The study was an analysis of genomic data from the ECOG-ACRIN E1910 study that examined the efficacy of randomised blinatumomab with chemotherapy for adults with Ph negative or BCR-ABL negative acute lymphoblastic leukaemia who achieved MRD negativity and could then be randomised to blinatumomab. The additional results of this study showing efficacy of blinatumomab were published last year in The New England Journal of Medicine and at that time we used a broad risk classification based on the subtype of leukaemia from prior studies of adult ALL that didn’t include MRD measurements and didn’t use blinatumomab. At that time it appeared that blinatumomab was effective across multiple groups of adult ALL, favourable, intermediate and poor risk, but we didn’t have clarity on several aspects and they included the overall genomic landscape of adult ALL, the genetic factors that determined the achievement of MRD negativity and the genetic factors that were associated with outcome with the use of blinatumomab. So the study that was presented at ASH provided a detailed update of the study including a comprehensive genomic portrait of the disease and associations with outcome.

What were the key results?

The results of the study were several. Firstly, turning to the overall genomic landscape there has been prior data looking at the different subtypes of adult ALL and some of the key drivers but this was the largest and most comprehensive genomic study of adult ALL that had thus far been performed. It showed that we could identify the 20 different subgroups and it showed that several subgroups were very common in adult ALL, including Ph+, Ph-like ALL, low hyperdiploid ALL and MLL-rearranged ALL.

We also identified several other groups and identified a group with diverse genomic alterations activating a family of transcription factors called the CEBP family, including CEPB alpha and CEPB beta. We presented some detailed integrated genomic and epigenomic data that described the nature of the alterations and the epigenetic drivers of deregulation of those genes. That provided a substory, if you like, of novelty about the nature of genomic alterations in adult ALL.

The next layer of findings were associations with MRD response and with response to blinatumomab. That showed that there were several groups that were associated with favourable response and several groups that were associated with relatively poor response, including Ph-like ALL and PAX5-altered ALL. These subgroup associations, again with achievement of MRD negativity, were statistically significant.

Then a key set of analyses were looking at the genetic features associated with blinatumomab response and we found that some groups had a very good response and some groups had an intermediate response to blinatumomab and some groups had a quite poor response. The quite striking finding was that cases that had low hyperdiploid ALL had a poor response with the addition of blinatumomab.

Those findings are, we feel, very important because they give potential insights into the types of patients that may benefit most from blinatumomab that will need to be tested in further studies. Biologically they also make sense, so specifically hyperdiploid ALL is associated with chromosomal losses, including the loss of the chromosome that encodes CD19, the target of blinatumomab. It’s known that failure of response to blinatumomab can in some patients be due to loss of CD19 and we expect that’s the basis for the poor response of at least some of these patients with that subtype of ALL.

In a final analysis we looked at the presence of what are termed myeloid mutations which are in other contexts often called clonal hematopoiesis of indeterminate potential mutations. These are mutations in genes such as P53, DNMT3A, ASXL1 and 2 and we showed that there was approximately 15% of adult ALL cases that had these myeloid mutations. Now, the majority of those were P53 mutations in patients with low hyperdiploid ALL which is a known association, a very important association, but, in addition, there were multiple other patients that had mutations of other genes. The interesting results there were that we found that the presence of these mutations was associated with an initial poor response but, in fact, the outcome of these patients with blinatumomab was quite favourable. It has been suggested that this might be the case and so this study provides, albeit in small numbers, but suggests that, indeed, patients with these myeloid mutations do indeed have a good response to blinatumomab.

So they were really the main key findings of the study that were presented at the meeting.

In what ways could these genomic insights guide more personalised treatment strategies in the future?

There may be important points that we shouldn’t assume that blinatumomab, as exciting as the data are in terms of response, we shouldn’t assume that it’s a cure-all and we shouldn’t ignore the nature of tumour-intrinsic alterations, specifically here looking at tumour genomics. So it strongly suggests that we must incorporate tumour genomics into future studies that are exploiting blinatumomab in the treatment of ALL across the age spectrum. How treatment might be tailored according to genomic alterations, those are questions for the future as future studies are designed. But, again, the key implication here is that we must consider genomic features as these trials are being designed.

Another point I would make is that this was a collaborative study directed by my colleagues, Mark Litsow and Selina Luger, and the team at the ECOG-ACRIN group and we co-ordinated the genomic analysis but it was a large team effort. So I recognise the many investigators that contributed to the study.