EACR 21, 26—29 June 2010, Oslo

Interview with Professor Carlos Caldas (University of Cambridge, UK)

Determinants of outcome in breast cancer patients

So what my group is interested in is understanding the molecular taxonomy of breast cancer, that is how distinct are different subtypes of breast cancer, how do they behave in the clinic and how they might respond to different therapies.

So we’re doing this work in two complementary approaches; one is taking advantage of samples from tumour banks from five university hospitals in the UK and in Canada for which we have more than five years of clinical follow-up and actually, for a lot of the patients, more than ten years. We’re profiling these exhaustively with array technology, looking at both copy number, gene expression, allelic information because we’re using SNP-based arrays and microRNA arrays. And then in compliment with that we’ve also very large collections of paraffin-embedded tumour blocks from patients, both from a population based collection in East Anglia and from clinical trials. These are close to 8,000 breast cancer cases and soon we’ll be close to 10,000, so we can ask important questions about survival and subtypes and, in particular, how survival changes with time. One of the observations that we have made is that breast cancers that have the better prognoses in the first five years after diagnosis, if women survive five years paradoxically the ones that have the worse prognosis subtypes, if they survive five years a lot of them are cured, whereas the oestrogen receptor positive breast cancers that are well differentiated and that have better prognoses in the first five years, those women continue to die of breast cancer thereafter. So paradoxically they have worse prognoses after the first five years.

Why do you think that is?

The biological implications of this are that oestrogen receptor positive breast cancers that are well differentiated maybe are an indolent disease and somehow the stem cells of these tumours divide less frequently and they can remain dormant for a long time and then wake up for whatever reason and divide again and eventually kill the patient.

Could this be related to patients receiving too much chemotherapy?

No, I don’t think so. What I meant by overtreatment, I should qualify that carefully, is the fact that with just good surgery and good local radiotherapy there is a good number of patients that are cured, we simply don’t know how to identify those patients. So what happens is that to save lives, and unquestionably we’re saving lives with adjuvant treatment, you need to treat groups of patients amongst which there will be patients that don’t benefit from it because they would be cured and other patients don’t benefit from it because they’re resistant to the adjuvant treatment. The problem we have is that we can’t identify those patients prospectively.

Is translational research starting to become an integral part of clinical trials?

We’re starting to see the beginnings of it but it’s very challenging. To do it properly requires infrastructure and therefore funding to do it. This is going to pose a great challenge, not just to academic endeavours but also to big pharma because until people can demonstrate clearly and unequivocally that there is clinical utility in doing this, in other words that by doing biomarkers that you’re really improving patient outcomes, but to get that evidence you need to do the studies in the first place so it’s a chicken and egg situation. Now we have been privileged to be in a healthcare system in the UK and with funding that has come into translational research over the last five years that has made these types of studies possible. I think the challenge is going to be, going forward, to continue to bring on that funding.

In the case of pharma, what they are worried is that these markers will further limit their market and slicing patients that they will not have as many of the blockbusters as they would hope to have. What we need to start to convince pharma is that actually it is to their advantage because if they would start selling drugs where as part of the prescription there is a diagnostic test, that diagnostic test indicates who is going to benefit or not, that that would be a benefit for patients and it would be a benefit for them because they would have less trouble with regulatory authorities in demonstrating activity and moreover, if you want, it opens a new business for them which is the molecular diagnostics business.