This symposium was dedicated to trying to find the next questions for immunotherapy. As we all know, immunotherapy is a revolution for the treatment of cancer; we have outstanding results for the treatment of certain tumour types including melanoma, lung cancer and many others. Still there are a lot of questions – why are the majority of patients not responding? How can we do better? This forum was dedicated to these questions and actually it was a point of view of a major academic researcher from a pharmaceutical company, from a [?? 0:36] representative and also basic researchers and in general researchers dedicated to immunotherapy in all its dimensions. The output was to try to find the next question to be addressed and there are several important questions which were pointed out – how to combine the immunotherapies and the traditional treatment; how to identify also those patients who seem not to respond at all. They have been qualified as hyper-progressors; nobody knows exactly if they exist but there may be something here which is intriguing, biologically speaking, and maybe you should do something else for these patients. This was what the symposium was all about.

Has this changed the outcomes for patients?

Yes, exactly, the outcome of patients has been completely changed. In these diseases which we were mentioning we now have a significant proportion of very, very long-term survivors. Patients who have been treated ten years ago, five years ago, with metastatic melanoma, with lung cancer, metastatic lung cancer, who are still alive and not progressing. Very interestingly sometimes immunotherapy can be of short duration and have still very long-term impact on the patient outcome. So this is something which has never happened ever in the field of oncology, this observation that we have such a significant and constant proportion of patients with long-term survival, patients who do not seem to relapse, patients who are tumour free or non-progressing. Of course it’s not enough, again, the percentage of patients who are doing very well is only a minority, it’s probably in the range of 10-20%. It seems to be better, by the way, with combinations, at least with some combinations but this is limited to a small number of diseases so far. This is all the fields that we have to explore.

What is quite clear is that we will need to explore that in a smart way because there are a lot of immune modulatory agents which are currently being developed: ICP [?] blockers, of course, but also stimulators of the immune system and targeted treatment, how to combine them. We have many, many different disease types, histological subtypes, molecular subtypes of histological subtypes, immune subtypes of molecular subtypes. The landscape is exploding in myriads of very rare diseases and we cannot explore each and every disease with any combination which is available because that would be thousands of clinical trials which are on the table to do. So we need to be smarter and to define rational ways to develop these new treatments. That was one of the remarks of the symposium yesterday which was the following: how can we be smart to develop the next combination? Everybody was agreeing that looking at preclinical models was the right way. We should not be empirical and explore the next combination or association which comes to our mind; we should be based on science and on experimental models in order to be as efficient as possible. But really that’s a question which is very open and maybe other types of scenarios will be emerging such as big data in trying to figure out these big databases of information on the immune and molecular subsets. That’s the future.

Any ideas of changes to the way that clinical trials are conducted in the future?

The clinical trials that we have been doing in the past, putting ten thousands of patients and comparing treatment A versus treatment B is no longer going to happen. It may be important for specific dedicated questions here but that’s going to be very much a minority of trials. The good trials are going to be trials which are focussed on a right population where we have understood what was going on and to which we will apply the appropriate combination of sequential treatment. Sometimes even trials as original or atypical such as n of 1 trials, where the patient is his own control for subsequent treatment, where you have so rare patients, patients are so unusual in the general population that we cannot do a multicentre clinical trial even in this situation, this is happening more and more. I think this fragmentation of disease, of nosology, of taxonomy of disease as we call it, is going to have a major impact on the design of clinical trials which is going to be more and more early phase and jumping immediately to the demonstration of the proof of concept.

Are there any other highlights from the conference?

The integration of big data and the extent to which molecular characterisation is penetrating the clinical field is very impressive. The conference has been very good for that in a broad sense. What we observe right now is that technologies that we are using for categorising the tumour of a given patient is changing almost every six months with cost decreasing , with improvement and refinement in how to analyse that, looking at liquid biopsies. That’s an unprecedented situation where the field is moving so quickly that the standards cannot cope with that. To elaborate, clinical practice guidelines need at least follow-up long-term or middle-term and to be able to have validation studies. This is not any more possible because it’s moving so quickly. That’s going also to be the case in the future. We were hearing a few minutes ago that probably the cost of whole genome sequencing may be as low as 200 euros or dollars in the coming five years. Then the problem will not be technical, the problem will be bioinformatics. So bioinformatics and penetration in the clinical area is also a major message of this conference.

The medical scientists used to do lab research, wet labs, I think in the future they will be doing more and more dry lab research, bioinformatics, to exploit what is ongoing and to exploit not only the molecular and immunological data but also the data which are relative to the patient – age, where he or she lives, treatment, medical history and so on. All these big data are going to change completely how we treat the patient and going to change also the patient’s life because not only improving patient life but the patient is going to use these tools to be better managed.

Is biology now becoming something between engineering and sociology?

Absolutely, I think that’s a good definition. Biology is much more integrated in the general landscape of sciences than what it used to be. Mathematics as well, possibly physics as well are going to come here in the landscape where nobody expected that. Mathematicians are crucial for us to help us cope with this big amount of data because nobody knows how to do that. But the social aspects are also very important and they are part of the big data so, yes, it’s much more integrated.

What does the future hold?

One of the discussions which took place in the symposium was whether a computer could replace doctors. One good answer was that actually it’s not going to happen ever because these are tools, this is improved reality, this is not a major change of practice. At the end of the day we need to have a careful analysis done by a human being, exploiting the best tools that we have at our disposition to make the best choices. But this will help us to avoid mistakes, this will help us, therefore, to improve the decision process, improve ultimately the outcome of patients and this will help patients as well to improve their way of handling their disease and their life in general.