Can a blood test diagnose cancer?

Naked Scientist Kat Arney, also works as a representative of the charity Cancer Research UK.

She has been at the National Cancer Research Institute's annual conference, which just took place in Liverpool. There she spoke with Warwick University scientist Ian Cree, who's working on a way to hunt for cancer using a blood sample...

Ian -  What we're hoping is that we can develop a blood test for cancer.

Kat -  This seems to be the dream, isn't it?  We see these stories popping up every so often, blood tests could detect all kinds of cancer.  How does yours work and is it, the key thing, reliable?

Ian -  Well at the moment, we don't know.  What we've done is to trawl the literature and we've actually discovered 816 markers for cancer that could be used in this way.

Kat -  So, what do we mean by markers?

Ian -  So, markers are things that circulate in the blood.  Cancers have various attributes and each of those attributes produces things, changes in the blood which we can measure and we call those biomarkers.

Kat -  Now with your test, if you could pick up different types of cancer, how would you know what cancer it was, where was it in the body because I guess this is a problem?  You can't just go, "I think you've got cancer, but we don't know where it is."

Ian -  Well, we actually think that's one of the problems with the previous work in this area that everyone has been looking for specific markers for specific cancers.  We actually think that a generic marker will work better and that we'll need to use multiple markers and the negatives to help the positives, that sort of thing.  So that we're able to detect the cancers as early as possible.  It doesn't actually matter terribly much if we don't know exactly where they are because radiology is now good enough that we can detect some very small changes.  We know where most of the changes are going to be by taking the generic test and then triaging the patients to reflex what we call 'reflex tests'.  Those reflex tests will allow us to determine a bit more about the cancer site and if there actually is one there.  So, we'll increase the specificity and we'll increase the knowledge of where it might be.

Kat -  So, you'd give someone this test and say, "We think you should go on for a CT scan or further investigations."

Ian -  Exactly, yes.  That's right.  we'd look to see whether there was anything other that they could do that would help to determine whether they actually had a problem.  The whole point of this is to detect the cancers really early so that patients can be treated and probably walk out the door ideally the same day without ever having known they've had one.

Kat -  It sounds wonderful that you could have a blood test that would pick up all sorts of different types of cancer, but how soon do you think this would be actually available say, in the clinic?

Ian -  Well, I'm afraid like most of these things, it's going to take some time.  at the moment, what we've done is to look through literature.  We've collected a very expert group of people together, so we've got a fabulous team which is very experienced in looking at all the different biomarkers that we're interested in, and in the ancillary things that we need to do.  So, there a big statistical questions to do around this.  There's a lot of other work that we need to do before we can put it into the clinic.  We're hoping to start testing this next year and as a concept, and then develop something which would be then submitted to a clinical trial probably, in about 3 or 4 years time.