there’s some combination. We just didn’t have the samples here to look at it.

Neil Osterweil
Great, thank you.

Prof Anton Hagenbeek
Further questions? Yes, over there.

Nancy Melville
Hi, Nancy Melville with Medscape Medical News. I just wanted to ask, I know this advanced sequencing is not available to clinicians, so are there other assays that can be used for this purpose?

Dr Christopher Hourigan
That’s a great question and one of the distinctions the European Leukaemia Network has been really stressing is that some of the genomic tests we use at diagnosis really aren’t applicable to this remission setting. So the settings, the use of genomics to profile leukaemia is now standard in risk stratification but it only works when you’ve got a large leukaemia burden. So some of these genomic tests that are currently available can’t be used in the residual disease setting, they’re just not sensitive enough to be used. So the European Leukaemia Network put out consensus guidelines which were published in Blood last year which recommend in the first instance quantitative PCR for tracking core binding factor and NPM1 mutated AML, which is probably a quarter of cases, and flow cytometry for other cases. Next generation sequencing is not part of those guidelines yet, I think it may be in the future, I’m part of that committee, but at the moment the clinical recommendation is quantitative PCR and flow cytometry.

Prof Anton Hagenbeek
Thank you. Further questions? So basically your study shows proof of a dose effect relationship, right? The more you give the better the chance that you eradicate both measurable MRD plus unmeasurable MRD. Do you agree with that?

Dr Christopher Hourigan
Yes, absolutely. So we’ve known for 20 years that detecting residual disease in someone in remission is associated with relapse, so the prognosis question. I think this is the first evidence I’ve seen in AML of that and this is why we really went after those samples from the randomised trial. The first evidence that we can do something about it in AML. We know in other diseases we can but this is the first time in AML we’ve seen it. As physicians it’s not as fatal, we can actually do something about it.

Prof Anton Hagenbeek
And could you say something about the sensitivity of your assay because MRD might range from one leukemic cell per hundred bone marrow cells that cannot be detected morphologically to one out of 106. So what is your sensitivity, what is your detection limit?

Dr Christopher Hourigan
Yes, absolutely, that’s a great question. One thing I’d stress and I always stress to new people who join our lab, we have clinicians joining our lab, is we’re only measuring what’s in the tube. So we never… with the highest possible sensitivity test we can only measure the sample we’re given and we’re only measuring probably 50 μL, 100μL of blood of the total person. So it’s actually test positive for MRD in the sample we receive, not in the patient and that’s an important distinction.

So our test with next generation sequencing, this is as deep as people have done de novo discovery with genomics so it’s one in a thousand sensitivity. Those methods that are much beyond that, quantitative PCR, would be several logs lower than that but even with this relatively low sensitivity, probably ten-fold or twenty-fold deeper than conventional medicine, we still see this nice stratification of people who are negative and positive.

Prof Anton Hagenbeek
And was there a difference in tumour load between both groups, the group that underwent myeloablative chemotherapy versus the reduced intensity?

Dr Christopher Hourigan
No, they were actually well matched beforehand. We don’t have any of the post samples.

Prof Anton Hagenbeek
That’s important, of course.