ASCO 2013

The problem of unnecessary treatment of prostate cancer

Dr Rosalind Eeles - Royal Marsden Hospital, London, UK

Ros, it’s great to get you here with us on ecancer.tv. A big problem in prostate cancer is unnecessary treatment and unnecessary biopsies, for instance, are quite a risk. So what have you been doing using genes to target biopsies a little bit better in prostate cancer?

Indeed, what we find is that if you screen the general population, just based on age, using a test called PSA or prostate-specific antigen, that’s a blood test to determine who should be biopsied, then you end up biopsying about 15% of men and detecting very few cancers, less than 5% of the total population will have cancer. Those prostate cancers that are detected are, in the main, quite indolent. Over two-thirds of them are indolent i.e. if you watch them they’re very unlikely to progress in that man’s lifetime. Whereas if you use genetics what we’re hoping to do is determine a man’s genetic make-up to determine whether he should be offered a prostate biopsy. We’ve been running a study called IMPACT since 2006 and to date we’ve published on the first 300 cases in the screening study. Men that have a mutation in BRCA1 or BRCA2 that normally increases the risk of breast cancer or ovarian cancer in women have an increased risk of cancer of the prostate. What we found in the pilot publication was that we don’t biopsy more men than you would in a population-based screening study but those that we do biopsy have a higher chance of having prostate cancer in their prostate biopsy.

Now with BRCA1 and 2, those are genetic markers that indicate a risk group for men, then?

Yes, they both increase the risk of cancer of the prostate and what we found in the pilot study was that if you biopsy a man, particularly if he has a BRCA2 mutation, then he has nearly a one in two chance of having prostate cancer on the biopsy result.

So you can be absolutely sure, really, that you have a good chance of biopsying the right person?

At the moment we only have pilot data and that’s on the first 300 men. What we found in the pilot as well was that instead of in the general population where over two-thirds of the cases are indolent, what we’re finding in the pilot data are that two-thirds of the men have disease that is aggressive and should be treated. So we see the complete flip-side from what we see in population-based screening. We’re detecting a number of cancers that are more aggressive and therefore would be significant for that man in terms of potentially shortening his life.

Now if you were depending on PSA to choose your men then how different would that be?

In the IMPACT study we biopsy a man if his PSA is greater than 3 and that’s the algorithm used in the general population screening studies that have been published from Europe. At the moment we’re currently analysing the data of the entire study so that’s just over 2,000 men in the entire study and we’re hoping to submit this paper for consideration for publication this year.

Just how big an improvement in the refinement of screening the correct people is the choice of the BRCA gene making, then, at the moment?

What it results in, according to the pilot data at least, in the first 300 men, is that you’re actually targeting those men that would most benefit, might potentially most benefit from treatment.

Under conventional targeting what is the norm, would you say then?

In population-based screening about a quarter of men might have disease that is clinically significant whereas what we’re finding in the targeting using the BRCA mutations is that over two-thirds of men have clinically significant disease.

What about the severity of that disease? Because many older men have prostate cancer; many of them might not need to worry about it.

The severity is based on accepted national guidelines, so they’re called the NICE guidelines, and the severity is what we call intermediate or high risk disease. We know that has a very high chance of progression within five years of a diagnosis if it’s left untreated.

The BRCA genes are one thing, do you have other genetic features that could be used to give you more refined diagnosis of high risk disease, for instance?

Indeed. The BRCA genes are relatively rare in the British population, however there are genetic variants that are much more common, they’re carried by, in the main, over half of men. However, each of these only confer a very slight increased risk of cancer of the prostate. However, they’re very important because they’re common and therefore if you carry a lot of them then their risks multiply together. So you can end up with quite substantial risks. To put that in context, if you have a room full of a hundred men and you profiled them for these common variants that we’ve now found all over the world, we’ve discovered 50 of these and there are 77 known. If you profiled the room with a hundred men then the man at the top 1% of the risk distribution in the room has a 4.7 fold risk, compared with the average of the men in the room. What we don’t know as yet is whether if you did screening based on that type of genetic profiling whether you would pick up disease that is indeed clinically significant. We’re hoping to start a study later this year that’s called the PROFILE study which will try to answer this question.

What should cancer doctors be making of this? As you say, there are certain unknowns about it at the moment too.

The main message, particularly from the BRCA work, is that it does look like, particularly with BRCA2, that prostate cancer associated with germline alterations in the BRCA2 gene that are harmful is a nasty disease. It’s more likely… we’ve recently had a publication in The Journal of Clinical Oncology, it’s more likely to be associated with nodal spread, three times more likely. We now have new drugs as well that can target tumours that have these specific alterations. At the moment they’re available in trial only, they’re called PARP inhibitors, but there are new drugs coming along that specifically target these types of alterations.

And how do you advise doctors to choose their patients for having this profiling?

What we’ve been finding is that individuals who have prostate cancer at the age of 65 or less have a very small chance of having a BRCA mutation, only about 1.5% chance of having BRCA1 or 2 when combined together. However, if a man has a family history of early-onset breast cancer and/or ovarian cancer, particularly multiple cases, then that man might want to consider having genetic testing for BRCA mutation.

Now when you move on to the full profile, taking into consideration all of the different molecules that could be involved, then do you simply profile everybody?

I think we’re a bit of a way from that time. However, I think it’s very important that we do the right research to be able to answer that question and indeed those types of studies are starting to be designed.

Is your feeling, you seem to have a very confident knowledge of what’s going on, which features genetically are influencing the disease, do you think that medicine is changing and we’re going towards genetics and away from purely clinical factors?

I think medicine is definitely changing. It’s changing from a generic treatment of, say, prostate cancer as one disease to what we call designer care or stratified medicine where you actually stratify patients both by their risk and also maybe by the types of changes within their tumours. We’ll be administering drugs based on genetic changes within tumours, rather than a generic classification of a tumour, say, as just a prostate cancer.

So what’s your take home message for the average doctor?

I think the take home message at the moment is if you have a patient in front of you that has prostate cancer diagnosed at less than the age of 65 and they have a strong family history of breast and/or ovarian cancer you might want to now consider referring them to a cancer geneticist for genetic testing for BRCA mutations. It looks like with good research studies that we might eventually get to the day where we can actually do targeted screening based on risk profiling based on genetic profiles.

And is it fairly easy to do these genetic tests?

At the moment the common variants are very easy to do. The BRCA mutations are easy to test for in terms of the technology is there but they are expensive. So many research labs are trying to develop now faster types of genetic test using a technology called next generation sequencing and that will bring the costs down. I predict that the costs will be far less than doing an MRI scan.

What sort of sample do you need?

A blood sample or a spit sample.

Fine. Well, Ros, it’s exciting to hear about all of these developments. Thank you for telling us about them on ecancer.tv.

Thank you.