01.12.12
Citizen power
Source: National Health Executive Nov/Dec 2012
Professor Andrew Hanby, Cancer Research UK scientist at the University of Leeds, talks to NHE about a project which lets citizens do the legwork on cancer research.
Drawing inspiration from the stars, cancer scientists have called on the power of the massed public and a “monster” level of IT infrastructure to create CellSlider, an innovative new approach to research.
One of the Cancer Research UK scientists behind the project is Prof Andrew Hanby from the University of Leeds. He explained how everyday people could volunteer their time and efforts to speed up the discovery of targeted treatments.
An astronomical task
Considering the beginning of their breakthrough, Prof Hanby described the time he spent looking at hundreds and hundreds of images of cancer tumours labelled with stains to pick out particular molecules.
Meanwhile, astronomers were already using the kind of technology that would be implemented for cancer research to help classify a huge number of galaxies, which would be impossible for any one person to achieve on their own.
Cancer drug trials can often contain over 2,000 participants, and researchers have to classify each image individually. This data can then help to build up a picture of how certain tumours respond to different drugs, to allow scientists to predict more accurately which patients to put through treatment.
Professor Hanby said: “It’s great for the patient because you don’t put some people through treatment they don’t need, and it’s great for the NHS because it’s important from the economic side of things.”
The data from all these samples is put into a grid, with each trial often done in triplicate, or in sets of sixes to take into account the possibility of variation across the tumour.
“But it really starts to become a monumental task,” Prof Hanby said. “I’m trying to fit it in between my other research work and I’m also an NHS consultant, as well as a university researcher so I’m doing my NHS work too.
“If we could find a way of getting the folks out there to help us score these things, we’d speed it all up and also free me up to do some other work which isn’t so amenable for the citizen to do, work in the lab, that kind of thing. It’s all about helping to find better ways to tailor things for patients, so it’s very innovative.”
Built-in systems
To ensure the results are reliable, it is important to build in certain checks to the system. Prof Hanby described the “quite complicated tasks” that people using the system have to complete.
“There’s no doubt that there are some people who fi nd these tasks very diffi cult. One of the ways around it: we already know what the score should be. In that way, we can recognise the people who are the better raters from the not so good raters, and place some reliance on the results we get.
“You can build in systems: there are ways of being able to see that they map nicely to the ones we do know the results for, and therefore we do think it’s more reliable than this set of people.”
The idea is for people to look at very large cohorts of tumours, and rank the intensity of staining, so if tumours that have a lot of X do better with drug Y, then that will give scientists information about the kind of things to look at in a more focused way, with regards to targeting treatment for patients.
Casting a wide net
However, he suggested that it was not about setting criteria for people to take part, and that the aim of the project required as large a pool of participants as possible.
“We have to be careful here because we don’t want to be saying to people, ‘oh, you’re rubbish’, or ‘you’re good’.”
Prof Hanby added that he didn’t think it would be right to turn anyone away, as it would be “disheartening”.
“We certainly wouldn’t say that to anyone; we want everybody to be involved and contributing.”
He went on to explain that in the future a wide range of different tasks would be available, and that people who performed best at one may not be right for another, providing another reason not to turn anyone away.
Yet critics could point out that cancer research is a serious and difficult profession, not a popularity contest.
If the public are to be involved in helping categorise data that could ultimately go on to affect treatment, surely we should be enforcing strict protocol to ensure only the most accurate are enlisted?
Prof Hanby explained that following the identifi cation of molecules which respond well to treatment, more rigorous testing would ensue, quality standards would be set up and guidelines enforced. He said: “But in the first instance, we need to cast our net wide and try to identify the good candidates.
“It’s not so much about patients, as things – that will help us discriminate what are the important things, what are the less important things."
Citizen science
The scale of the project means application from research to treatment will not be easy, although it holds huge potential.
“Going from a study like this to using it in daily practice would be a large task, but we can now, assuming it all goes well, more quickly identify molecules which might be helpful in targeting drugs,” Prof Hanby said.
The use of citizen science in other forms of cancer, and even other types of disease is a possibility, although this is thought to be the first implementation of its kind in the field of healthcare.
“Logistically it’s quite a big task. Not only do you need thousands of images, you need to have the cloud-type web infrastructure to be able to support it all and allow thousands of hits.
“It’s not a trivial task and I suspect no-one is actually doing this. I think this really is a first, as far as healthcare goes. It’s an infrastructural monster basically. It could save literally thousands of man hours in terms of cancer researcher time.”
The speed of the project, and the time before it could be conceivably applied to other cancers, “depends on the citizens”, with over 9,000 unique visitors to the site and 92,000 classifications made at the time Prof Hanby spoke to NHE. “Its consequences could be immense,” he concluded.