Of course retinoic acid is a therapy that is particularly successful in APL where there is a strong mechanistic rationale for using it based on the fact that the initiating oncogene affects the receptor for retinoic acid. Now a presentation here at ASH is showing that the addition of retinoic acid to non-APL types of AML, in particular those associated with mutations in NPM may benefit from the addition of retinoic acid.

Why hasn’t retinoic acid been used in AML sooner?

It’s a bit hard to understand, also it’s not entirely understandable how the [?? 0:48] is working in these types of AMLs because for APL it’s pretty clear, it’s directly acting on the initiating oncogene. For these other types of disease it could be a more generalised pro-differentiating effect which may be helping the effect that standard chemotherapy already has. So probably in the past it was harder to identify a specific sub-group in which this drug would have had some effect. Now, with the advent of increased genetic characterisation of AML, this is becoming better understood that there are specific subtypes that maybe benefitting of this therapy although through as yet not entirely well understood mechanisms.

Is it tricky to no longer use chemotherapy?

It is but it’s also, of course, extremely toxic and not always as well tolerated. Also the paradoxical problem that we have with chemotherapy is that very often, since a lot of haematological patients are young patients, there is a relatively high incidence of secondary tumours in chemotherapy treated individuals after a few years. So getting rid of chemotherapy could be, if you want, the next big step in oncology. We have already seen that happening in APL, for instance.

Now, genetic sequencing is really high on the agenda at most medical meetings these days, there has been a lot of talk about biology but I gather that in a lecture Professor Donner was talking about bench to bedside, actually getting it, really applying it. What do you make of this?

Yes, absolutely. Last year marked, if you want, the turning point in genetics into clinics because all the data of big genetic efforts, genetic screening sequencing efforts, have been published, mainly the TGCA programme. But a number of other consortia have published large scale genetic characterisation of several malignancies, including myeloid malignancies. Up to last year all this was an effort mostly directed to better understanding the biology of the disease but now we’re seeing that a number of concepts arising from all these efforts are making their way into the clinics. They’re certainly helping a lot to stratify patients according to different risk groups, probably defying what were the standard clinically based prognostic scores and maybe in the long run they’re going to also lead to the identification of precise targets that we can then attack with the specific therapies.

The best place to get this genetic information is during prognosis?

At the moment it certainly is the most immediately payable, if you want, benefit, absolutely. So it’s helping us to identify patients that would look, for instance, non-treatable or should, according to standard and classical prognostic scores, should be treated in a certain way and they are helping us to modify our treatment. For instance this is particularly evident in some diseases where you have patients where you are usually unwilling to carry out a certain treatment, for instance myelodysplastic syndromes in which a lot of very unexpected findings came from genetic screens, for instance mutations in splicing factors and epigenetic factors being extremely common and being differentially associated with a prognosis. So now you have a patient, as Professor Donner pointed out yesterday, that according to classical scores would not have been treated because his disease was thought to move on with a very slow pace but now you know that, for instance, there’s a p53 mutation and you are willing to treat that patient because it may evolve more easily into an AML.

Do you think doctors will be willing to contradict traditional prognostic features?

It’s going to happen for sure; we’ve seen that happening with pretty much everything in medicine. At the beginning it’s mostly probably based on the ease with which we are able to carry out this type of testing and up to now the real bottleneck has been cost and the complexity of the test but now this is becoming increasingly easier for routine labs to carry out. So it’s probably not going to encounter too much resistance.

Genetic sequencing is a fascinating subject here at the EHA, but just one more question on this, though, we tend to think that with sequencing, next generation sequencing, this detailed information of the genome we could find targets and perhaps find one really good target and a drug or perhaps a collection of targets and a collection of drugs which will really make huge steps forward. Is that your thought?

I think there has been a tendency towards thinking about next gen sequencing as a way to identifying mutated targets to attack. But I think we should move past this way of thinking because a lot of druggable targets are not actually mutated in these cancers. So we’ve seen it today, for instance, in the AML session with a number of drugs entering phase I trial that are not actually targeting specific mutations like Wnt inhibitors or NET inhibitors, they are generalised inhibitors. They inhibit specific biological pathways but these are not necessarily the pathways that are mutated in these patients. But what the genetic screen could allow us to understand is which patients are more susceptible to specific targeting, even if the drug does not target that specific mutation.

How do you think these genetic factors can optimise treatments?

Absolutely. So what’s going to happen is that we are going to slice each disease in smaller and smaller pieces and each piece is going to be highly sensitive to one specific drug. This is going to help us to really assign the best treatment in the least toxic way to each patient.

These could be conventional treatments that are targeted more accurately to the patient group?

That is correct. I was giving the example of a p53 mutated MDS patient that would not be treated, necessarily with an aggressive approach nowadays and now we know that those patients will have a much higher chance of shifting their MDS into an AML and so they will require a more aggressive treatment.