WIN 2016
Models and critical preclinical data shaping personal therapy
Prof Brian Leyland-Jones - Avera McKennan Cancer Institute, Sioux-Falls, USA
I think the one thing that has become abundantly clear from this conference is sequencing is going to be routine at initial diagnosis within a very short period of time. The second thing is because of feedback loops and parallel pathways what you are going to see is combination therapies in everyone and WIN is the model of this because it brings together triple therapies at a minimum. So WIN is unique in being the global organisation which covers this. The third thing, as Leroy Hood elegantly pointed out, he started the P4 initiative in 2012, sequencing is going to be routine at birth and instead of focus more on disease it is going to be focussed on maintaining wellness. So the whole of the sequencing of the genome will utterly transform medicine.
How do you feel about the costs of sequencing and the use of big data?
If you just think of a metastatic disease patient at this point in time it costs approximately $1 million per patient to treat those. If you think, if you refocussed all of that money and concentrated on, for cancer, initial aggressive treatment and longer term on completely maintaining wellness those should be our foci. Put all the money early in diagnosis, in prevention and you will avoid those huge costs at the end of life.
What else from WIN have you found interesting?
The cell business of organoids is fascinating. We’ve been trying to do in vitro sensitivity assays ourselves in our own lab and are very frustrated. What Rob showed is that these organoids are feasible, they reflect the complete clonal populations of each of the tumour. So it looks fantastic for drug sensitivity. It still has to be tested out, one of the key issues at the moment is cost. It’s costing around €10,000 per organoid but they seem to have got the timeframe down perfectly now; they can grow these up in four weeks and then do drug sensitivity testing. So in reality you could probably feedback to your patient in five or six weeks and offer them therapies which have been proven to work on their own tumours ex vivo.
Do you think you could offer completely personalised therapy by combining sequencing data and organoids?
You’ve hit on the nail exactly. So if you think you could do a whole genome sequence in the patient itself, you could actually do some proteomics, work out all of the networks that are regulating that tumour, that could be used for matching your drugs in a drug screening model in the organoids, still you would have a complete picture of the sequencing from the tumour plus in vitro ex vivo sensitivity testing and combine those and you have a truly individualised patient therapy.
What are your thoughts on Dr Livio Trusolino’s talk on precision medicine in colorectal cancer?
The PDXs were fascinating in that they had reflected, at least in their colon models, they’d reflected about ten different drivers amongst their PDX organoids. So what I had asked from the podium is whether you could actually take those ten or eleven genomic drivers in the colorectal population and then run a basket or umbrella trial and run drugs individualised to those. The only limitation on this, of course, is it would have to be combinations of drugs so you would have to in each of those PDXs identify at least a couple of drivers so as to give maximal gain to the patient.
What are your thoughts on the SHIVA trial and combination therapies?
The SHIVA trial had the fault, forgive me, of TAPER and MATCH with matching single drugs to the dominant driver. That is fantastic for pharma, it’s less good for the patient because of these whole feedback pathways that always keep the cancer alive. The cancer will do anything to evolve to stay alive. So all of these single agent trials, whether it’s going to be SHIVA, TAPER, NCI-MATCH, will not have dramatic results, one will only have dramatic results from the WIN motto that you need triple therapies to basically drive the tumour. So the future basket trials will have to be focussed on N-of-1 where you give triple combos to each individual patient. We have sequenced 500 patients in the recent past, we have usually treated these with double or triple therapies. At this moment we have a 60% RECIST response rate; in our fourth line breast cancer patients we unbelievably have a 57% RECIST response rate and in our neoadjuvant triple negatives, where we are adding targeted therapies onto standard therapy, we have at this moment a 79% pathological complete response rate. So they are small numbers, it’s early days, our drug matching is not perfect but this gives great hope when you apply triple therapies matched to the genomic drivers.
