Taking complex genomic analysis into practice

Share :
Published: 26 Sep 2016
Views: 1705
Rating:
Save
Dr Philip Beer - Sanger Institute, Hinxton, UK

Dr Beer speaks with ecancertv at IBCD 2016 about making use of genomic data in a clinically actionable setting.

He opens with comments on the work of Dr Moritz Gerstung, who spoke with ecancer here, and considers the issues of storage, access and regulation in genomic libraries.

Dr Beer describes how unregulated access and off-label use of precision medicine is damaging precision medicine in American clinics, compared to hurdles in data access and funding facing European researchers.

A webcast of his presentation is available here.

 

IBCD 2016 - Innovation and biomarkers in cancer drug development 

Taking complex genomic analysis into practice

Dr Philip Beer - Sanger Institute, Hinxton, UK


I followed an excellent talk from an investigator at the EBI Institute in Cambridge who has analysed a large leukaemia dataset and has shown that by adding in complex genetic information you can potentially impact the patient care pathway. This was an excellent study but performed in a research setting so my talk really followed on from that to say how can we take this kind of analysis into the clinic and specifically some of the hurdles involved in terms of data sharing, how you get enough patients together. The project presented had 1,500 patients in but work they have done on power has suggested you might need to analyse up to 10,000 patients of a given tumour type in order to pull all the relevant information out of this complex sequencing. So this has to involve people getting together, sharing data. That in itself is an issue, it needs to be resourced, properly managed etc., and the data need to be of the right quality, particularly clinical data which, outside of clinical trials, is often hard to get hold of really good quality clinical data. The patient records in the UK and many other countries just don’t contain the quality of data that’s required for these kinds of studies.

Another big issue relates to the regulatory side – how do you actually regulate the use, the generation and the use, of big datasets like that in the clinic. So even if you can construct them to the right quality standards, who are the competent authorities, what kind of quality standards do you need to implement when generating these datasets and performing the analysis and how will they be signed off and regulated? So there has been some interesting discussion in other sessions again around how that is being done across the UK, across Europe, across America. There’s a particular issue around the complex bioinformatics analysis that needs to be performed, even to core variance in next generation sequencing data, and then beyond that to construct complex algorithms to make patient predictions. At the moment there’s a lot of variability, there isn’t one standard that people are following or adhering to and regulatory bodies are not imposing a single standard. So there was a talk from a representative from Foundation Medicine in America and they have a good system to regulate their algorithm development but it’s not following an internationally recognised standard. So I advocated using something a little bit more regulated that could be used by everybody who is following the same standards essentially.

Do you think the debate on personalised medicine will spark a lot of conversations afterwards?

It was an interesting debate in that it was framed as really a black and white, yes or no, question. The think that struck me was that the speakers seemed to agree with each other most of the time in that we all agree, or most people agree, that there is something in precision medicine, that this will have some impact on patient care. We also agree that there’s a lot more work that needs to be done, we’re really taking the first steps, that this is not being enacted now. What is clear is that there is a problem which is pretty much exclusively an American problem is the extensive off-label use of targeted therapies being guided by genomic tests that are being generated in a unregulated fashion. So unfortunately the bigger problem in the UK and across Europe is actually getting access to drugs for their on-label use, for their authorised use, getting stuff funded. We don’t seem to have this big unregulated market. So I would agree with the speakers that this isn’t a good thing, this isn’t a good place to be.

But really the key thing, we kind of went through this with the sequencing, the whole genome, that there was a big build up, then it was completed, we had the first copy of the human genome. Within weeks, months, people are saying, well, what good did that do us? But of course it was just the first step and we still haven’t unlocked all of the exciting discoveries that will come from that. Really it’s the early stages for precision medicine, that this is a long path and we’re not there yet, it isn’t changing the way that cancer is being treated yet but it is changing mind-sets and how people are approaching drug development and research etc. But what we do need, we have a lot of genetic information now about cancer, a lot of big sequencing studies. We know what the mutations are and a next big challenge is to work out what they mean. So it’s linking genomic data from tumours to what happens to the patients and particularly how they respond to therapy and how they respond to different types of therapy. That will really then begin to unlock the power of the complex genomic analyses.

Do you have any other thoughts on the conference?

Another key thing that has come out of this conference which has had representation from patient advocates as well is the need to have better sharing of information and better involvement with patient groups. So the long-term aim of precision medicine is that patients are treated as individuals and that treatment is tailored for specific patients with specific tumours in cancer. But the pathway to that will generate uncertainty because the amount of data that will be generated is difficult to assimilate. If you go beyond the realms of, OK, you’ve got lung cancer, it’s not curable, we’ve got one chemotherapy regime that might buy you some time. Suddenly you’re into the realm of you’ve got lung cancer and you’ve got a complex genetic profile, there are some drugs that we can use to match up with this, we don’t know quite how they all work together yet, there is going to be some risks, benefits with all of these. That the precision actually creates a lot of complexity and it raises a lot of issues for us as physicians in how we deal with that. So there needs to be more training in terms of helping people to understand the genomic analysis. Most people here went to medical school at a time where this kind of thing was not taught, it wasn’t appreciated. So education for physicians in the genomics but also in the statistical side, dealing with probabilities – how do you explain 5%, 10%, 15% differences to patients? How do you match up different probabilities from different trials? For many of us there will be a big time impact on that, there’s a lot of pressure on healthcare delivery and this kind of information takes a long time to discuss, to put across for people to understand.

So I think there is something that has come out of this conference for me going forward is that I think there is an increasing gap in connecting with the patient groups.